Merge branch 'mesa_7_7_branch'
authorBrian Paul <brianp@vmware.com>
Sun, 27 Dec 2009 17:24:21 +0000 (10:24 -0700)
committerBrian Paul <brianp@vmware.com>
Sun, 27 Dec 2009 17:24:21 +0000 (10:24 -0700)
866 files changed:
Makefile
SConstruct
bin/mklib
common.py
configs/darwin
configs/default
configs/linux-cell
configs/linux-dri
configs/linux-i965 [new file with mode: 0644]
configure.ac
docs/relnotes-7.8.html [new file with mode: 0644]
docs/relnotes.html
progs/demos/arbfplight.c
progs/demos/arbfslight.c
progs/demos/arbocclude.c
progs/demos/bounce.c
progs/demos/clearspd.c
progs/demos/copypix.c
progs/demos/cubemap.c
progs/demos/drawpix.c
progs/demos/engine.c
progs/demos/fbo_firecube.c
progs/demos/fire.c
progs/demos/fogcoord.c
progs/demos/fplight.c
progs/demos/fslight.c
progs/demos/gamma.c
progs/demos/gearbox.c
progs/demos/gears.c
progs/demos/geartrain.c
progs/demos/gloss.c
progs/demos/gltestperf.c
progs/demos/ipers.c
progs/demos/isosurf.c
progs/demos/lodbias.c
progs/demos/morph3d.c
progs/demos/multiarb.c
progs/demos/paltex.c
progs/demos/pointblast.c
progs/demos/projtex.c
progs/demos/ray.c
progs/demos/readpix.c
progs/demos/renormal.c
progs/demos/shadowtex.c
progs/demos/spectex.c
progs/demos/spriteblast.c
progs/demos/teapot.c
progs/demos/terrain.c
progs/demos/tessdemo.c
progs/demos/texcyl.c
progs/demos/texenv.c
progs/demos/textures.c
progs/demos/trispd.c
progs/demos/tunnel.c
progs/demos/tunnel2.c
progs/demos/vao_demo.c
progs/demos/winpos.c
progs/fp/fp-tri.c
progs/fp/mov-imm.txt [new file with mode: 0644]
progs/fp/mov-param.txt [new file with mode: 0644]
progs/glsl/CH11-bumpmaptex.frag [new file with mode: 0644]
progs/glsl/bump.c
progs/glsl/deriv.c
progs/osdemos/Makefile
progs/rbug/bin_to_bmp.c
progs/rbug/tex_dump.c
progs/tests/Makefile
progs/tests/texleak.c [new file with mode: 0644]
progs/trivial/.gitignore
progs/trivial/Makefile
progs/trivial/tri-orig.c
progs/trivial/vbo-tri.c [new file with mode: 0644]
progs/vp/add-param-imm.txt [new file with mode: 0644]
scons/gallium.py
src/SConscript
src/gallium/auxiliary/cso_cache/cso_context.c
src/gallium/auxiliary/cso_cache/cso_context.h
src/gallium/auxiliary/draw/Makefile
src/gallium/auxiliary/draw/SConscript
src/gallium/auxiliary/draw/draw_context.c
src/gallium/auxiliary/draw/draw_context.h
src/gallium/auxiliary/draw/draw_gs.c [new file with mode: 0644]
src/gallium/auxiliary/draw/draw_gs.h [new file with mode: 0644]
src/gallium/auxiliary/draw/draw_pipe_aaline.c
src/gallium/auxiliary/draw/draw_pipe_aapoint.c
src/gallium/auxiliary/draw/draw_pipe_clip.c
src/gallium/auxiliary/draw/draw_pipe_cull.c
src/gallium/auxiliary/draw/draw_pipe_offset.c
src/gallium/auxiliary/draw/draw_pipe_pstipple.c
src/gallium/auxiliary/draw/draw_pipe_stipple.c
src/gallium/auxiliary/draw/draw_pipe_wide_line.c
src/gallium/auxiliary/draw/draw_pipe_wide_point.c
src/gallium/auxiliary/draw/draw_private.h
src/gallium/auxiliary/draw/draw_pt.c
src/gallium/auxiliary/draw/draw_pt.h
src/gallium/auxiliary/draw/draw_pt_fetch.c
src/gallium/auxiliary/draw/draw_pt_fetch_shade_pipeline.c
src/gallium/auxiliary/draw/draw_pt_post_vs.c
src/gallium/auxiliary/draw/draw_pt_util.c
src/gallium/auxiliary/draw/draw_pt_varray_tmp_linear.h
src/gallium/auxiliary/draw/draw_pt_vcache.c
src/gallium/auxiliary/draw/draw_vs.c
src/gallium/auxiliary/draw/draw_vs.h
src/gallium/auxiliary/draw/draw_vs_aos.c
src/gallium/auxiliary/draw/draw_vs_varient.c
src/gallium/auxiliary/gallivm/tgsitollvm.cpp
src/gallium/auxiliary/pipebuffer/pb_buffer.h
src/gallium/auxiliary/pipebuffer/pb_buffer_fenced.c
src/gallium/auxiliary/pipebuffer/pb_bufmgr_cache.c
src/gallium/auxiliary/tgsi/tgsi-instruction-set.txt
src/gallium/auxiliary/tgsi/tgsi_build.c
src/gallium/auxiliary/tgsi/tgsi_build.h
src/gallium/auxiliary/tgsi/tgsi_dump.c
src/gallium/auxiliary/tgsi/tgsi_dump.h
src/gallium/auxiliary/tgsi/tgsi_dump_c.c
src/gallium/auxiliary/tgsi/tgsi_exec.c
src/gallium/auxiliary/tgsi/tgsi_exec.h
src/gallium/auxiliary/tgsi/tgsi_iterate.c
src/gallium/auxiliary/tgsi/tgsi_iterate.h
src/gallium/auxiliary/tgsi/tgsi_parse.c
src/gallium/auxiliary/tgsi/tgsi_parse.h
src/gallium/auxiliary/tgsi/tgsi_ppc.c
src/gallium/auxiliary/tgsi/tgsi_sanity.c
src/gallium/auxiliary/tgsi/tgsi_scan.c
src/gallium/auxiliary/tgsi/tgsi_scan.h
src/gallium/auxiliary/tgsi/tgsi_sse2.c
src/gallium/auxiliary/tgsi/tgsi_text.c
src/gallium/auxiliary/tgsi/tgsi_transform.c
src/gallium/auxiliary/tgsi/tgsi_transform.h
src/gallium/auxiliary/tgsi/tgsi_ureg.c
src/gallium/auxiliary/tgsi/tgsi_ureg.h
src/gallium/auxiliary/tgsi/tgsi_util.c
src/gallium/auxiliary/util/Makefile
src/gallium/auxiliary/util/SConscript
src/gallium/auxiliary/util/u_blit.c
src/gallium/auxiliary/util/u_blitter.c [new file with mode: 0644]
src/gallium/auxiliary/util/u_blitter.h [new file with mode: 0644]
src/gallium/auxiliary/util/u_clear.h
src/gallium/auxiliary/util/u_debug.c
src/gallium/auxiliary/util/u_dl.c [new file with mode: 0644]
src/gallium/auxiliary/util/u_dl.h [new file with mode: 0644]
src/gallium/auxiliary/util/u_format.c
src/gallium/auxiliary/util/u_format.csv
src/gallium/auxiliary/util/u_format.h
src/gallium/auxiliary/util/u_format_access.py
src/gallium/auxiliary/util/u_format_table.py
src/gallium/auxiliary/util/u_gen_mipmap.c
src/gallium/auxiliary/util/u_linear.c
src/gallium/auxiliary/util/u_linear.h
src/gallium/auxiliary/util/u_math.h
src/gallium/auxiliary/util/u_pack_color.h
src/gallium/auxiliary/util/u_prim.h
src/gallium/auxiliary/util/u_rect.c
src/gallium/auxiliary/util/u_rect.h
src/gallium/auxiliary/util/u_simple_shaders.c
src/gallium/auxiliary/util/u_simple_shaders.h
src/gallium/auxiliary/util/u_surface.c
src/gallium/auxiliary/util/u_texture.c [new file with mode: 0644]
src/gallium/auxiliary/util/u_texture.h [new file with mode: 0644]
src/gallium/auxiliary/util/u_tile.c
src/gallium/auxiliary/util/u_upload_mgr.h
src/gallium/auxiliary/vl/vl_compositor.c
src/gallium/auxiliary/vl/vl_mpeg12_mc_renderer.c
src/gallium/auxiliary/vl/vl_shader_build.c
src/gallium/drivers/cell/ppu/cell_clear.c
src/gallium/drivers/cell/ppu/cell_draw_arrays.c
src/gallium/drivers/cell/ppu/cell_gen_fp.c
src/gallium/drivers/cell/ppu/cell_pipe_state.c
src/gallium/drivers/cell/ppu/cell_state_derived.c
src/gallium/drivers/cell/ppu/cell_state_emit.c
src/gallium/drivers/cell/ppu/cell_texture.c
src/gallium/drivers/cell/spu/spu_exec.c
src/gallium/drivers/cell/spu/spu_util.c
src/gallium/drivers/failover/fo_context.h
src/gallium/drivers/failover/fo_state.c
src/gallium/drivers/failover/fo_state_emit.c
src/gallium/drivers/i915/i915_context.c
src/gallium/drivers/i915/i915_fpc_translate.c
src/gallium/drivers/i915/i915_state.c
src/gallium/drivers/i915/i915_state_derived.c
src/gallium/drivers/i915/i915_state_sampler.c
src/gallium/drivers/i915/i915_surface.c
src/gallium/drivers/i915/i915_texture.c
src/gallium/drivers/i965/Makefile [new file with mode: 0644]
src/gallium/drivers/i965/SConscript [new file with mode: 0644]
src/gallium/drivers/i965/brw_batchbuffer.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_batchbuffer.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_cc.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_clip.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_clip.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_clip_line.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_clip_point.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_clip_state.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_clip_tri.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_clip_unfilled.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_clip_util.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_context.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_context.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_curbe.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_debug.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_defines.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_disasm.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_disasm.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_draw.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_draw.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_draw_upload.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_eu.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_eu.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_eu_debug.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_eu_emit.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_eu_util.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_gs.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_gs.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_gs_emit.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_gs_state.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_misc_state.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_blend.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_clear.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_depth.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_fb.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_flush.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_misc.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_query.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_rast.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_rast.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_sampler.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_shader.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_pipe_vertex.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_reg.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_screen.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_screen.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_screen_buffers.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_screen_surface.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_screen_tex_layout.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_screen_texture.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_sf.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_sf.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_sf_emit.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_sf_state.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_state.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_state_batch.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_state_cache.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_state_debug.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_state_upload.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_structs.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_structs_dump.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_structs_dump.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_structs_dump.py [new file with mode: 0755]
src/gallium/drivers/i965/brw_swtnl.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_types.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_urb.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_util.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_util.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_vs.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_vs.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_vs_emit.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_vs_state.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_vs_surface_state.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_winsys.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_winsys_debug.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm.h [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_constant_buffer.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_debug.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_emit.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_fp.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_glsl.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_iz.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_pass0.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_pass1.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_pass2.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_sampler_state.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_state.c [new file with mode: 0644]
src/gallium/drivers/i965/brw_wm_surface_state.c [new file with mode: 0644]
src/gallium/drivers/i965/intel_decode.c [new file with mode: 0644]
src/gallium/drivers/i965/intel_decode.h [new file with mode: 0644]
src/gallium/drivers/i965/intel_structs.h [new file with mode: 0644]
src/gallium/drivers/identity/id_context.c
src/gallium/drivers/identity/id_objects.c
src/gallium/drivers/identity/id_objects.h
src/gallium/drivers/identity/id_public.h
src/gallium/drivers/identity/id_screen.c
src/gallium/drivers/llvmpipe/README
src/gallium/drivers/llvmpipe/SConscript
src/gallium/drivers/llvmpipe/lp_bld_arit.c
src/gallium/drivers/llvmpipe/lp_bld_debug.c
src/gallium/drivers/llvmpipe/lp_bld_depth.c
src/gallium/drivers/llvmpipe/lp_bld_interp.c
src/gallium/drivers/llvmpipe/lp_bld_sample.c
src/gallium/drivers/llvmpipe/lp_bld_tgsi_soa.c
src/gallium/drivers/llvmpipe/lp_clear.c
src/gallium/drivers/llvmpipe/lp_context.c
src/gallium/drivers/llvmpipe/lp_context.h
src/gallium/drivers/llvmpipe/lp_debug.h [new file with mode: 0644]
src/gallium/drivers/llvmpipe/lp_draw_arrays.c
src/gallium/drivers/llvmpipe/lp_jit.c
src/gallium/drivers/llvmpipe/lp_screen.c
src/gallium/drivers/llvmpipe/lp_setup.c
src/gallium/drivers/llvmpipe/lp_state.h
src/gallium/drivers/llvmpipe/lp_state_blend.c
src/gallium/drivers/llvmpipe/lp_state_derived.c
src/gallium/drivers/llvmpipe/lp_state_fs.c
src/gallium/drivers/llvmpipe/lp_state_rasterizer.c
src/gallium/drivers/llvmpipe/lp_state_sampler.c
src/gallium/drivers/llvmpipe/lp_state_surface.c
src/gallium/drivers/llvmpipe/lp_state_vs.c
src/gallium/drivers/llvmpipe/lp_tex_cache.c
src/gallium/drivers/llvmpipe/lp_tex_sample_c.c
src/gallium/drivers/llvmpipe/lp_texture.c
src/gallium/drivers/llvmpipe/lp_tile_cache.c
src/gallium/drivers/nouveau/nouveau_stateobj.h
src/gallium/drivers/nv04/nv04_context.c
src/gallium/drivers/nv04/nv04_fragtex.c
src/gallium/drivers/nv04/nv04_miptree.c
src/gallium/drivers/nv04/nv04_state.c
src/gallium/drivers/nv04/nv04_state.h
src/gallium/drivers/nv04/nv04_surface_2d.c
src/gallium/drivers/nv04/nv04_transfer.c
src/gallium/drivers/nv04/nv04_vbo.c
src/gallium/drivers/nv10/nv10_context.c
src/gallium/drivers/nv10/nv10_fragtex.c
src/gallium/drivers/nv10/nv10_miptree.c
src/gallium/drivers/nv10/nv10_state.c
src/gallium/drivers/nv10/nv10_state.h
src/gallium/drivers/nv10/nv10_transfer.c
src/gallium/drivers/nv10/nv10_vbo.c
src/gallium/drivers/nv20/nv20_context.c
src/gallium/drivers/nv20/nv20_fragtex.c
src/gallium/drivers/nv20/nv20_miptree.c
src/gallium/drivers/nv20/nv20_state.c
src/gallium/drivers/nv20/nv20_state.h
src/gallium/drivers/nv20/nv20_state_emit.c
src/gallium/drivers/nv20/nv20_transfer.c
src/gallium/drivers/nv20/nv20_vbo.c
src/gallium/drivers/nv20/nv20_vertprog.c
src/gallium/drivers/nv30/nv30_context.c
src/gallium/drivers/nv30/nv30_context.h
src/gallium/drivers/nv30/nv30_fragprog.c
src/gallium/drivers/nv30/nv30_fragtex.c
src/gallium/drivers/nv30/nv30_miptree.c
src/gallium/drivers/nv30/nv30_state.c
src/gallium/drivers/nv30/nv30_state.h
src/gallium/drivers/nv30/nv30_state_emit.c
src/gallium/drivers/nv30/nv30_transfer.c
src/gallium/drivers/nv30/nv30_vbo.c
src/gallium/drivers/nv30/nv30_vertprog.c
src/gallium/drivers/nv40/nv40_context.c
src/gallium/drivers/nv40/nv40_context.h
src/gallium/drivers/nv40/nv40_draw.c
src/gallium/drivers/nv40/nv40_fragprog.c
src/gallium/drivers/nv40/nv40_fragtex.c
src/gallium/drivers/nv40/nv40_miptree.c
src/gallium/drivers/nv40/nv40_state.c
src/gallium/drivers/nv40/nv40_state.h
src/gallium/drivers/nv40/nv40_state_emit.c
src/gallium/drivers/nv40/nv40_transfer.c
src/gallium/drivers/nv40/nv40_vbo.c
src/gallium/drivers/nv40/nv40_vertprog.c
src/gallium/drivers/nv50/nv50_context.c
src/gallium/drivers/nv50/nv50_context.h
src/gallium/drivers/nv50/nv50_miptree.c
src/gallium/drivers/nv50/nv50_program.c
src/gallium/drivers/nv50/nv50_program.h
src/gallium/drivers/nv50/nv50_query.c
src/gallium/drivers/nv50/nv50_screen.c
src/gallium/drivers/nv50/nv50_state.c
src/gallium/drivers/nv50/nv50_state_validate.c
src/gallium/drivers/nv50/nv50_surface.c
src/gallium/drivers/nv50/nv50_tex.c
src/gallium/drivers/nv50/nv50_texture.h
src/gallium/drivers/nv50/nv50_transfer.c
src/gallium/drivers/nv50/nv50_vbo.c
src/gallium/drivers/r300/Makefile
src/gallium/drivers/r300/SConscript
src/gallium/drivers/r300/r300_blit.c [new file with mode: 0644]
src/gallium/drivers/r300/r300_blit.h [new file with mode: 0644]
src/gallium/drivers/r300/r300_clear.c [deleted file]
src/gallium/drivers/r300/r300_clear.h [deleted file]
src/gallium/drivers/r300/r300_context.c
src/gallium/drivers/r300/r300_context.h
src/gallium/drivers/r300/r300_cs.h
src/gallium/drivers/r300/r300_emit.c
src/gallium/drivers/r300/r300_emit.h
src/gallium/drivers/r300/r300_fs.c
src/gallium/drivers/r300/r300_fs.h
src/gallium/drivers/r300/r300_reg.h
src/gallium/drivers/r300/r300_render.c
src/gallium/drivers/r300/r300_screen.c
src/gallium/drivers/r300/r300_screen.h
src/gallium/drivers/r300/r300_shader_semantics.h [new file with mode: 0644]
src/gallium/drivers/r300/r300_state.c
src/gallium/drivers/r300/r300_state_derived.c
src/gallium/drivers/r300/r300_state_inlines.h
src/gallium/drivers/r300/r300_state_invariant.c
src/gallium/drivers/r300/r300_texture.c
src/gallium/drivers/r300/r300_tgsi_to_rc.c
src/gallium/drivers/r300/r300_vbo.c [deleted file]
src/gallium/drivers/r300/r300_vbo.h [deleted file]
src/gallium/drivers/r300/r300_vs.c
src/gallium/drivers/r300/r300_vs.h
src/gallium/drivers/r300/r300_winsys.h
src/gallium/drivers/softpipe/sp_clear.c
src/gallium/drivers/softpipe/sp_context.c
src/gallium/drivers/softpipe/sp_context.h
src/gallium/drivers/softpipe/sp_draw_arrays.c
src/gallium/drivers/softpipe/sp_flush.c
src/gallium/drivers/softpipe/sp_fs_exec.c
src/gallium/drivers/softpipe/sp_screen.c
src/gallium/drivers/softpipe/sp_setup.c
src/gallium/drivers/softpipe/sp_state.h
src/gallium/drivers/softpipe/sp_state_blend.c
src/gallium/drivers/softpipe/sp_state_derived.c
src/gallium/drivers/softpipe/sp_state_fs.c
src/gallium/drivers/softpipe/sp_state_rasterizer.c
src/gallium/drivers/softpipe/sp_state_sampler.c
src/gallium/drivers/softpipe/sp_state_surface.c
src/gallium/drivers/softpipe/sp_tex_sample.c
src/gallium/drivers/softpipe/sp_tex_tile_cache.c
src/gallium/drivers/softpipe/sp_texture.c
src/gallium/drivers/softpipe/sp_tile_cache.c
src/gallium/drivers/softpipe/sp_winsys.h
src/gallium/drivers/svga/svga_context.h
src/gallium/drivers/svga/svga_pipe_clear.c
src/gallium/drivers/svga/svga_pipe_sampler.c
src/gallium/drivers/svga/svga_pipe_vertex.c
src/gallium/drivers/svga/svga_pipe_vs.c
src/gallium/drivers/svga/svga_screen_buffer.c
src/gallium/drivers/svga/svga_screen_texture.c
src/gallium/drivers/svga/svga_screen_texture.h
src/gallium/drivers/svga/svga_state_constants.c
src/gallium/drivers/svga/svga_state_need_swtnl.c
src/gallium/drivers/svga/svga_state_vs.c
src/gallium/drivers/svga/svga_swtnl_draw.c
src/gallium/drivers/svga/svga_swtnl_state.c
src/gallium/drivers/svga/svga_tgsi_decl_sm20.c
src/gallium/drivers/svga/svga_tgsi_decl_sm30.c
src/gallium/drivers/svga/svga_tgsi_insn.c
src/gallium/drivers/svga/svgadump/svga_dump.c
src/gallium/drivers/svga/svgadump/svga_dump.h
src/gallium/drivers/svga/svgadump/svga_dump.py
src/gallium/drivers/svga/svgadump/svga_shader_dump.c
src/gallium/drivers/trace/tr_context.c
src/gallium/drivers/trace/tr_context.h
src/gallium/drivers/trace/tr_dump_state.c
src/gallium/drivers/trace/tr_dump_state.h
src/gallium/drivers/trace/tr_rbug.c
src/gallium/drivers/trace/tr_screen.c
src/gallium/include/pipe/p_compiler.h
src/gallium/include/pipe/p_config.h
src/gallium/include/pipe/p_context.h
src/gallium/include/pipe/p_defines.h
src/gallium/include/pipe/p_format.h
src/gallium/include/pipe/p_refcnt.h
src/gallium/include/pipe/p_shader_tokens.h
src/gallium/include/pipe/p_state.h
src/gallium/include/pipe/p_video_state.h
src/gallium/state_trackers/dri/dri_drawable.c
src/gallium/state_trackers/egl/egl_surface.c
src/gallium/state_trackers/python/gallium.i
src/gallium/state_trackers/python/p_context.i
src/gallium/state_trackers/python/p_device.i
src/gallium/state_trackers/python/p_format.i
src/gallium/state_trackers/python/p_texture.i
src/gallium/state_trackers/python/retrace/interpreter.py
src/gallium/state_trackers/python/samples/gs.py [new file with mode: 0644]
src/gallium/state_trackers/python/samples/tri.py
src/gallium/state_trackers/python/st_device.c
src/gallium/state_trackers/python/st_device.h
src/gallium/state_trackers/python/st_sample.c
src/gallium/state_trackers/python/st_sample.h
src/gallium/state_trackers/python/st_softpipe_winsys.c
src/gallium/state_trackers/python/tests/base.py
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-abs.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-add.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-dp3.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-dp4.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-dst.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-ex2.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-flr.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-frc.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-lg2.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-lit.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-lrp.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-mad.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-max.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-min.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-mov.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-mul.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-rcp.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-rsq.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-sge.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-srcmod-abs.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-srcmod-absneg.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-srcmod-neg.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-srcmod-swz.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-sub.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/frag-xpd.sh
src/gallium/state_trackers/python/tests/regress/fragment-shader/fragment-shader.py
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-abs.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-add.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-arl.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-arr.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-dp3.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-dp4.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-dst.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-ex2.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-flr.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-frc.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-lg2.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-lit.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-lrp.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-mad.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-max.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-min.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-mov.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-mul.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-rcp.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-rsq.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-sge.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-slt.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-srcmod-abs.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-srcmod-absneg.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-srcmod-neg.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-srcmod-swz.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-sub.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vert-xpd.sh
src/gallium/state_trackers/python/tests/regress/vertex-shader/vertex-shader.py
src/gallium/state_trackers/python/tests/surface_copy.py
src/gallium/state_trackers/python/tests/texture_render.py
src/gallium/state_trackers/python/tests/texture_sample.py
src/gallium/state_trackers/python/tests/texture_transfer.py
src/gallium/state_trackers/vega/api_filters.c
src/gallium/state_trackers/vega/asm_filters.h
src/gallium/state_trackers/vega/asm_util.h
src/gallium/state_trackers/vega/image.c
src/gallium/state_trackers/vega/mask.c
src/gallium/state_trackers/vega/paint.c
src/gallium/state_trackers/vega/renderer.c
src/gallium/state_trackers/vega/shaders_cache.c
src/gallium/state_trackers/vega/vg_tracker.c
src/gallium/state_trackers/vega/vg_translate.c
src/gallium/state_trackers/wgl/stw_framebuffer.c
src/gallium/state_trackers/wgl/stw_pixelformat.c
src/gallium/state_trackers/xorg/xorg_composite.c
src/gallium/state_trackers/xorg/xorg_crtc.c
src/gallium/state_trackers/xorg/xorg_dri2.c
src/gallium/state_trackers/xorg/xorg_exa.c
src/gallium/state_trackers/xorg/xorg_exa_tgsi.c
src/gallium/state_trackers/xorg/xorg_renderer.c
src/gallium/state_trackers/xorg/xorg_xv.c
src/gallium/state_trackers/xorg/xvmc/surface.c
src/gallium/winsys/drm/SConscript
src/gallium/winsys/drm/i965/Makefile [new file with mode: 0644]
src/gallium/winsys/drm/i965/SConscript [new file with mode: 0644]
src/gallium/winsys/drm/i965/dri/Makefile [new file with mode: 0644]
src/gallium/winsys/drm/i965/dri/SConscript [new file with mode: 0644]
src/gallium/winsys/drm/i965/egl/Makefile [new file with mode: 0644]
src/gallium/winsys/drm/i965/gem/Makefile [new file with mode: 0644]
src/gallium/winsys/drm/i965/gem/SConscript [new file with mode: 0644]
src/gallium/winsys/drm/i965/gem/i965_drm_api.c [new file with mode: 0644]
src/gallium/winsys/drm/i965/gem/i965_drm_buffer.c [new file with mode: 0644]
src/gallium/winsys/drm/i965/gem/i965_drm_winsys.h [new file with mode: 0644]
src/gallium/winsys/drm/i965/xlib/Makefile [new file with mode: 0644]
src/gallium/winsys/drm/i965/xlib/xlib_i965.c [new file with mode: 0644]
src/gallium/winsys/drm/i965/xorg/Makefile [new file with mode: 0644]
src/gallium/winsys/drm/i965/xorg/intel_xorg.c [new file with mode: 0644]
src/gallium/winsys/drm/intel/dri/Makefile
src/gallium/winsys/drm/intel/gem/intel_drm_fence.c
src/gallium/winsys/drm/nouveau/drm/nouveau_drm_api.c
src/gallium/winsys/drm/radeon/core/Makefile
src/gallium/winsys/drm/radeon/core/SConscript
src/gallium/winsys/drm/radeon/core/radeon_buffer.c
src/gallium/winsys/drm/radeon/core/radeon_buffer.h
src/gallium/winsys/drm/radeon/core/radeon_drm.c
src/gallium/winsys/drm/radeon/core/radeon_drm.h
src/gallium/winsys/drm/radeon/core/radeon_r300.c
src/gallium/winsys/drm/radeon/core/radeon_r300.h
src/gallium/winsys/drm/radeon/core/radeon_winsys.h [new file with mode: 0644]
src/gallium/winsys/drm/radeon/core/radeon_winsys_softpipe.c [deleted file]
src/gallium/winsys/drm/radeon/core/radeon_winsys_softpipe.h [deleted file]
src/gallium/winsys/drm/radeon/xorg/Makefile
src/gallium/winsys/drm/radeon/xorg/radeon_xorg.c
src/gallium/winsys/drm/vmware/core/vmw_surface.c
src/gallium/winsys/egl_xlib/egl_xlib.c
src/gallium/winsys/egl_xlib/sw_winsys.c
src/gallium/winsys/g3dvl/xlib/xsp_winsys.c
src/gallium/winsys/gdi/SConscript
src/gallium/winsys/gdi/gdi_llvmpipe_winsys.c
src/gallium/winsys/gdi/gdi_softpipe_winsys.c
src/gallium/winsys/xlib/xlib_cell.c
src/gallium/winsys/xlib/xlib_llvmpipe.c
src/gallium/winsys/xlib/xlib_softpipe.c
src/glsl/Makefile [new file with mode: 0644]
src/glsl/Makefile.template [new file with mode: 0644]
src/glsl/SConscript [new file with mode: 0644]
src/glsl/apps/Makefile [new file with mode: 0644]
src/glsl/apps/compile.c [new file with mode: 0644]
src/glsl/apps/process.c [new file with mode: 0644]
src/glsl/apps/purify.c [new file with mode: 0644]
src/glsl/apps/tokenise.c [new file with mode: 0644]
src/glsl/apps/version.c [new file with mode: 0644]
src/glsl/cl/Makefile [new file with mode: 0644]
src/glsl/cl/sl_cl_parse.c [new file with mode: 0644]
src/glsl/cl/sl_cl_parse.h [new file with mode: 0644]
src/glsl/pp/Makefile [new file with mode: 0644]
src/glsl/pp/sl_pp_context.c [new file with mode: 0644]
src/glsl/pp/sl_pp_context.h [new file with mode: 0644]
src/glsl/pp/sl_pp_define.c [new file with mode: 0644]
src/glsl/pp/sl_pp_dict.c [new file with mode: 0644]
src/glsl/pp/sl_pp_dict.h [new file with mode: 0644]
src/glsl/pp/sl_pp_error.c [new file with mode: 0644]
src/glsl/pp/sl_pp_expression.c [new file with mode: 0644]
src/glsl/pp/sl_pp_expression.h [new file with mode: 0644]
src/glsl/pp/sl_pp_extension.c [new file with mode: 0644]
src/glsl/pp/sl_pp_if.c [new file with mode: 0644]
src/glsl/pp/sl_pp_line.c [new file with mode: 0644]
src/glsl/pp/sl_pp_macro.c [new file with mode: 0644]
src/glsl/pp/sl_pp_macro.h [new file with mode: 0644]
src/glsl/pp/sl_pp_pragma.c [new file with mode: 0644]
src/glsl/pp/sl_pp_process.c [new file with mode: 0644]
src/glsl/pp/sl_pp_process.h [new file with mode: 0644]
src/glsl/pp/sl_pp_public.h [new file with mode: 0644]
src/glsl/pp/sl_pp_purify.c [new file with mode: 0644]
src/glsl/pp/sl_pp_purify.h [new file with mode: 0644]
src/glsl/pp/sl_pp_token.c [new file with mode: 0644]
src/glsl/pp/sl_pp_token.h [new file with mode: 0644]
src/glsl/pp/sl_pp_token_util.c [new file with mode: 0644]
src/glsl/pp/sl_pp_token_util.h [new file with mode: 0644]
src/glsl/pp/sl_pp_version.c [new file with mode: 0644]
src/mesa/Makefile
src/mesa/Makefile.mgw
src/mesa/SConscript
src/mesa/drivers/common/driverfuncs.c
src/mesa/drivers/dri/ffb/ffb_vbtmp.h
src/mesa/drivers/dri/gamma/gamma_render.c
src/mesa/drivers/dri/i915/i830_state.c
src/mesa/drivers/dri/i915/i830_texstate.c
src/mesa/drivers/dri/i915/i830_vtbl.c
src/mesa/drivers/dri/i915/i915_context.c
src/mesa/drivers/dri/i915/i915_fragprog.c
src/mesa/drivers/dri/i915/i915_state.c
src/mesa/drivers/dri/i915/i915_texstate.c
src/mesa/drivers/dri/i915/i915_vtbl.c
src/mesa/drivers/dri/i915/intel_tris.c
src/mesa/drivers/dri/i965/brw_clip.c
src/mesa/drivers/dri/i965/brw_clip.h
src/mesa/drivers/dri/i965/brw_clip_line.c
src/mesa/drivers/dri/i965/brw_clip_state.c
src/mesa/drivers/dri/i965/brw_clip_tri.c
src/mesa/drivers/dri/i965/brw_clip_util.c
src/mesa/drivers/dri/i965/brw_context.c
src/mesa/drivers/dri/i965/brw_context.h
src/mesa/drivers/dri/i965/brw_defines.h
src/mesa/drivers/dri/i965/brw_disasm.c
src/mesa/drivers/dri/i965/brw_draw.c
src/mesa/drivers/dri/i965/brw_draw_upload.c
src/mesa/drivers/dri/i965/brw_eu_emit.c
src/mesa/drivers/dri/i965/brw_fallback.c
src/mesa/drivers/dri/i965/brw_gs.c
src/mesa/drivers/dri/i965/brw_gs.h
src/mesa/drivers/dri/i965/brw_gs_emit.c
src/mesa/drivers/dri/i965/brw_gs_state.c
src/mesa/drivers/dri/i965/brw_misc_state.c
src/mesa/drivers/dri/i965/brw_sf_emit.c
src/mesa/drivers/dri/i965/brw_sf_state.c
src/mesa/drivers/dri/i965/brw_tex_layout.c
src/mesa/drivers/dri/i965/brw_urb.c
src/mesa/drivers/dri/i965/brw_vs_emit.c
src/mesa/drivers/dri/i965/brw_vs_state.c
src/mesa/drivers/dri/i965/brw_vtbl.c
src/mesa/drivers/dri/i965/brw_wm.h
src/mesa/drivers/dri/i965/brw_wm_emit.c
src/mesa/drivers/dri/i965/brw_wm_fp.c
src/mesa/drivers/dri/i965/brw_wm_glsl.c
src/mesa/drivers/dri/i965/brw_wm_state.c
src/mesa/drivers/dri/i965/brw_wm_surface_state.c
src/mesa/drivers/dri/intel/intel_batchbuffer.c
src/mesa/drivers/dri/intel/intel_blit.c
src/mesa/drivers/dri/intel/intel_context.c
src/mesa/drivers/dri/intel/intel_context.h
src/mesa/drivers/dri/intel/intel_extensions.c
src/mesa/drivers/dri/intel/intel_fbo.c
src/mesa/drivers/dri/intel/intel_mipmap_tree.c
src/mesa/drivers/dri/intel/intel_pixel_bitmap.c
src/mesa/drivers/dri/intel/intel_pixel_draw.c
src/mesa/drivers/dri/intel/intel_regions.c
src/mesa/drivers/dri/intel/intel_screen.c
src/mesa/drivers/dri/intel/intel_screen.h
src/mesa/drivers/dri/intel/intel_span.c
src/mesa/drivers/dri/intel/intel_state.c
src/mesa/drivers/dri/intel/intel_swapbuffers.c
src/mesa/drivers/dri/intel/intel_tex_format.c
src/mesa/drivers/dri/mach64/mach64_native_vb.c
src/mesa/drivers/dri/mach64/mach64_native_vbtmp.h
src/mesa/drivers/dri/mach64/mach64_tex.c
src/mesa/drivers/dri/mach64/mach64_vbtmp.h
src/mesa/drivers/dri/r128/r128_tris.c
src/mesa/drivers/dri/r200/r200_maos_arrays.c
src/mesa/drivers/dri/r200/r200_swtcl.c
src/mesa/drivers/dri/r200/r200_texstate.c
src/mesa/drivers/dri/r300/Makefile
src/mesa/drivers/dri/r300/r300_blit.c [new file with mode: 0644]
src/mesa/drivers/dri/r300/r300_blit.h [new file with mode: 0644]
src/mesa/drivers/dri/r300/r300_cmdbuf.c
src/mesa/drivers/dri/r300/r300_cmdbuf.h
src/mesa/drivers/dri/r300/r300_context.c
src/mesa/drivers/dri/r300/r300_context.h
src/mesa/drivers/dri/r300/r300_draw.c
src/mesa/drivers/dri/r300/r300_emit.c
src/mesa/drivers/dri/r300/r300_ioctl.c [deleted file]
src/mesa/drivers/dri/r300/r300_ioctl.h [deleted file]
src/mesa/drivers/dri/r300/r300_render.c
src/mesa/drivers/dri/r300/r300_state.c
src/mesa/drivers/dri/r300/r300_swtcl.c
src/mesa/drivers/dri/r300/r300_tex.c
src/mesa/drivers/dri/r300/r300_tex.h
src/mesa/drivers/dri/r300/r300_texcopy.c [new file with mode: 0644]
src/mesa/drivers/dri/r300/r300_texstate.c
src/mesa/drivers/dri/r300/r300_vertprog.c
src/mesa/drivers/dri/r600/r600_context.c
src/mesa/drivers/dri/r600/r600_texstate.c
src/mesa/drivers/dri/r600/r700_assembler.c
src/mesa/drivers/dri/r600/r700_assembler.h
src/mesa/drivers/dri/r600/r700_chip.c
src/mesa/drivers/dri/r600/r700_clear.c
src/mesa/drivers/dri/r600/r700_fragprog.c
src/mesa/drivers/dri/r600/r700_fragprog.h
src/mesa/drivers/dri/r600/r700_render.c
src/mesa/drivers/dri/r600/r700_shader.c
src/mesa/drivers/dri/r600/r700_shader.h
src/mesa/drivers/dri/r600/r700_shaderinst.h
src/mesa/drivers/dri/r600/r700_vertprog.c
src/mesa/drivers/dri/radeon/radeon_common.c
src/mesa/drivers/dri/radeon/radeon_common.h
src/mesa/drivers/dri/radeon/radeon_common_context.c
src/mesa/drivers/dri/radeon/radeon_common_context.h
src/mesa/drivers/dri/radeon/radeon_dma.c
src/mesa/drivers/dri/radeon/radeon_fbo.c
src/mesa/drivers/dri/radeon/radeon_ioctl.c
src/mesa/drivers/dri/radeon/radeon_maos_arrays.c
src/mesa/drivers/dri/radeon/radeon_maos_vbtmp.h
src/mesa/drivers/dri/radeon/radeon_maos_verts.c
src/mesa/drivers/dri/radeon/radeon_mipmap_tree.c
src/mesa/drivers/dri/radeon/radeon_span.c
src/mesa/drivers/dri/radeon/radeon_swtcl.c
src/mesa/drivers/dri/radeon/radeon_texstate.c
src/mesa/drivers/dri/savage/savagerender.c
src/mesa/drivers/dri/savage/savagetris.c
src/mesa/drivers/dri/sis/sis_tris.c
src/mesa/drivers/dri/tdfx/tdfx_vb.c
src/mesa/drivers/dri/tdfx/tdfx_vbtmp.h
src/mesa/drivers/dri/unichrome/via_tris.c
src/mesa/drivers/glide/fxddtex.c
src/mesa/drivers/glide/fxvb.c
src/mesa/drivers/glide/fxvbtmp.h
src/mesa/drivers/osmesa/Makefile
src/mesa/drivers/windows/gldirect/dx7/gld_primitive_dx7.c
src/mesa/drivers/windows/gldirect/dx7/gld_vb_d3d_render_dx7.c
src/mesa/drivers/windows/gldirect/dx8/gld_primitive_dx8.c
src/mesa/drivers/windows/gldirect/dx8/gld_vb_d3d_render_dx8.c
src/mesa/drivers/windows/gldirect/dx9/gld_primitive_dx9.c
src/mesa/drivers/windows/gldirect/dx9/gld_vb_d3d_render_dx9.c
src/mesa/main/config.h
src/mesa/main/context.c
src/mesa/main/dd.h
src/mesa/main/ffvertex_prog.c
src/mesa/main/get.c
src/mesa/main/get_gen.py
src/mesa/main/imports.c
src/mesa/main/imports.h
src/mesa/main/macros.h
src/mesa/main/mtypes.h
src/mesa/main/texenvprogram.c
src/mesa/main/texgetimage.c
src/mesa/main/texobj.c
src/mesa/main/texrender.c
src/mesa/main/version.h
src/mesa/math/m_debug_norm.c
src/mesa/shader/arbprogparse.c
src/mesa/shader/atifragshader.c
src/mesa/shader/descrip.mms
src/mesa/shader/grammar/grammar.c [deleted file]
src/mesa/shader/grammar/grammar.h [deleted file]
src/mesa/shader/grammar/grammar.syn [deleted file]
src/mesa/shader/grammar/grammar_crt.c [deleted file]
src/mesa/shader/grammar/grammar_crt.h [deleted file]
src/mesa/shader/grammar/grammar_mesa.c [deleted file]
src/mesa/shader/grammar/grammar_mesa.h [deleted file]
src/mesa/shader/grammar/grammar_syn.h [deleted file]
src/mesa/shader/prog_execute.c
src/mesa/shader/prog_instruction.h
src/mesa/shader/slang/descrip.mms
src/mesa/shader/slang/library/.gitignore [new file with mode: 0644]
src/mesa/shader/slang/library/Makefile
src/mesa/shader/slang/library/SConscript [new file with mode: 0644]
src/mesa/shader/slang/library/gc_to_bin.c [deleted file]
src/mesa/shader/slang/library/slang_120_core_gc.h [deleted file]
src/mesa/shader/slang/library/slang_builtin_120_common_gc.h [deleted file]
src/mesa/shader/slang/library/slang_builtin_120_fragment_gc.h [deleted file]
src/mesa/shader/slang/library/slang_common_builtin_gc.h [deleted file]
src/mesa/shader/slang/library/slang_core_gc.h [deleted file]
src/mesa/shader/slang/library/slang_fragment_builtin_gc.h [deleted file]
src/mesa/shader/slang/library/slang_pp_directives.syn [deleted file]
src/mesa/shader/slang/library/slang_pp_directives_syn.h [deleted file]
src/mesa/shader/slang/library/slang_pp_expression.syn [deleted file]
src/mesa/shader/slang/library/slang_pp_expression_syn.h [deleted file]
src/mesa/shader/slang/library/slang_pp_version.syn [deleted file]
src/mesa/shader/slang/library/slang_pp_version_syn.h [deleted file]
src/mesa/shader/slang/library/slang_shader.syn [deleted file]
src/mesa/shader/slang/library/slang_shader_syn.h [deleted file]
src/mesa/shader/slang/library/slang_version.syn [deleted file]
src/mesa/shader/slang/library/slang_vertex_builtin_gc.h [deleted file]
src/mesa/shader/slang/library/syn_to_c.c [deleted file]
src/mesa/shader/slang/slang_compile.c
src/mesa/shader/slang/slang_emit.c
src/mesa/shader/slang/slang_link.c
src/mesa/shader/slang/slang_preprocess.c [deleted file]
src/mesa/shader/slang/slang_preprocess.h [deleted file]
src/mesa/sources.mak
src/mesa/state_tracker/st_atom.c
src/mesa/state_tracker/st_atom.h
src/mesa/state_tracker/st_atom_framebuffer.c
src/mesa/state_tracker/st_atom_pixeltransfer.c
src/mesa/state_tracker/st_atom_sampler.c
src/mesa/state_tracker/st_atom_shader.c
src/mesa/state_tracker/st_atom_texture.c
src/mesa/state_tracker/st_cb_bitmap.c
src/mesa/state_tracker/st_cb_clear.c
src/mesa/state_tracker/st_cb_drawpixels.c
src/mesa/state_tracker/st_cb_fbo.c
src/mesa/state_tracker/st_cb_program.c
src/mesa/state_tracker/st_cb_readpixels.c
src/mesa/state_tracker/st_cb_texture.c
src/mesa/state_tracker/st_context.h
src/mesa/state_tracker/st_debug.c
src/mesa/state_tracker/st_draw.c
src/mesa/state_tracker/st_draw_feedback.c
src/mesa/state_tracker/st_extensions.c
src/mesa/state_tracker/st_format.c
src/mesa/state_tracker/st_gen_mipmap.c
src/mesa/state_tracker/st_mesa_to_tgsi.c
src/mesa/state_tracker/st_mesa_to_tgsi.h
src/mesa/state_tracker/st_program.c
src/mesa/state_tracker/st_program.h
src/mesa/state_tracker/st_texture.c
src/mesa/swrast_setup/ss_tritmp.h
src/mesa/tnl/t_context.h
src/mesa/tnl/t_draw.c
src/mesa/tnl/t_pipeline.c
src/mesa/tnl/t_vb_fog.c
src/mesa/tnl/t_vb_light.c
src/mesa/tnl/t_vb_lighttmp.h
src/mesa/tnl/t_vb_normals.c
src/mesa/tnl/t_vb_program.c
src/mesa/tnl/t_vb_texgen.c
src/mesa/tnl/t_vb_texmat.c
src/mesa/tnl/t_vb_vertex.c
src/mesa/tnl/t_vertex_generic.c
src/mesa/tnl_dd/t_dd_dmatmp.h
src/mesa/tnl_dd/t_dd_tritmp.h
src/mesa/tnl_dd/t_dd_vb.c
src/mesa/tnl_dd/t_dd_vbtmp.h
src/mesa/vf/vf_generic.c
src/mesa/x86/gen_matypes.c
windows/VC7/mesa/mesa/mesa.vcproj
windows/VC8/mesa/mesa/mesa.vcproj

index e927cb96851a485419ab05fe802b2cfcd44c9481..eb8dcc1235a0c93b8c856a47057fbc9439f3d1be 100644 (file)
--- a/Makefile
+++ b/Makefile
@@ -105,6 +105,7 @@ irix6-n32-static \
 irix6-o32 \
 irix6-o32-static \
 linux \
+linux-i965 \
 linux-alpha \
 linux-alpha-static \
 linux-cell \
@@ -225,6 +226,10 @@ MAIN_FILES = \
        $(DIRECTORY)/include/GL/vms_x_fix.h                             \
        $(DIRECTORY)/include/GL/wglext.h                                \
        $(DIRECTORY)/include/GL/wmesa.h                                 \
+       $(DIRECTORY)/src/glsl/Makefile                                  \
+       $(DIRECTORY)/src/glsl/*/Makefile                                \
+       $(DIRECTORY)/src/glsl/*/SConscript                              \
+       $(DIRECTORY)/src/glsl/*/*.[ch]                                  \
        $(DIRECTORY)/src/Makefile                                       \
        $(DIRECTORY)/src/mesa/Makefile*                                 \
        $(DIRECTORY)/src/mesa/sources.mak                               \
@@ -240,7 +245,6 @@ MAIN_FILES = \
        $(DIRECTORY)/src/mesa/shader/*.[chly]                           \
        $(DIRECTORY)/src/mesa/shader/Makefile                           \
        $(DIRECTORY)/src/mesa/shader/descrip.mms                        \
-       $(DIRECTORY)/src/mesa/shader/grammar/*.[ch]                     \
        $(DIRECTORY)/src/mesa/shader/slang/*.[ch]                       \
        $(DIRECTORY)/src/mesa/shader/slang/descrip.mms                  \
        $(DIRECTORY)/src/mesa/shader/slang/library/*.[ch]               \
index f43c10cecb3fef5eb98aff5029f5c66a4164ee36..8e063e280780baeb32f2009a40f6849a037efccb 100644 (file)
@@ -32,10 +32,10 @@ import common
 default_statetrackers = 'mesa'
 
 if common.default_platform in ('linux', 'freebsd', 'darwin'):
-       default_drivers = 'softpipe,failover,svga,i915,trace,identity,llvmpipe'
+       default_drivers = 'softpipe,failover,svga,i915,i965,trace,identity,llvmpipe'
        default_winsys = 'xlib'
 elif common.default_platform in ('winddk',):
-       default_drivers = 'softpipe,svga,i915,trace,identity'
+       default_drivers = 'softpipe,svga,i915,i965,trace,identity'
        default_winsys = 'all'
 else:
        default_drivers = 'all'
@@ -46,9 +46,9 @@ common.AddOptions(opts)
 opts.Add(ListVariable('statetrackers', 'state trackers to build', default_statetrackers,
                      ['mesa', 'python', 'xorg']))
 opts.Add(ListVariable('drivers', 'pipe drivers to build', default_drivers,
-                     ['softpipe', 'failover', 'svga', 'i915', 'cell', 'trace', 'r300', 'identity', 'llvmpipe']))
+                     ['softpipe', 'failover', 'svga', 'i915', 'i965', 'cell', 'trace', 'r300', 'identity', 'llvmpipe']))
 opts.Add(ListVariable('winsys', 'winsys drivers to build', default_winsys,
-                     ['xlib', 'vmware', 'intel', 'gdi', 'radeon']))
+                     ['xlib', 'vmware', 'intel', 'i965', 'gdi', 'radeon']))
 
 opts.Add(EnumVariable('MSVS_VERSION', 'MS Visual C++ version', None, allowed_values=('7.1', '8.0', '9.0')))
 
@@ -160,6 +160,25 @@ Export('env')
 # TODO: Build several variants at the same time?
 # http://www.scons.org/wiki/SimultaneousVariantBuilds
 
+if env['platform'] != common.default_platform:
+    # GLSL code has to be built twice -- one for the host OS, another for the target OS...
+
+    host_env = Environment(
+        # options are ignored
+        # default tool is used
+        toolpath = ['#scons'], 
+        ENV = os.environ,
+    )
+
+    host_env['platform'] = common.default_platform
+
+    SConscript(
+        'src/glsl/SConscript',
+        variant_dir = env['build'] + '/host',
+        duplicate = 0, # http://www.scons.org/doc/0.97/HTML/scons-user/x2261.html
+        exports={'env':host_env},
+    )
+
 SConscript(
        'src/SConscript',
        variant_dir = env['build'],
index 3bec160b40cff6a230c886a201ad5e0e6c6c7678..fb160a2aae92e9dc56ce613971e0be24979ba743 100755 (executable)
--- a/bin/mklib
+++ b/bin/mklib
 # CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 
+# Given a list of files, look for .a archives and unpack them.
+# Return the original list of files minus the .a files plus the unpacked files.
+expand_archives() {
+    FILES=$@
+    NEWFILES=""
+    for FILE in $FILES ; do
+        case $FILE in
+            *.a)
+                # extract the .o files from this .a archive
+                MEMBERS=`ar t $FILE`
+                ar x $FILE
+                NEWFILES="$NEWFILES $MEMBERS"
+                ;;
+            *)
+                # other file type, just add to list
+                NEWFILES="$NEWFILES $FILE"
+                ;;
+        esac
+    done
+    echo $NEWFILES
+}
+
+
+# Given a list of files, look for .a archives and return a list of all objects
+# in the .a archives.
+contents_of_archives() {
+    FILES=$@
+    NEWFILES=""
+    for FILE in $FILES ; do
+        case $FILE in
+            *.a)
+                # get list of members in this .a archive
+                MEMBERS=`ar t $FILE`
+                NEWFILES="$NEWFILES $MEMBERS"
+                ;;
+            *)
+                # skip other file types
+                ;;
+        esac
+    done
+    echo $NEWFILES
+}
+
+
+# Print usage info.
+usage() {
+    echo 'Usage: mklib [options] objects'
+    echo 'Create a shared library from object files.'
+    echo '  -o LIBRARY    specifies the name of the resulting library, without'
+    echo '                the leading "lib" or any suffix.'
+    echo '                (eg: "-o GL" might result in "libGL.so" being made)'
+    echo '  -major N      specifies major version number (default is 1)'
+    echo '  -minor N      specifies minor version number (default is 0)'
+    echo '  -patch N      specifies patch version number (default is 0)'
+    echo '  -lLIBRARY     specifies a dependency on LIBRARY'
+    echo '  -LDIR         search in DIR for library dependencies at build time'
+    echo '  -RDIR         search in DIR for library dependencies at run time'
+    echo '  -linker L     explicity specify the linker program to use (eg: gcc, g++)'
+    echo '                Not observed on all systems at this time.'
+    echo '  -ldflags OPT  specify any additional linker flags in OPT'
+    echo '  -cplusplus    link with C++ runtime'
+    echo '  -static       make a static library (default is dynamic/shared)'
+    echo '  -dlopen       make a shared library suitable for dynamic loading'
+    echo '  -install DIR  put resulting library file(s) in DIR'
+    echo '  -arch ARCH    override using `uname` to determine host system'
+    echo '  -archopt OPT  specify an extra achitecture-specific option OPT'
+    echo '  -altopts OPTS alternate options to override all others'
+    echo "  -noprefix     don't prefix library name with 'lib' nor add any suffix"
+    echo '  -exports FILE only export the symbols listed in FILE'
+    echo '  -id NAME      Sets the id of the dylib (Darwin)'
+    echo '  -h, --help    display this information and exit'
+}
+
+
 #
 # Option defaults
 #
@@ -52,31 +126,7 @@ while true
 do
     case $1 in
        '-h' | '--help')
-           echo 'Usage: mklib [options] objects'
-           echo 'Create a shared library from object files.'
-           echo '  -o LIBRARY    specifies the name of the resulting library, without'
-           echo '                the leading "lib" or any suffix.'
-           echo '                (eg: "-o GL" might result in "libGL.so" being made)'
-           echo '  -major N      specifies major version number (default is 1)'
-           echo '  -minor N      specifies minor version number (default is 0)'
-           echo '  -patch N      specifies patch version number (default is 0)'
-           echo '  -lLIBRARY     specifies a dependency on LIBRARY'
-           echo '  -LDIR         search in DIR for library dependencies at build time'
-           echo '  -RDIR         search in DIR for library dependencies at run time'
-           echo '  -linker L     explicity specify the linker program to use (eg: gcc, g++)'
-           echo '                Not observed on all systems at this time.'
-           echo '  -ldflags OPT  specify any additional linker flags in OPT'
-           echo '  -cplusplus    link with C++ runtime'
-           echo '  -static       make a static library (default is dynamic/shared)'
-           echo '  -dlopen       make a shared library suitable for dynamic loading'
-           echo '  -install DIR  put resulting library file(s) in DIR'
-           echo '  -arch ARCH    override using `uname` to determine host system'
-           echo '  -archopt OPT  specify an extra achitecture-specific option OPT'
-           echo '  -altopts OPTS alternate options to override all others'
-           echo "  -noprefix     don't prefix library name with 'lib' nor add any suffix"
-           echo '  -exports FILE only export the symbols listed in FILE'
-           echo '  -id NAME      Sets the id of the dylib (Darwin)'
-           echo '  -h, --help    display this information and exit'
+           usage
            exit 1
            ;;
        '-o')
@@ -197,11 +247,11 @@ fi
 # Error checking
 #
 if [ "x${LIBNAME}" = "x" ] ; then
-    echo "mklib: Error: no library name specified"
+    echo "mklib: Error: no library name specified (-h for help)"
     exit 1
 fi
 if [ "x${OBJECTS}" = "x" ] ; then
-    echo "mklib: Error: no object files specified"
+    echo "mklib: Error: no object files specified (-h for help)"
     exit 1
 fi
 
@@ -269,45 +319,29 @@ case $ARCH in
             # finish up
             FINAL_LIBS="${LIBNAME}"
         elif [ $STATIC = 1 ] ; then
+           # make a static .a library
             LIBNAME="lib${LIBNAME}.a"     # prefix with "lib", suffix with ".a"
             echo "mklib: Making" $ARCH "static library: " ${LIBNAME}
-            LINK="ar"
             OPTS="-ru"
             if [ "${ALTOPTS}" ] ; then
                 OPTS=${ALTOPTS}
             fi
-            rm -f ${LIBNAME}
 
-           # expand any .a objects into constituent .o files.
-           NEWOBJECTS=""
-           DELETIA=""
-           for OBJ in $OBJECTS ; do
-               case $OBJ in
-                   *.a)
-                       # extract the .o files from this .a archive
-                       FILES=`ar t $OBJ`
-                       ar x $OBJ
-                       NEWOBJECTS="$NEWOBJECTS $FILES"
-                       # keep track of temporary .o files and delete them below
-                       DELETIA="$DELETIA $FILES"
-                       ;;
-                   *)
-                       # ordinary .o file
-                       NEWOBJECTS="$NEWOBJECTS $OBJ"
-                       ;;
-               esac
-           done
+           # expand .a into .o files
+           NEW_OBJECTS=`expand_archives $OBJECTS`
 
-            # make lib
-            ${LINK} ${OPTS} ${LIBNAME} ${NEWOBJECTS}
+            # make static lib
+            rm -f ${LIBNAME}
+            ar ${OPTS} ${LIBNAME} ${NEW_OBJECTS}
             ranlib ${LIBNAME}
 
            # remove temporary extracted .o files
-           rm -f ${DELETIA}
+           rm -f `contents_of_archives $OBJECTS`
 
             # finish up
             FINAL_LIBS=${LIBNAME}
         else
+           # make dynamic library
            LIBNAME="lib${LIBNAME}"     # prefix with "lib"
            case $ARCH in 'Linux' | 'GNU' | GNU/*)
                OPTS="-Xlinker -Bsymbolic -shared -Wl,-soname,${LIBNAME}.so.${MAJOR}"
@@ -489,13 +523,24 @@ case $ARCH in
            ${LINK} ${OPTS} ${LDFLAGS} -o ${LIBNAME} ${OBJECTS} ${DEPS}
            FINAL_LIBS=${LIBNAME}
         elif [ $STATIC = 1 ] ; then
+           # make a static .a library
            STLIB="lib${LIBNAME}.a"
            echo "mklib: Making FreeBSD static library: " ${STLIB}
+
+           # expand .a into .o files
+           NEW_OBJECTS=`expand_archives $OBJECTS`
+
+            # make static lib
            rm -f ${STLIB}
-           ar cq ${STLIB} ${OBJECTS}
+           ar cq ${STLIB} ${NEW_OBJECTS}
            ranlib ${STLIB}
+
+           # remove temporary extracted .o files
+           rm -f `contents_of_archives $OBJECTS`
+
            FINAL_LIBS=${STLIB}
        else
+           # make dynamic library
            SHLIB="lib${LIBNAME}.so.${MAJOR}"
            OPTS="-shared -Wl,-soname,${SHLIB}"
             if [ "${ALTOPTS}" ] ; then
index 3b6bf52c0354e8435a9471362759e03caf5c97f1..101fc558f40f276a4a55a895fa7c8c8c27ed42bb 100644 (file)
--- a/common.py
+++ b/common.py
@@ -12,7 +12,7 @@ import platform as _platform
 
 _platform_map = {
        'linux2': 'linux',
-       'win32': 'winddk',
+       'win32': 'windows',
 }
 
 default_platform = sys.platform
index 824f71c0cc1bbaed3f82e0370da67539d44547ec..25e8d057a929aa1f6398c7edd1a6b32ff0ee4192 100644 (file)
@@ -49,7 +49,7 @@ GLW_LIB_DEPS = -L$(TOP)/$(LIB_DIR) -l$(GL_LIB) -L$(INSTALL_DIR)/$(LIB_DIR) -L$(X
 APP_LIB_DEPS = -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -l$(GLU_LIB) -l$(GL_LIB) -L$(INSTALL_DIR)/$(LIB_DIR) -L$(X11_DIR)/$(LIB_DIR) -lX11 -lXmu -lXt -lXi -lm
 
 # omit glw lib for now:
-SRC_DIRS = glx/x11 mesa glu glut/glx glew
+SRC_DIRS = glsl glx/x11 mesa glu glut/glx glew
 GLU_DIRS = sgi
 DRIVER_DIRS = osmesa
 #DRIVER_DIRS = dri
index 9af816cf9bc811ee98f51e48fa24ee84061b8109..8a975982a8b1ac814b681ee3199bef86579b35b1 100644 (file)
@@ -9,7 +9,7 @@ CONFIG_NAME = default
 
 # Version info
 MESA_MAJOR=7
-MESA_MINOR=7
+MESA_MINOR=8
 MESA_TINY=0
 MESA_VERSION = $(MESA_MAJOR).$(MESA_MINOR).$(MESA_TINY)
 
@@ -83,7 +83,7 @@ MOTIF_CFLAGS = -I/usr/include/Motif1.2
 
 # Directories to build
 LIB_DIR = lib
-SRC_DIRS = mesa gallium egl gallium/winsys glu glut/glx glew glw
+SRC_DIRS = glsl mesa gallium egl gallium/winsys glu glut/glx glew glw
 GLU_DIRS = sgi
 DRIVER_DIRS = x11 osmesa
 # Which subdirs under $(TOP)/progs/ to enter:
@@ -96,7 +96,7 @@ EGL_DRIVERS_DIRS = demo
 GALLIUM_DIRS = auxiliary drivers state_trackers
 GALLIUM_AUXILIARY_DIRS = rbug draw translate cso_cache pipebuffer tgsi sct rtasm util indices vl
 GALLIUM_AUXILIARIES = $(foreach DIR,$(GALLIUM_AUXILIARY_DIRS),$(TOP)/src/gallium/auxiliary/$(DIR)/lib$(DIR).a)
-GALLIUM_DRIVERS_DIRS = softpipe failover svga i915 trace identity
+GALLIUM_DRIVERS_DIRS = softpipe failover svga i915 i965 trace identity
 GALLIUM_DRIVERS = $(foreach DIR,$(GALLIUM_DRIVERS_DIRS),$(TOP)/src/gallium/drivers/$(DIR)/lib$(DIR).a)
 GALLIUM_WINSYS_DIRS = xlib egl_xlib
 GALLIUM_WINSYS_DRM_DIRS =
index 19d435fecce8c542128a2f02e3797a66904ed1c9..221655c35dc0c6b51f2217ddb43714245466ca3e 100644 (file)
@@ -33,7 +33,7 @@ CFLAGS = $(OPT_FLAGS) -Wall -Winline -Wmissing-prototypes \
 CXXFLAGS = $(CFLAGS)
 
 # Omitting glw here:
-SRC_DIRS = gallium mesa gallium/winsys glu glut/glx glew
+SRC_DIRS = glsl mesa gallium gallium/winsys glu glut/glx glew
 
 # Build no traditional Mesa drivers:
 DRIVER_DIRS =
index 0802543347a1af9326ede85d65b3e88a07d503c5..cf1f4e1983397aabbbd5d63c22c309ea995863be 100644 (file)
@@ -60,7 +60,7 @@ EGL_DRIVERS_DIRS = demo glx
 DRIVER_DIRS = dri
 WINDOW_SYSTEM = dri
 GALLIUM_WINSYS_DIRS = drm
-GALLIUM_WINSYS_DRM_DIRS = vmware intel
+GALLIUM_WINSYS_DRM_DIRS = vmware intel i965
 GALLIUM_STATE_TRACKERS_DIRS = egl
 
 DRI_DIRS = i810 i915 i965 mach64 mga r128 r200 r300 radeon \
diff --git a/configs/linux-i965 b/configs/linux-i965
new file mode 100644 (file)
index 0000000..e66abc3
--- /dev/null
@@ -0,0 +1,8 @@
+# Configuration for standalone mode i965 debug
+
+include $(TOP)/configs/linux-debug
+
+CONFIG_NAME = linux-i965
+
+GALLIUM_DRIVER_DIRS = i965
+GALLIUM_WINSYS_DIRS = drm/i965/xlib
index a15ca111e273a42b1e9c299cdbee0cfb87c57564..d83dd4315b7d2c1131b1c06b983d89be0c127575 100644 (file)
@@ -414,7 +414,7 @@ esac
 dnl
 dnl Driver specific build directories
 dnl
-SRC_DIRS="mesa glew"
+SRC_DIRS="glsl mesa glew"
 GLU_DIRS="sgi"
 WINDOW_SYSTEM=""
 GALLIUM_DIRS="auxiliary drivers state_trackers"
@@ -1221,10 +1221,10 @@ AC_ARG_ENABLE([gallium-intel],
     [enable_gallium_intel="$enableval"],
     [enable_gallium_intel=auto])
 if test "x$enable_gallium_intel" = xyes; then
-    GALLIUM_WINSYS_DRM_DIRS="$GALLIUM_WINSYS_DRM_DIRS intel"
-    GALLIUM_DRIVERS_DIRS="$GALLIUM_DRIVERS_DIRS i915"
+    GALLIUM_WINSYS_DRM_DIRS="$GALLIUM_WINSYS_DRM_DIRS intel i965"
+    GALLIUM_DRIVERS_DIRS="$GALLIUM_DRIVERS_DIRS i915 i965"
 elif test "x$enable_gallium_intel" = xauto; then
-    GALLIUM_DRIVERS_DIRS="$GALLIUM_DRIVERS_DIRS i915"
+    GALLIUM_DRIVERS_DIRS="$GALLIUM_DRIVERS_DIRS i915 i965"
 fi
 
 dnl
diff --git a/docs/relnotes-7.8.html b/docs/relnotes-7.8.html
new file mode 100644 (file)
index 0000000..dc00356
--- /dev/null
@@ -0,0 +1,53 @@
+<HTML>
+
+<TITLE>Mesa Release Notes</TITLE>
+
+<head><link rel="stylesheet" type="text/css" href="mesa.css"></head>
+
+<BODY>
+
+<body bgcolor="#eeeeee">
+
+<H1>Mesa 7.8 Release Notes / date TBD</H1>
+
+<p>
+Mesa 7.8 is a new development release.
+People who are concerned with stability and reliability should stick
+with a previous release or wait for Mesa 7.8.1.
+</p>
+<p>
+Mesa 7.8 implements the OpenGL 2.1 API, but the version reported by
+glGetString(GL_VERSION) depends on the particular driver being used.
+Some drivers don't support all the features required in OpenGL 2.1.
+</p>
+<p>
+See the <a href="install.html">Compiling/Installing page</a> for prerequisites
+for DRI hardware acceleration.
+</p>
+
+
+<h2>MD5 checksums</h2>
+<pre>
+tbd
+</pre>
+
+
+<h2>New features</h2>
+<ul>
+<li>TBD
+</ul>
+
+
+<h2>Bug fixes</h2>
+<ul>
+<li>TBD
+</ul>
+
+
+<h2>Changes</h2>
+<ul>
+<li>TBD
+</ul>
+
+</body>
+</html>
index b301ee046a70606e6e93a042e46d05f4e43fe9f1..d0d9b6e5b98f3d7f6325b5e632aa33a90bf930f5 100644 (file)
@@ -13,6 +13,7 @@ The release notes summarize what's new or changed in each Mesa release.
 </p>
 
 <UL>
+<LI><A HREF="relnotes-7.8.html">7.8 release notes</A>
 <LI><A HREF="relnotes-7.7.html">7.7 release notes</A>
 <LI><A HREF="relnotes-7.6.1.html">7.6.1 release notes</A>
 <LI><A HREF="relnotes-7.6.html">7.6 release notes</A>
index 7b7a12bf88dc6a1af79b2e90167fd365e5a51482..861b3b72522cb7f3b9603dd9e340c771222c7222 100644 (file)
@@ -388,9 +388,8 @@ static void Init( void )
 
 int main( int argc, char *argv[] )
 {
-   glutInit( &argc, argv );
-   glutInitWindowPosition( 0, 0 );
    glutInitWindowSize( 200, 200 );
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH );
    Win = glutCreateWindow(argv[0]);
    glutReshapeFunc( Reshape );
index 275c85105ede20ba13bd7d7744dd434f964314bc..a0ce7f3f5bce3de68597f0d293b04b86e8d500e6 100644 (file)
@@ -311,9 +311,8 @@ static void Init (void)
 
 int main (int argc, char *argv[])
 {
-       glutInit (&argc, argv);
-       glutInitWindowPosition ( 0, 0);
        glutInitWindowSize (200, 200);
+       glutInit (&argc, argv);
        glutInitDisplayMode (GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH);
        glutCreateWindow (argv[0]);
        glutReshapeFunc (Reshape);
index 9188ad5a5658743c1fb3e3edaaffe0dd004e8f74..f669a1f26ba328d49904dd2738ccc5f684da7415 100644 (file)
@@ -268,9 +268,8 @@ static void Init( void )
 
 int main( int argc, char *argv[] )
 {
-   glutInit( &argc, argv );
-   glutInitWindowPosition( 0, 0 );
    glutInitWindowSize( 400, 400 );
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH );
    Win = glutCreateWindow(argv[0]);
    glewInit();
index 436bc7d1fb880813b711aba663f5ee5f0ce19262..a9a291e0a6f573ae1849eb544090fab6992d7d41 100644 (file)
@@ -192,10 +192,8 @@ visible(int vis)
 
 int main(int argc, char *argv[])
 {
-  glutInit(&argc, argv);
-  glutInitWindowPosition(0, 0);
   glutInitWindowSize(600, 450);
-
+  glutInit(&argc, argv);
 
   IndexMode = argc > 1 && strcmp(argv[1], "-ci") == 0;
   if (IndexMode)
index 42953f6675e01bd5fa3458f19d5e1515925c6437..c40f118a64bf5cf72521d8610bc269352d4979a0 100644 (file)
@@ -187,9 +187,8 @@ int main( int argc, char *argv[] )
 
    Init( argc, argv );
 
-   glutInit( &argc, argv );
    glutInitWindowSize( (int) Width, (int) Height );
-   glutInitWindowPosition( 0, 0 );
+   glutInit( &argc, argv );
 
    mode = ColorMode | GLUT_DOUBLE;
    if (BufferMask & GL_STENCIL_BUFFER_BIT)
index a13339ea62f86a803f6cd017a31ab92c92045e99..98c835f44f83f3ce21811aca054c5def22e177ef 100644 (file)
@@ -237,6 +237,9 @@ int main( int argc, char *argv[] )
    const char *filename = IMAGE_FILE;
    int i = 1;
 
+   glutInitWindowSize( WinWidth, WinHeight );
+   glutInit( &argc, argv );
+
    if (argc > i && strcmp(argv[i], "-ci")==0) {
       ciMode = GL_TRUE;
       i++;
@@ -245,10 +248,6 @@ int main( int argc, char *argv[] )
       filename = argv[i];
    }
 
-   glutInit( &argc, argv );
-   glutInitWindowPosition( 0, 0 );
-   glutInitWindowSize( WinWidth, WinHeight );
-
    if (ciMode)
       glutInitDisplayMode( GLUT_INDEX | GLUT_DOUBLE );
    else
index 20332b1d96026a7f6c96f698a6d35512ca748132..3e79d6a55890ce513ee6a0425a459c7642b3196c 100644 (file)
@@ -613,9 +613,8 @@ static void parse_args(int argc, char *argv[])
 
 int main( int argc, char *argv[] )
 {
-   glutInit(&argc, argv);
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(600, 500);
+   glutInit(&argc, argv);
    glutInitDisplayMode( GLUT_RGB | GLUT_DEPTH | GLUT_DOUBLE );
    glutCreateWindow("Texture Cube Mapping");
    glewInit();
index 5490bcc6355fd486fa451fce9e70f9105837da06..9bc7086e04b4cfcdb4bd051200ffaba073c9ebab 100644 (file)
@@ -328,6 +328,9 @@ int main( int argc, char *argv[] )
    const char *filename = IMAGE_FILE;
    int i = 1;
 
+   glutInitWindowSize( 500, 400 );
+   glutInit( &argc, argv );
+
    if (argc > i && strcmp(argv[i], "-ci")==0) {
       ciMode = GL_TRUE;
       i++;
@@ -336,10 +339,6 @@ int main( int argc, char *argv[] )
       filename = argv[i];
    }
 
-   glutInit( &argc, argv );
-   glutInitWindowPosition( 0, 0 );
-   glutInitWindowSize( 500, 400 );
-
    if (ciMode)
       glutInitDisplayMode( GLUT_INDEX | GLUT_DOUBLE );
    else
index 3cf311e7782ce139d56c65fa4fd3fef5018231a0..c54e3b8fb80f01c76ee5be62744ed58d62b23e31 100644 (file)
@@ -1308,8 +1308,8 @@ Init(void)
 int
 main(int argc, char *argv[])
 {
-   glutInit(&argc, argv);
    glutInitWindowSize(WinWidth, WinHeight);
+   glutInit(&argc, argv);
    glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH);
    glutCreateWindow("OpenGL Engine Demo");
    glewInit();
index b3f7e00e5afecbc9b915b667eccb20d8d9f6a322..17767a148f9efb9097b9c2966f4c7385fac1bfc8 100644 (file)
@@ -1051,11 +1051,11 @@ visible(int vis)
 int
 main(int argc, char *argv[])
 {
+   glutInitWindowSize(WinWidth, WinHeight);
    glutInit(&argc, argv);
 
    glutInitDisplayMode(GLUT_RGB | GLUT_DEPTH | GLUT_DOUBLE);
 
-   glutInitWindowSize(WinWidth, WinHeight);
    Win = glutCreateWindow("fbo_firecube");
    glewInit();
    init(argc, argv);
index dd70f6141dba4817452073aeb7a55cbf45fff744..3db45418fadfc559fd5761ea1fd5ab4bd3238307 100644 (file)
@@ -733,7 +733,6 @@ main(int ac, char **av)
       HEIGHT = atoi(av[3]);
    }
 
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(WIDTH, HEIGHT);
    glutInit(&ac, av);
 
index 7d5c11aabf3ff88ed5c0bbc4da3267b3cd3a90c3..336aa580755f4456ff688408dd1d4b8d22cd6015 100644 (file)
@@ -404,8 +404,8 @@ Init(void)
 int
 main( int argc, char *argv[] )
 {
-   glutInit( &argc, argv );
    glutInitWindowSize( 600, 600 );
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH );
    glutCreateWindow(argv[0]);
    glewInit();
index c297354103fedefa7c7599098bb7f3192edb2329..f52a4f7868669f3b7377e85649c52f9fdc570e59 100644 (file)
@@ -268,9 +268,8 @@ static void Init( void )
 
 int main( int argc, char *argv[] )
 {
-   glutInit( &argc, argv );
-   glutInitWindowPosition( 0, 0 );
    glutInitWindowSize( 200, 200 );
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH );
    Win = glutCreateWindow(argv[0]);
    glewInit();
index acba3e9583f15c62f8d3827cf7e1919a60f16a92..395b7caa2c1355f29c74b5b3b09488eabf4663c0 100644 (file)
@@ -600,9 +600,8 @@ ParseOptions(int argc, char *argv[])
 int
 main(int argc, char *argv[])
 {
-   glutInit(&argc, argv);
-   glutInitWindowPosition( 0, 0);
    glutInitWindowSize(200, 200);
+   glutInit(&argc, argv);
    glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH);
    win = glutCreateWindow(argv[0]);
    glutReshapeFunc(Reshape);
index 9b2c3e1052b148abd2a88455e6368016eeaffc75..61c6d125fbe294888c8149cc96c4a1c6a3bef2f0 100644 (file)
@@ -146,12 +146,9 @@ display(void)
 int
 main(int argc, char **argv)
 {
+  glutInitWindowSize(500, 400);
   glutInit(&argc, argv);
   glutInitDisplayMode(GLUT_RGB | GLUT_SINGLE);
-
-  glutInitWindowPosition(50, 50);
-  glutInitWindowSize(500, 400);
-
   glutCreateWindow("gamma test patterns");
   glutReshapeFunc(Reshape);
   glutDisplayFunc(display);
index 2dcf32f92ff64460414506c6d171c8c86df72625..71d0281904be08eb170d63ea1b918ff355893763 100644 (file)
@@ -474,10 +474,10 @@ visible(int vis)
 int
 main(int argc, char *argv[])
 {
+   glutInitWindowSize(WinWidth, WinHeight);
    glutInit(&argc, argv);
    glutInitDisplayMode(GLUT_RGB | GLUT_DEPTH | GLUT_DOUBLE);
 
-   glutInitWindowSize(WinWidth, WinHeight);
    Win = glutCreateWindow("gearbox");
    init(argc, argv);
 
index 6016162d6f74c6973bf80113269a7af87c587c4e..31a5b79eab969ec6bece56f90350a343f473b908 100644 (file)
@@ -385,11 +385,9 @@ visible(int vis)
 
 int main(int argc, char *argv[])
 {
+  glutInitWindowSize(300, 300);
   glutInit(&argc, argv);
   glutInitDisplayMode(GLUT_RGB | GLUT_DEPTH | GLUT_DOUBLE);
-
-  glutInitWindowPosition(0, 0);
-  glutInitWindowSize(300, 300);
   win = glutCreateWindow("Gears");
   init(argc, argv);
 
index 8fe405e80760e32d8c1d7c9f4330223b1f0d900f..d2a195f39a214ea212d68b559205214847ec6477 100644 (file)
@@ -1049,19 +1049,18 @@ main (int argc, char *argv[])
 {
     char *file;
 
-    if (argc < 2)
-       file = "geartrain.dat";
-    else
-       file = argv[1];
-
-    glutInit(&argc, argv); 
-    glutInitWindowPosition (0, 0);
     glutInitWindowSize(640,480);
+    glutInit(&argc, argv); 
     glutInitDisplayMode (GLUT_RGB | GLUT_DEPTH | GLUT_DOUBLE );
 
     if (glutCreateWindow ("Gear Train Simulation") == GL_FALSE)
       exit (1);
 
+    if (argc < 2)
+       file = "geartrain.dat";
+    else
+       file = argv[1];
+
     getdata (file);
     process ();
     init ();
index 69694b23a09bc5cb44782a435da30761a39214fa..578736b4e27c0e654ea8f1f13cbd17a8366c3f29 100644 (file)
@@ -436,8 +436,8 @@ static void Init( int argc, char *argv[] )
 
 int main( int argc, char *argv[] )
 {
-   glutInit( &argc, argv );
    glutInitWindowSize(WinWidth, WinHeight);
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH );
    glutCreateWindow(argv[0] );
    glewInit();
index 2188b0241924034f42f545c3f45237df9b8b1462..3658a395988a0ac33eb9fbed6fd9f71db5b4972a 100644 (file)
@@ -569,10 +569,9 @@ main(int ac, char **av)
    if (ac == 2)
       frontbuffer = 0;
 
+   glutInitWindowSize(640, 480);
    glutInit(&ac, av);
    glutInitDisplayMode(GLUT_DOUBLE | GLUT_RGB | GLUT_DEPTH);
-   glutInitWindowPosition(0, 0);
-   glutInitWindowSize(640, 480);
    glutCreateWindow("OpenGL/Mesa Performances");
    glutDisplayFunc(display);
    glutMainLoop();
index ed03673cb62112a1849d78ead398d87f2ab8bb80..265378b90a9b883b4c2770c165859ca5a84c8cad 100644 (file)
@@ -682,7 +682,6 @@ main(int ac, char **av)
    fprintf(stderr,
           "IperS V1.0\nWritten by David Bucciarelli (tech.hmw@plus.it)\n");
 
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(WIDTH, HEIGHT);
    glutInit(&ac, av);
 
index 2e9dff172613b1da5688a74d2dadf740a498ac27..56781f68189fb420015a8dfe403e27d1a960c19f 100644 (file)
@@ -1061,9 +1061,8 @@ int main(int argc, char **argv)
 
    read_surface( "isosurf.dat" );
 
-   glutInit( &argc, argv);
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(400, 400);
+   glutInit( &argc, argv);
 
    type = GLUT_DEPTH;
    type |= GLUT_RGB;
index 8d39bd605a7da472958b3656b7de3a5495e7b26a..23488b1616e0feeccef0f827b11be738510870f3 100644 (file)
@@ -282,9 +282,8 @@ static void Init( void )
 
 int main( int argc, char *argv[] )
 {
-   glutInit( &argc, argv );
-   glutInitWindowPosition( 0, 0 );
    glutInitWindowSize( 350, 350 );
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE );
    win = glutCreateWindow(argv[0]);
    glutReshapeFunc( Reshape );
index 01a06aba0ff77dae88da7827017cc2bd91b9e0aa..0f8ac426f33d4238d4db0fc7a96e029d9ff88cfc 100644 (file)
@@ -841,9 +841,8 @@ int main(int argc, char **argv)
 
   object=1;
 
-  glutInit(&argc, argv);
-  glutInitWindowPosition(0,0);
   glutInitWindowSize(640,480);
+  glutInit(&argc, argv);
 
   glutInitDisplayMode( GLUT_DEPTH | GLUT_DOUBLE | GLUT_RGB );
 
index 3d89d3a13e9ba809a3a6447a7696e42c65c6b70f..82796a0c3f4f4341c96b2ebac2d5cfe5575886d8 100644 (file)
@@ -338,9 +338,8 @@ int main( int argc, char *argv[] )
 {
    GLint i;
 
-   glutInit( &argc, argv );
    glutInitWindowSize( 300, 300 );
-   glutInitWindowPosition( 0, 0 );
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE );
    glutCreateWindow(argv[0] );
    glewInit();
index d0cbcfb5e9a3b130de18850a2228e601e59ee686..5d4b8662b5cb8d050f9522be2ec9bfc9943eca02 100644 (file)
@@ -248,12 +248,9 @@ static void Init2( void )
 
 int main( int argc, char *argv[] )
 {
-   glutInit( &argc, argv );
-   glutInitWindowPosition( 0, 0 );
    glutInitWindowSize( 400, 300 );
-
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE );
-
    glutCreateWindow(argv[0]);
    glewInit();
 
index 2d70b72589418fdd389642fc6ba8f4ecb7aadca7..b4d0a67f9930d075327603d76a52eb895b4f9ec5 100644 (file)
@@ -427,10 +427,10 @@ int
 main(int argc, char **argv)
 {
   int i;
+
+  glutInitWindowSize(300, 300);
   glutInit(&argc, argv);
   glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH | GLUT_MULTISAMPLE);
-  glutInitWindowPosition(0, 0);
-  glutInitWindowSize(300, 300);
 
   for (i=1; i<argc; i++) {
     if(!strcmp("-noms", argv[i])) {
index 503cf5de08884faaaae39f305472cf496929fc9e..d162568b9f0913720d3558ddfaa11aa47c0fcb97 100644 (file)
@@ -998,18 +998,17 @@ menu(int selection)
 int
 main(int argc, char **argv)
 {
+  glutInitWindowSize(500,500);
   glutInit(&argc, argv);
+  glutInitDisplayMode(GLUT_RGBA | GLUT_DEPTH | GLUT_DOUBLE);
+  (void) glutCreateWindow("projtex");
+  glewInit();
 
   if (argc > 1) {
      NumTextures = atoi(argv[1]);
   }
   assert(NumTextures <= MAX_TEX);
 
-  glutInitDisplayMode(GLUT_RGBA | GLUT_DEPTH | GLUT_DOUBLE);
-  glutInitWindowSize(500,500);
-  (void) glutCreateWindow("projtex");
-  glewInit();
-
   loadTexture = loadImageTextures;
   drawObject = drawCube;
   initialize();
index e9211aa3399b5a934b6b8be76c5de6eaf35756f3..9cc464d2390a5722b812d830e7f2924a0aa0ab18 100644 (file)
@@ -856,7 +856,6 @@ main(int ac, char **av)
       }
     */
 
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(WIDTH, HEIGHT);
    glutInit(&ac, av);
 
index 182b01d7ffca72e7363e04736405d76e457d1320..cc4e4902694d0681a61dbe13c657dad4fcc9290f 100644 (file)
@@ -383,12 +383,11 @@ int
 main( int argc, char *argv[] )
 {
    GLboolean ciMode = GL_FALSE;
+   glutInitWindowSize( 750, 250 );
+   glutInit( &argc, argv );
    if (argc > 1 && strcmp(argv[1], "-ci")==0) {
       ciMode = GL_TRUE;
    }
-   glutInit( &argc, argv );
-   glutInitWindowPosition( 0, 0 );
-   glutInitWindowSize( 750, 250 );
    if (ciMode)
       glutInitDisplayMode( GLUT_INDEX | GLUT_DOUBLE );
    else
index 9e5da95484c810dabde1df372b424daf59c82e59..61dd860ddbe29ce92407e02b6432b393da8c81a7 100644 (file)
@@ -112,12 +112,9 @@ key(unsigned char k, int x, int y)
 
 int main( int argc, char *argv[] )
 {
-   glutInit( &argc, argv );
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize( 400, 400 );
-
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE );
-
    glutCreateWindow(argv[0]);
 
    Init();
index dc5a4bbc480b8e4a4c780af0a7b62958ed449e48..036f73d40b6650b563cbe94b47ccb05fb2393989 100644 (file)
@@ -1022,9 +1022,8 @@ PrintHelp(void)
 int
 main(int argc, char *argv[])
 {
-   glutInit(&argc, argv);
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(WindowWidth, WindowHeight);
+   glutInit(&argc, argv);
    glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH | GLUT_STENCIL);
    glutCreateWindow(argv[0]);
    glewInit();
index 6ab1191579c3be6d41e43329c40e84c8e2a057b7..c1dada9d634784b16f14bb931008a7d1644d772b 100644 (file)
@@ -239,13 +239,9 @@ static void ModeMenu(int entry)
 
 int main( int argc, char *argv[] )
 {
-
-   glutInit( &argc, argv );
-   glutInitWindowPosition( 0, 0 );
    glutInitWindowSize( 300, 300 );
-
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH );
-
    glutCreateWindow( "spectex" );
 
    Init();
index d73b680b791dfd568a0ef5718a46e17eead88e68..36e226e1d07929ae7befb6700b8fb7ec6250e207 100644 (file)
@@ -493,6 +493,8 @@ int
 main(int argc, char **argv)
 {
   int i;
+
+  glutInitWindowSize(600,300);
   glutInit(&argc, argv);
   glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH | GLUT_MULTISAMPLE);
 
@@ -506,8 +508,6 @@ main(int argc, char **argv)
       linearFiltering = 0;
     }
   }
-  glutInitWindowPosition(0, 0);
-  glutInitWindowSize(600,300);
   glutCreateWindow("sprite blast");
   glewInit();
   glutReshapeFunc(reshape);
index 6bf6e06409513d688e53c2c2b109f2646adb6b3a..04a675f36e615123e16da90ef4fc101b26b05c0f 100644 (file)
@@ -638,7 +638,6 @@ int main(int ac, char **av)
     }
     */
 
-  glutInitWindowPosition(0,0);
   glutInitWindowSize(WIDTH,HEIGHT);
   glutInit(&ac,av);
 
index be78ea4950f7a959096de6fd1564993aa368b95a..627c3bfb57a8ef3813ba970e2e4eac93202439de 100644 (file)
@@ -623,7 +623,6 @@ init(void)
 int
 main(int ac, char **av)
 {
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(WIDTH, HEIGHT);
    glutInit(&ac, av);
 
index f71cea127448ca09e15ad2c416950db21969cbc1..8b988e36c136202b3fe15806f0f5e2de25d0e68e 100644 (file)
@@ -503,10 +503,9 @@ int main( int argc, char **argv )
 
    usage();
 
+   glutInitWindowSize( 400, 400 );
    glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_SINGLE | GLUT_RGB );
-   glutInitWindowPosition(0, 0);
-   glutInitWindowSize( 400, 400 );
    glutCreateWindow( argv[0] );
 
    /* GH: Bit of a hack...
index 0e6089bced54e977eefff53845d16e28b5cc9686..4df97ec4f502593f44518e067b0c58c4bf2db26b 100644 (file)
@@ -261,10 +261,8 @@ static void Init( int argc, char *argv[] )
 
 int main( int argc, char *argv[] )
 {
-   glutInit( &argc, argv );
    glutInitWindowSize( 400, 400 );
-   glutInitWindowPosition( 0, 0 );
-
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE );
 
    Win = glutCreateWindow(argv[0] );
index c5a8b13f35fbcc692988daadaf8f52b4f55f4bb0..260ca8718f0e3be0675bfe1e94f2569c79a8ee29 100644 (file)
@@ -765,6 +765,7 @@ int main( int argc, char *argv[] )
    GLboolean info = GL_FALSE;
    int i;
 
+   glutInitWindowSize( winWidth, winHeight );
    glutInit( &argc, argv );
 
    for ( i = 1 ; i < argc ; i++ ) {
@@ -786,8 +787,6 @@ int main( int argc, char *argv[] )
       glutInitDisplayMode( GLUT_RGB | GLUT_SINGLE );
    }
 
-   glutInitWindowSize( winWidth, winHeight );
-   glutInitWindowPosition( 0, 0 );
    Win = glutCreateWindow( "Texture Environment Test" );
 
    initialize();
index 31e1bbb3db0631b91fc340c0fbfc199a663c7862..1415ef1c43cfb5b05b4e85d0b28a5190e45c4986 100644 (file)
@@ -356,9 +356,8 @@ Usage(void)
 int
 main(int argc, char *argv[])
 {
-   glutInit(&argc, argv);
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(700, 700);
+   glutInit(&argc, argv);
    glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH);
    Win = glutCreateWindow(argv[0]);
    glutReshapeFunc(Reshape);
index 165d088e3d306bb245192288c97ae25ded11fb64..c07c3409472d4414c071ce9ed824eb4fc4b2c5d5 100644 (file)
@@ -227,15 +227,12 @@ static void Help( const char *program )
 
 int main( int argc, char *argv[] )
 {
-   printf("For options:  %s -help\n", argv[0]);
-   glutInit( &argc, argv );
    glutInitWindowSize( (int) Width, (int) Height );
-   glutInitWindowPosition( 0, 0 );
-
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH );
-
    glutCreateWindow( argv[0] );
 
+   printf("For options:  %s -help\n", argv[0]);
    if (argc==2 && strcmp(argv[1],"-help")==0) {
       Help(argv[0]);
       return 0;
index 6981da32988db505d92f2b112a6423c7418a4b50..efc007c79aa5af29cc46f6d97b887f24a11423a5 100644 (file)
@@ -491,7 +491,6 @@ main(int ac, char **av)
    fprintf(stderr,
           "Tunnel V1.5\nWritten by David Bucciarelli (tech.hmw@plus.it)\n");
 
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(WIDTH, HEIGHT);
    glutInit(&ac, av);
 
index 0288ea0f8ceaede6de5f4e5aee1d99f825ec1a55..f15da95b18bea25edd7f377d39c445571a68b66d 100644 (file)
@@ -559,7 +559,6 @@ main(int ac, char **av)
    fprintf(stderr,
           "Tunnel2 V1.0\nWritten by David Bucciarelli (tech.hmw@plus.it)\n");
 
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(WIDTHC0, HEIGHTC0);
    glutInit(&ac, av);
 
index 206e06fc6c7141d255e3c06c07b7551985893e2c..5ce07ecc63b8ceba64db64f9e7044ce1567ff105 100644 (file)
@@ -317,9 +317,8 @@ static void Init( void )
 
 int main( int argc, char *argv[] )
 {
-   glutInit( &argc, argv );
-   glutInitWindowPosition( 0, 0 );
    glutInitWindowSize( Width, Height );
+   glutInit( &argc, argv );
    glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH );
    Win = glutCreateWindow( "GL_APPLE_vertex_array_object demo" );
    glutReshapeFunc( Reshape );
index 13a9c7e9a8f7e1382661db85805ececf7145c5f7..f935f9bee0f1ae829fb67b9743bc755f0cd3d1fa 100644 (file)
@@ -96,9 +96,8 @@ static void init( void )
 
 int main( int argc, char *argv[] )
 {
-   glutInit(&argc, argv);
-   glutInitWindowPosition(0, 0);
    glutInitWindowSize(500, 500);
+   glutInit(&argc, argv);
    glutInitDisplayMode( GLUT_RGB );
 
    if (glutCreateWindow("winpos") <= 0) {
index 26af66ad84e72979e41dace04a1d24f4d8b10470..ed29a2d683dde77c094853fea831b040cf44aaec 100644 (file)
@@ -176,6 +176,17 @@ static void Init( void )
    }
 
 
+   {
+      const float Ambient[4] = { 0.0, 1.0, 0.0, 0.0 };
+      const float Diffuse[4] = { 1.0, 0.0, 0.0, 0.0 };
+      const float Specular[4] = { 0.0, 0.0, 1.0, 0.0 };
+      const float Emission[4] = { 0.0, 0.0, 0.0, 1.0 };
+      glMaterialfv(GL_FRONT_AND_BACK, GL_AMBIENT, Ambient);
+      glMaterialfv(GL_FRONT_AND_BACK, GL_DIFFUSE, Diffuse);
+      glMaterialfv(GL_FRONT_AND_BACK, GL_SPECULAR, Specular);
+      glMaterialfv(GL_FRONT_AND_BACK, GL_EMISSION, Emission);
+   }
+
    glClearColor(.1, .3, .5, 0);
 }
 
diff --git a/progs/fp/mov-imm.txt b/progs/fp/mov-imm.txt
new file mode 100644 (file)
index 0000000..38e4807
--- /dev/null
@@ -0,0 +1,3 @@
+!!ARBfp1.0
+MOV result.color, {0.5, 0.8, 0.3, 1.0};
+END
diff --git a/progs/fp/mov-param.txt b/progs/fp/mov-param.txt
new file mode 100644 (file)
index 0000000..13d82fe
--- /dev/null
@@ -0,0 +1,4 @@
+!!ARBfp1.0
+PARAM Diffuse = state.material.diffuse; 
+MOV result.color, Diffuse;
+END
diff --git a/progs/glsl/CH11-bumpmaptex.frag b/progs/glsl/CH11-bumpmaptex.frag
new file mode 100644 (file)
index 0000000..b1f93b7
--- /dev/null
@@ -0,0 +1,47 @@
+//
+// Fragment shader for procedural bumps
+//
+// Authors: John Kessenich, Randi Rost
+//
+// Copyright (c) 2002-2006 3Dlabs Inc. Ltd. 
+//
+// See 3Dlabs-License.txt for license information
+//
+// Texture mapping/modulation added by Brian Paul
+//
+
+varying vec3 LightDir;
+varying vec3 EyeDir;
+
+uniform float BumpDensity;     // = 16.0
+uniform float BumpSize;        // = 0.15
+uniform float SpecularFactor;  // = 0.5
+
+sampler2D Tex;
+
+void main()
+{
+    vec3 ambient = vec3(0.25);
+    vec3 litColor;
+    vec2 c = BumpDensity * gl_TexCoord[0].st;
+    vec2 p = fract(c) - vec2(0.5);
+
+    float d, f;
+    d = p.x * p.x + p.y * p.y;
+    f = inversesqrt(d + 1.0);
+
+    if (d >= BumpSize)
+        { p = vec2(0.0); f = 1.0; }
+
+    vec3 SurfaceColor = texture2D(Tex, gl_TexCoord[0].st).xyz;
+
+    vec3 normDelta = vec3(p.x, p.y, 1.0) * f;
+    litColor = SurfaceColor * (ambient + max(dot(normDelta, LightDir), 0.0));
+    vec3 reflectDir = reflect(LightDir, normDelta);
+    
+    float spec = max(dot(EyeDir, reflectDir), 0.0);
+    spec *= SpecularFactor;
+    litColor = min(litColor + spec, vec3(1.0));
+
+    gl_FragColor = vec4(litColor, 1.0);
+}
index 50a0900f1c796d36a0c49fb738df68795b7b7b0d..e31afab9392e4f78388446cc42306764661aa112 100644 (file)
 #include <GL/glew.h>
 #include <GL/glut.h>
 #include "shaderutil.h"
+#include "readtex.h"
 
 
 static char *FragProgFile = "CH11-bumpmap.frag";
+static char *FragTexProgFile = "CH11-bumpmaptex.frag";
 static char *VertProgFile = "CH11-bumpmap.vert";
+static char *TextureFile = "../images/tile.rgb";
 
 /* program/shader objects */
 static GLuint fragShader;
+static GLuint fragTexShader;
 static GLuint vertShader;
 static GLuint program;
+static GLuint texProgram;
 
 
 static struct uniform_info Uniforms[] = {
@@ -32,13 +37,26 @@ static struct uniform_info Uniforms[] = {
    END_OF_UNIFORMS
 };
 
+static struct uniform_info TexUniforms[] = {
+   { "LightPosition",  1, GL_FLOAT_VEC3, { 0.57737, 0.57735, 0.57735, 0.0 }, -1 },
+   { "Tex",            1, GL_INT,   { 0, 0, 0, 0 }, -1 },
+   { "BumpDensity",    1, GL_FLOAT, { 10.0, 0, 0, 0 }, -1 },
+   { "BumpSize",       1, GL_FLOAT, { 0.125, 0, 0, 0 }, -1 },
+   { "SpecularFactor", 1, GL_FLOAT, { 0.5, 0, 0, 0 }, -1 },
+   END_OF_UNIFORMS
+};
+
 static GLint win = 0;
 
 static GLfloat xRot = 20.0f, yRot = 0.0f, zRot = 0.0f;
 
 static GLint tangentAttrib;
+static GLint tangentAttribTex;
+
+static GLuint Texture;
 
 static GLboolean Anim = GL_FALSE;
+static GLboolean Textured = GL_FALSE;
 
 
 static void
@@ -135,6 +153,11 @@ Redisplay(void)
    glRotatef(yRot, 0.0f, 1.0f, 0.0f);
    glRotatef(zRot, 0.0f, 0.0f, 1.0f);
 
+   if (Textured)
+      glUseProgram(texProgram);
+   else
+      glUseProgram(program);
+
    Cube(1.5);
 
    glPopMatrix();
@@ -163,8 +186,10 @@ static void
 CleanUp(void)
 {
    glDeleteShader(fragShader);
+   glDeleteShader(fragTexShader);
    glDeleteShader(vertShader);
    glDeleteProgram(program);
+   glDeleteProgram(texProgram);
    glutDestroyWindow(win);
 }
 
@@ -181,6 +206,9 @@ Key(unsigned char key, int x, int y)
       Anim = !Anim;
       glutIdleFunc(Anim ? Idle : NULL);
       break;
+   case 't':
+      Textured = !Textured;
+      break;
    case 'z':
       zRot += step;
       break;
@@ -254,6 +282,26 @@ Init(void)
 
    CheckError(__LINE__);
 
+
+   /*
+    * As above, but fragment shader also uses a texture map.
+    */
+   fragTexShader = CompileShaderFile(GL_FRAGMENT_SHADER, FragTexProgFile);
+   texProgram = LinkShaders(vertShader, fragTexShader);
+   glUseProgram(texProgram);
+   assert(glIsProgram(texProgram));
+   assert(glIsShader(fragTexShader));
+   SetUniformValues(texProgram, TexUniforms);
+   PrintUniforms(TexUniforms);
+
+   /*
+    * Load tex image.
+    */
+   glGenTextures(1, &Texture);
+   glBindTexture(GL_TEXTURE_2D, Texture);
+   LoadRGBMipmaps(TextureFile, GL_RGB);
+
+
    glClearColor(0.4f, 0.4f, 0.8f, 0.0f);
 
    glEnable(GL_DEPTH_TEST);
@@ -268,10 +316,13 @@ ParseOptions(int argc, char *argv[])
    int i;
    for (i = 1; i < argc; i++) {
       if (strcmp(argv[i], "-fs") == 0) {
-         FragProgFile = argv[i+1];
+         FragProgFile = argv[++i];
       }
       else if (strcmp(argv[i], "-vs") == 0) {
-         VertProgFile = argv[i+1];
+         VertProgFile = argv[++i];
+      }
+      else if (strcmp(argv[i], "-t") == 0) {
+         TextureFile = argv[++i];
       }
    }
 }
index 30f2b75fef7c4c93b6c58bba6725dfb6530129fc..588246b71ac6b2079fe4eaa2ad62ac4de1ed623b 100644 (file)
@@ -27,11 +27,15 @@ static GLuint SphereList, RectList, CurList;
 static GLint win = 0;
 static GLboolean anim = GL_TRUE;
 static GLfloat xRot = 0.0f, yRot = 0.0f;
+static GLint WinSize[2];
+static GLint WinSizeUniform = -1;
 
 
 static void
 Redisplay(void)
 {
+   glUniform2iv(WinSizeUniform, 1, WinSize);
+
    glClear(GL_COLOR_BUFFER_BIT | GL_DEPTH_BUFFER_BIT);
 
    glPushMatrix();
@@ -55,6 +59,8 @@ Idle(void)
 static void
 Reshape(int width, int height)
 {
+   WinSize[0] = width;
+   WinSize[1] = height;
    glViewport(0, 0, width, height);
    glMatrixMode(GL_PROJECTION);
    glLoadIdentity();
@@ -163,8 +169,10 @@ static void
 Init(void)
 {
    static const char *fragShaderText =
+      "uniform ivec2 WinSize; \n"
       "void main() {\n"
-      "   gl_FragColor = abs(dFdy(gl_TexCoord[0])) * 50.0;\n"
+      "   vec2 d = dFdy(gl_TexCoord[0].xy) * vec2(WinSize); \n"
+      "   gl_FragColor =  vec4(d.x, d.y, 0.0, 1.0);\n"
       "  // gl_FragColor = gl_TexCoord[0];\n"
       "}\n";
    static const char *vertShaderText =
@@ -181,6 +189,7 @@ Init(void)
    program = LinkShaders(vertShader, fragShader);
 
    glUseProgram(program);
+   WinSizeUniform = glGetUniformLocation(program, "WinSize");
 
    /*assert(glGetError() == 0);*/
 
@@ -220,8 +229,10 @@ ParseOptions(int argc, char *argv[])
 int
 main(int argc, char *argv[])
 {
+   WinSize[0] = WinSize[1] = 200;
+
    glutInit(&argc, argv);
-   glutInitWindowSize(200, 200);
+   glutInitWindowSize(WinSize[0], WinSize[1]);
    glutInitDisplayMode(GLUT_RGB | GLUT_DOUBLE | GLUT_DEPTH);
    win = glutCreateWindow(argv[0]);
    glewInit();
index 023ea02ae27021abbf97d81fda0556ac8d73beb9..f53515cb0a76397c9d605a7b22f91571b1e0ec97 100644 (file)
@@ -5,7 +5,7 @@ include $(TOP)/configs/current
 
 INCDIR = $(TOP)/include
 
-OSMESA_LIBS = -L$(TOP)/$(LIB_DIR) -l$(GLU_LIB) -l$(OSMESA_LIB) $(APP_LIB_DEPS)
+OSMESA_LIBS = -L$(TOP)/$(LIB_DIR) -l$(GLU_LIB) -l$(GL_LIB) -l$(OSMESA_LIB) $(APP_LIB_DEPS)
 
 OSMESA16_LIBS = -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -lOSMesa16 -l$(GLU_LIB) \
        -l$(GL_LIB) $(APP_LIB_DEPS)
@@ -13,12 +13,6 @@ OSMESA16_LIBS = -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -lOSMesa16 -l$(GLU_LIB) \
 OSMESA32_LIBS = -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -lOSMesa32 -l$(GLU_LIB) \
        -l$(GL_LIB) $(APP_LIB_DEPS)
 
-LIB_DEP = $(TOP)/$(LIB_DIR)/$(GL_LIB_NAME) $(TOP)/$(LIB_DIR)/$(GLU_LIB_NAME) \
-       $(TOP)/$(LIB_DIR)/$(GLUT_LIB_NAME)
-
-LIBS = -L$(TOP)/$(LIB_DIR) -l$(GLUT_LIB) -l$(GLU_LIB) -l$(GL_LIB) \
-       $(APP_LIB_DEPS)
-
 PROGS = \
        osdemo \
        ostest1
@@ -30,11 +24,6 @@ PROGS = \
 .SUFFIXES: .c
 
 
-# make executable from .c file:
-.c: $(LIB_DEP) readtex.o
-       $(CC) -I$(INCDIR) $(CFLAGS) $(LDFLAGS) $< readtex.o $(LIBS) -o $@
-
-
 ##### TARGETS #####
 
 default: readtex.o $(PROGS)
index cdae3486ce008e80cc2d9bb277673872ab0a16e0..49a5416787a6d7c1726281028467bc81931eeb22 100644 (file)
 #include "pipe/p_compiler.h"
 #include "pipe/p_format.h"
 #include "pipe/p_state.h"
+#include "util/u_format.h"
 #include "util/u_memory.h"
 #include "util/u_debug.h"
+#include "util/u_format.h"
 #include "util/u_network.h"
 #include "util/u_tile.h"
 
@@ -54,10 +56,7 @@ static void dump(unsigned width, unsigned height,
                  unsigned src_stride, enum pipe_format src_format,
                  uint8_t *data, unsigned src_size)
 {
-   struct pipe_format_block src_block;
-
    enum pipe_format dst_format = PIPE_FORMAT_R32G32B32A32_FLOAT;
-   struct pipe_format_block dst_block;
    unsigned dst_stride;
    unsigned dst_size;
    float *rgba;
@@ -65,20 +64,17 @@ static void dump(unsigned width, unsigned height,
    char filename[512];
 
    {
-      pf_get_block(src_format, &src_block);
-      assert(src_stride >= pf_get_stride(&src_block, width));
-      assert(src_size >= pf_get_2d_size(&src_block, src_stride, width));
+      assert(src_stride >= util_format_get_stride(src_format, width));
    }
    {
-      pf_get_block(dst_format, &dst_block);
-      dst_stride = pf_get_stride(&dst_block, width);
-      dst_size = pf_get_2d_size(&dst_block, dst_stride, width);
+      dst_stride = util_format_get_stride(dst_format, width);
+      dst_size = util_format_get_2d_size(dst_format, dst_stride, width);
       rgba = MALLOC(dst_size);
    }
 
    util_snprintf(filename, 512, "%s.bmp", pf_name(src_format));
 
-   if (pf_is_compressed(src_format)) {
+   if (util_format_is_compressed(src_format)) {
       debug_printf("skipping: %s\n", filename);
       return;
    }
index f9e06ee994aa030c82d6ba3db49fc8e7590eb406..963f8eeedea684bfa4ba8666e6fe0207c4e7412e 100644 (file)
@@ -27,6 +27,7 @@
 #include "pipe/p_state.h"
 #include "util/u_memory.h"
 #include "util/u_debug.h"
+#include "util/u_format.h"
 #include "util/u_network.h"
 #include "util/u_tile.h"
 #include "rbug/rbug.h"
@@ -49,7 +50,7 @@ static void dump(rbug_texture_t tex,
    util_snprintf(filename, 512, "%llu_%s_%u.bmp",
                  (unsigned long long)tex, pf_name(info->format), mip);
 
-   if (pf_is_compressed(info->format)) {
+   if (util_format_is_compressed(info->format)) {
       debug_printf("skipping: %s\n", filename);
       return;
    }
index 197e14d5b00a0cc1f7425f57a769d28c7166b9a6..3e2541186b12b45409d0507d06cc94dfb7742869 100644 (file)
@@ -98,6 +98,7 @@ SOURCES = \
        texdown \
        texfilt.c \
        texgenmix.c \
+       texleak.c \
        texline.c \
        texobj.c \
        texobjshare.c \
diff --git a/progs/tests/texleak.c b/progs/tests/texleak.c
new file mode 100644 (file)
index 0000000..5cf4ff3
--- /dev/null
@@ -0,0 +1,140 @@
+/*
+ * 'Texture leak' test
+ *
+ * Allocates and uses an additional texture of the maximum supported size for
+ * each frame. This tests the system's ability to cope with using increasing
+ * amounts of texture memory.
+ *
+ * Michel Dänzer  July 2009  This program is in the public domain.
+ */
+
+
+#include <math.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <sys/time.h>
+#include <unistd.h>
+#include <GL/glew.h>
+#include <GL/glut.h>
+
+
+GLint size;
+GLvoid *image;
+static GLuint numTexObj;
+static GLuint *texObj;
+
+
+static void Idle( void )
+{
+   glutPostRedisplay();
+}
+
+
+static void DrawObject(void)
+{
+   static const GLfloat tex_coords[] = {  0.0,  0.0,  1.0,  1.0,  0.0 };
+   static const GLfloat vtx_coords[] = { -1.0, -1.0,  1.0,  1.0, -1.0 };
+   GLint i, j;
+
+   glEnable(GL_TEXTURE_2D);
+
+   for (i = 0; i < numTexObj; i++) {
+      glBindTexture(GL_TEXTURE_2D, texObj[i]);
+      glBegin(GL_QUADS);
+      for (j = 0; j < 4; j++ ) {
+         glTexCoord2f(tex_coords[j], tex_coords[j+1]);
+         glVertex2f( vtx_coords[j], vtx_coords[j+1] );
+      }
+      glEnd();
+   }
+}
+
+
+static void Display( void )
+{
+   struct timeval start, end;
+
+   texObj = realloc(texObj, ++numTexObj * sizeof(*texObj));
+
+   /* allocate a texture object */
+   glGenTextures(1, texObj + (numTexObj - 1));
+
+   glBindTexture(GL_TEXTURE_2D, texObj[numTexObj - 1]);
+   glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MIN_FILTER, GL_NEAREST);
+   glTexParameteri(GL_TEXTURE_2D, GL_TEXTURE_MAG_FILTER, GL_NEAREST);
+   memset(image, (16 * numTexObj) & 0xff, 4 * size * size);
+   glTexImage2D(GL_TEXTURE_2D, 0, GL_RGBA, size, size, 0,
+                GL_RGBA, GL_UNSIGNED_BYTE, image);
+
+   gettimeofday(&start, NULL);
+
+   glClear( GL_COLOR_BUFFER_BIT );
+
+   glPushMatrix();
+      glScalef(5.0, 5.0, 5.0);
+      DrawObject();
+   glPopMatrix();
+
+   glutSwapBuffers();
+
+   glFinish();
+   gettimeofday(&end, NULL);
+   printf("Rendering frame took %lu ms using %u MB of textures\n",
+         end.tv_sec * 1000 + end.tv_usec / 1000 - start.tv_sec * 1000 -
+         start.tv_usec / 1000, numTexObj * 4 * size / 1024 * size / 1024);
+
+   sleep(1);
+}
+
+
+static void Reshape( int width, int height )
+{
+   glViewport( 0, 0, width, height );
+   glMatrixMode( GL_PROJECTION );
+   glLoadIdentity();
+   glOrtho( -6.0, 6.0, -6.0, 6.0, 10.0, 100.0 );
+   glMatrixMode( GL_MODELVIEW );
+   glLoadIdentity();
+   glTranslatef( 0.0, 0.0, -70.0 );
+}
+
+
+static void Init( int argc, char *argv[] )
+{
+   glGetIntegerv(GL_MAX_TEXTURE_SIZE, &size);
+   printf("%d x %d max texture size\n", size, size);
+
+   image = malloc(4 * size * size);
+   if (!image) {
+      fprintf(stderr, "Failed to allocate %u bytes of memory\n", 4 * size * size);
+      exit(1);
+   }
+
+   glPixelStorei(GL_UNPACK_ALIGNMENT, 1);
+
+   glShadeModel(GL_FLAT);
+   glClearColor(0.3, 0.3, 0.4, 1.0);
+
+   Idle();
+}
+
+
+int main( int argc, char *argv[] )
+{
+   glutInit( &argc, argv );
+   glutInitWindowSize( 300, 300 );
+   glutInitWindowPosition( 0, 0 );
+   glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE );
+   glutCreateWindow(argv[0] );
+   glewInit();
+
+   Init( argc, argv );
+
+   glutReshapeFunc( Reshape );
+   glutDisplayFunc( Display );
+   glutIdleFunc(Idle);
+
+   glutMainLoop();
+   return 0;
+}
index 4d6e405c500e394f7071d91848411502b1d089bd..4317eb607feef093f322572b5ac0123b81acde6a 100644 (file)
@@ -147,6 +147,7 @@ vbo-drawarrays
 vbo-drawelements
 vbo-drawrange
 vbo-noninterleaved
+vbo-tri
 vp-array
 vp-array-int
 vp-clip
index 70728616d28b187ac1783178d651949e9efefd4c..e15ec33ab5922ef8bda76994f4dce7f99031fdc5 100644 (file)
@@ -153,6 +153,7 @@ SOURCES = \
        tristrip-clip.c \
        tristrip-flat.c \
        tristrip.c \
+       vbo-tri.c \
        vbo-drawarrays.c \
        vbo-noninterleaved.c \
        vbo-drawelements.c \
index d86d34c39dee63a786fd79cf726e7f2cce57e250..f86ac52a02649fdfca003f8e765ba0268abbcb5a 100644 (file)
@@ -51,7 +51,7 @@ static void Reshape(int width, int height)
 
     glMatrixMode(GL_PROJECTION);
     glLoadIdentity();
-/*     glOrtho(-1.0, 1.0, -1.0, 1.0, -0.5, 1000.0); */
+    glOrtho(-1.0, 1.0, -1.0, 1.0, -0.5, 1000.0);
     glMatrixMode(GL_MODELVIEW);
 }
 
@@ -74,11 +74,11 @@ static void Draw(void)
 
    glBegin(GL_TRIANGLES);
    glColor3f(0,0,.7); 
-   glVertex3f( 0.9, -0.9, -0.0);
+   glVertex3f( 0.9, -0.9, -30.0);
    glColor3f(.8,0,0); 
-   glVertex3f( 0.9,  0.9, -0.0);
+   glVertex3f( 0.9,  0.9, -30.0);
    glColor3f(0,.9,0); 
-   glVertex3f(-0.9,  0.0, -0.0);
+   glVertex3f(-0.9,  0.0, -30.0);
    glEnd();
 
    glFlush();
@@ -119,7 +119,7 @@ int main(int argc, char **argv)
 
     glutInitWindowPosition(0, 0); glutInitWindowSize( 250, 250);
 
-    type = GLUT_RGB | GLUT_ALPHA;
+    type = GLUT_RGB;
     type |= (doubleBuffer) ? GLUT_DOUBLE : GLUT_SINGLE;
     glutInitDisplayMode(type);
 
diff --git a/progs/trivial/vbo-tri.c b/progs/trivial/vbo-tri.c
new file mode 100644 (file)
index 0000000..d4cba14
--- /dev/null
@@ -0,0 +1,131 @@
+/* Even simpler for many drivers than trivial/tri -- pass-through
+ * vertex shader and vertex data in a VBO.
+ */
+
+#include <assert.h>
+#include <string.h>
+#include <stdio.h>
+#include <stdlib.h>
+#include <math.h>
+#include <GL/glew.h>
+#include <GL/glut.h>
+
+
+struct {
+   GLfloat pos[4];
+   GLfloat color[4];
+} verts[] =  
+{
+   { {  -0.9, -0.9, 0.0, 1.0 },
+     {.8,0,0, 1}, 
+   },
+
+   { {  0.9,  -0.9, 0.0, 1.0 },
+     { 0, .9, 0, 1 },
+   },
+
+   { { 0,  0.9, 0.0, 1.0 },
+     {0,0,.7, 1}, 
+   },
+};
+
+GLuint arrayObj;
+
+static void Init( void )
+{
+   GLint errno;
+   GLuint prognum;
+   
+   static const char *prog1 =
+      "!!ARBvp1.0\n"
+      "MOV  result.color, vertex.color;\n"
+      "MOV  result.position, vertex.position;\n"
+      "END\n";
+
+
+   glGenProgramsARB(1, &prognum);
+
+   glBindProgramARB(GL_VERTEX_PROGRAM_ARB, prognum);
+   glProgramStringARB(GL_VERTEX_PROGRAM_ARB, GL_PROGRAM_FORMAT_ASCII_ARB,
+                     strlen(prog1), (const GLubyte *) prog1);
+
+   assert(glIsProgramARB(prognum));
+   errno = glGetError();
+   printf("glGetError = %d\n", errno);
+   if (errno != GL_NO_ERROR)
+   {
+      GLint errorpos;
+
+      glGetIntegerv(GL_PROGRAM_ERROR_POSITION_ARB, &errorpos);
+      printf("errorpos: %d\n", errorpos);
+      printf("%s\n", (char *)glGetString(GL_PROGRAM_ERROR_STRING_ARB));
+   }
+
+
+   glEnableClientState( GL_VERTEX_ARRAY );
+   glEnableClientState( GL_COLOR_ARRAY );
+
+   glGenBuffersARB(1, &arrayObj);
+   glBindBufferARB(GL_ARRAY_BUFFER_ARB, arrayObj);
+   glBufferDataARB(GL_ARRAY_BUFFER_ARB, sizeof(verts), verts, GL_STATIC_DRAW_ARB);
+
+   glVertexPointer( 4, GL_FLOAT, sizeof(verts[0]), 0 );
+   glColorPointer( 4, GL_FLOAT, sizeof(verts[0]), (void *)(4*sizeof(float)) );
+}
+
+
+
+static void Display( void )
+{
+   glClearColor(0.3, 0.3, 0.3, 1);
+   glClear( GL_COLOR_BUFFER_BIT );
+
+   glEnable(GL_VERTEX_PROGRAM_NV);
+   glDrawArrays( GL_TRIANGLES, 0, 3 );
+
+   glutSwapBuffers();
+}
+
+
+static void Reshape( int width, int height )
+{
+   glViewport( 0, 0, width, height );
+   glMatrixMode( GL_PROJECTION );
+   glLoadIdentity();
+   glOrtho(-1.0, 1.0, -1.0, 1.0, -0.5, 1000.0);
+   glMatrixMode( GL_MODELVIEW );
+   glLoadIdentity();
+   /*glTranslatef( 0.0, 0.0, -15.0 );*/
+}
+
+
+static void Key( unsigned char key, int x, int y )
+{
+   (void) x;
+   (void) y;
+   switch (key) {
+      case 27:
+         exit(0);
+         break;
+   }
+   glutPostRedisplay();
+}
+
+
+
+
+int main( int argc, char *argv[] )
+{
+   glutInit( &argc, argv );
+   glutInitWindowPosition( 0, 0 );
+   glutInitWindowSize( 250, 250 );
+   glutInitDisplayMode( GLUT_RGB | GLUT_DOUBLE );
+   glutCreateWindow(argv[0]);
+   glewInit();
+   glutReshapeFunc( Reshape );
+   glutKeyboardFunc( Key );
+   glutDisplayFunc( Display );
+   Init();
+   glutMainLoop();
+   return 0;
+}
diff --git a/progs/vp/add-param-imm.txt b/progs/vp/add-param-imm.txt
new file mode 100644 (file)
index 0000000..90bcf96
--- /dev/null
@@ -0,0 +1,7 @@
+!!ARBvp1.0
+TEMP R0;
+PARAM Emission = state.material.emission; 
+ADD R0, vertex.color, {-0.5}.x;
+ADD result.color, R0, Emission.w;
+MOV  result.position, vertex.position;
+END
index f4e82e8e0a58c046d3c36db99eb18426bc2372a6..1f9cd3d9203485ea45d6e82ee584397e2839281f 100644 (file)
@@ -156,6 +156,12 @@ def symlink(target, source, env):
         os.remove(target)
     os.symlink(os.path.basename(source), target)
 
+def install_program(env, source):
+    source = str(source[0])
+    target_dir =  os.path.join(env.Dir('#.').srcnode().abspath, env['build'], 'bin')
+    target_name = str(source)
+    env.InstallAs(os.path.join(target_dir, target_name), source)
+
 def install_shared_library(env, source, version = ()):
     source = str(source[0])
     version = tuple(map(str, version))
@@ -169,6 +175,7 @@ def install_shared_library(env, source, version = ()):
         last = env.Command(os.path.join(target_dir, target_name), last, action) 
 
 def createInstallMethods(env):
+    env.AddMethod(install_program, 'InstallProgram')
     env.AddMethod(install_shared_library, 'InstallSharedLibrary')
 
 
@@ -236,6 +243,8 @@ def generate(env):
     env['build'] = build_dir
     env.SConsignFile(os.path.join(build_dir, '.sconsign'))
     env.CacheDir('build/cache')
+    env['CONFIGUREDIR'] = os.path.join(build_dir, 'conf')
+    env['CONFIGURELOG'] = os.path.join(os.path.abspath(build_dir), 'config.log')
 
     # Parallel build
     if env.GetOption('num_jobs') <= 1:
@@ -370,7 +379,6 @@ def generate(env):
             '-Wno-long-long',
             '-ffast-math',
             '-fmessage-length=0', # be nice to Eclipse
-            '-fno-strict-aliasing', # we violate strict pointer aliasing rules
         ]
         cflags += [
             '-Werror=declaration-after-statement',
index 5440acd1351f57de3064b6890346f456d2fcf49f..6083fcbec98cc903930696e7a5572eedf0067f19 100644 (file)
@@ -1,5 +1,6 @@
 Import('*')
 
+SConscript('glsl/SConscript')
 SConscript('gallium/SConscript')
 
 if 'mesa' in env['statetrackers']:
index 4f13b3e2baddf4935c0e74c5bd3350a4e2902844..2b16332e14324e1fed46ff50e28e4a823bc557a4 100644 (file)
@@ -42,6 +42,7 @@
 #include "cso_cache/cso_context.h"
 #include "cso_cache/cso_cache.h"
 #include "cso_cache/cso_hash.h"
+#include "cso_context.h"
 
 struct cso_context {
    struct pipe_context *pipe;
@@ -50,28 +51,43 @@ struct cso_context {
    struct {
       void *samplers[PIPE_MAX_SAMPLERS];
       unsigned nr_samplers;
+
+      void *vertex_samplers[PIPE_MAX_VERTEX_SAMPLERS];
+      unsigned nr_vertex_samplers;
    } hw;
 
    void *samplers[PIPE_MAX_SAMPLERS];
    unsigned nr_samplers;
 
+   void *vertex_samplers[PIPE_MAX_VERTEX_SAMPLERS];
+   unsigned nr_vertex_samplers;
+
    unsigned nr_samplers_saved;
    void *samplers_saved[PIPE_MAX_SAMPLERS];
 
+   unsigned nr_vertex_samplers_saved;
+   void *vertex_samplers_saved[PIPE_MAX_VERTEX_SAMPLERS];
+
    struct pipe_texture *textures[PIPE_MAX_SAMPLERS];
    uint nr_textures;
 
+   struct pipe_texture *vertex_textures[PIPE_MAX_VERTEX_SAMPLERS];
+   uint nr_vertex_textures;
+
    uint nr_textures_saved;
    struct pipe_texture *textures_saved[PIPE_MAX_SAMPLERS];
 
+   uint nr_vertex_textures_saved;
+   struct pipe_texture *vertex_textures_saved[PIPE_MAX_SAMPLERS];
+
    /** Current and saved state.
     * The saved state is used as a 1-deep stack.
     */
    void *blend, *blend_saved;
    void *depth_stencil, *depth_stencil_saved;
    void *rasterizer, *rasterizer_saved;
-   void *fragment_shader, *fragment_shader_saved;
-   void *vertex_shader, *vertex_shader_saved;
+   void *fragment_shader, *fragment_shader_saved, *geometry_shader;
+   void *vertex_shader, *vertex_shader_saved, *geometry_shader_saved;
 
    struct pipe_framebuffer_state fb, fb_saved;
    struct pipe_viewport_state vp, vp_saved;
@@ -244,7 +260,9 @@ void cso_release_all( struct cso_context *ctx )
    if (ctx->pipe) {
       ctx->pipe->bind_blend_state( ctx->pipe, NULL );
       ctx->pipe->bind_rasterizer_state( ctx->pipe, NULL );
-      ctx->pipe->bind_sampler_states( ctx->pipe, 0, NULL );
+      ctx->pipe->bind_fragment_sampler_states( ctx->pipe, 0, NULL );
+      if (ctx->pipe->bind_vertex_sampler_states)
+         ctx->pipe->bind_vertex_sampler_states(ctx->pipe, 0, NULL);
       ctx->pipe->bind_depth_stencil_alpha_state( ctx->pipe, NULL );
       ctx->pipe->bind_fs_state( ctx->pipe, NULL );
       ctx->pipe->bind_vs_state( ctx->pipe, NULL );
@@ -255,6 +273,11 @@ void cso_release_all( struct cso_context *ctx )
       pipe_texture_reference(&ctx->textures_saved[i], NULL);
    }
 
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      pipe_texture_reference(&ctx->vertex_textures[i], NULL);
+      pipe_texture_reference(&ctx->vertex_textures_saved[i], NULL);
+   }
+
    free_framebuffer_state(&ctx->fb);
    free_framebuffer_state(&ctx->fb_saved);
 
@@ -378,6 +401,46 @@ enum pipe_error cso_single_sampler(struct cso_context *ctx,
    return PIPE_OK;
 }
 
+enum pipe_error
+cso_single_vertex_sampler(struct cso_context *ctx,
+                          unsigned idx,
+                          const struct pipe_sampler_state *templ)
+{
+   void *handle = NULL;
+
+   if (templ != NULL) {
+      unsigned hash_key = cso_construct_key((void*)templ, sizeof(struct pipe_sampler_state));
+      struct cso_hash_iter iter = cso_find_state_template(ctx->cache,
+                                                          hash_key, CSO_SAMPLER,
+                                                          (void*)templ);
+
+      if (cso_hash_iter_is_null(iter)) {
+         struct cso_sampler *cso = MALLOC(sizeof(struct cso_sampler));
+         if (!cso)
+            return PIPE_ERROR_OUT_OF_MEMORY;
+
+         memcpy(&cso->state, templ, sizeof(*templ));
+         cso->data = ctx->pipe->create_sampler_state(ctx->pipe, &cso->state);
+         cso->delete_state = (cso_state_callback)ctx->pipe->delete_sampler_state;
+         cso->context = ctx->pipe;
+
+         iter = cso_insert_state(ctx->cache, hash_key, CSO_SAMPLER, cso);
+         if (cso_hash_iter_is_null(iter)) {
+            FREE(cso);
+            return PIPE_ERROR_OUT_OF_MEMORY;
+         }
+
+         handle = cso->data;
+      }
+      else {
+         handle = ((struct cso_sampler *)cso_hash_iter_data(iter))->data;
+      }
+   }
+
+   ctx->vertex_samplers[idx] = handle;
+   return PIPE_OK;
+}
+
 void cso_single_sampler_done( struct cso_context *ctx )
 {
    unsigned i;
@@ -398,7 +461,36 @@ void cso_single_sampler_done( struct cso_context *ctx )
       memcpy(ctx->hw.samplers, ctx->samplers, ctx->nr_samplers * sizeof(void *));
       ctx->hw.nr_samplers = ctx->nr_samplers;
 
-      ctx->pipe->bind_sampler_states(ctx->pipe, ctx->nr_samplers, ctx->samplers);
+      ctx->pipe->bind_fragment_sampler_states(ctx->pipe, ctx->nr_samplers, ctx->samplers);
+   }
+}
+
+void
+cso_single_vertex_sampler_done(struct cso_context *ctx)
+{
+   unsigned i;
+
+   /* find highest non-null sampler */
+   for (i = PIPE_MAX_VERTEX_SAMPLERS; i > 0; i--) {
+      if (ctx->vertex_samplers[i - 1] != NULL)
+         break;
+   }
+
+   ctx->nr_vertex_samplers = i;
+
+   if (ctx->hw.nr_vertex_samplers != ctx->nr_vertex_samplers ||
+       memcmp(ctx->hw.vertex_samplers,
+              ctx->vertex_samplers,
+              ctx->nr_vertex_samplers * sizeof(void *)) != 0) 
+   {
+      memcpy(ctx->hw.vertex_samplers,
+             ctx->vertex_samplers,
+             ctx->nr_vertex_samplers * sizeof(void *));
+      ctx->hw.nr_vertex_samplers = ctx->nr_vertex_samplers;
+
+      ctx->pipe->bind_vertex_sampler_states(ctx->pipe,
+                                            ctx->nr_vertex_samplers,
+                                            ctx->vertex_samplers);
    }
 }
 
@@ -447,6 +539,21 @@ void cso_restore_samplers(struct cso_context *ctx)
    cso_single_sampler_done( ctx );
 }
 
+void
+cso_save_vertex_samplers(struct cso_context *ctx)
+{
+   ctx->nr_vertex_samplers_saved = ctx->nr_vertex_samplers;
+   memcpy(ctx->vertex_samplers_saved, ctx->vertex_samplers, sizeof(ctx->vertex_samplers));
+}
+
+void
+cso_restore_vertex_samplers(struct cso_context *ctx)
+{
+   ctx->nr_vertex_samplers = ctx->nr_vertex_samplers_saved;
+   memcpy(ctx->vertex_samplers, ctx->vertex_samplers_saved, sizeof(ctx->vertex_samplers));
+   cso_single_vertex_sampler_done(ctx);
+}
+
 
 enum pipe_error cso_set_sampler_textures( struct cso_context *ctx,
                                           uint count,
@@ -461,7 +568,7 @@ enum pipe_error cso_set_sampler_textures( struct cso_context *ctx,
    for ( ; i < PIPE_MAX_SAMPLERS; i++)
       pipe_texture_reference(&ctx->textures[i], NULL);
 
-   ctx->pipe->set_sampler_textures(ctx->pipe, count, textures);
+   ctx->pipe->set_fragment_sampler_textures(ctx->pipe, count, textures);
 
    return PIPE_OK;
 }
@@ -491,13 +598,71 @@ void cso_restore_sampler_textures( struct cso_context *ctx )
    for ( ; i < PIPE_MAX_SAMPLERS; i++)
       pipe_texture_reference(&ctx->textures[i], NULL);
 
-   ctx->pipe->set_sampler_textures(ctx->pipe, ctx->nr_textures, ctx->textures);
+   ctx->pipe->set_fragment_sampler_textures(ctx->pipe, ctx->nr_textures, ctx->textures);
 
    ctx->nr_textures_saved = 0;
 }
 
 
 
+enum pipe_error
+cso_set_vertex_sampler_textures(struct cso_context *ctx,
+                                uint count,
+                                struct pipe_texture **textures)
+{
+   uint i;
+
+   ctx->nr_vertex_textures = count;
+
+   for (i = 0; i < count; i++) {
+      pipe_texture_reference(&ctx->vertex_textures[i], textures[i]);
+   }
+   for ( ; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      pipe_texture_reference(&ctx->vertex_textures[i], NULL);
+   }
+
+   ctx->pipe->set_vertex_sampler_textures(ctx->pipe, count, textures);
+
+   return PIPE_OK;
+}
+
+void
+cso_save_vertex_sampler_textures(struct cso_context *ctx)
+{
+   uint i;
+
+   ctx->nr_vertex_textures_saved = ctx->nr_vertex_textures;
+   for (i = 0; i < ctx->nr_vertex_textures; i++) {
+      assert(!ctx->vertex_textures_saved[i]);
+      pipe_texture_reference(&ctx->vertex_textures_saved[i], ctx->vertex_textures[i]);
+   }
+}
+
+void
+cso_restore_vertex_sampler_textures(struct cso_context *ctx)
+{
+   uint i;
+
+   ctx->nr_vertex_textures = ctx->nr_vertex_textures_saved;
+
+   for (i = 0; i < ctx->nr_vertex_textures; i++) {
+      pipe_texture_reference(&ctx->vertex_textures[i], NULL);
+      ctx->vertex_textures[i] = ctx->vertex_textures_saved[i];
+      ctx->vertex_textures_saved[i] = NULL;
+   }
+   for ( ; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      pipe_texture_reference(&ctx->vertex_textures[i], NULL);
+   }
+
+   ctx->pipe->set_vertex_sampler_textures(ctx->pipe,
+                                          ctx->nr_vertex_textures,
+                                          ctx->vertex_textures);
+
+   ctx->nr_vertex_textures_saved = 0;
+}
+
+
+
 enum pipe_error cso_set_depth_stencil_alpha(struct cso_context *ctx,
                                             const struct pipe_depth_stencil_alpha_state *templ)
 {
@@ -863,3 +1028,38 @@ enum pipe_error cso_set_blend_color(struct cso_context *ctx,
    }
    return PIPE_OK;
 }
+
+enum pipe_error cso_set_geometry_shader_handle(struct cso_context *ctx,
+                                               void *handle)
+{
+   if (ctx->geometry_shader != handle) {
+      ctx->geometry_shader = handle;
+      ctx->pipe->bind_gs_state(ctx->pipe, handle);
+   }
+   return PIPE_OK;
+}
+
+void cso_delete_geometry_shader(struct cso_context *ctx, void *handle)
+{
+    if (handle == ctx->geometry_shader) {
+      /* unbind before deleting */
+      ctx->pipe->bind_gs_state(ctx->pipe, NULL);
+      ctx->geometry_shader = NULL;
+   }
+   ctx->pipe->delete_gs_state(ctx->pipe, handle);
+}
+
+void cso_save_geometry_shader(struct cso_context *ctx)
+{
+   assert(!ctx->geometry_shader_saved);
+   ctx->geometry_shader_saved = ctx->geometry_shader;
+}
+
+void cso_restore_geometry_shader(struct cso_context *ctx)
+{
+   if (ctx->geometry_shader_saved != ctx->geometry_shader) {
+      ctx->pipe->bind_gs_state(ctx->pipe, ctx->geometry_shader_saved);
+      ctx->geometry_shader = ctx->geometry_shader_saved;
+   }
+   ctx->geometry_shader_saved = NULL;
+}
index 69630e98bae151073a21e3f8c5a4c0784e04c4a0..b9e313e32d6cff726c975a6469d862b16b6836b9 100644 (file)
@@ -84,6 +84,20 @@ enum pipe_error cso_single_sampler( struct cso_context *cso,
 
 void cso_single_sampler_done( struct cso_context *cso );
 
+void
+cso_save_vertex_samplers(struct cso_context *cso);
+
+void
+cso_restore_vertex_samplers(struct cso_context *cso);
+
+enum pipe_error
+cso_single_vertex_sampler(struct cso_context *cso,
+                          unsigned nr,
+                          const struct pipe_sampler_state *states);
+
+void
+cso_single_vertex_sampler_done(struct cso_context *cso);
+
 
 
 enum pipe_error cso_set_sampler_textures( struct cso_context *cso,
@@ -94,6 +108,17 @@ void cso_restore_sampler_textures( struct cso_context *cso );
 
 
 
+enum pipe_error
+cso_set_vertex_sampler_textures(struct cso_context *cso,
+                                uint count,
+                                struct pipe_texture **textures);
+void
+cso_save_vertex_sampler_textures(struct cso_context *cso);
+void
+cso_restore_vertex_sampler_textures(struct cso_context *cso);
+
+
+
 /* These aren't really sensible -- most of the time the api provides
  * object semantics for shaders anyway, and the cases where it doesn't
  * (eg mesa's internall-generated texenv programs), it will be up to
@@ -121,6 +146,13 @@ void cso_save_vertex_shader(struct cso_context *cso);
 void cso_restore_vertex_shader(struct cso_context *cso);
 
 
+enum pipe_error cso_set_geometry_shader_handle(struct cso_context *ctx,
+                                               void *handle);
+void cso_delete_geometry_shader(struct cso_context *ctx, void *handle);
+void cso_save_geometry_shader(struct cso_context *cso);
+void cso_restore_geometry_shader(struct cso_context *cso);
+
+
 
 enum pipe_error cso_set_framebuffer(struct cso_context *cso,
                                     const struct pipe_framebuffer_state *fb);
index 5041dcc072bfd7a2e350af91254447e977a8cc7d..248167465fe9f11a1f81b0f6267c929fa3b64c1e 100644 (file)
@@ -5,6 +5,7 @@ LIBNAME = draw
 
 C_SOURCES = \
        draw_context.c \
+        draw_gs.c \
        draw_pipe.c \
        draw_pipe_aaline.c \
        draw_pipe_aapoint.c \
index 5f05aa324a5bcf7099f1ce34974c8e2b71786e6e..a022c145e9076db5a2f5a6f677d9c837a810e4bb 100644 (file)
@@ -40,7 +40,8 @@ draw = env.ConvenienceLibrary(
                'draw_vs_llvm.c',
                'draw_vs_ppc.c',
                'draw_vs_sse.c',
-               'draw_vs_varient.c'
+               'draw_vs_varient.c',
+                'draw_gs.c'
        ])
 
 auxiliaries.insert(0, draw)
index a4f1fcddc1a0cd6f9de25be4bd0c21c791f5da4e..667aa46b208ce3f4bc90fcc56b0d0087c053ff13 100644 (file)
@@ -36,6 +36,7 @@
 #include "draw_context.h"
 #include "draw_vbuf.h"
 #include "draw_vs.h"
+#include "draw_gs.h"
 #include "draw_pt.h"
 #include "draw_pipe.h"
 
@@ -67,6 +68,9 @@ struct draw_context *draw_create( void )
    if (!draw_vs_init( draw ))
       goto fail;
 
+   if (!draw_gs_init( draw ))
+      goto fail;
+
    return draw;
 
 fail:
@@ -231,11 +235,19 @@ draw_set_mapped_vertex_buffer(struct draw_context *draw,
 
 void
 draw_set_mapped_constant_buffer(struct draw_context *draw,
-                                const void *buffer, 
+                                unsigned shader_type,
+                                const void *buffer,
                                 unsigned size )
 {
-   draw->pt.user.constants = buffer;
-   draw_vs_set_constants( draw, (const float (*)[4])buffer, size );
+   debug_assert(shader_type == PIPE_SHADER_VERTEX ||
+                shader_type == PIPE_SHADER_GEOMETRY);
+   if (shader_type == PIPE_SHADER_VERTEX) {
+      draw->pt.user.vs_constants = buffer;
+      draw_vs_set_constants( draw, (const float (*)[4])buffer, size );
+   } else if (shader_type == PIPE_SHADER_GEOMETRY) {
+      draw->pt.user.gs_constants = buffer;
+      draw_gs_set_constants( draw, (const float (*)[4])buffer, size );
+   }
 }
 
 
@@ -298,7 +310,7 @@ draw_set_force_passthrough( struct draw_context *draw, boolean enable )
  * a post-transformed vertex.
  *
  * With this function, drivers that use the draw module should have no reason
- * to track the current vertex shader.
+ * to track the current vertex/geometry shader.
  *
  * Note that the draw module may sometimes generate vertices with extra
  * attributes (such as texcoords for AA lines).  The driver can call this
@@ -309,43 +321,59 @@ draw_set_force_passthrough( struct draw_context *draw, boolean enable )
  * work for the drivers.
  */
 int
-draw_find_vs_output(const struct draw_context *draw,
-                    uint semantic_name, uint semantic_index)
+draw_find_shader_output(const struct draw_context *draw,
+                        uint semantic_name, uint semantic_index)
 {
    const struct draw_vertex_shader *vs = draw->vs.vertex_shader;
+   const struct draw_geometry_shader *gs = draw->gs.geometry_shader;
    uint i;
-   for (i = 0; i < vs->info.num_outputs; i++) {
-      if (vs->info.output_semantic_name[i] == semantic_name &&
-          vs->info.output_semantic_index[i] == semantic_index)
+   const struct tgsi_shader_info *info = &vs->info;
+
+   if (gs)
+      info = &gs->info;
+
+   for (i = 0; i < info->num_outputs; i++) {
+      if (info->output_semantic_name[i] == semantic_name &&
+          info->output_semantic_index[i] == semantic_index)
          return i;
    }
 
    /* XXX there may be more than one extra vertex attrib.
     * For example, simulated gl_FragCoord and gl_PointCoord.
     */
-   if (draw->extra_vp_outputs.semantic_name == semantic_name &&
-       draw->extra_vp_outputs.semantic_index == semantic_index) {
-      return draw->extra_vp_outputs.slot;
+   if (draw->extra_shader_outputs.semantic_name == semantic_name &&
+       draw->extra_shader_outputs.semantic_index == semantic_index) {
+      return draw->extra_shader_outputs.slot;
    }
+
    return 0;
 }
 
 
 /**
- * Return number of vertex shader outputs.
+ * Return number of the shader outputs.
+ *
+ * If geometry shader is present, its output will be returned,
+ * if not vertex shader is used.
  */
 uint
-draw_num_vs_outputs(const struct draw_context *draw)
+draw_num_shader_outputs(const struct draw_context *draw)
 {
    uint count = draw->vs.vertex_shader->info.num_outputs;
-   if (draw->extra_vp_outputs.slot > 0)
+
+   /* if geometry shader is present, its outputs go to te
+    * driver, not the vertex shaders */
+   if (draw->gs.geometry_shader)
+      count = draw->gs.geometry_shader->info.num_outputs;
+
+   if (draw->extra_shader_outputs.slot > 0)
       count++;
    return count;
 }
 
 
 /**
- * Provide TGSI sampler objects for vertex shaders that use texture fetches.
+ * Provide TGSI sampler objects for vertex/geometry shaders that use texture fetches.
  * This might only be used by software drivers for the time being.
  */
 void
@@ -355,6 +383,8 @@ draw_texture_samplers(struct draw_context *draw,
 {
    draw->vs.num_samplers = num_samplers;
    draw->vs.samplers = samplers;
+   draw->gs.num_samplers = num_samplers;
+   draw->gs.samplers = samplers;
 }
 
 
@@ -366,13 +396,6 @@ void draw_set_render( struct draw_context *draw,
    draw->render = render;
 }
 
-void draw_set_edgeflags( struct draw_context *draw,
-                         const unsigned *edgeflag )
-{
-   draw->pt.user.edgeflag = edgeflag;
-}
-
-
 
 
 /**
@@ -428,3 +451,18 @@ void draw_do_flush( struct draw_context *draw, unsigned flags )
       draw->flushing = FALSE;
    }
 }
+
+
+int draw_current_shader_outputs(struct draw_context *draw)
+{
+   if (draw->gs.geometry_shader)
+      return draw->gs.num_gs_outputs;
+   return draw->vs.num_vs_outputs;
+}
+
+int draw_current_shader_position_output(struct draw_context *draw)
+{
+   if (draw->gs.geometry_shader)
+      return draw->gs.position_output;
+   return draw->vs.position_output;
+}
index d529e4e9a27f7e3f35c7eb24b0732f6fff1cc722..b716209df299019b0c309ff94730507b235522c0 100644 (file)
@@ -45,6 +45,7 @@ struct pipe_context;
 struct draw_context;
 struct draw_stage;
 struct draw_vertex_shader;
+struct draw_geometry_shader;
 struct tgsi_sampler;
 
 
@@ -85,11 +86,11 @@ draw_install_pstipple_stage(struct draw_context *draw, struct pipe_context *pipe
 
 
 int
-draw_find_vs_output(const struct draw_context *draw,
-                    uint semantic_name, uint semantic_index);
+draw_find_shader_output(const struct draw_context *draw,
+                        uint semantic_name, uint semantic_index);
 
 uint
-draw_num_vs_outputs(const struct draw_context *draw);
+draw_num_shader_outputs(const struct draw_context *draw);
 
 
 void
@@ -112,6 +113,17 @@ void draw_delete_vertex_shader(struct draw_context *draw,
                                struct draw_vertex_shader *dvs);
 
 
+/*
+ * Geometry shader functions
+ */
+struct draw_geometry_shader *
+draw_create_geometry_shader(struct draw_context *draw,
+                            const struct pipe_shader_state *shader);
+void draw_bind_geometry_shader(struct draw_context *draw,
+                               struct draw_geometry_shader *dvs);
+void draw_delete_geometry_shader(struct draw_context *draw,
+                                 struct draw_geometry_shader *dvs);
+
 
 /*
  * Vertex data functions
@@ -140,12 +152,10 @@ void draw_set_mapped_vertex_buffer(struct draw_context *draw,
                                    unsigned attr, const void *buffer);
 
 void draw_set_mapped_constant_buffer(struct draw_context *draw,
+                                     unsigned shader_type,
                                      const void *buffer,
                                      unsigned size );
 
-void draw_set_edgeflags( struct draw_context *draw,
-                         const unsigned *edgeflag );
-
 
 /***********************************************************************
  * draw_prim.c 
diff --git a/src/gallium/auxiliary/draw/draw_gs.c b/src/gallium/auxiliary/draw/draw_gs.c
new file mode 100644 (file)
index 0000000..5db2e75
--- /dev/null
@@ -0,0 +1,338 @@
+/**************************************************************************
+ *
+ * Copyright 2009 VMWare Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+#include "draw_gs.h"
+
+#include "draw_private.h"
+#include "draw_context.h"
+
+#include "tgsi/tgsi_parse.h"
+#include "tgsi/tgsi_exec.h"
+
+#include "pipe/p_shader_tokens.h"
+
+#include "util/u_math.h"
+#include "util/u_memory.h"
+
+#define MAX_PRIM_VERTICES 6
+/* fixme: move it from here */
+#define MAX_PRIMITIVES 64
+
+boolean
+draw_gs_init( struct draw_context *draw )
+{
+   draw->gs.machine = tgsi_exec_machine_create();
+   if (!draw->gs.machine)
+      return FALSE;
+
+   draw->gs.machine->Primitives = align_malloc(
+      MAX_PRIMITIVES * sizeof(struct tgsi_exec_vector), 16);
+   if (!draw->gs.machine->Primitives)
+      return FALSE;
+   memset(draw->gs.machine->Primitives, 0,
+          MAX_PRIMITIVES * sizeof(struct tgsi_exec_vector));
+
+   return TRUE;
+}
+
+
+void draw_gs_set_constants( struct draw_context *draw,
+                            const float (*constants)[4],
+                            unsigned size )
+{
+}
+
+
+struct draw_geometry_shader *
+draw_create_geometry_shader(struct draw_context *draw,
+                            const struct pipe_shader_state *state)
+{
+   struct draw_geometry_shader *gs;
+   int i;
+
+   gs = CALLOC_STRUCT(draw_geometry_shader);
+
+   if (!gs)
+      return NULL;
+
+   gs->state = *state;
+   gs->state.tokens = tgsi_dup_tokens(state->tokens);
+   if (!gs->state.tokens) {
+      FREE(gs);
+      return NULL;
+   }
+
+   tgsi_scan_shader(state->tokens, &gs->info);
+
+   /* setup the defaults */
+   gs->input_primitive = PIPE_PRIM_TRIANGLES;
+   gs->output_primitive = PIPE_PRIM_TRIANGLE_STRIP;
+   gs->max_output_vertices = 32;
+
+   for (i = 0; i < gs->info.num_properties; ++i) {
+      if (gs->info.properties[i].name ==
+          TGSI_PROPERTY_GS_INPUT_PRIM)
+         gs->input_primitive = gs->info.properties[i].data[0];
+      else if (gs->info.properties[i].name ==
+               TGSI_PROPERTY_GS_OUTPUT_PRIM)
+         gs->output_primitive = gs->info.properties[i].data[0];
+      else if (gs->info.properties[i].name ==
+               TGSI_PROPERTY_GS_MAX_VERTICES)
+         gs->max_output_vertices = gs->info.properties[i].data[0];
+   }
+
+   gs->machine = draw->gs.machine;
+
+   if (gs)
+   {
+      uint i;
+      for (i = 0; i < gs->info.num_outputs; i++) {
+         if (gs->info.output_semantic_name[i] == TGSI_SEMANTIC_POSITION &&
+             gs->info.output_semantic_index[i] == 0)
+            gs->position_output = i;
+      }
+   }
+
+   return gs;
+}
+
+void draw_bind_geometry_shader(struct draw_context *draw,
+                               struct draw_geometry_shader *dgs)
+{
+   draw_do_flush(draw, DRAW_FLUSH_STATE_CHANGE);
+
+   if (dgs) {
+      draw->gs.geometry_shader = dgs;
+      draw->gs.num_gs_outputs = dgs->info.num_outputs;
+      draw->gs.position_output = dgs->position_output;
+      draw_geometry_shader_prepare(dgs, draw);
+   }
+   else {
+      draw->gs.geometry_shader = NULL;
+      draw->gs.num_gs_outputs = 0;
+   }
+}
+
+void draw_delete_geometry_shader(struct draw_context *draw,
+                                 struct draw_geometry_shader *dgs)
+{
+   FREE(dgs);
+}
+
+static INLINE int num_vertices_for_prim(int prim)
+{
+   switch(prim) {
+   case PIPE_PRIM_POINTS:
+      return 1;
+   case PIPE_PRIM_LINES:
+      return 2;
+   case PIPE_PRIM_LINE_LOOP:
+      return 2;
+   case PIPE_PRIM_LINE_STRIP:
+      return 2;
+   case PIPE_PRIM_TRIANGLES:
+      return 3;
+   case PIPE_PRIM_TRIANGLE_STRIP:
+      return 3;
+   case PIPE_PRIM_TRIANGLE_FAN:
+      return 3;
+   case PIPE_PRIM_LINES_ADJACENCY:
+   case PIPE_PRIM_LINE_STRIP_ADJACENCY:
+      return 4;
+   case PIPE_PRIM_TRIANGLES_ADJACENCY:
+   case PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY:
+      return 6;
+   default:
+      assert(!"Bad geometry shader input");
+      return 0;
+   }
+}
+
+static void draw_fetch_geometry_input(struct draw_geometry_shader *shader,
+                                      int start_primitive,
+                                      int num_primitives,
+                                      const float (*input_ptr)[4],
+                                      unsigned input_vertex_stride,
+                                      unsigned inputs_from_vs)
+{
+   struct tgsi_exec_machine *machine = shader->machine;
+   unsigned slot, vs_slot, k, j;
+   unsigned num_vertices = num_vertices_for_prim(shader->input_primitive);
+   int idx = 0;
+
+   for (slot = 0, vs_slot = 0; slot < shader->info.num_inputs; slot++) {
+      /*debug_printf("Slot = %d (semantic = %d)\n", slot,
+        shader->info.input_semantic_name[slot]);*/
+      if (shader->info.input_semantic_name[slot] ==
+          TGSI_SEMANTIC_PRIMID) {
+         for (j = 0; j < num_primitives; ++j) {
+            machine->Inputs[idx].xyzw[0].f[j] = (float)start_primitive + j;
+            machine->Inputs[idx].xyzw[1].f[j] = (float)start_primitive + j;
+            machine->Inputs[idx].xyzw[2].f[j] = (float)start_primitive + j;
+            machine->Inputs[idx].xyzw[3].f[j] = (float)start_primitive + j;
+         }
+         ++idx;
+      } else {
+         for (j = 0; j < num_primitives; ++j) {
+            int vidx = idx;
+            const float (*prim_ptr)[4];
+            /*debug_printf("    %d) Prim (num_verts = %d)\n", start_primitive + j,
+              num_vertices);*/
+            prim_ptr = (const float (*)[4])(
+               (const char *)input_ptr +
+               (j * num_vertices * input_vertex_stride));
+
+            for (k = 0; k < num_vertices; ++k, ++vidx) {
+               const float (*input)[4];
+               input = (const float (*)[4])(
+                  (const char *)prim_ptr + (k * input_vertex_stride));
+               vidx = k * TGSI_EXEC_MAX_INPUT_ATTRIBS + slot;
+               /*debug_printf("\t%d)(%d) Input vert:\n", vidx, k);*/
+#if 1
+               assert(!util_is_inf_or_nan(input[vs_slot][0]));
+               assert(!util_is_inf_or_nan(input[vs_slot][1]));
+               assert(!util_is_inf_or_nan(input[vs_slot][2]));
+               assert(!util_is_inf_or_nan(input[vs_slot][3]));
+#endif
+               machine->Inputs[vidx].xyzw[0].f[j] = input[vs_slot][0];
+               machine->Inputs[vidx].xyzw[1].f[j] = input[vs_slot][1];
+               machine->Inputs[vidx].xyzw[2].f[j] = input[vs_slot][2];
+               machine->Inputs[vidx].xyzw[3].f[j] = input[vs_slot][3];
+#if 0
+               debug_printf("\t\t%d %f %f %f %f\n", slot,
+                            machine->Inputs[vidx].xyzw[0].f[j],
+                            machine->Inputs[vidx].xyzw[1].f[j],
+                            machine->Inputs[vidx].xyzw[2].f[j],
+                            machine->Inputs[vidx].xyzw[3].f[j]);
+#endif
+            }
+         }
+         ++vs_slot;
+         idx += num_vertices;
+      }
+   }
+}
+
+static INLINE void
+draw_geometry_fetch_outputs(struct draw_geometry_shader *shader,
+                            int num_primitives,
+                            float (*output)[4],
+                            unsigned vertex_size)
+{
+   struct tgsi_exec_machine *machine = shader->machine;
+   unsigned prim_idx, j, slot;
+
+   /* Unswizzle all output results.
+    */
+   /* FIXME: handle all the primitives produced by the gs, not just
+    * the first one
+    unsigned prim_count =
+    mach->Temps[TEMP_PRIMITIVE_I].xyzw[TEMP_PRIMITIVE_C].u[0];*/
+   for (prim_idx = 0; prim_idx < num_primitives; ++prim_idx) {
+      unsigned num_verts_per_prim = machine->Primitives[0];
+      for (j = 0; j < num_verts_per_prim; j++) {
+         int idx = (prim_idx * num_verts_per_prim + j) *
+                   shader->info.num_outputs;
+#ifdef DEBUG_OUTPUTS
+         debug_printf("%d) Output vert:\n", idx);
+#endif
+         for (slot = 0; slot < shader->info.num_outputs; slot++) {
+            output[slot][0] = machine->Outputs[idx + slot].xyzw[0].f[prim_idx];
+            output[slot][1] = machine->Outputs[idx + slot].xyzw[1].f[prim_idx];
+            output[slot][2] = machine->Outputs[idx + slot].xyzw[2].f[prim_idx];
+            output[slot][3] = machine->Outputs[idx + slot].xyzw[3].f[prim_idx];
+#ifdef DEBUG_OUTPUTS
+            debug_printf("\t%d: %f %f %f %f\n", slot,
+                         output[slot][0],
+                         output[slot][1],
+                         output[slot][2],
+                         output[slot][3]);
+#endif
+            debug_assert(!util_is_inf_or_nan(output[slot][0]));
+         }
+         output = (float (*)[4])((char *)output + vertex_size);
+      }
+   }
+}
+
+void draw_geometry_shader_run(struct draw_geometry_shader *shader,
+                              const float (*input)[4],
+                              float (*output)[4],
+                              const float (*constants)[4],
+                              unsigned count,
+                              unsigned input_stride,
+                              unsigned vertex_size)
+{
+   struct tgsi_exec_machine *machine = shader->machine;
+   unsigned int i;
+   unsigned num_vertices = num_vertices_for_prim(shader->input_primitive);
+   unsigned num_primitives = count/num_vertices;
+   unsigned inputs_from_vs = 0;
+
+   machine->Consts = constants;
+
+   for (i = 0; i < shader->info.num_inputs; ++i) {
+      if (shader->info.input_semantic_name[i] != TGSI_SEMANTIC_PRIMID)
+         ++inputs_from_vs;
+   }
+
+   for (i = 0; i < num_primitives; ++i) {
+      unsigned int max_primitives = 1;
+
+      draw_fetch_geometry_input(shader, i, max_primitives, input,
+                                input_stride, inputs_from_vs);
+
+      tgsi_set_exec_mask(machine,
+                         1,
+                         max_primitives > 1,
+                         max_primitives > 2,
+                         max_primitives > 3);
+
+      /* run interpreter */
+      tgsi_exec_machine_run(machine);
+
+      draw_geometry_fetch_outputs(shader, max_primitives,
+                                  output, vertex_size);
+   }
+}
+
+void draw_geometry_shader_delete(struct draw_geometry_shader *shader)
+{
+   FREE((void*) shader->state.tokens);
+   FREE(shader);
+}
+
+void draw_geometry_shader_prepare(struct draw_geometry_shader *shader,
+                                  struct draw_context *draw)
+{
+    if (shader->machine->Tokens != shader->state.tokens) {
+       tgsi_exec_machine_bind_shader(shader->machine,
+                                     shader->state.tokens,
+                                     draw->gs.num_samplers,
+                                     draw->gs.samplers);
+    }
+}
diff --git a/src/gallium/auxiliary/draw/draw_gs.h b/src/gallium/auxiliary/draw/draw_gs.h
new file mode 100644 (file)
index 0000000..d6a97d9
--- /dev/null
@@ -0,0 +1,76 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMWare Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef DRAW_GS_H
+#define DRAW_GS_H
+
+#include "draw_context.h"
+#include "draw_private.h"
+
+
+#define MAX_TGSI_PRIMITIVES 4
+
+struct draw_context;
+
+/**
+ * Private version of the compiled geometry shader
+ */
+struct draw_geometry_shader {
+   struct draw_context *draw;
+
+   struct tgsi_exec_machine *machine;
+
+   /* This member will disappear shortly:*/
+   struct pipe_shader_state state;
+
+   struct tgsi_shader_info info;
+   unsigned position_output;
+
+   unsigned max_output_vertices;
+   unsigned input_primitive;
+   unsigned output_primitive;
+
+   /* Extracted from shader:
+    */
+   const float (*immediates)[4];
+};
+
+void draw_geometry_shader_run(struct draw_geometry_shader *shader,
+                              const float (*input)[4],
+                              float (*output)[4],
+                              const float (*constants)[4],
+                              unsigned count,
+                              unsigned input_stride,
+                              unsigned output_stride);
+
+void draw_geometry_shader_prepare(struct draw_geometry_shader *shader,
+                                  struct draw_context *draw);
+
+void draw_geometry_shader_delete(struct draw_geometry_shader *shader);
+
+
+#endif
index 9f956715a228135df0459358ef0ff8242cf789e3..4585dcdb48ab8f39dc911d868e0d4fb5a67db27f 100644 (file)
@@ -35,6 +35,8 @@
 #include "pipe/p_context.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_shader_tokens.h"
+
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -139,29 +141,29 @@ aa_transform_decl(struct tgsi_transform_context *ctx,
    struct aa_transform_context *aactx = (struct aa_transform_context *) ctx;
 
    if (decl->Declaration.File == TGSI_FILE_OUTPUT &&
-       decl->Semantic.SemanticName == TGSI_SEMANTIC_COLOR &&
-       decl->Semantic.SemanticIndex == 0) {
-      aactx->colorOutput = decl->DeclarationRange.First;
+       decl->Semantic.Name == TGSI_SEMANTIC_COLOR &&
+       decl->Semantic.Index == 0) {
+      aactx->colorOutput = decl->Range.First;
    }
    else if (decl->Declaration.File == TGSI_FILE_SAMPLER) {
       uint i;
-      for (i = decl->DeclarationRange.First;
-           i <= decl->DeclarationRange.Last; i++) {
+      for (i = decl->Range.First;
+           i <= decl->Range.Last; i++) {
          aactx->samplersUsed |= 1 << i;
       }
    }
    else if (decl->Declaration.File == TGSI_FILE_INPUT) {
-      if ((int) decl->DeclarationRange.Last > aactx->maxInput)
-         aactx->maxInput = decl->DeclarationRange.Last;
-      if (decl->Semantic.SemanticName == TGSI_SEMANTIC_GENERIC &&
-           (int) decl->Semantic.SemanticIndex > aactx->maxGeneric) {
-         aactx->maxGeneric = decl->Semantic.SemanticIndex;
+      if ((int) decl->Range.Last > aactx->maxInput)
+         aactx->maxInput = decl->Range.Last;
+      if (decl->Semantic.Name == TGSI_SEMANTIC_GENERIC &&
+           (int) decl->Semantic.Index > aactx->maxGeneric) {
+         aactx->maxGeneric = decl->Semantic.Index;
       }
    }
    else if (decl->Declaration.File == TGSI_FILE_TEMPORARY) {
       uint i;
-      for (i = decl->DeclarationRange.First;
-           i <= decl->DeclarationRange.Last; i++) {
+      for (i = decl->Range.First;
+           i <= decl->Range.Last; i++) {
          aactx->tempsUsed |= (1 << i);
       }
    }
@@ -228,30 +230,30 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       /* XXX this could be linear... */
       decl.Declaration.Interpolate = TGSI_INTERPOLATE_PERSPECTIVE;
       decl.Declaration.Semantic = 1;
-      decl.Semantic.SemanticName = TGSI_SEMANTIC_GENERIC;
-      decl.Semantic.SemanticIndex = aactx->maxGeneric + 1;
-      decl.DeclarationRange.First = 
-      decl.DeclarationRange.Last = aactx->maxInput + 1;
+      decl.Semantic.Name = TGSI_SEMANTIC_GENERIC;
+      decl.Semantic.Index = aactx->maxGeneric + 1;
+      decl.Range.First = 
+      decl.Range.Last = aactx->maxInput + 1;
       ctx->emit_declaration(ctx, &decl);
 
       /* declare new sampler */
       decl = tgsi_default_full_declaration();
       decl.Declaration.File = TGSI_FILE_SAMPLER;
-      decl.DeclarationRange.First = 
-      decl.DeclarationRange.Last = aactx->freeSampler;
+      decl.Range.First = 
+      decl.Range.Last = aactx->freeSampler;
       ctx->emit_declaration(ctx, &decl);
 
       /* declare new temp regs */
       decl = tgsi_default_full_declaration();
       decl.Declaration.File = TGSI_FILE_TEMPORARY;
-      decl.DeclarationRange.First = 
-      decl.DeclarationRange.Last = aactx->texTemp;
+      decl.Range.First = 
+      decl.Range.Last = aactx->texTemp;
       ctx->emit_declaration(ctx, &decl);
 
       decl = tgsi_default_full_declaration();
       decl.Declaration.File = TGSI_FILE_TEMPORARY;
-      decl.DeclarationRange.First = 
-      decl.DeclarationRange.Last = aactx->colorTemp;
+      decl.Range.First = 
+      decl.Range.Last = aactx->colorTemp;
       ctx->emit_declaration(ctx, &decl);
 
       aactx->firstInstruction = FALSE;
@@ -265,14 +267,15 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_TEX;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = aactx->texTemp;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = aactx->texTemp;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.InstructionExtTexture.Texture = TGSI_TEXTURE_2D;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_INPUT;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = aactx->maxInput + 1;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_SAMPLER;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = aactx->freeSampler;
+      newInst.Instruction.Texture = TRUE;
+      newInst.Texture.Texture = TGSI_TEXTURE_2D;
+      newInst.Src[0].Register.File = TGSI_FILE_INPUT;
+      newInst.Src[0].Register.Index = aactx->maxInput + 1;
+      newInst.Src[1].Register.File = TGSI_FILE_SAMPLER;
+      newInst.Src[1].Register.Index = aactx->freeSampler;
 
       ctx->emit_instruction(ctx, &newInst);
 
@@ -280,26 +283,26 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_MOV;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_OUTPUT;
-      newInst.FullDstRegisters[0].DstRegister.Index = aactx->colorOutput;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_XYZ;
+      newInst.Dst[0].Register.File = TGSI_FILE_OUTPUT;
+      newInst.Dst[0].Register.Index = aactx->colorOutput;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_XYZ;
       newInst.Instruction.NumSrcRegs = 1;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = aactx->colorTemp;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = aactx->colorTemp;
       ctx->emit_instruction(ctx, &newInst);
 
       /* MUL alpha */
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_MUL;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_OUTPUT;
-      newInst.FullDstRegisters[0].DstRegister.Index = aactx->colorOutput;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_W;
+      newInst.Dst[0].Register.File = TGSI_FILE_OUTPUT;
+      newInst.Dst[0].Register.Index = aactx->colorOutput;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_W;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = aactx->colorTemp;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = aactx->texTemp;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = aactx->colorTemp;
+      newInst.Src[1].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[1].Register.Index = aactx->texTemp;
       ctx->emit_instruction(ctx, &newInst);
 
       /* END */
@@ -316,11 +319,11 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       uint i;
 
       for (i = 0; i < inst->Instruction.NumDstRegs; i++) {
-         struct tgsi_full_dst_register *dst = &inst->FullDstRegisters[i];
-         if (dst->DstRegister.File == TGSI_FILE_OUTPUT &&
-             dst->DstRegister.Index == aactx->colorOutput) {
-            dst->DstRegister.File = TGSI_FILE_TEMPORARY;
-            dst->DstRegister.Index = aactx->colorTemp;
+         struct tgsi_full_dst_register *dst = &inst->Dst[i];
+         if (dst->Register.File == TGSI_FILE_OUTPUT &&
+             dst->Register.Index == aactx->colorOutput) {
+            dst->Register.File = TGSI_FILE_TEMPORARY;
+            dst->Register.Index = aactx->colorTemp;
          }
       }
 
@@ -398,10 +401,9 @@ aaline_create_texture(struct aaline_stage *aaline)
    texTemp.target = PIPE_TEXTURE_2D;
    texTemp.format = PIPE_FORMAT_A8_UNORM; /* XXX verify supported by driver! */
    texTemp.last_level = MAX_TEXTURE_LEVEL;
-   texTemp.width[0] = 1 << MAX_TEXTURE_LEVEL;
-   texTemp.height[0] = 1 << MAX_TEXTURE_LEVEL;
-   texTemp.depth[0] = 1;
-   pf_get_block(texTemp.format, &texTemp.block);
+   texTemp.width0 = 1 << MAX_TEXTURE_LEVEL;
+   texTemp.height0 = 1 << MAX_TEXTURE_LEVEL;
+   texTemp.depth0 = 1;
 
    aaline->texture = screen->texture_create(screen, &texTemp);
    if (!aaline->texture)
@@ -413,11 +415,11 @@ aaline_create_texture(struct aaline_stage *aaline)
     */
    for (level = 0; level <= MAX_TEXTURE_LEVEL; level++) {
       struct pipe_transfer *transfer;
-      const uint size = aaline->texture->width[level];
+      const uint size = u_minify(aaline->texture->width0, level);
       ubyte *data;
       uint i, j;
 
-      assert(aaline->texture->width[level] == aaline->texture->height[level]);
+      assert(aaline->texture->width0 == aaline->texture->height0);
 
       /* This texture is new, no need to flush. 
        */
@@ -658,13 +660,13 @@ aaline_first_line(struct draw_stage *stage, struct prim_header *header)
    }
 
    /* update vertex attrib info */
-   aaline->tex_slot = draw->vs.num_vs_outputs;
-   aaline->pos_slot = draw->vs.position_output;
+   aaline->tex_slot = draw_current_shader_outputs(draw);
+   aaline->pos_slot = draw_current_shader_position_output(draw);;
 
    /* advertise the extra post-transformed vertex attribute */
-   draw->extra_vp_outputs.semantic_name = TGSI_SEMANTIC_GENERIC;
-   draw->extra_vp_outputs.semantic_index = aaline->fs->generic_attrib;
-   draw->extra_vp_outputs.slot = aaline->tex_slot;
+   draw->extra_shader_outputs.semantic_name = TGSI_SEMANTIC_GENERIC;
+   draw->extra_shader_outputs.semantic_index = aaline->fs->generic_attrib;
+   draw->extra_shader_outputs.slot = aaline->tex_slot;
 
    /* how many samplers? */
    /* we'll use sampler/texture[pstip->sampler_unit] for the stipple */
@@ -705,7 +707,7 @@ aaline_flush(struct draw_stage *stage, unsigned flags)
                                        aaline->state.texture);
    draw->suspend_flushing = FALSE;
 
-   draw->extra_vp_outputs.slot = 0;
+   draw->extra_shader_outputs.slot = 0;
 }
 
 
@@ -896,16 +898,16 @@ draw_install_aaline_stage(struct draw_context *draw, struct pipe_context *pipe)
    aaline->driver_bind_fs_state = pipe->bind_fs_state;
    aaline->driver_delete_fs_state = pipe->delete_fs_state;
 
-   aaline->driver_bind_sampler_states = pipe->bind_sampler_states;
-   aaline->driver_set_sampler_textures = pipe->set_sampler_textures;
+   aaline->driver_bind_sampler_states = pipe->bind_fragment_sampler_states;
+   aaline->driver_set_sampler_textures = pipe->set_fragment_sampler_textures;
 
    /* override the driver's functions */
    pipe->create_fs_state = aaline_create_fs_state;
    pipe->bind_fs_state = aaline_bind_fs_state;
    pipe->delete_fs_state = aaline_delete_fs_state;
 
-   pipe->bind_sampler_states = aaline_bind_sampler_states;
-   pipe->set_sampler_textures = aaline_set_sampler_textures;
+   pipe->bind_fragment_sampler_states = aaline_bind_sampler_states;
+   pipe->set_fragment_sampler_textures = aaline_set_sampler_textures;
    
    /* Install once everything is known to be OK:
     */
index ae1712fe122a901bcfafda05c98d32b927299c28..d86717e5182bbe5e23207f0bfcfe4381dfed04c3 100644 (file)
@@ -131,22 +131,22 @@ aa_transform_decl(struct tgsi_transform_context *ctx,
    struct aa_transform_context *aactx = (struct aa_transform_context *) ctx;
 
    if (decl->Declaration.File == TGSI_FILE_OUTPUT &&
-       decl->Semantic.SemanticName == TGSI_SEMANTIC_COLOR &&
-       decl->Semantic.SemanticIndex == 0) {
-      aactx->colorOutput = decl->DeclarationRange.First;
+       decl->Semantic.Name == TGSI_SEMANTIC_COLOR &&
+       decl->Semantic.Index == 0) {
+      aactx->colorOutput = decl->Range.First;
    }
    else if (decl->Declaration.File == TGSI_FILE_INPUT) {
-      if ((int) decl->DeclarationRange.Last > aactx->maxInput)
-         aactx->maxInput = decl->DeclarationRange.Last;
-      if (decl->Semantic.SemanticName == TGSI_SEMANTIC_GENERIC &&
-           (int) decl->Semantic.SemanticIndex > aactx->maxGeneric) {
-         aactx->maxGeneric = decl->Semantic.SemanticIndex;
+      if ((int) decl->Range.Last > aactx->maxInput)
+         aactx->maxInput = decl->Range.Last;
+      if (decl->Semantic.Name == TGSI_SEMANTIC_GENERIC &&
+           (int) decl->Semantic.Index > aactx->maxGeneric) {
+         aactx->maxGeneric = decl->Semantic.Index;
       }
    }
    else if (decl->Declaration.File == TGSI_FILE_TEMPORARY) {
       uint i;
-      for (i = decl->DeclarationRange.First;
-           i <= decl->DeclarationRange.Last; i++) {
+      for (i = decl->Range.First;
+           i <= decl->Range.Last; i++) {
          aactx->tempsUsed |= (1 << i);
       }
    }
@@ -198,23 +198,23 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       /* XXX this could be linear... */
       decl.Declaration.Interpolate = TGSI_INTERPOLATE_PERSPECTIVE;
       decl.Declaration.Semantic = 1;
-      decl.Semantic.SemanticName = TGSI_SEMANTIC_GENERIC;
-      decl.Semantic.SemanticIndex = aactx->maxGeneric + 1;
-      decl.DeclarationRange.First = 
-      decl.DeclarationRange.Last = texInput;
+      decl.Semantic.Name = TGSI_SEMANTIC_GENERIC;
+      decl.Semantic.Index = aactx->maxGeneric + 1;
+      decl.Range.First = 
+      decl.Range.Last = texInput;
       ctx->emit_declaration(ctx, &decl);
 
       /* declare new temp regs */
       decl = tgsi_default_full_declaration();
       decl.Declaration.File = TGSI_FILE_TEMPORARY;
-      decl.DeclarationRange.First = 
-      decl.DeclarationRange.Last = tmp0;
+      decl.Range.First = 
+      decl.Range.Last = tmp0;
       ctx->emit_declaration(ctx, &decl);
 
       decl = tgsi_default_full_declaration();
       decl.Declaration.File = TGSI_FILE_TEMPORARY;
-      decl.DeclarationRange.First = 
-      decl.DeclarationRange.Last = aactx->colorTemp;
+      decl.Range.First = 
+      decl.Range.Last = aactx->colorTemp;
       ctx->emit_declaration(ctx, &decl);
 
       aactx->firstInstruction = FALSE;
@@ -234,30 +234,30 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_MUL;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_XY;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_XY;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_INPUT;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = texInput;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_INPUT;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = texInput;
+      newInst.Src[0].Register.File = TGSI_FILE_INPUT;
+      newInst.Src[0].Register.Index = texInput;
+      newInst.Src[1].Register.File = TGSI_FILE_INPUT;
+      newInst.Src[1].Register.Index = texInput;
       ctx->emit_instruction(ctx, &newInst);
 
       /* ADD t0.x, t0.x, t0.y;  # x^2 + y^2 */
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_ADD;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_X;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_X;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleX = TGSI_SWIZZLE_X;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[1].SrcRegister.SwizzleX = TGSI_SWIZZLE_Y;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = tmp0;
+      newInst.Src[0].Register.SwizzleX = TGSI_SWIZZLE_X;
+      newInst.Src[1].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[1].Register.Index = tmp0;
+      newInst.Src[1].Register.SwizzleX = TGSI_SWIZZLE_Y;
       ctx->emit_instruction(ctx, &newInst);
 
 #if NORMALIZE  /* OPTIONAL normalization of length */
@@ -265,24 +265,24 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_RSQ;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_X;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_X;
       newInst.Instruction.NumSrcRegs = 1;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = tmp0;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = tmp0;
       ctx->emit_instruction(ctx, &newInst);
 
       /* RCP t0.x, t0.x; */
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_RCP;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_X;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_X;
       newInst.Instruction.NumSrcRegs = 1;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = tmp0;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = tmp0;
       ctx->emit_instruction(ctx, &newInst);
 #endif
 
@@ -290,16 +290,16 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_SGT;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_Y;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_Y;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleY = TGSI_SWIZZLE_X;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_INPUT;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = texInput;
-      newInst.FullSrcRegisters[1].SrcRegister.SwizzleY = TGSI_SWIZZLE_W;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = tmp0;
+      newInst.Src[0].Register.SwizzleY = TGSI_SWIZZLE_X;
+      newInst.Src[1].Register.File = TGSI_FILE_INPUT;
+      newInst.Src[1].Register.Index = texInput;
+      newInst.Src[1].Register.SwizzleY = TGSI_SWIZZLE_W;
       ctx->emit_instruction(ctx, &newInst);
 
       /* KIL -tmp0.yyyy;   # if -tmp0.y < 0, KILL */
@@ -307,13 +307,13 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       newInst.Instruction.Opcode = TGSI_OPCODE_KIL;
       newInst.Instruction.NumDstRegs = 0;
       newInst.Instruction.NumSrcRegs = 1;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleX = TGSI_SWIZZLE_Y;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleY = TGSI_SWIZZLE_Y;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleZ = TGSI_SWIZZLE_Y;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleW = TGSI_SWIZZLE_Y;
-      newInst.FullSrcRegisters[0].SrcRegister.Negate = 1;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = tmp0;
+      newInst.Src[0].Register.SwizzleX = TGSI_SWIZZLE_Y;
+      newInst.Src[0].Register.SwizzleY = TGSI_SWIZZLE_Y;
+      newInst.Src[0].Register.SwizzleZ = TGSI_SWIZZLE_Y;
+      newInst.Src[0].Register.SwizzleW = TGSI_SWIZZLE_Y;
+      newInst.Src[0].Register.Negate = 1;
       ctx->emit_instruction(ctx, &newInst);
 
 
@@ -323,77 +323,77 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_SUB;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_Z;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_Z;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_INPUT;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = texInput;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleZ = TGSI_SWIZZLE_W;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_INPUT;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = texInput;
-      newInst.FullSrcRegisters[1].SrcRegister.SwizzleZ = TGSI_SWIZZLE_Z;
+      newInst.Src[0].Register.File = TGSI_FILE_INPUT;
+      newInst.Src[0].Register.Index = texInput;
+      newInst.Src[0].Register.SwizzleZ = TGSI_SWIZZLE_W;
+      newInst.Src[1].Register.File = TGSI_FILE_INPUT;
+      newInst.Src[1].Register.Index = texInput;
+      newInst.Src[1].Register.SwizzleZ = TGSI_SWIZZLE_Z;
       ctx->emit_instruction(ctx, &newInst);
 
       /* RCP t0.z, t0.z;  # t0.z = 1 / m */
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_RCP;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_Z;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_Z;
       newInst.Instruction.NumSrcRegs = 1;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleX = TGSI_SWIZZLE_Z;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = tmp0;
+      newInst.Src[0].Register.SwizzleX = TGSI_SWIZZLE_Z;
       ctx->emit_instruction(ctx, &newInst);
 
       /* SUB t0.y, 1, t0.x;  # d = 1 - d */
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_SUB;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_Y;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_Y;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_INPUT;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = texInput;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleY = TGSI_SWIZZLE_W;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[1].SrcRegister.SwizzleY = TGSI_SWIZZLE_X;
+      newInst.Src[0].Register.File = TGSI_FILE_INPUT;
+      newInst.Src[0].Register.Index = texInput;
+      newInst.Src[0].Register.SwizzleY = TGSI_SWIZZLE_W;
+      newInst.Src[1].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[1].Register.Index = tmp0;
+      newInst.Src[1].Register.SwizzleY = TGSI_SWIZZLE_X;
       ctx->emit_instruction(ctx, &newInst);
 
       /* MUL t0.w, t0.y, t0.z;   # coverage = d * m */
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_MUL;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_W;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_W;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleW = TGSI_SWIZZLE_Y;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[1].SrcRegister.SwizzleW = TGSI_SWIZZLE_Z;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = tmp0;
+      newInst.Src[0].Register.SwizzleW = TGSI_SWIZZLE_Y;
+      newInst.Src[1].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[1].Register.Index = tmp0;
+      newInst.Src[1].Register.SwizzleW = TGSI_SWIZZLE_Z;
       ctx->emit_instruction(ctx, &newInst);
 
       /* SLE t0.y, t0.x, tex.z;  # bool b = distance <= k */
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_SLE;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_Y;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_Y;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleY = TGSI_SWIZZLE_X;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_INPUT;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = texInput;
-      newInst.FullSrcRegisters[1].SrcRegister.SwizzleY = TGSI_SWIZZLE_Z;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = tmp0;
+      newInst.Src[0].Register.SwizzleY = TGSI_SWIZZLE_X;
+      newInst.Src[1].Register.File = TGSI_FILE_INPUT;
+      newInst.Src[1].Register.Index = texInput;
+      newInst.Src[1].Register.SwizzleY = TGSI_SWIZZLE_Z;
       ctx->emit_instruction(ctx, &newInst);
 
       /* CMP t0.w, -t0.y, tex.w, t0.w;
@@ -405,29 +405,29 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_CMP;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = tmp0;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_W;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = tmp0;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_W;
       newInst.Instruction.NumSrcRegs = 3;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleX = TGSI_SWIZZLE_Y;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleY = TGSI_SWIZZLE_Y;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleZ = TGSI_SWIZZLE_Y;
-      newInst.FullSrcRegisters[0].SrcRegister.SwizzleW = TGSI_SWIZZLE_Y;
-      newInst.FullSrcRegisters[0].SrcRegister.Negate = 1;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_INPUT;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = texInput;
-      newInst.FullSrcRegisters[1].SrcRegister.SwizzleX = TGSI_SWIZZLE_W;
-      newInst.FullSrcRegisters[1].SrcRegister.SwizzleY = TGSI_SWIZZLE_W;
-      newInst.FullSrcRegisters[1].SrcRegister.SwizzleZ = TGSI_SWIZZLE_W;
-      newInst.FullSrcRegisters[1].SrcRegister.SwizzleW = TGSI_SWIZZLE_W;
-      newInst.FullSrcRegisters[2].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[2].SrcRegister.Index = tmp0;
-      newInst.FullSrcRegisters[2].SrcRegister.SwizzleX = TGSI_SWIZZLE_W;
-      newInst.FullSrcRegisters[2].SrcRegister.SwizzleY = TGSI_SWIZZLE_W;
-      newInst.FullSrcRegisters[2].SrcRegister.SwizzleZ = TGSI_SWIZZLE_W;
-      newInst.FullSrcRegisters[2].SrcRegister.SwizzleW = TGSI_SWIZZLE_W;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = tmp0;
+      newInst.Src[0].Register.SwizzleX = TGSI_SWIZZLE_Y;
+      newInst.Src[0].Register.SwizzleY = TGSI_SWIZZLE_Y;
+      newInst.Src[0].Register.SwizzleZ = TGSI_SWIZZLE_Y;
+      newInst.Src[0].Register.SwizzleW = TGSI_SWIZZLE_Y;
+      newInst.Src[0].Register.Negate = 1;
+      newInst.Src[1].Register.File = TGSI_FILE_INPUT;
+      newInst.Src[1].Register.Index = texInput;
+      newInst.Src[1].Register.SwizzleX = TGSI_SWIZZLE_W;
+      newInst.Src[1].Register.SwizzleY = TGSI_SWIZZLE_W;
+      newInst.Src[1].Register.SwizzleZ = TGSI_SWIZZLE_W;
+      newInst.Src[1].Register.SwizzleW = TGSI_SWIZZLE_W;
+      newInst.Src[2].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[2].Register.Index = tmp0;
+      newInst.Src[2].Register.SwizzleX = TGSI_SWIZZLE_W;
+      newInst.Src[2].Register.SwizzleY = TGSI_SWIZZLE_W;
+      newInst.Src[2].Register.SwizzleZ = TGSI_SWIZZLE_W;
+      newInst.Src[2].Register.SwizzleW = TGSI_SWIZZLE_W;
       ctx->emit_instruction(ctx, &newInst);
 
    }
@@ -439,26 +439,26 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_MOV;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_OUTPUT;
-      newInst.FullDstRegisters[0].DstRegister.Index = aactx->colorOutput;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_XYZ;
+      newInst.Dst[0].Register.File = TGSI_FILE_OUTPUT;
+      newInst.Dst[0].Register.Index = aactx->colorOutput;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_XYZ;
       newInst.Instruction.NumSrcRegs = 1;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = aactx->colorTemp;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = aactx->colorTemp;
       ctx->emit_instruction(ctx, &newInst);
 
       /* MUL result.color.w, colorTemp, tmp0.w; */
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_MUL;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_OUTPUT;
-      newInst.FullDstRegisters[0].DstRegister.Index = aactx->colorOutput;
-      newInst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_W;
+      newInst.Dst[0].Register.File = TGSI_FILE_OUTPUT;
+      newInst.Dst[0].Register.Index = aactx->colorOutput;
+      newInst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_W;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = aactx->colorTemp;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = aactx->tmp0;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = aactx->colorTemp;
+      newInst.Src[1].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[1].Register.Index = aactx->tmp0;
       ctx->emit_instruction(ctx, &newInst);
    }
    else {
@@ -468,11 +468,11 @@ aa_transform_inst(struct tgsi_transform_context *ctx,
       uint i;
 
       for (i = 0; i < inst->Instruction.NumDstRegs; i++) {
-         struct tgsi_full_dst_register *dst = &inst->FullDstRegisters[i];
-         if (dst->DstRegister.File == TGSI_FILE_OUTPUT &&
-             dst->DstRegister.Index == aactx->colorOutput) {
-            dst->DstRegister.File = TGSI_FILE_TEMPORARY;
-            dst->DstRegister.Index = aactx->colorTemp;
+         struct tgsi_full_dst_register *dst = &inst->Dst[i];
+         if (dst->Register.File == TGSI_FILE_OUTPUT &&
+             dst->Register.Index == aactx->colorOutput) {
+            dst->Register.File = TGSI_FILE_TEMPORARY;
+            dst->Register.Index = aactx->colorTemp;
          }
       }
    }
@@ -687,14 +687,14 @@ aapoint_first_point(struct draw_stage *stage, struct prim_header *header)
    bind_aapoint_fragment_shader(aapoint);
 
    /* update vertex attrib info */
-   aapoint->tex_slot = draw->vs.num_vs_outputs;
+   aapoint->tex_slot = draw_current_shader_outputs(draw);
    assert(aapoint->tex_slot > 0); /* output[0] is vertex pos */
 
-   aapoint->pos_slot = draw->vs.position_output;
+   aapoint->pos_slot = draw_current_shader_position_output(draw);
 
-   draw->extra_vp_outputs.semantic_name = TGSI_SEMANTIC_GENERIC;
-   draw->extra_vp_outputs.semantic_index = aapoint->fs->generic_attrib;
-   draw->extra_vp_outputs.slot = aapoint->tex_slot;
+   draw->extra_shader_outputs.semantic_name = TGSI_SEMANTIC_GENERIC;
+   draw->extra_shader_outputs.semantic_index = aapoint->fs->generic_attrib;
+   draw->extra_shader_outputs.slot = aapoint->tex_slot;
 
    /* find psize slot in post-transform vertex */
    aapoint->psize_slot = -1;
@@ -731,7 +731,7 @@ aapoint_flush(struct draw_stage *stage, unsigned flags)
    aapoint->driver_bind_fs_state(pipe, aapoint->fs->driver_fs);
    draw->suspend_flushing = FALSE;
 
-   draw->extra_vp_outputs.slot = 0;
+   draw->extra_shader_outputs.slot = 0;
 }
 
 
index 0670268a196f2c6d933dfa32fb5654bafbc06ca2..205cda5eabed89e8976ad0faeed4799ac959b069 100644 (file)
@@ -114,8 +114,8 @@ static void interp( const struct clipper *clip,
                    const struct vertex_header *out, 
                    const struct vertex_header *in )
 {
-   const unsigned nr_attrs = clip->stage.draw->vs.num_vs_outputs;
-   const unsigned pos_attr = clip->stage.draw->vs.position_output;
+   const unsigned nr_attrs = draw_current_shader_outputs(clip->stage.draw);
+   const unsigned pos_attr = draw_current_shader_position_output(clip->stage.draw);
    unsigned j;
 
    /* Vertex header.
index 0a70483858cdff525e4b071ed7bdffb7d9982461..11b39db5990b2541ca71fe0b7e41450322cb2127 100644 (file)
@@ -55,7 +55,7 @@ static INLINE struct cull_stage *cull_stage( struct draw_stage *stage )
 static void cull_tri( struct draw_stage *stage,
                      struct prim_header *header )
 {
-   const unsigned pos = stage->draw->vs.position_output;
+   const unsigned pos = draw_current_shader_position_output(stage->draw);
 
    /* Window coords: */
    const float *v0 = header->v[0]->data[pos];
index 40798a5d6e7b394b25fe9eb2595067c83e040b77..e829492423e3b0e9275a0f13c498ebf9acc4858c 100644 (file)
@@ -63,7 +63,7 @@ static INLINE struct offset_stage *offset_stage( struct draw_stage *stage )
 static void do_offset_tri( struct draw_stage *stage,
                           struct prim_header *header )
 {
-   const unsigned pos = stage->draw->vs.position_output;
+   const unsigned pos = draw_current_shader_position_output(stage->draw);
    struct offset_stage *offset = offset_stage(stage);   
    float inv_det = 1.0f / header->det;
 
index 283502cdf3e3bd31afd3114df5b0a854fe6b13ec..0cc2b7186416dce314a29d1a3f984360612b0298 100644 (file)
@@ -38,6 +38,7 @@
 #include "pipe/p_defines.h"
 #include "pipe/p_shader_tokens.h"
 
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -133,20 +134,20 @@ pstip_transform_decl(struct tgsi_transform_context *ctx,
 
    if (decl->Declaration.File == TGSI_FILE_SAMPLER) {
       uint i;
-      for (i = decl->DeclarationRange.First;
-           i <= decl->DeclarationRange.Last; i++) {
+      for (i = decl->Range.First;
+           i <= decl->Range.Last; i++) {
          pctx->samplersUsed |= 1 << i;
       }
    }
    else if (decl->Declaration.File == TGSI_FILE_INPUT) {
-      pctx->maxInput = MAX2(pctx->maxInput, (int) decl->DeclarationRange.Last);
-      if (decl->Semantic.SemanticName == TGSI_SEMANTIC_POSITION)
-         pctx->wincoordInput = (int) decl->DeclarationRange.First;
+      pctx->maxInput = MAX2(pctx->maxInput, (int) decl->Range.Last);
+      if (decl->Semantic.Name == TGSI_SEMANTIC_POSITION)
+         pctx->wincoordInput = (int) decl->Range.First;
    }
    else if (decl->Declaration.File == TGSI_FILE_TEMPORARY) {
       uint i;
-      for (i = decl->DeclarationRange.First;
-           i <= decl->DeclarationRange.Last; i++) {
+      for (i = decl->Range.First;
+           i <= decl->Range.Last; i++) {
          pctx->tempsUsed |= (1 << i);
       }
    }
@@ -226,25 +227,25 @@ pstip_transform_inst(struct tgsi_transform_context *ctx,
          decl.Declaration.File = TGSI_FILE_INPUT;
          decl.Declaration.Interpolate = TGSI_INTERPOLATE_LINEAR; /* XXX? */
          decl.Declaration.Semantic = 1;
-         decl.Semantic.SemanticName = TGSI_SEMANTIC_POSITION;
-         decl.Semantic.SemanticIndex = 0;
-         decl.DeclarationRange.First = 
-            decl.DeclarationRange.Last = wincoordInput;
+         decl.Semantic.Name = TGSI_SEMANTIC_POSITION;
+         decl.Semantic.Index = 0;
+         decl.Range.First = 
+            decl.Range.Last = wincoordInput;
          ctx->emit_declaration(ctx, &decl);
       }
 
       /* declare new sampler */
       decl = tgsi_default_full_declaration();
       decl.Declaration.File = TGSI_FILE_SAMPLER;
-      decl.DeclarationRange.First = 
-      decl.DeclarationRange.Last = pctx->freeSampler;
+      decl.Range.First = 
+      decl.Range.Last = pctx->freeSampler;
       ctx->emit_declaration(ctx, &decl);
 
       /* declare new temp regs */
       decl = tgsi_default_full_declaration();
       decl.Declaration.File = TGSI_FILE_TEMPORARY;
-      decl.DeclarationRange.First = 
-      decl.DeclarationRange.Last = pctx->texTemp;
+      decl.Range.First = 
+      decl.Range.Last = pctx->texTemp;
       ctx->emit_declaration(ctx, &decl);
 
       /* emit immediate = {1/32, 1/32, 1, 1}
@@ -280,27 +281,28 @@ pstip_transform_inst(struct tgsi_transform_context *ctx,
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_MUL;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = pctx->texTemp;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = pctx->texTemp;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_INPUT;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = wincoordInput;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_IMMEDIATE;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = pctx->numImmed;
+      newInst.Src[0].Register.File = TGSI_FILE_INPUT;
+      newInst.Src[0].Register.Index = wincoordInput;
+      newInst.Src[1].Register.File = TGSI_FILE_IMMEDIATE;
+      newInst.Src[1].Register.Index = pctx->numImmed;
       ctx->emit_instruction(ctx, &newInst);
 
       /* TEX texTemp, texTemp, sampler; */
       newInst = tgsi_default_full_instruction();
       newInst.Instruction.Opcode = TGSI_OPCODE_TEX;
       newInst.Instruction.NumDstRegs = 1;
-      newInst.FullDstRegisters[0].DstRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullDstRegisters[0].DstRegister.Index = pctx->texTemp;
+      newInst.Dst[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Dst[0].Register.Index = pctx->texTemp;
       newInst.Instruction.NumSrcRegs = 2;
-      newInst.InstructionExtTexture.Texture = TGSI_TEXTURE_2D;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = pctx->texTemp;
-      newInst.FullSrcRegisters[1].SrcRegister.File = TGSI_FILE_SAMPLER;
-      newInst.FullSrcRegisters[1].SrcRegister.Index = pctx->freeSampler;
+      newInst.Instruction.Texture = TRUE;
+      newInst.Texture.Texture = TGSI_TEXTURE_2D;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = pctx->texTemp;
+      newInst.Src[1].Register.File = TGSI_FILE_SAMPLER;
+      newInst.Src[1].Register.Index = pctx->freeSampler;
       ctx->emit_instruction(ctx, &newInst);
 
       /* KIL -texTemp;   # if -texTemp < 0, KILL fragment */
@@ -308,9 +310,9 @@ pstip_transform_inst(struct tgsi_transform_context *ctx,
       newInst.Instruction.Opcode = TGSI_OPCODE_KIL;
       newInst.Instruction.NumDstRegs = 0;
       newInst.Instruction.NumSrcRegs = 1;
-      newInst.FullSrcRegisters[0].SrcRegister.File = TGSI_FILE_TEMPORARY;
-      newInst.FullSrcRegisters[0].SrcRegister.Index = pctx->texTemp;
-      newInst.FullSrcRegisters[0].SrcRegister.Negate = 1;
+      newInst.Src[0].Register.File = TGSI_FILE_TEMPORARY;
+      newInst.Src[0].Register.Index = pctx->texTemp;
+      newInst.Src[0].Register.Negate = 1;
       ctx->emit_instruction(ctx, &newInst);
    }
 
@@ -427,10 +429,9 @@ pstip_create_texture(struct pstip_stage *pstip)
    texTemp.target = PIPE_TEXTURE_2D;
    texTemp.format = PIPE_FORMAT_A8_UNORM; /* XXX verify supported by driver! */
    texTemp.last_level = 0;
-   texTemp.width[0] = 32;
-   texTemp.height[0] = 32;
-   texTemp.depth[0] = 1;
-   pf_get_block(texTemp.format, &texTemp.block);
+   texTemp.width0 = 32;
+   texTemp.height0 = 32;
+   texTemp.depth0 = 1;
 
    pstip->texture = screen->texture_create(screen, &texTemp);
    if (pstip->texture == NULL)
@@ -754,8 +755,8 @@ draw_install_pstipple_stage(struct draw_context *draw,
    pstip->driver_bind_fs_state = pipe->bind_fs_state;
    pstip->driver_delete_fs_state = pipe->delete_fs_state;
 
-   pstip->driver_bind_sampler_states = pipe->bind_sampler_states;
-   pstip->driver_set_sampler_textures = pipe->set_sampler_textures;
+   pstip->driver_bind_sampler_states = pipe->bind_fragment_sampler_states;
+   pstip->driver_set_sampler_textures = pipe->set_fragment_sampler_textures;
    pstip->driver_set_polygon_stipple = pipe->set_polygon_stipple;
 
    /* override the driver's functions */
@@ -763,8 +764,8 @@ draw_install_pstipple_stage(struct draw_context *draw,
    pipe->bind_fs_state = pstip_bind_fs_state;
    pipe->delete_fs_state = pstip_delete_fs_state;
 
-   pipe->bind_sampler_states = pstip_bind_sampler_states;
-   pipe->set_sampler_textures = pstip_set_sampler_textures;
+   pipe->bind_fragment_sampler_states = pstip_bind_sampler_states;
+   pipe->set_fragment_sampler_textures = pstip_set_sampler_textures;
    pipe->set_polygon_stipple = pstip_set_polygon_stipple;
 
    return TRUE;
index 6e921bac27858f84ef01a055c7a2f2767bf32f56..70fbab9ea7673dca3cbdf1a1162dd4fb0a6caaff 100644 (file)
@@ -73,7 +73,8 @@ screen_interp( struct draw_context *draw,
                const struct vertex_header *v1 )
 {
    uint attr;
-   for (attr = 0; attr < draw->vs.num_vs_outputs; attr++) {
+   int num_outputs = draw_current_shader_outputs(draw);
+   for (attr = 0; attr < num_outputs; attr++) {
       const float *val0 = v0->data[attr];
       const float *val1 = v1->data[attr];
       float *newv = dst->data[attr];
@@ -121,7 +122,7 @@ stipple_line(struct draw_stage *stage, struct prim_header *header)
    struct stipple_stage *stipple = stipple_stage(stage);
    struct vertex_header *v0 = header->v[0];
    struct vertex_header *v1 = header->v[1];
-   const unsigned pos = stage->draw->vs.position_output;
+   const unsigned pos = draw_current_shader_position_output(stage->draw);
    const float *pos0 = v0->data[pos];
    const float *pos1 = v1->data[pos];
    float start = 0;
index f32cbef983d71aa70a2c9d9dcab0682cfe8d1bd5..3073c8708251e0ee521c516b5b18c6301ab16e19 100644 (file)
@@ -59,7 +59,7 @@ static void wideline_line( struct draw_stage *stage,
                            struct prim_header *header )
 {
    /*const struct wideline_stage *wide = wideline_stage(stage);*/
-   const unsigned pos = stage->draw->vs.position_output;
+   const unsigned pos = draw_current_shader_position_output(stage->draw);
    const float half_width = 0.5f * stage->draw->rasterizer->line_width;
 
    struct prim_header tri;
index 7d76a7dbf39e32cfc7d385e59826453a23e21f42..8dc50c0ab431b7fc1f6d4a04db5b06bcc12a96b9 100644 (file)
@@ -112,7 +112,7 @@ static void set_texcoords(const struct widepoint_stage *wide,
 
    if (wide->point_coord_fs_input >= 0) {
       /* put gl_PointCoord into the extra vertex slot */
-      uint slot = wide->stage.draw->extra_vp_outputs.slot;
+      uint slot = wide->stage.draw->extra_shader_outputs.slot;
       v->data[slot][0] = tc[0];
       v->data[slot][1] = tc[1];
       v->data[slot][2] = 0.0F;
@@ -130,7 +130,7 @@ static void widepoint_point( struct draw_stage *stage,
                              struct prim_header *header )
 {
    const struct widepoint_stage *wide = widepoint_stage(stage);
-   const unsigned pos = stage->draw->vs.position_output;
+   const unsigned pos = draw_current_shader_position_output(stage->draw);
    const boolean sprite = (boolean) stage->draw->rasterizer->point_sprite;
    float half_size;
    float left_adj, right_adj, bot_adj, top_adj;
@@ -257,13 +257,13 @@ static void widepoint_first_point( struct draw_stage *stage,
       wide->point_coord_fs_input = find_pntc_input_attrib(draw);
 
       /* setup extra vp output (point coord implemented as a texcoord) */
-      draw->extra_vp_outputs.semantic_name = TGSI_SEMANTIC_GENERIC;
-      draw->extra_vp_outputs.semantic_index = 0;
-      draw->extra_vp_outputs.slot = draw->vs.num_vs_outputs;
+      draw->extra_shader_outputs.semantic_name = TGSI_SEMANTIC_GENERIC;
+      draw->extra_shader_outputs.semantic_index = 0;
+      draw->extra_shader_outputs.slot = draw_current_shader_outputs(draw);
    }
    else {
       wide->point_coord_fs_input = -1;
-      draw->extra_vp_outputs.slot = 0;
+      draw->extra_shader_outputs.slot = 0;
    }
 
    wide->psize_slot = -1;
@@ -287,7 +287,7 @@ static void widepoint_flush( struct draw_stage *stage, unsigned flags )
 {
    stage->point = widepoint_first_point;
    stage->next->flush( stage->next, flags );
-   stage->draw->extra_vp_outputs.slot = 0;
+   stage->draw->extra_shader_outputs.slot = 0;
 }
 
 
index 41fcb16a0a5a1df71c3de27bbafaec358fd5b61c..e49041556bd5d1c386e5f78448b19e5f306aad30 100644 (file)
@@ -142,8 +142,6 @@ struct draw_context
 
       /* user-space vertex data, buffers */
       struct {
-         const unsigned *edgeflag;
-
          /** vertex element/index buffer (ex: glDrawElements) */
          const void *elts;
          /** bytes per index (0, 1, 2 or 4) */
@@ -154,8 +152,9 @@ struct draw_context
          /** vertex arrays */
          const void *vbuffer[PIPE_MAX_ATTRIBS];
          
-         /** constant buffer (for vertex shader) */
-         const void *constants;
+         /** constant buffer (for vertex/geometry shader) */
+         const void *vs_constants;
+         const void *gs_constants;
       } user;
 
       boolean test_fse;         /* enable FSE even though its not correct (eg for softpipe) */
@@ -184,6 +183,7 @@ struct draw_context
       struct draw_vertex_shader *vertex_shader;
       uint num_vs_outputs;  /**< convenience, from vertex_shader */
       uint position_output;
+      uint edgeflag_output;
 
       /** TGSI program interpreter runtime state */
       struct tgsi_exec_machine *machine;
@@ -212,6 +212,18 @@ struct draw_context
       struct translate_cache *emit_cache;
    } vs;
 
+   struct {
+      struct draw_geometry_shader *geometry_shader;
+      uint num_gs_outputs;  /**< convenience, from geometry_shader */
+      uint position_output;
+
+      /** TGSI program interpreter runtime state */
+      struct tgsi_exec_machine *machine;
+
+      uint num_samplers;
+      struct tgsi_sampler **samplers;
+   } gs;
+
    /* Clip derived state:
     */
    float plane[12][4];
@@ -223,7 +235,7 @@ struct draw_context
       uint semantic_name;
       uint semantic_index;
       int slot;
-   } extra_vp_outputs;
+   } extra_shader_outputs;
 
    unsigned reduced_prim;
 
@@ -246,6 +258,19 @@ void draw_vs_set_constants( struct draw_context *,
 
 
 
+/*******************************************************************************
+ * Geometry shading code:
+ */
+boolean draw_gs_init( struct draw_context *draw );
+void draw_gs_set_constants( struct draw_context *,
+                            const float (*constants)[4],
+                            unsigned size );
+
+/*******************************************************************************
+ * Common shading code:
+ */
+int draw_current_shader_outputs(struct draw_context *draw);
+int draw_current_shader_position_output(struct draw_context *draw);
 
 /*******************************************************************************
  * Vertex processing (was passthrough) code:
index 4865a2d85421cc28f74fc753629bce7db1503d8c..2801dbafe4735b44246d406b310492b0bb190110 100644 (file)
@@ -314,12 +314,3 @@ draw_arrays(struct draw_context *draw, unsigned prim,
    /* drawing done here: */
    draw_pt_arrays(draw, prim, start, count);
 }
-
-boolean draw_pt_get_edgeflag( struct draw_context *draw,
-                              unsigned idx )
-{
-   if (draw->pt.user.edgeflag)
-      return (draw->pt.user.edgeflag[idx/32] & (1 << (idx%32))) != 0;
-   else
-      return 1;
-}
index 7a17a9fb6b2818013966b6b13c7a78be0849a6a8..20edf7a227ef32f4b8673954ec51b75f9447be3c 100644 (file)
@@ -149,11 +149,6 @@ struct draw_pt_middle_end *draw_pt_middle_fse( struct draw_context *draw );
 struct draw_pt_middle_end *draw_pt_fetch_pipeline_or_emit(struct draw_context *draw);
 
 
-/* More helpers:
- */
-boolean draw_pt_get_edgeflag( struct draw_context *draw,
-                              unsigned idx );
-
 
 /*******************************************************************************
  * HW vertex emit:
@@ -217,7 +212,8 @@ boolean draw_pt_post_vs_run( struct pt_post_vs *pvs,
 void draw_pt_post_vs_prepare( struct pt_post_vs *pvs,
                              boolean bypass_clipping,
                              boolean bypass_viewport,
-                             boolean opengl );
+                             boolean opengl,
+                             boolean need_edgeflags );
 
 struct pt_post_vs *draw_pt_post_vs_create( struct draw_context *draw );
 
index 65c3a34c3474a664fbac65042117800174eda9db..305bfef4352b002c60aa06879878d697e20b33b8 100644 (file)
@@ -42,11 +42,11 @@ struct pt_fetch {
    struct translate *translate;
 
    unsigned vertex_size;
-   boolean need_edgeflags;
 
    struct translate_cache *cache;
 };
 
+
 /* Perform the fetch from API vertex elements & vertex buffers, to a
  * contiguous set of float[4] attributes as required for the
  * vertex_shader->run_linear() method.
@@ -120,7 +120,12 @@ void draw_pt_fetch_prepare( struct pt_fetch *fetch,
       fetch->translate = translate_cache_find(fetch->cache, &key);
 
       {
-         static struct vertex_header vh = { 0, 1, 0, UNDEFINED_VERTEX_ID, { .0f, .0f, .0f, .0f } };
+         static struct vertex_header vh = { 0,
+                                            1,
+                                            0,
+                                            UNDEFINED_VERTEX_ID,
+                                            { .0f, .0f, .0f, .0f } };
+
         fetch->translate->set_buffer(fetch->translate,
                                      draw->pt.nr_vertex_buffers,
                                      &vh,
@@ -128,9 +133,6 @@ void draw_pt_fetch_prepare( struct pt_fetch *fetch,
       }
    }
 
-   fetch->need_edgeflags = ((draw->rasterizer->fill_cw != PIPE_POLYGON_MODE_FILL ||
-                             draw->rasterizer->fill_ccw != PIPE_POLYGON_MODE_FILL) &&
-                            draw->pt.user.edgeflag);
 }
 
 
@@ -158,17 +160,6 @@ void draw_pt_fetch_run( struct pt_fetch *fetch,
                        count,
                        verts );
 
-   /* Edgeflags are hard to fit into a translate program, populate
-    * them separately if required.  In the setup above they are
-    * defaulted to one, so only need this if there is reason to change
-    * that default:
-    */
-   if (fetch->need_edgeflags) {
-      for (i = 0; i < count; i++) {
-         struct vertex_header *vh = (struct vertex_header *)(verts + i * fetch->vertex_size);
-         vh->edgeflag = draw_pt_get_edgeflag( draw, elts[i] );
-      }
-   }
 }
 
 
@@ -193,18 +184,6 @@ void draw_pt_fetch_run_linear( struct pt_fetch *fetch,
                    start,
                    count,
                    verts );
-
-   /* Edgeflags are hard to fit into a translate program, populate
-    * them separately if required.  In the setup above they are
-    * defaulted to one, so only need this if there is reason to change
-    * that default:
-    */
-   if (fetch->need_edgeflags) {
-      for (i = 0; i < count; i++) {
-         struct vertex_header *vh = (struct vertex_header *)(verts + i * fetch->vertex_size);
-         vh->edgeflag = draw_pt_get_edgeflag( draw, start + i );
-      }
-   }
 }
 
 
index df6c265b7ec9026a95cd6383b1d70adb242efcbb..1a9df4cac5d5269723996da3535f7b9c1d455d59 100644 (file)
@@ -32,6 +32,7 @@
 #include "draw/draw_vertex.h"
 #include "draw/draw_pt.h"
 #include "draw/draw_vs.h"
+#include "draw/draw_gs.h"
 #include "translate/translate.h"
 
 
@@ -85,9 +86,9 @@ static void fetch_pipeline_prepare( struct draw_pt_middle_end *middle,
    draw_pt_post_vs_prepare( fpme->post_vs,
                            (boolean)draw->bypass_clipping,
                            (boolean)(draw->identity_viewport ||
-                                      draw->rasterizer->bypass_vs_clip_and_viewport),
-                           (boolean)draw->rasterizer->gl_rasterization_rules );
-                           
+                           draw->rasterizer->bypass_vs_clip_and_viewport),
+                           (boolean)draw->rasterizer->gl_rasterization_rules,
+                           (draw->vs.edgeflag_output ? true : false) );    
 
    if (!(opt & PT_PIPELINE)) {
       draw_pt_emit_prepare( fpme->emit, 
@@ -119,7 +120,8 @@ static void fetch_pipeline_run( struct draw_pt_middle_end *middle,
 {
    struct fetch_pipeline_middle_end *fpme = (struct fetch_pipeline_middle_end *)middle;
    struct draw_context *draw = fpme->draw;
-   struct draw_vertex_shader *shader = draw->vs.vertex_shader;
+   struct draw_vertex_shader *vshader = draw->vs.vertex_shader;
+   struct draw_geometry_shader *gshader = draw->gs.geometry_shader;
    unsigned opt = fpme->opt;
    unsigned alloc_count = align( fetch_count, 4 );
 
@@ -147,13 +149,21 @@ static void fetch_pipeline_run( struct draw_pt_middle_end *middle,
     */
    if (opt & PT_SHADE)
    {
-      shader->run_linear(shader, 
-                        (const float (*)[4])pipeline_verts->data,
-                        (      float (*)[4])pipeline_verts->data,
-                        (const float (*)[4])draw->pt.user.constants,
-                        fetch_count,
-                        fpme->vertex_size,
-                        fpme->vertex_size);
+      vshader->run_linear(vshader,
+                          (const float (*)[4])pipeline_verts->data,
+                          (      float (*)[4])pipeline_verts->data,
+                          (const float (*)[4])draw->pt.user.vs_constants,
+                          fetch_count,
+                          fpme->vertex_size,
+                          fpme->vertex_size);
+      if (gshader)
+         draw_geometry_shader_run(gshader,
+                                  (const float (*)[4])pipeline_verts->data,
+                                  (      float (*)[4])pipeline_verts->data,
+                                  (const float (*)[4])draw->pt.user.gs_constants,
+                                  fetch_count,
+                                  fpme->vertex_size,
+                                  fpme->vertex_size);
    }
 
    if (draw_pt_post_vs_run( fpme->post_vs,
@@ -196,6 +206,7 @@ static void fetch_pipeline_linear_run( struct draw_pt_middle_end *middle,
    struct fetch_pipeline_middle_end *fpme = (struct fetch_pipeline_middle_end *)middle;
    struct draw_context *draw = fpme->draw;
    struct draw_vertex_shader *shader = draw->vs.vertex_shader;
+   struct draw_geometry_shader *geometry_shader = draw->gs.geometry_shader;
    unsigned opt = fpme->opt;
    unsigned alloc_count = align( count, 4 );
 
@@ -226,10 +237,19 @@ static void fetch_pipeline_linear_run( struct draw_pt_middle_end *middle,
       shader->run_linear(shader,
                         (const float (*)[4])pipeline_verts->data,
                         (      float (*)[4])pipeline_verts->data,
-                        (const float (*)[4])draw->pt.user.constants,
+                        (const float (*)[4])draw->pt.user.vs_constants,
                         count,
                         fpme->vertex_size,
                         fpme->vertex_size);
+
+      if (geometry_shader)
+         draw_geometry_shader_run(geometry_shader,
+                                  (const float (*)[4])pipeline_verts->data,
+                                  (      float (*)[4])pipeline_verts->data,
+                                  (const float (*)[4])draw->pt.user.gs_constants,
+                                  count,
+                                  fpme->vertex_size,
+                                  fpme->vertex_size);
    }
 
    if (draw_pt_post_vs_run( fpme->post_vs,
@@ -270,6 +290,7 @@ static boolean fetch_pipeline_linear_run_elts( struct draw_pt_middle_end *middle
    struct fetch_pipeline_middle_end *fpme = (struct fetch_pipeline_middle_end *)middle;
    struct draw_context *draw = fpme->draw;
    struct draw_vertex_shader *shader = draw->vs.vertex_shader;
+   struct draw_geometry_shader *geometry_shader = draw->gs.geometry_shader;
    unsigned opt = fpme->opt;
    unsigned alloc_count = align( count, 4 );
 
@@ -296,10 +317,19 @@ static boolean fetch_pipeline_linear_run_elts( struct draw_pt_middle_end *middle
       shader->run_linear(shader,
                         (const float (*)[4])pipeline_verts->data,
                         (      float (*)[4])pipeline_verts->data,
-                        (const float (*)[4])draw->pt.user.constants,
+                        (const float (*)[4])draw->pt.user.vs_constants,
                         count,
                         fpme->vertex_size,
                         fpme->vertex_size);
+
+      if (geometry_shader)
+         draw_geometry_shader_run(geometry_shader,
+                                  (const float (*)[4])pipeline_verts->data,
+                                  (      float (*)[4])pipeline_verts->data,
+                                  (const float (*)[4])draw->pt.user.gs_constants,
+                                  count,
+                                  fpme->vertex_size,
+                                  fpme->vertex_size);
    }
 
    if (draw_pt_post_vs_run( fpme->post_vs,
index 6c1cb48e8b82202da1638d2a7a9cdf9cbd40c247..55151823a14e4071bd539ccd208506ddc6feb6ff 100644 (file)
@@ -100,7 +100,7 @@ static boolean post_vs_cliptest_viewport_gl( struct pt_post_vs *pvs,
    struct vertex_header *out = vertices;
    const float *scale = pvs->draw->viewport.scale;
    const float *trans = pvs->draw->viewport.translate;
-   const unsigned pos = pvs->draw->vs.position_output;
+   const unsigned pos = draw_current_shader_position_output(pvs->draw);
    unsigned clipped = 0;
    unsigned j;
 
@@ -147,6 +147,39 @@ static boolean post_vs_cliptest_viewport_gl( struct pt_post_vs *pvs,
 
 
 
+/* As above plus edgeflags
+ */
+static boolean 
+post_vs_cliptest_viewport_gl_edgeflag(struct pt_post_vs *pvs,
+                                      struct vertex_header *vertices,
+                                      unsigned count,
+                                      unsigned stride )
+{
+   unsigned j;
+   boolean needpipe;
+
+   needpipe = post_vs_cliptest_viewport_gl( pvs, vertices, count, stride);
+
+   /* If present, copy edgeflag VS output into vertex header.
+    * Otherwise, leave header as is.
+    */
+   if (pvs->draw->vs.edgeflag_output) {
+      struct vertex_header *out = vertices;
+      int ef = pvs->draw->vs.edgeflag_output;
+
+      for (j = 0; j < count; j++) {
+         const float *edgeflag = out->data[ef];
+         out->edgeflag = !(edgeflag[0] != 1.0f);
+         needpipe |= !out->edgeflag;
+         out = (struct vertex_header *)( (char *)out + stride );
+      }
+   }
+   return needpipe;
+}
+
+
+
+
 /* If bypass_clipping is set, skip cliptest and rhw divide.
  */
 static boolean post_vs_viewport( struct pt_post_vs *pvs,
@@ -157,7 +190,7 @@ static boolean post_vs_viewport( struct pt_post_vs *pvs,
    struct vertex_header *out = vertices;
    const float *scale = pvs->draw->viewport.scale;
    const float *trans = pvs->draw->viewport.translate;
-   const unsigned pos = pvs->draw->vs.position_output;
+   const unsigned pos = draw_current_shader_position_output(pvs->draw);
    unsigned j;
 
    if (0) debug_printf("%s\n", __FUNCTION__);
@@ -201,17 +234,29 @@ boolean draw_pt_post_vs_run( struct pt_post_vs *pvs,
 void draw_pt_post_vs_prepare( struct pt_post_vs *pvs,
                              boolean bypass_clipping,
                              boolean bypass_viewport,
-                             boolean opengl )
+                             boolean opengl,
+                             boolean need_edgeflags )
 {
-   if (bypass_clipping) {
-      if (bypass_viewport)
-        pvs->run = post_vs_none;
-      else
-        pvs->run = post_vs_viewport;
+   if (!need_edgeflags) {
+      if (bypass_clipping) {
+         if (bypass_viewport)
+            pvs->run = post_vs_none;
+         else
+            pvs->run = post_vs_viewport;
+      }
+      else {
+         /* if (opengl) */
+         pvs->run = post_vs_cliptest_viewport_gl;
+      }
    }
    else {
-      /* if (opengl) */
-      pvs->run = post_vs_cliptest_viewport_gl;
+      /* If we need to copy edgeflags to the vertex header, it should
+       * mean we're running the primitive pipeline.  Hence the bypass
+       * flags should be false.
+       */
+      assert(!bypass_clipping);
+      assert(!bypass_viewport);
+      pvs->run = post_vs_cliptest_viewport_gl_edgeflag;
    }
 }
 
index b61fa291436c4a79abf9c954ec809df0f71dc03a..17c3b8cec269accbbcc8db9ea8fada7731fb2dcd 100644 (file)
@@ -50,16 +50,32 @@ void draw_pt_split_prim(unsigned prim, unsigned *first, unsigned *incr)
       *first = 2;
       *incr = 1;
       break;
+   case PIPE_PRIM_LINES_ADJACENCY:
+      *first = 4;
+      *incr = 2;
+      break;
+   case PIPE_PRIM_LINE_STRIP_ADJACENCY:
+      *first = 4;
+      *incr = 1;
+      break;
    case PIPE_PRIM_TRIANGLES:
       *first = 3;
       *incr = 3;
       break;
+   case PIPE_PRIM_TRIANGLES_ADJACENCY:
+      *first = 6;
+      *incr = 3;
+      break;
    case PIPE_PRIM_TRIANGLE_STRIP:
    case PIPE_PRIM_TRIANGLE_FAN:
    case PIPE_PRIM_POLYGON:
       *first = 3;
       *incr = 1;
       break;
+   case PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY:
+      *first = 6;
+      *incr = 1;
+      break;
    case PIPE_PRIM_QUADS:
       *first = 4;
       *incr = 4;
index 010c7a18a7c93e0c0b35546728ca3487f6285a72..f0aec5febabdb0b518cd626c5f28661c08a713d7 100644 (file)
@@ -36,6 +36,10 @@ static void FUNC(struct draw_pt_front_end *frontend,
    case PIPE_PRIM_TRIANGLE_STRIP:
    case PIPE_PRIM_QUADS:
    case PIPE_PRIM_QUAD_STRIP:
+   case PIPE_PRIM_LINES_ADJACENCY:
+   case PIPE_PRIM_LINE_STRIP_ADJACENCY:
+   case PIPE_PRIM_TRIANGLES_ADJACENCY:
+   case PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY:
       for (j = 0; j < count;) {
          unsigned remaining = count - j;
          unsigned nr = trim( MIN2(varray->driver_fetch_max, remaining), first, incr );
index d3f179ced18bb2431d912048825fad787b1f2914..757c48745456d59db25a07084cb1a691d36f07df 100644 (file)
@@ -346,7 +346,8 @@ vcache_check_run( struct draw_pt_front_end *frontend,
                        vcache->fetch_max,
                        draw_count);
       
-   if (max_index == 0xffffffff ||
+   if (max_index >= DRAW_PIPE_MAX_VERTICES ||
+       fetch_count >= UNDEFINED_VERTEX_ID ||
        fetch_count > draw_count) {
       if (0) debug_printf("fail\n");
       goto fail;
index 790e89ed820af7e7ab8d3c036b8f242a2c31931c..3553689532687ffaa326a5c93863d91f1e7f7c35 100644 (file)
@@ -101,6 +101,9 @@ draw_create_vertex_shader(struct draw_context *draw,
          if (vs->info.output_semantic_name[i] == TGSI_SEMANTIC_POSITION &&
              vs->info.output_semantic_index[i] == 0)
             vs->position_output = i;
+         else if (vs->info.output_semantic_name[i] == TGSI_SEMANTIC_EDGEFLAG &&
+             vs->info.output_semantic_index[i] == 0)
+            vs->edgeflag_output = i;
       }
    }
 
@@ -120,6 +123,7 @@ draw_bind_vertex_shader(struct draw_context *draw,
       draw->vs.vertex_shader = dvs;
       draw->vs.num_vs_outputs = dvs->info.num_outputs;
       draw->vs.position_output = dvs->position_output;
+      draw->vs.edgeflag_output = dvs->edgeflag_output;
       dvs->prepare( dvs, draw );
    }
    else {
index 89ae158751a931edc5b3c49ed50ebf0dcb3d0b38..e3b807ebd0e1df50ffdf6fa3c4aec523fb9e0b68 100644 (file)
@@ -107,6 +107,7 @@ struct draw_vertex_shader {
 
    struct tgsi_shader_info info;
    unsigned position_output;
+   unsigned edgeflag_output;
 
    /* Extracted from shader:
     */
index 88bc790b6211b79fae74455c16250fbc4e050eae..1aaae4ab7a4eed74924f4953d83557cb1ad82086 100644 (file)
@@ -361,8 +361,8 @@ static struct x86_reg aos_get_shader_reg_ptr( struct aos_compilation *cp,
 static struct x86_reg get_dst_ptr( struct aos_compilation *cp, 
                                    const struct tgsi_full_dst_register *dst )
 {
-   unsigned file = dst->DstRegister.File;
-   unsigned idx = dst->DstRegister.Index;
+   unsigned file = dst->Register.File;
+   unsigned idx = dst->Register.Index;
    unsigned i;
    
 
@@ -529,8 +529,8 @@ static struct x86_reg fetch_src( struct aos_compilation *cp,
                                  const struct tgsi_full_src_register *src ) 
 {
    struct x86_reg arg0 = aos_get_shader_reg(cp, 
-                                            src->SrcRegister.File, 
-                                            src->SrcRegister.Index);
+                                            src->Register.File, 
+                                            src->Register.Index);
    unsigned i;
    ubyte swz = 0;
    unsigned negs = 0;
@@ -620,8 +620,8 @@ static void x87_fld_src( struct aos_compilation *cp,
                          unsigned channel ) 
 {
    struct x86_reg arg0 = aos_get_shader_reg_ptr(cp, 
-                                                src->SrcRegister.File, 
-                                                src->SrcRegister.Index);
+                                                src->Register.File, 
+                                                src->Register.Index);
 
    unsigned swizzle = tgsi_util_get_full_src_register_swizzle( src, channel );
    unsigned neg = tgsi_util_get_full_src_register_sign_mode( src, channel );
@@ -669,15 +669,15 @@ static void store_dest( struct aos_compilation *cp,
 {
    struct x86_reg dst;
 
-   switch (reg->DstRegister.WriteMask) {
+   switch (reg->Register.WriteMask) {
    case 0:
       return;
    
    case TGSI_WRITEMASK_XYZW:
       aos_adopt_xmm_reg(cp, 
                         get_xmm_writable(cp, result), 
-                        reg->DstRegister.File,
-                        reg->DstRegister.Index,
+                        reg->Register.File,
+                        reg->Register.Index,
                         TRUE);
       return;
    default: 
@@ -685,10 +685,10 @@ static void store_dest( struct aos_compilation *cp,
    }
 
    dst = aos_get_shader_reg_xmm(cp, 
-                                reg->DstRegister.File,
-                                reg->DstRegister.Index);
+                                reg->Register.File,
+                                reg->Register.Index);
 
-   switch (reg->DstRegister.WriteMask) {
+   switch (reg->Register.WriteMask) {
    case TGSI_WRITEMASK_X:
       sse_movss(cp->func, dst, get_xmm(cp, result));
       break;
@@ -710,14 +710,14 @@ static void store_dest( struct aos_compilation *cp,
       break;
 
    default:
-      mask_write(cp, dst, result, reg->DstRegister.WriteMask);
+      mask_write(cp, dst, result, reg->Register.WriteMask);
       break;
    }
 
    aos_adopt_xmm_reg(cp, 
                      dst, 
-                     reg->DstRegister.File,
-                     reg->DstRegister.Index,
+                     reg->Register.File,
+                     reg->Register.Index,
                      TRUE);
 
 }
@@ -737,7 +737,7 @@ static void store_scalar_dest( struct aos_compilation *cp,
                                const struct tgsi_full_dst_register *reg,
                                struct x86_reg result )
 {
-   unsigned writemask = reg->DstRegister.WriteMask;
+   unsigned writemask = reg->Register.WriteMask;
    struct x86_reg dst;
 
    if (writemask != TGSI_WRITEMASK_X &&
@@ -754,12 +754,12 @@ static void store_scalar_dest( struct aos_compilation *cp,
 
    result = get_xmm(cp, result);
    dst = aos_get_shader_reg_xmm(cp, 
-                                reg->DstRegister.File,
-                                reg->DstRegister.Index);
+                                reg->Register.File,
+                                reg->Register.Index);
 
 
 
-   switch (reg->DstRegister.WriteMask) {
+   switch (reg->Register.WriteMask) {
    case TGSI_WRITEMASK_X:
       sse_movss(cp->func, dst, result);
       break;
@@ -782,8 +782,8 @@ static void store_scalar_dest( struct aos_compilation *cp,
 
    aos_adopt_xmm_reg(cp, 
                      dst, 
-                     reg->DstRegister.File,
-                     reg->DstRegister.Index,
+                     reg->Register.File,
+                     reg->Register.Index,
                      TRUE);
 }
    
@@ -819,7 +819,7 @@ static void x87_fstp_dest4( struct aos_compilation *cp,
                             const struct tgsi_full_dst_register *dst )
 {
    struct x86_reg ptr = get_dst_ptr(cp, dst); 
-   unsigned writemask = dst->DstRegister.WriteMask;
+   unsigned writemask = dst->Register.WriteMask;
 
    x87_fst_or_nop(cp->func, writemask, 0, ptr);
    x87_fst_or_nop(cp->func, writemask, 1, ptr);
@@ -956,7 +956,7 @@ static void emit_print( struct aos_compilation *cp,
 
 static boolean emit_ABS( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
    struct x86_reg neg = aos_get_internal(cp, IMM_NEGS);
    struct x86_reg tmp = aos_get_xmm_reg(cp);
 
@@ -964,27 +964,27 @@ static boolean emit_ABS( struct aos_compilation *cp, const struct tgsi_full_inst
    sse_mulps(cp->func, tmp, neg);
    sse_maxps(cp->func, tmp, arg0);
    
-   store_dest(cp, &op->FullDstRegisters[0], tmp);
+   store_dest(cp, &op->Dst[0], tmp);
    return TRUE;
 }
 
 static boolean emit_ADD( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg dst = get_xmm_writable(cp, arg0);
 
    sse_addps(cp->func, dst, arg1);
 
-   store_dest(cp, &op->FullDstRegisters[0], dst);
+   store_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 static boolean emit_COS( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
-   x87_fld_src(cp, &op->FullSrcRegisters[0], 0);
+   x87_fld_src(cp, &op->Src[0], 0);
    x87_fcos(cp->func);
-   x87_fstp_dest4(cp, &op->FullDstRegisters[0]);
+   x87_fstp_dest4(cp, &op->Dst[0]);
    return TRUE;
 }
 
@@ -993,8 +993,8 @@ static boolean emit_COS( struct aos_compilation *cp, const struct tgsi_full_inst
  */
 static boolean emit_DP3( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg tmp = aos_get_xmm_reg(cp); 
    struct x86_reg dst = get_xmm_writable(cp, arg0);
 
@@ -1007,14 +1007,14 @@ static boolean emit_DP3( struct aos_compilation *cp, const struct tgsi_full_inst
    sse_addss(cp->func, dst, tmp);
    
    aos_release_xmm_reg(cp, tmp.idx);
-   store_scalar_dest(cp, &op->FullDstRegisters[0], dst);
+   store_scalar_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 static boolean emit_DP4( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg tmp = aos_get_xmm_reg(cp);      
    struct x86_reg dst = get_xmm_writable(cp, arg0);
 
@@ -1028,14 +1028,14 @@ static boolean emit_DP4( struct aos_compilation *cp, const struct tgsi_full_inst
    sse_addss(cp->func, dst, tmp);
 
    aos_release_xmm_reg(cp, tmp.idx);
-   store_scalar_dest(cp, &op->FullDstRegisters[0], dst);
+   store_scalar_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 static boolean emit_DPH( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg tmp = aos_get_xmm_reg(cp);
    struct x86_reg dst = get_xmm_writable(cp, arg0);
 
@@ -1051,14 +1051,14 @@ static boolean emit_DPH( struct aos_compilation *cp, const struct tgsi_full_inst
    sse_addss(cp->func, dst, tmp);
 
    aos_release_xmm_reg(cp, tmp.idx);
-   store_scalar_dest(cp, &op->FullDstRegisters[0], dst);
+   store_scalar_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 static boolean emit_DST( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-    struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-    struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+    struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+    struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
     struct x86_reg dst = aos_get_xmm_reg(cp);
     struct x86_reg tmp = aos_get_xmm_reg(cp);
     struct x86_reg ones = aos_get_internal(cp, IMM_ONES);
@@ -1073,25 +1073,25 @@ static boolean emit_DST( struct aos_compilation *cp, const struct tgsi_full_inst
     sse_mulps(cp->func, dst, tmp);
 
     aos_release_xmm_reg(cp, tmp.idx);
-    store_dest(cp, &op->FullDstRegisters[0], dst);
+    store_dest(cp, &op->Dst[0], dst);
     return TRUE;
 }
 
 static boolean emit_LG2( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
    x87_fld1(cp->func);         /* 1 */
-   x87_fld_src(cp, &op->FullSrcRegisters[0], 0);       /* a0 1 */
+   x87_fld_src(cp, &op->Src[0], 0);    /* a0 1 */
    x87_fyl2x(cp->func);        /* log2(a0) */
-   x87_fstp_dest4(cp, &op->FullDstRegisters[0]);
+   x87_fstp_dest4(cp, &op->Dst[0]);
    return TRUE;
 }
 
 #if 0
 static boolean emit_EX2( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
-   x87_fld_src(cp, &op->FullSrcRegisters[0], 0);
+   x87_fld_src(cp, &op->Src[0], 0);
    x87_emit_ex2(cp);
-   x87_fstp_dest4(cp, &op->FullDstRegisters[0]);
+   x87_fstp_dest4(cp, &op->Dst[0]);
    return TRUE;
 }
 #endif
@@ -1099,8 +1099,8 @@ static boolean emit_EX2( struct aos_compilation *cp, const struct tgsi_full_inst
 
 static boolean emit_FLR( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
-   struct x86_reg dst = get_dst_ptr(cp, &op->FullDstRegisters[0]); 
-   unsigned writemask = op->FullDstRegisters[0].DstRegister.WriteMask;
+   struct x86_reg dst = get_dst_ptr(cp, &op->Dst[0]); 
+   unsigned writemask = op->Dst[0].Register.WriteMask;
    int i;
 
    set_fpu_round_neg_inf( cp );
@@ -1109,7 +1109,7 @@ static boolean emit_FLR( struct aos_compilation *cp, const struct tgsi_full_inst
     */
    for (i = 3; i >= 0; i--) {
       if (writemask & (1<<i)) {
-         x87_fld_src(cp, &op->FullSrcRegisters[0], i);   
+         x87_fld_src(cp, &op->Src[0], i);   
       }
    }
 
@@ -1126,8 +1126,8 @@ static boolean emit_FLR( struct aos_compilation *cp, const struct tgsi_full_inst
 
 static boolean emit_RND( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
-   struct x86_reg dst = get_dst_ptr(cp, &op->FullDstRegisters[0]); 
-   unsigned writemask = op->FullDstRegisters[0].DstRegister.WriteMask;
+   struct x86_reg dst = get_dst_ptr(cp, &op->Dst[0]); 
+   unsigned writemask = op->Dst[0].Register.WriteMask;
    int i;
 
    set_fpu_round_nearest( cp );
@@ -1136,7 +1136,7 @@ static boolean emit_RND( struct aos_compilation *cp, const struct tgsi_full_inst
     */
    for (i = 3; i >= 0; i--) {
       if (writemask & (1<<i)) {
-         x87_fld_src(cp, &op->FullSrcRegisters[0], i);   
+         x87_fld_src(cp, &op->Src[0], i);   
       }
    }
 
@@ -1153,10 +1153,10 @@ static boolean emit_RND( struct aos_compilation *cp, const struct tgsi_full_inst
 
 static boolean emit_FRC( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
-   struct x86_reg dst = get_dst_ptr(cp, &op->FullDstRegisters[0]); 
+   struct x86_reg dst = get_dst_ptr(cp, &op->Dst[0]); 
    struct x86_reg st0 = x86_make_reg(file_x87, 0);
    struct x86_reg st1 = x86_make_reg(file_x87, 1);
-   unsigned writemask = op->FullDstRegisters[0].DstRegister.WriteMask;
+   unsigned writemask = op->Dst[0].Register.WriteMask;
    int i;
 
    set_fpu_round_neg_inf( cp );
@@ -1166,7 +1166,7 @@ static boolean emit_FRC( struct aos_compilation *cp, const struct tgsi_full_inst
     */
    for (i = 3; i >= 0; i--) {
       if (writemask & (1<<i)) {
-         x87_fld_src(cp, &op->FullSrcRegisters[0], i);   
+         x87_fld_src(cp, &op->Src[0], i);   
       }
    }
 
@@ -1190,7 +1190,7 @@ static boolean emit_FRC( struct aos_compilation *cp, const struct tgsi_full_inst
 static boolean emit_LIT( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
    struct x86_reg ecx = x86_make_reg( file_REG32, reg_CX );
-   unsigned writemask = op->FullDstRegisters[0].DstRegister.WriteMask;
+   unsigned writemask = op->Dst[0].Register.WriteMask;
    unsigned lit_count = cp->lit_count++;
    struct x86_reg result, arg0;
    unsigned i;
@@ -1209,10 +1209,10 @@ static boolean emit_LIT( struct aos_compilation *cp, const struct tgsi_full_inst
    if (writemask != TGSI_WRITEMASK_XYZW) 
       result = x86_make_disp(cp->machine_EDX, Offset(struct aos_machine, tmp[0]));
    else 
-      result = get_dst_ptr(cp, &op->FullDstRegisters[0]);    
+      result = get_dst_ptr(cp, &op->Dst[0]);    
 
    
-   arg0 = fetch_src( cp, &op->FullSrcRegisters[0] );
+   arg0 = fetch_src( cp, &op->Src[0] );
    if (arg0.file == file_XMM) {
       struct x86_reg tmp = x86_make_disp(cp->machine_EDX, 
                                          Offset(struct aos_machine, tmp[1]));
@@ -1259,7 +1259,7 @@ static boolean emit_LIT( struct aos_compilation *cp, const struct tgsi_full_inst
 
    if (writemask != TGSI_WRITEMASK_XYZW) {
       store_dest( cp, 
-                  &op->FullDstRegisters[0],
+                  &op->Dst[0],
                   get_xmm_writable( cp, result ) );
    }
 
@@ -1269,8 +1269,8 @@ static boolean emit_LIT( struct aos_compilation *cp, const struct tgsi_full_inst
 #if 0   
 static boolean emit_inline_LIT( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg dst = get_dst_ptr(cp, &op->FullDstRegisters[0]); 
-   unsigned writemask = op->FullDstRegisters[0].DstRegister.WriteMask;
+   struct x86_reg dst = get_dst_ptr(cp, &op->Dst[0]); 
+   unsigned writemask = op->Dst[0].Register.WriteMask;
 
    if (writemask & TGSI_WRITEMASK_YZ) {
       struct x86_reg st1 = x86_make_reg(file_x87, 1);
@@ -1286,13 +1286,13 @@ static boolean emit_inline_LIT( struct aos_compilation *cp, const struct tgsi_fu
        */
       x87_fldz(cp->func);                           /* 1 0  */
 #endif
-      x87_fld_src(cp, &op->FullSrcRegisters[0], 1); /* a1 1 0  */
+      x87_fld_src(cp, &op->Src[0], 1); /* a1 1 0  */
       x87_fcomi(cp->func, st2);                            /* a1 1 0  */
       x87_fcmovb(cp->func, st1);                    /* a1' 1 0  */
       x87_fstp(cp->func, st1);                      /* a1' 0  */
       x87_fstp(cp->func, st1);                      /* a1'  */
 
-      x87_fld_src(cp, &op->FullSrcRegisters[0], 3); /* a3 a1'  */
+      x87_fld_src(cp, &op->Src[0], 3); /* a3 a1'  */
       x87_fxch(cp->func, st1);                      /* a1' a3  */
       
 
@@ -1305,7 +1305,7 @@ static boolean emit_inline_LIT( struct aos_compilation *cp, const struct tgsi_fu
       /* a0' = max2(a0, 0):
        */
       x87_fldz(cp->func);                           /* 0 r2 */
-      x87_fld_src(cp, &op->FullSrcRegisters[0], 0); /* a0 0 r2 */
+      x87_fld_src(cp, &op->Src[0], 0); /* a0 0 r2 */
       x87_fcomi(cp->func, st1);        
       x87_fcmovb(cp->func, st1);                    /* a0' 0 r2 */
 
@@ -1333,58 +1333,58 @@ static boolean emit_inline_LIT( struct aos_compilation *cp, const struct tgsi_fu
 
 static boolean emit_MAX( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg dst = get_xmm_writable(cp, arg0);
 
    sse_maxps(cp->func, dst, arg1);
 
-   store_dest(cp, &op->FullDstRegisters[0], dst);
+   store_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 
 static boolean emit_MIN( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg dst = get_xmm_writable(cp, arg0);
 
    sse_minps(cp->func, dst, arg1);
 
-   store_dest(cp, &op->FullDstRegisters[0], dst);
+   store_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 static boolean emit_MOV( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
    struct x86_reg dst = get_xmm_writable(cp, arg0);
 
    /* potentially nothing to do */
 
-   store_dest(cp, &op->FullDstRegisters[0], dst);
+   store_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 static boolean emit_MUL( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg dst = get_xmm_writable(cp, arg0);
 
    sse_mulps(cp->func, dst, arg1);
 
-   store_dest(cp, &op->FullDstRegisters[0], dst);
+   store_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 
 static boolean emit_MAD( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
-   struct x86_reg arg2 = fetch_src(cp, &op->FullSrcRegisters[2]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
+   struct x86_reg arg2 = fetch_src(cp, &op->Src[2]);
 
    /* If we can't clobber old contents of arg0, get a temporary & copy
     * it there, then clobber it...
@@ -1393,7 +1393,7 @@ static boolean emit_MAD( struct aos_compilation *cp, const struct tgsi_full_inst
 
    sse_mulps(cp->func, arg0, arg1);
    sse_addps(cp->func, arg0, arg2);
-   store_dest(cp, &op->FullDstRegisters[0], arg0);
+   store_dest(cp, &op->Dst[0], arg0);
    return TRUE;
 }
 
@@ -1425,13 +1425,13 @@ static float PIPE_CDECL _exp2(float x)
 static boolean emit_POW( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
 #if 0
-   x87_fld_src(cp, &op->FullSrcRegisters[1], 0);  /* a1.x */
-   x87_fld_src(cp, &op->FullSrcRegisters[0], 0);       /* a0.x a1.x */
+   x87_fld_src(cp, &op->Src[1], 0);  /* a1.x */
+   x87_fld_src(cp, &op->Src[0], 0);    /* a0.x a1.x */
    x87_fyl2x(cp->func);                                        /* a1*log2(a0) */
 
    x87_emit_ex2( cp );         /* 2^(a1*log2(a0)) */
 
-   x87_fstp_dest4(cp, &op->FullDstRegisters[0]);
+   x87_fstp_dest4(cp, &op->Dst[0]);
 #else
    uint i;
 
@@ -1450,9 +1450,9 @@ static boolean emit_POW( struct aos_compilation *cp, const struct tgsi_full_inst
 
    x86_lea( cp->func, cp->stack_ESP, x86_make_disp(cp->stack_ESP, -8) );
 
-   x87_fld_src( cp, &op->FullSrcRegisters[1], 0 );
+   x87_fld_src( cp, &op->Src[1], 0 );
    x87_fstp( cp->func, x86_make_disp( cp->stack_ESP, 4 ) );
-   x87_fld_src( cp, &op->FullSrcRegisters[0], 0 );
+   x87_fld_src( cp, &op->Src[0], 0 );
    x87_fstp( cp->func, x86_make_disp( cp->stack_ESP, 0 ) );
 
    /* tmp_EAX has been pushed & will be restored below */
@@ -1467,7 +1467,7 @@ static boolean emit_POW( struct aos_compilation *cp, const struct tgsi_full_inst
     */
    cp->func->x87_stack++;
 
-   x87_fstp_dest4( cp, &op->FullDstRegisters[0] );
+   x87_fstp_dest4( cp, &op->Dst[0] );
 #endif
    return TRUE;
 }
@@ -1493,7 +1493,7 @@ static boolean emit_EXPBASE2( struct aos_compilation *cp, const struct tgsi_full
 
    x86_lea( cp->func, cp->stack_ESP, x86_make_disp(cp->stack_ESP, -4) );
 
-   x87_fld_src( cp, &op->FullSrcRegisters[0], 0 );
+   x87_fld_src( cp, &op->Src[0], 0 );
    x87_fstp( cp->func, x86_make_disp( cp->stack_ESP, 0 ) );
 
    /* tmp_EAX has been pushed & will be restored below */
@@ -1508,7 +1508,7 @@ static boolean emit_EXPBASE2( struct aos_compilation *cp, const struct tgsi_full
     */
    cp->func->x87_stack++;
 
-   x87_fstp_dest4( cp, &op->FullDstRegisters[0] );
+   x87_fstp_dest4( cp, &op->Dst[0] );
 
    return TRUE;
 }
@@ -1517,7 +1517,7 @@ static boolean emit_EXPBASE2( struct aos_compilation *cp, const struct tgsi_full
 
 static boolean emit_RCP( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
    struct x86_reg dst = aos_get_xmm_reg(cp);
 
    if (cp->have_sse2) {
@@ -1531,7 +1531,7 @@ static boolean emit_RCP( struct aos_compilation *cp, const struct tgsi_full_inst
       sse_divss(cp->func, dst, arg0);
    }
 
-   store_scalar_dest(cp, &op->FullDstRegisters[0], dst);
+   store_scalar_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
@@ -1551,14 +1551,14 @@ static boolean emit_RCP( struct aos_compilation *cp, const struct tgsi_full_inst
 static boolean emit_RSQ( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
    if (0) {
-      struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
+      struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
       struct x86_reg r = aos_get_xmm_reg(cp);
       sse_rsqrtss(cp->func, r, arg0);
-      store_scalar_dest(cp, &op->FullDstRegisters[0], r);
+      store_scalar_dest(cp, &op->Dst[0], r);
       return TRUE;
    }
    else {
-      struct x86_reg arg0           = fetch_src(cp, &op->FullSrcRegisters[0]);
+      struct x86_reg arg0           = fetch_src(cp, &op->Src[0]);
       struct x86_reg r              = aos_get_xmm_reg(cp);
 
       struct x86_reg neg_half       = get_reg_ptr( cp, AOS_FILE_INTERNAL, IMM_RSQ );
@@ -1578,7 +1578,7 @@ static boolean emit_RSQ( struct aos_compilation *cp, const struct tgsi_full_inst
       sse_addss(   cp->func, tmp, one_point_five ); /* 1.5 - .5 * a * r * r */
       sse_mulss(   cp->func, r,  tmp );             /* r * (1.5 - .5 * a * r * r) */
 
-      store_scalar_dest(cp, &op->FullDstRegisters[0], r);
+      store_scalar_dest(cp, &op->Dst[0], r);
 
       aos_release_xmm_reg(cp, tmp.idx);
 
@@ -1589,23 +1589,23 @@ static boolean emit_RSQ( struct aos_compilation *cp, const struct tgsi_full_inst
 
 static boolean emit_SGE( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg ones = aos_get_internal(cp, IMM_ONES);
    struct x86_reg dst = get_xmm_writable(cp, arg0);
 
    sse_cmpps(cp->func, dst, arg1, cc_NotLessThan);
    sse_andps(cp->func, dst, ones);
 
-   store_dest(cp, &op->FullDstRegisters[0], dst);
+   store_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 static boolean emit_SIN( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
-   x87_fld_src(cp, &op->FullSrcRegisters[0], 0);
+   x87_fld_src(cp, &op->Src[0], 0);
    x87_fsin(cp->func);
-   x87_fstp_dest4(cp, &op->FullDstRegisters[0]);
+   x87_fstp_dest4(cp, &op->Dst[0]);
    return TRUE;
 }
 
@@ -1613,46 +1613,46 @@ static boolean emit_SIN( struct aos_compilation *cp, const struct tgsi_full_inst
 
 static boolean emit_SLT( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg ones = aos_get_internal(cp, IMM_ONES);
    struct x86_reg dst = get_xmm_writable(cp, arg0);
    
    sse_cmpps(cp->func, dst, arg1, cc_LessThan);
    sse_andps(cp->func, dst, ones);
 
-   store_dest(cp, &op->FullDstRegisters[0], dst);
+   store_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 static boolean emit_SUB( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg dst = get_xmm_writable(cp, arg0);
 
    sse_subps(cp->func, dst, arg1);
 
-   store_dest(cp, &op->FullDstRegisters[0], dst);
+   store_dest(cp, &op->Dst[0], dst);
    return TRUE;
 }
 
 static boolean emit_TRUNC( struct aos_compilation *cp, const struct tgsi_full_instruction *op )
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
    struct x86_reg tmp0 = aos_get_xmm_reg(cp);
 
    sse2_cvttps2dq(cp->func, tmp0, arg0);
    sse2_cvtdq2ps(cp->func, tmp0, tmp0);
 
-   store_dest(cp, &op->FullDstRegisters[0], tmp0);
+   store_dest(cp, &op->Dst[0], tmp0);
    return TRUE;
 }
 
 static boolean emit_XPD( struct aos_compilation *cp, const struct tgsi_full_instruction *op ) 
 {
-   struct x86_reg arg0 = fetch_src(cp, &op->FullSrcRegisters[0]);
-   struct x86_reg arg1 = fetch_src(cp, &op->FullSrcRegisters[1]);
+   struct x86_reg arg0 = fetch_src(cp, &op->Src[0]);
+   struct x86_reg arg1 = fetch_src(cp, &op->Src[1]);
    struct x86_reg tmp0 = aos_get_xmm_reg(cp);
    struct x86_reg tmp1 = aos_get_xmm_reg(cp);
 
@@ -1670,7 +1670,7 @@ static boolean emit_XPD( struct aos_compilation *cp, const struct tgsi_full_inst
 
 
    aos_release_xmm_reg(cp, tmp0.idx);
-   store_dest(cp, &op->FullDstRegisters[0], tmp1);
+   store_dest(cp, &op->Dst[0], tmp1);
    return TRUE;
 }
 
@@ -1897,10 +1897,10 @@ static void find_last_write_outputs( struct aos_compilation *cp )
          continue;
 
       for (i = 0; i < TGSI_FULL_MAX_DST_REGISTERS; i++) {
-         if (parse.FullToken.FullInstruction.FullDstRegisters[i].DstRegister.File ==
+         if (parse.FullToken.FullInstruction.Dst[i].Register.File ==
              TGSI_FILE_OUTPUT) 
          {
-            unsigned idx = parse.FullToken.FullInstruction.FullDstRegisters[i].DstRegister.Index;
+            unsigned idx = parse.FullToken.FullInstruction.Dst[i].Register.Index;
             cp->output_last_write[idx] = this_instruction;
          }
       }
index 7ee567d4789823fb1a0a645e8aa87cb997e85d0e..d16692584e5236fc2874c99c37954c534693da21 100644 (file)
@@ -147,11 +147,12 @@ static void PIPE_CDECL vsvg_run_elts( struct draw_vs_varient *varient,
    vsvg->base.vs->run_linear( vsvg->base.vs, 
                               temp_buffer,
                               temp_buffer,
-                              (const float (*)[4])vsvg->base.vs->draw->pt.user.constants,
+                              (const float (*)[4])vsvg->base.vs->draw->pt.user.vs_constants,
                               count,
                               temp_vertex_stride, 
                               temp_vertex_stride);
 
+   /* FIXME: geometry shading? */
 
    if (vsvg->base.key.clip) {
       /* not really handling clipping, just do the rhw so we can
@@ -207,7 +208,7 @@ static void PIPE_CDECL vsvg_run_linear( struct draw_vs_varient *varient,
    vsvg->base.vs->run_linear( vsvg->base.vs, 
                               temp_buffer,
                               temp_buffer,
-                              (const float (*)[4])vsvg->base.vs->draw->pt.user.constants,
+                              (const float (*)[4])vsvg->base.vs->draw->pt.user.vs_constants,
                               count,
                               temp_vertex_stride, 
                               temp_vertex_stride);
index bf84401e11213425c77fe40724a4120b9cee3799..5cafe8c3f0c3ec2e02bdd948ea745de38cfe27a0 100644 (file)
@@ -94,8 +94,8 @@ translate_declaration(struct gallivm_ir *prog,
       unsigned first, last, mask;
       uint interp_method;
 
-      first = decl->DeclarationRange.First;
-      last = decl->DeclarationRange.Last;
+      first = decl->Range.First;
+      last = decl->Range.Last;
       mask = decl->Declaration.UsageMask;
 
       /* Do not touch WPOS.xy */
@@ -149,7 +149,7 @@ translate_declarationir(struct gallivm_ir *,
                       struct tgsi_full_declaration *)
 {
    if (decl->Declaration.File == TGSI_FILE_ADDRESS) {
-      int idx = decl->DeclarationRange.First;
+      int idx = decl->Range.First;
       storage->addAddress(idx);
    }
 }
@@ -234,26 +234,26 @@ translate_instruction(llvm::Module *module,
    inputs[3] = 0;
 
    for (int i = 0; i < inst->Instruction.NumSrcRegs; ++i) {
-      struct tgsi_full_src_register *src = &inst->FullSrcRegisters[i];
+      struct tgsi_full_src_register *src = &inst->Src[i];
       llvm::Value *val = 0;
       llvm::Value *indIdx = 0;
 
-      if (src->SrcRegister.Indirect) {
-         indIdx = storage->addrElement(src->SrcRegisterInd.Index);
+      if (src->Register.Indirect) {
+         indIdx = storage->addrElement(src->Indirect.Index);
          indIdx = storage->extractIndex(indIdx);
       }
-      if (src->SrcRegister.File == TGSI_FILE_CONSTANT) {
-         val = storage->constElement(src->SrcRegister.Index, indIdx);
-      } else if (src->SrcRegister.File == TGSI_FILE_INPUT) {
-         val = storage->inputElement(src->SrcRegister.Index, indIdx);
-      } else if (src->SrcRegister.File == TGSI_FILE_TEMPORARY) {
-         val = storage->tempElement(src->SrcRegister.Index);
-      } else if (src->SrcRegister.File == TGSI_FILE_OUTPUT) {
-         val = storage->outputElement(src->SrcRegister.Index, indIdx);
-      } else if (src->SrcRegister.File == TGSI_FILE_IMMEDIATE) {
-         val = storage->immediateElement(src->SrcRegister.Index);
+      if (src->Register.File == TGSI_FILE_CONSTANT) {
+         val = storage->constElement(src->Register.Index, indIdx);
+      } else if (src->Register.File == TGSI_FILE_INPUT) {
+         val = storage->inputElement(src->Register.Index, indIdx);
+      } else if (src->Register.File == TGSI_FILE_TEMPORARY) {
+         val = storage->tempElement(src->Register.Index);
+      } else if (src->Register.File == TGSI_FILE_OUTPUT) {
+         val = storage->outputElement(src->Register.Index, indIdx);
+      } else if (src->Register.File == TGSI_FILE_IMMEDIATE) {
+         val = storage->immediateElement(src->Register.Index);
       } else {
-         fprintf(stderr, "ERROR: not supported llvm source %d\n", src->SrcRegister.File);
+         fprintf(stderr, "ERROR: not supported llvm source %d\n", src->Register.File);
          return;
       }
 
@@ -656,14 +656,14 @@ translate_instruction(llvm::Module *module,
 
    /* store results  */
    for (int i = 0; i < inst->Instruction.NumDstRegs; ++i) {
-      struct tgsi_full_dst_register *dst = &inst->FullDstRegisters[i];
-
-      if (dst->DstRegister.File == TGSI_FILE_OUTPUT) {
-         storage->setOutputElement(dst->DstRegister.Index, out, dst->DstRegister.WriteMask);
-      } else if (dst->DstRegister.File == TGSI_FILE_TEMPORARY) {
-         storage->setTempElement(dst->DstRegister.Index, out, dst->DstRegister.WriteMask);
-      } else if (dst->DstRegister.File == TGSI_FILE_ADDRESS) {
-         storage->setAddrElement(dst->DstRegister.Index, out, dst->DstRegister.WriteMask);
+      struct tgsi_full_dst_register *dst = &inst->Dst[i];
+
+      if (dst->Register.File == TGSI_FILE_OUTPUT) {
+         storage->setOutputElement(dst->Register.Index, out, dst->Register.WriteMask);
+      } else if (dst->Register.File == TGSI_FILE_TEMPORARY) {
+         storage->setTempElement(dst->Register.Index, out, dst->Register.WriteMask);
+      } else if (dst->Register.File == TGSI_FILE_ADDRESS) {
+         storage->setAddrElement(dst->Register.Index, out, dst->Register.WriteMask);
       } else {
          fprintf(stderr, "ERROR: unsupported LLVM destination!");
          assert(!"wrong destination");
@@ -683,16 +683,16 @@ translate_instructionir(llvm::Module *module,
    std::vector< std::vector<llvm::Value*> > inputs(inst->Instruction.NumSrcRegs);
 
    for (int i = 0; i < inst->Instruction.NumSrcRegs; ++i) {
-      struct tgsi_full_src_register *src = &inst->FullSrcRegisters[i];
+      struct tgsi_full_src_register *src = &inst->Src[i];
       std::vector<llvm::Value*> val;
       llvm::Value *indIdx = 0;
       int swizzle = swizzleInt(src);
 
-      if (src->SrcRegister.Indirect) {
-         indIdx = storage->addrElement(src->SrcRegisterInd.Index);
+      if (src->Register.Indirect) {
+         indIdx = storage->addrElement(src->Indirect.Index);
       }
-      val = storage->load((enum tgsi_file_type)src->SrcRegister.File,
-                          src->SrcRegister.Index, swizzle, instr->getIRBuilder(), indIdx);
+      val = storage->load((enum tgsi_file_type)src->Register.File,
+                          src->Register.Index, swizzle, instr->getIRBuilder(), indIdx);
 
       inputs[i] = val;
    }
@@ -993,9 +993,9 @@ translate_instructionir(llvm::Module *module,
 
    /* store results  */
    for (int i = 0; i < inst->Instruction.NumDstRegs; ++i) {
-      struct tgsi_full_dst_register *dst = &inst->FullDstRegisters[i];
-      storage->store((enum tgsi_file_type)dst->DstRegister.File,
-                     dst->DstRegister.Index, out, dst->DstRegister.WriteMask,
+      struct tgsi_full_dst_register *dst = &inst->Dst[i];
+      storage->store((enum tgsi_file_type)dst->Register.File,
+                     dst->Register.Index, out, dst->Register.WriteMask,
                     instr->getIRBuilder() );
    }
 }
index 4ef372233f0dbd50db920959aa042f293833a1b6..eb7e84be84863c1a3336ccde8dacd92bde9cd294 100644 (file)
@@ -237,8 +237,9 @@ pb_reference(struct pb_buffer **dst,
 {
    struct pb_buffer *old = *dst;
 
-   if (pipe_reference((struct pipe_reference**)dst, &src->base.reference))
+   if (pipe_reference(&(*dst)->base.reference, &src->base.reference))
       pb_destroy( old );
+   *dst = src;
 }
 
 
index 2ef4293d4d7bfb9529018ed2b62ecbd9b22eb2dd..a9375abd21880e52fef1cec4910ee8f38adfc578 100644 (file)
@@ -1,6 +1,6 @@
 /**************************************************************************
  *
- * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * Copyright 2007-2009 VMware, Inc.
  * All Rights Reserved.
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
@@ -18,7 +18,7 @@
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
- * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * IN NO EVENT SHALL VMWARE AND/OR ITS SUPPLIERS BE LIABLE FOR
  * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
  * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
  * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
@@ -80,11 +80,27 @@ struct fenced_buffer_list
  */
 struct fenced_buffer
 {
+   /* 
+    * Immutable members.
+    */
+
    struct pb_buffer base;
-   
    struct pb_buffer *buffer;
+   struct fenced_buffer_list *list;
+
+   /**
+    * Protected by fenced_buffer_list::mutex
+    */
+   struct list_head head;
 
-   /* FIXME: protect access with mutex */
+   /**
+    * Following members are mutable and protected by this mutex.
+    * 
+    * You may lock this mutex alone, or lock it with fenced_buffer_list::mutex
+    * held, but in order to prevent deadlocks you must never lock 
+    * fenced_buffer_list::mutex with this mutex held.
+    */
+   pipe_mutex mutex;
 
    /**
     * A bitmask of PIPE_BUFFER_USAGE_CPU/GPU_READ/WRITE describing the current
@@ -96,9 +112,6 @@ struct fenced_buffer
    struct pb_validate *vl;
    unsigned validation_flags;
    struct pipe_fence_handle *fence;
-
-   struct list_head head;
-   struct fenced_buffer_list *list;
 };
 
 
@@ -110,15 +123,24 @@ fenced_buffer(struct pb_buffer *buf)
 }
 
 
+/**
+ * Add the buffer to the fenced list.
+ * 
+ * fenced_buffer_list::mutex and fenced_buffer::mutex must be held, in this 
+ * order before calling this function.
+ * 
+ * Reference count should be incremented before calling this function.
+ */
 static INLINE void
-_fenced_buffer_add(struct fenced_buffer *fenced_buf)
+fenced_buffer_add_locked(struct fenced_buffer_list *fenced_list, 
+                         struct fenced_buffer *fenced_buf)
 {
-   struct fenced_buffer_list *fenced_list = fenced_buf->list;
-
    assert(pipe_is_referenced(&fenced_buf->base.base.reference));
    assert(fenced_buf->flags & PIPE_BUFFER_USAGE_GPU_READ_WRITE);
    assert(fenced_buf->fence);
 
+   /* TODO: Move the reference count increment here */
+   
 #ifdef DEBUG
    LIST_DEL(&fenced_buf->head);
    assert(fenced_list->numUnfenced);
@@ -130,32 +152,16 @@ _fenced_buffer_add(struct fenced_buffer *fenced_buf)
 
 
 /**
- * Actually destroy the buffer.
+ * Remove the buffer from the fenced list.
+ * 
+ * fenced_buffer_list::mutex and fenced_buffer::mutex must be held, in this 
+ * order before calling this function.
+ * 
+ * Reference count should be decremented after calling this function.
  */
 static INLINE void
-_fenced_buffer_destroy(struct fenced_buffer *fenced_buf)
-{
-   struct fenced_buffer_list *fenced_list = fenced_buf->list;
-   
-   assert(!pipe_is_referenced(&fenced_buf->base.base.reference));
-   assert(!fenced_buf->fence);
-#ifdef DEBUG
-   assert(fenced_buf->head.prev);
-   assert(fenced_buf->head.next);
-   LIST_DEL(&fenced_buf->head);
-   assert(fenced_list->numUnfenced);
-   --fenced_list->numUnfenced;
-#else
-   (void)fenced_list;
-#endif
-   pb_reference(&fenced_buf->buffer, NULL);
-   FREE(fenced_buf);
-}
-
-
-static INLINE void
-_fenced_buffer_remove(struct fenced_buffer_list *fenced_list,
-                      struct fenced_buffer *fenced_buf)
+fenced_buffer_remove_locked(struct fenced_buffer_list *fenced_list,
+                            struct fenced_buffer *fenced_buf)
 {
    struct pb_fence_ops *ops = fenced_list->ops;
 
@@ -177,37 +183,53 @@ _fenced_buffer_remove(struct fenced_buffer_list *fenced_list,
    ++fenced_list->numUnfenced;
 #endif
    
-   /**
-    * FIXME!!!
-    */
-
-   if(!pipe_is_referenced(&fenced_buf->base.base.reference))
-      _fenced_buffer_destroy(fenced_buf);
+   /* TODO: Move the reference count decrement and destruction here */
 }
 
 
+/**
+ * Wait for the fence to expire, and remove it from the fenced list.
+ * 
+ * fenced_buffer::mutex must be held. fenced_buffer_list::mutex must not be 
+ * held -- it will
+ */
 static INLINE enum pipe_error
-_fenced_buffer_finish(struct fenced_buffer *fenced_buf)
+fenced_buffer_finish_locked(struct fenced_buffer_list *fenced_list,
+                              struct fenced_buffer *fenced_buf)
 {
-   struct fenced_buffer_list *fenced_list = fenced_buf->list;
    struct pb_fence_ops *ops = fenced_list->ops;
+   enum pipe_error ret = PIPE_ERROR;
 
 #if 0
    debug_warning("waiting for GPU");
 #endif
 
+   assert(pipe_is_referenced(&fenced_buf->base.base.reference));
    assert(fenced_buf->fence);
+
+   /* Acquire the global lock */
+   pipe_mutex_unlock(fenced_buf->mutex);
+   pipe_mutex_lock(fenced_list->mutex);
+   pipe_mutex_lock(fenced_buf->mutex);
+
    if(fenced_buf->fence) {
-      if(ops->fence_finish(ops, fenced_buf->fence, 0) != 0) {
-        return PIPE_ERROR;
+      if(ops->fence_finish(ops, fenced_buf->fence, 0) == 0) {
+         /* Remove from the fenced list */
+         /* TODO: remove consequents */
+         fenced_buffer_remove_locked(fenced_list, fenced_buf);
+         
+         p_atomic_dec(&fenced_buf->base.base.reference.count);
+         assert(pipe_is_referenced(&fenced_buf->base.base.reference));
+
+         fenced_buf->flags &= ~PIPE_BUFFER_USAGE_GPU_READ_WRITE;
+
+         ret = PIPE_OK;
       }
-      /* Remove from the fenced list */
-      /* TODO: remove consequents */
-      _fenced_buffer_remove(fenced_list, fenced_buf);
    }
 
-   fenced_buf->flags &= ~PIPE_BUFFER_USAGE_GPU_READ_WRITE;
-   return PIPE_OK;
+   pipe_mutex_unlock(fenced_list->mutex);
+
+   return ret;
 }
 
 
@@ -215,12 +237,13 @@ _fenced_buffer_finish(struct fenced_buffer *fenced_buf)
  * Free as many fenced buffers from the list head as possible. 
  */
 static void
-_fenced_buffer_list_check_free(struct fenced_buffer_list *fenced_list, 
+fenced_buffer_list_check_free_locked(struct fenced_buffer_list *fenced_list, 
                                int wait)
 {
    struct pb_fence_ops *ops = fenced_list->ops;
    struct list_head *curr, *next;
    struct fenced_buffer *fenced_buf;
+   struct pb_buffer *pb_buf;
    struct pipe_fence_handle *prev_fence = NULL;
 
    curr = fenced_list->delayed.next;
@@ -228,21 +251,30 @@ _fenced_buffer_list_check_free(struct fenced_buffer_list *fenced_list,
    while(curr != &fenced_list->delayed) {
       fenced_buf = LIST_ENTRY(struct fenced_buffer, curr, head);
 
+      pipe_mutex_lock(fenced_buf->mutex);
+
       if(fenced_buf->fence != prev_fence) {
         int signaled;
         if (wait)
            signaled = ops->fence_finish(ops, fenced_buf->fence, 0);
         else
            signaled = ops->fence_signalled(ops, fenced_buf->fence, 0);
-        if (signaled != 0)
+        if (signaled != 0) {
+            pipe_mutex_unlock(fenced_buf->mutex);
            break;
+         }
         prev_fence = fenced_buf->fence;
       }
       else {
         assert(ops->fence_signalled(ops, fenced_buf->fence, 0) == 0);
       }
 
-      _fenced_buffer_remove(fenced_list, fenced_buf);
+      fenced_buffer_remove_locked(fenced_list, fenced_buf);
+      pipe_mutex_unlock(fenced_buf->mutex);
+
+      pb_buf = &fenced_buf->base;
+      pb_reference(&pb_buf, NULL);
+      
 
       curr = next; 
       next = curr->next;
@@ -256,30 +288,25 @@ fenced_buffer_destroy(struct pb_buffer *buf)
    struct fenced_buffer *fenced_buf = fenced_buffer(buf);   
    struct fenced_buffer_list *fenced_list = fenced_buf->list;
 
-   pipe_mutex_lock(fenced_list->mutex);
    assert(!pipe_is_referenced(&fenced_buf->base.base.reference));
-   if (fenced_buf->fence) {
-      struct pb_fence_ops *ops = fenced_list->ops;
-      if(ops->fence_signalled(ops, fenced_buf->fence, 0) == 0) {
-        struct list_head *curr, *prev;
-        curr = &fenced_buf->head;
-        prev = curr->prev;
-        do {
-           fenced_buf = LIST_ENTRY(struct fenced_buffer, curr, head);
-           assert(ops->fence_signalled(ops, fenced_buf->fence, 0) == 0);
-           _fenced_buffer_remove(fenced_list, fenced_buf);
-           curr = prev;
-           prev = curr->prev;
-        } while (curr != &fenced_list->delayed);
-      }          
-      else {
-        /* delay destruction */
-      }
-   }
-   else {
-      _fenced_buffer_destroy(fenced_buf);
-   }
+   assert(!fenced_buf->fence);
+
+#ifdef DEBUG
+   pipe_mutex_lock(fenced_list->mutex);
+   assert(fenced_buf->head.prev);
+   assert(fenced_buf->head.next);
+   LIST_DEL(&fenced_buf->head);
+   assert(fenced_list->numUnfenced);
+   --fenced_list->numUnfenced;
    pipe_mutex_unlock(fenced_list->mutex);
+#else
+   (void)fenced_list;
+#endif
+
+   pb_reference(&fenced_buf->buffer, NULL);
+
+   pipe_mutex_destroy(fenced_buf->mutex);
+   FREE(fenced_buf);
 }
 
 
@@ -290,24 +317,23 @@ fenced_buffer_map(struct pb_buffer *buf,
    struct fenced_buffer *fenced_buf = fenced_buffer(buf);
    struct fenced_buffer_list *fenced_list = fenced_buf->list;
    struct pb_fence_ops *ops = fenced_list->ops;
-   void *map;
+   void *map = NULL;
+
+   pipe_mutex_lock(fenced_buf->mutex);
 
    assert(!(flags & PIPE_BUFFER_USAGE_GPU_READ_WRITE));
    
    /* Serialize writes */
    if((fenced_buf->flags & PIPE_BUFFER_USAGE_GPU_WRITE) ||
       ((fenced_buf->flags & PIPE_BUFFER_USAGE_GPU_READ) && (flags & PIPE_BUFFER_USAGE_CPU_WRITE))) {
-      if(flags & PIPE_BUFFER_USAGE_DONTBLOCK) {
+      if((flags & PIPE_BUFFER_USAGE_DONTBLOCK) &&
+          ops->fence_signalled(ops, fenced_buf->fence, 0) == 0) {
          /* Don't wait for the GPU to finish writing */
-         if(ops->fence_signalled(ops, fenced_buf->fence, 0) == 0)
-            _fenced_buffer_remove(fenced_list, fenced_buf);
-         else
-            return NULL;
-      }
-      else {
-         /* Wait for the GPU to finish writing */
-         _fenced_buffer_finish(fenced_buf);
+         goto finish;
       }
+
+      /* Wait for the GPU to finish writing */
+      fenced_buffer_finish_locked(fenced_list, fenced_buf);
    }
 
 #if 0
@@ -324,6 +350,9 @@ fenced_buffer_map(struct pb_buffer *buf,
       fenced_buf->flags |= flags & PIPE_BUFFER_USAGE_CPU_READ_WRITE;
    }
 
+finish:
+   pipe_mutex_unlock(fenced_buf->mutex);
+   
    return map;
 }
 
@@ -332,6 +361,9 @@ static void
 fenced_buffer_unmap(struct pb_buffer *buf)
 {
    struct fenced_buffer *fenced_buf = fenced_buffer(buf);
+   
+   pipe_mutex_lock(fenced_buf->mutex);
+   
    assert(fenced_buf->mapcount);
    if(fenced_buf->mapcount) {
       pb_unmap(fenced_buf->buffer);
@@ -339,6 +371,8 @@ fenced_buffer_unmap(struct pb_buffer *buf)
       if(!fenced_buf->mapcount)
         fenced_buf->flags &= ~PIPE_BUFFER_USAGE_CPU_READ_WRITE;
    }
+   
+   pipe_mutex_unlock(fenced_buf->mutex);
 }
 
 
@@ -350,11 +384,14 @@ fenced_buffer_validate(struct pb_buffer *buf,
    struct fenced_buffer *fenced_buf = fenced_buffer(buf);
    enum pipe_error ret;
    
+   pipe_mutex_lock(fenced_buf->mutex);
+
    if(!vl) {
       /* invalidate */
       fenced_buf->vl = NULL;
       fenced_buf->validation_flags = 0;
-      return PIPE_OK;
+      ret = PIPE_OK;
+      goto finish;
    }
    
    assert(flags & PIPE_BUFFER_USAGE_GPU_READ_WRITE);
@@ -362,14 +399,17 @@ fenced_buffer_validate(struct pb_buffer *buf,
    flags &= PIPE_BUFFER_USAGE_GPU_READ_WRITE;
 
    /* Buffer cannot be validated in two different lists */ 
-   if(fenced_buf->vl && fenced_buf->vl != vl)
-      return PIPE_ERROR_RETRY;
+   if(fenced_buf->vl && fenced_buf->vl != vl) {
+      ret = PIPE_ERROR_RETRY;
+      goto finish;
+   }
    
 #if 0
    /* Do not validate if buffer is still mapped */
    if(fenced_buf->flags & PIPE_BUFFER_USAGE_CPU_READ_WRITE) {
       /* TODO: wait for the thread that mapped the buffer to unmap it */
-      return PIPE_ERROR_RETRY;
+      ret = PIPE_ERROR_RETRY;
+      goto finish;
    }
    /* Final sanity checking */
    assert(!(fenced_buf->flags & PIPE_BUFFER_USAGE_CPU_READ_WRITE));
@@ -379,17 +419,21 @@ fenced_buffer_validate(struct pb_buffer *buf,
    if(fenced_buf->vl == vl &&
       (fenced_buf->validation_flags & flags) == flags) {
       /* Nothing to do -- buffer already validated */
-      return PIPE_OK;
+      ret = PIPE_OK;
+      goto finish;
    }
    
    ret = pb_validate(fenced_buf->buffer, vl, flags);
    if (ret != PIPE_OK)
-      return ret;
+      goto finish;
    
    fenced_buf->vl = vl;
    fenced_buf->validation_flags |= flags;
    
-   return PIPE_OK;
+finish:
+   pipe_mutex_unlock(fenced_buf->mutex);
+
+   return ret;
 }
 
 
@@ -404,29 +448,36 @@ fenced_buffer_fence(struct pb_buffer *buf,
    fenced_buf = fenced_buffer(buf);
    fenced_list = fenced_buf->list;
    ops = fenced_list->ops;
-   
-   if(fence == fenced_buf->fence) {
-      /* Nothing to do */
-      return;
-   }
 
-   assert(fenced_buf->vl);
-   assert(fenced_buf->validation_flags);
-   
    pipe_mutex_lock(fenced_list->mutex);
-   if (fenced_buf->fence)
-      _fenced_buffer_remove(fenced_list, fenced_buf);
-   if (fence) {
-      ops->fence_reference(ops, &fenced_buf->fence, fence);
-      fenced_buf->flags |= fenced_buf->validation_flags;
-      _fenced_buffer_add(fenced_buf);
-   }
-   pipe_mutex_unlock(fenced_list->mutex);
+   pipe_mutex_lock(fenced_buf->mutex);
+
+   assert(pipe_is_referenced(&fenced_buf->base.base.reference));
+
+   if(fence != fenced_buf->fence) {
+      assert(fenced_buf->vl);
+      assert(fenced_buf->validation_flags);
+      
+      if (fenced_buf->fence) {
+         fenced_buffer_remove_locked(fenced_list, fenced_buf);
+         p_atomic_dec(&fenced_buf->base.base.reference.count);
+         assert(pipe_is_referenced(&fenced_buf->base.base.reference));
+      }
+      if (fence) {
+         ops->fence_reference(ops, &fenced_buf->fence, fence);
+         fenced_buf->flags |= fenced_buf->validation_flags;
+         p_atomic_inc(&fenced_buf->base.base.reference.count);
+         fenced_buffer_add_locked(fenced_list, fenced_buf);
+      }
+
+      pb_fence(fenced_buf->buffer, fence);
    
-   pb_fence(fenced_buf->buffer, fence);
+      fenced_buf->vl = NULL;
+      fenced_buf->validation_flags = 0;
+   }
 
-   fenced_buf->vl = NULL;
-   fenced_buf->validation_flags = 0;
+   pipe_mutex_unlock(fenced_buf->mutex);
+   pipe_mutex_unlock(fenced_list->mutex);
 }
 
 
@@ -436,6 +487,7 @@ fenced_buffer_get_base_buffer(struct pb_buffer *buf,
                               pb_size *offset)
 {
    struct fenced_buffer *fenced_buf = fenced_buffer(buf);
+   /* NOTE: accesses immutable members only -- mutex not necessary */
    pb_get_base_buffer(fenced_buf->buffer, base_buf, offset);
 }
 
@@ -475,6 +527,8 @@ fenced_buffer_create(struct fenced_buffer_list *fenced_list,
    buf->buffer = buffer;
    buf->list = fenced_list;
    
+   pipe_mutex_init(buf->mutex);
+
 #ifdef DEBUG
    pipe_mutex_lock(fenced_list->mutex);
    LIST_ADDTAIL(&buf->head, &fenced_list->unfenced);
@@ -516,7 +570,7 @@ fenced_buffer_list_check_free(struct fenced_buffer_list *fenced_list,
                               int wait)
 {
    pipe_mutex_lock(fenced_list->mutex);
-   _fenced_buffer_list_check_free(fenced_list, wait);
+   fenced_buffer_list_check_free_locked(fenced_list, wait);
    pipe_mutex_unlock(fenced_list->mutex);
 }
 
@@ -538,11 +592,13 @@ fenced_buffer_list_dump(struct fenced_buffer_list *fenced_list)
    next = curr->next;
    while(curr != &fenced_list->unfenced) {
       fenced_buf = LIST_ENTRY(struct fenced_buffer, curr, head);
+      pipe_mutex_lock(fenced_buf->mutex);
       assert(!fenced_buf->fence);
       debug_printf("%10p %7u %7u\n",
                    (void *) fenced_buf,
                    fenced_buf->base.base.size,
                    p_atomic_read(&fenced_buf->base.base.reference.count));
+      pipe_mutex_unlock(fenced_buf->mutex);
       curr = next; 
       next = curr->next;
    }
@@ -552,6 +608,7 @@ fenced_buffer_list_dump(struct fenced_buffer_list *fenced_list)
    while(curr != &fenced_list->delayed) {
       int signaled;
       fenced_buf = LIST_ENTRY(struct fenced_buffer, curr, head);
+      pipe_mutex_lock(fenced_buf->mutex);
       signaled = ops->fence_signalled(ops, fenced_buf->fence, 0);
       debug_printf("%10p %7u %7u %10p %s\n",
                    (void *) fenced_buf,
@@ -559,6 +616,7 @@ fenced_buffer_list_dump(struct fenced_buffer_list *fenced_list)
                    p_atomic_read(&fenced_buf->base.base.reference.count),
                    (void *) fenced_buf->fence,
                    signaled == 0 ? "y" : "n");
+      pipe_mutex_unlock(fenced_buf->mutex);
       curr = next; 
       next = curr->next;
    }
@@ -579,8 +637,8 @@ fenced_buffer_list_destroy(struct fenced_buffer_list *fenced_list)
 #if defined(PIPE_OS_LINUX) || defined(PIPE_OS_BSD) || defined(PIPE_OS_SOLARIS)
       sched_yield();
 #endif
-      _fenced_buffer_list_check_free(fenced_list, 1);
       pipe_mutex_lock(fenced_list->mutex);
+      fenced_buffer_list_check_free_locked(fenced_list, 1);
    }
 
 #ifdef DEBUG
@@ -588,6 +646,7 @@ fenced_buffer_list_destroy(struct fenced_buffer_list *fenced_list)
 #endif
       
    pipe_mutex_unlock(fenced_list->mutex);
+   pipe_mutex_destroy(fenced_list->mutex);
    
    fenced_list->ops->destroy(fenced_list->ops);
    
index 57d1ede45a44d90a98692286abb3ccdcefa7d63e..7b34c8e35782aac1a69e8da68952973c827e1f5f 100644 (file)
@@ -294,7 +294,7 @@ pb_cache_manager_create_buffer(struct pb_manager *_mgr,
       LIST_DEL(&buf->head);
       pipe_mutex_unlock(mgr->mutex);
       /* Increase refcount */
-      pb_reference((struct pb_buffer**)&buf, &buf->base);
+      pipe_reference(NULL, &buf->base.base.reference);
       return &buf->base;
    }
    
index eb492076b7d2dcaf935d5bdc30406fac0327d762..080fd4c7310dc53647aaf6d837ba40706ee97527 100644 (file)
@@ -1129,3 +1129,35 @@ TGSI Instruction Specification
 
   target            Label of target instruction.
 
+
+3  Other tokens
+===============
+
+
+3.1  Declaration Semantic
+-------------------------
+
+
+  Follows Declaration token if Semantic bit is set.
+
+  Since its purpose is to link a shader with other stages of the pipeline,
+  it is valid to follow only those Declaration tokens that declare a register
+  either in INPUT or OUTPUT file.
+
+  SemanticName field contains the semantic name of the register being declared.
+  There is no default value.
+
+  SemanticIndex is an optional subscript that can be used to distinguish
+  different register declarations with the same semantic name. The default value
+  is 0.
+
+  The meanings of the individual semantic names are explained in the following
+  sections.
+
+
+3.1.1  FACE
+
+  Valid only in a fragment shader INPUT declaration.
+
+  FACE.x is negative when the primitive is back facing. FACE.x is positive
+  when the primitive is front facing.
index 4fa10e2f7e3920cabcf69775f1a0a52a5f35e493..de9cbc863051f7d2495fe3f61cb9cd95865bfa33 100644 (file)
 #include "tgsi_build.h"
 #include "tgsi_parse.h"
 
-/*
- * version
- */
-
-struct tgsi_version
-tgsi_build_version( void )
-{
-   struct tgsi_version  version;
-
-   version.MajorVersion = 1;
-   version.MinorVersion = 1;
-   version.Padding = 0;
-
-   return version;
-}
 
 /*
  * header
@@ -122,7 +107,6 @@ tgsi_default_declaration( void )
    declaration.Centroid = 0;
    declaration.Invariant = 0;
    declaration.Padding = 0;
-   declaration.Extended = 0;
 
    return declaration;
 }
@@ -173,7 +157,7 @@ tgsi_default_full_declaration( void )
    struct tgsi_full_declaration  full_declaration;
 
    full_declaration.Declaration  = tgsi_default_declaration();
-   full_declaration.DeclarationRange = tgsi_default_declaration_range();
+   full_declaration.Range = tgsi_default_declaration_range();
    full_declaration.Semantic = tgsi_default_declaration_semantic();
 
    return full_declaration;
@@ -210,8 +194,8 @@ tgsi_build_full_declaration(
    size++;
 
    *dr = tgsi_build_declaration_range(
-      full_decl->DeclarationRange.First,
-      full_decl->DeclarationRange.Last,
+      full_decl->Range.First,
+      full_decl->Range.Last,
       declaration,
       header );
 
@@ -224,8 +208,8 @@ tgsi_build_full_declaration(
       size++;
 
       *ds = tgsi_build_declaration_semantic(
-         full_decl->Semantic.SemanticName,
-         full_decl->Semantic.SemanticIndex,
+         full_decl->Semantic.Name,
+         full_decl->Semantic.Index,
          declaration,
          header );
    }
@@ -270,8 +254,8 @@ tgsi_default_declaration_semantic( void )
 {
    struct tgsi_declaration_semantic ds;
 
-   ds.SemanticName = TGSI_SEMANTIC_POSITION;
-   ds.SemanticIndex = 0;
+   ds.Name = TGSI_SEMANTIC_POSITION;
+   ds.Index = 0;
    ds.Padding = 0;
 
    return ds;
@@ -290,8 +274,8 @@ tgsi_build_declaration_semantic(
    assert( semantic_index <= 0xFFFF );
 
    ds = tgsi_default_declaration_semantic();
-   ds.SemanticName = semantic_name;
-   ds.SemanticIndex = semantic_index;
+   ds.Name = semantic_name;
+   ds.Index = semantic_index;
 
    declaration_grow( declaration, header );
 
@@ -311,7 +295,6 @@ tgsi_default_immediate( void )
    immediate.NrTokens = 1;
    immediate.DataType = TGSI_IMM_FLOAT32;
    immediate.Padding = 0;
-   immediate.Extended = 0;
 
    return immediate;
 }
@@ -416,24 +399,26 @@ tgsi_default_instruction( void )
    struct tgsi_instruction instruction;
 
    instruction.Type = TGSI_TOKEN_TYPE_INSTRUCTION;
-   instruction.NrTokens = 1;
+   instruction.NrTokens = 0;
    instruction.Opcode = TGSI_OPCODE_MOV;
    instruction.Saturate = TGSI_SAT_NONE;
+   instruction.Predicate = 0;
    instruction.NumDstRegs = 1;
    instruction.NumSrcRegs = 1;
+   instruction.Label = 0;
+   instruction.Texture = 0;
    instruction.Padding  = 0;
-   instruction.Extended = 0;
 
    return instruction;
 }
 
 struct tgsi_instruction
-tgsi_build_instruction(
-   unsigned opcode,
-   unsigned saturate,
-   unsigned num_dst_regs,
-   unsigned num_src_regs,
-   struct tgsi_header *header )
+tgsi_build_instruction(unsigned opcode,
+                       unsigned saturate,
+                       unsigned predicate,
+                       unsigned num_dst_regs,
+                       unsigned num_src_regs,
+                       struct tgsi_header *header)
 {
    struct tgsi_instruction instruction;
 
@@ -445,6 +430,7 @@ tgsi_build_instruction(
    instruction = tgsi_default_instruction();
    instruction.Opcode = opcode;
    instruction.Saturate = saturate;
+   instruction.Predicate = predicate;
    instruction.NumDstRegs = num_dst_regs;
    instruction.NumSrcRegs = num_src_regs;
 
@@ -472,18 +458,16 @@ tgsi_default_full_instruction( void )
    unsigned i;
 
    full_instruction.Instruction = tgsi_default_instruction();
-   full_instruction.InstructionExtLabel = tgsi_default_instruction_ext_label();
-   full_instruction.InstructionExtTexture = tgsi_default_instruction_ext_texture();
-   full_instruction.InstructionExtPredicate = tgsi_default_instruction_ext_predicate();
+   full_instruction.Predicate = tgsi_default_instruction_predicate();
+   full_instruction.Label = tgsi_default_instruction_label();
+   full_instruction.Texture = tgsi_default_instruction_texture();
    for( i = 0;  i < TGSI_FULL_MAX_DST_REGISTERS; i++ ) {
-      full_instruction.FullDstRegisters[i] = tgsi_default_full_dst_register();
+      full_instruction.Dst[i] = tgsi_default_full_dst_register();
    }
    for( i = 0;  i < TGSI_FULL_MAX_SRC_REGISTERS; i++ ) {
-      full_instruction.FullSrcRegisters[i] = tgsi_default_full_src_register();
+      full_instruction.Src[i] = tgsi_default_full_src_register();
    }
 
-   full_instruction.Flags = 0x0;
-
    return full_instruction;
 }
 
@@ -504,77 +488,70 @@ tgsi_build_full_instruction(
    instruction = (struct tgsi_instruction *) &tokens[size];
    size++;
 
-   *instruction = tgsi_build_instruction(
-      full_inst->Instruction.Opcode,
-      full_inst->Instruction.Saturate,
-      full_inst->Instruction.NumDstRegs,
-      full_inst->Instruction.NumSrcRegs,
-      header );
+   *instruction = tgsi_build_instruction(full_inst->Instruction.Opcode,
+                                         full_inst->Instruction.Saturate,
+                                         full_inst->Instruction.Predicate,
+                                         full_inst->Instruction.NumDstRegs,
+                                         full_inst->Instruction.NumSrcRegs,
+                                         header);
    prev_token = (struct tgsi_token  *) instruction;
 
-   if( tgsi_compare_instruction_ext_label(
-         full_inst->InstructionExtLabel,
-         tgsi_default_instruction_ext_label() ) ) {
-      struct tgsi_instruction_ext_label *instruction_ext_label;
+   if (full_inst->Instruction.Predicate) {
+      struct tgsi_instruction_predicate *instruction_predicate;
 
-      if( maxsize <= size )
+      if (maxsize <= size) {
          return 0;
-      instruction_ext_label =
-         (struct  tgsi_instruction_ext_label *) &tokens[size];
+      }
+      instruction_predicate = (struct tgsi_instruction_predicate *)&tokens[size];
       size++;
 
-      *instruction_ext_label = tgsi_build_instruction_ext_label(
-         full_inst->InstructionExtLabel.Label,
-         prev_token,
-         instruction,
-         header );
-      prev_token = (struct tgsi_token  *) instruction_ext_label;
+      *instruction_predicate =
+         tgsi_build_instruction_predicate(full_inst->Predicate.Index,
+                                          full_inst->Predicate.Negate,
+                                          full_inst->Predicate.SwizzleX,
+                                          full_inst->Predicate.SwizzleY,
+                                          full_inst->Predicate.SwizzleZ,
+                                          full_inst->Predicate.SwizzleW,
+                                          instruction,
+                                          header);
    }
 
-   if( tgsi_compare_instruction_ext_texture(
-         full_inst->InstructionExtTexture,
-         tgsi_default_instruction_ext_texture() ) ) {
-      struct tgsi_instruction_ext_texture *instruction_ext_texture;
+   if (full_inst->Instruction.Label) {
+      struct tgsi_instruction_label *instruction_label;
 
       if( maxsize <= size )
          return 0;
-      instruction_ext_texture =
-         (struct  tgsi_instruction_ext_texture *) &tokens[size];
+      instruction_label =
+         (struct  tgsi_instruction_label *) &tokens[size];
       size++;
 
-      *instruction_ext_texture = tgsi_build_instruction_ext_texture(
-         full_inst->InstructionExtTexture.Texture,
+      *instruction_label = tgsi_build_instruction_label(
+         full_inst->Label.Label,
          prev_token,
          instruction,
-         header   );
-      prev_token = (struct tgsi_token  *) instruction_ext_texture;
+         header );
+      prev_token = (struct tgsi_token  *) instruction_label;
    }
 
-   if (tgsi_compare_instruction_ext_predicate(full_inst->InstructionExtPredicate,
-                                              tgsi_default_instruction_ext_predicate())) {
-      struct tgsi_instruction_ext_predicate *instruction_ext_predicate;
+   if (full_inst->Instruction.Texture) {
+      struct tgsi_instruction_texture *instruction_texture;
 
-      if (maxsize <= size) {
+      if( maxsize <= size )
          return 0;
-      }
-      instruction_ext_predicate = (struct tgsi_instruction_ext_predicate *)&tokens[size];
+      instruction_texture =
+         (struct  tgsi_instruction_texture *) &tokens[size];
       size++;
 
-      *instruction_ext_predicate =
-         tgsi_build_instruction_ext_predicate(full_inst->InstructionExtPredicate.SrcIndex,
-                                              full_inst->InstructionExtPredicate.Negate,
-                                              full_inst->InstructionExtPredicate.SwizzleX,
-                                              full_inst->InstructionExtPredicate.SwizzleY,
-                                              full_inst->InstructionExtPredicate.SwizzleZ,
-                                              full_inst->InstructionExtPredicate.SwizzleW,
-                                              prev_token,
-                                              instruction,
-                                              header);
-      prev_token = (struct tgsi_token *)instruction_ext_predicate;
+      *instruction_texture = tgsi_build_instruction_texture(
+         full_inst->Texture.Texture,
+         prev_token,
+         instruction,
+         header   );
+      prev_token = (struct tgsi_token  *) instruction_texture;
    }
 
    for( i = 0;  i <   full_inst->Instruction.NumDstRegs; i++ ) {
-      const struct tgsi_full_dst_register *reg = &full_inst->FullDstRegisters[i];
+      const struct tgsi_full_dst_register *reg = &full_inst->Dst[i];
       struct tgsi_dst_register *dst_register;
       struct tgsi_token *prev_token;
 
@@ -584,34 +561,15 @@ tgsi_build_full_instruction(
       size++;
 
       *dst_register = tgsi_build_dst_register(
-         reg->DstRegister.File,
-         reg->DstRegister.WriteMask,
-         reg->DstRegister.Indirect,
-         reg->DstRegister.Index,
+         reg->Register.File,
+         reg->Register.WriteMask,
+         reg->Register.Indirect,
+         reg->Register.Index,
          instruction,
          header );
       prev_token = (struct tgsi_token  *) dst_register;
 
-      if( tgsi_compare_dst_register_ext_modulate(
-            reg->DstRegisterExtModulate,
-            tgsi_default_dst_register_ext_modulate() ) ) {
-         struct tgsi_dst_register_ext_modulate *dst_register_ext_modulate;
-
-         if( maxsize <= size )
-            return 0;
-         dst_register_ext_modulate =
-            (struct  tgsi_dst_register_ext_modulate *) &tokens[size];
-         size++;
-
-         *dst_register_ext_modulate = tgsi_build_dst_register_ext_modulate(
-            reg->DstRegisterExtModulate.Modulate,
-            prev_token,
-            instruction,
-            header );
-         prev_token = (struct tgsi_token  *) dst_register_ext_modulate;
-      }
-
-      if( reg->DstRegister.Indirect ) {
+      if( reg->Register.Indirect ) {
          struct tgsi_src_register *ind;
 
          if( maxsize <= size )
@@ -620,22 +578,23 @@ tgsi_build_full_instruction(
          size++;
 
          *ind = tgsi_build_src_register(
-            reg->DstRegisterInd.File,
-            reg->DstRegisterInd.SwizzleX,
-            reg->DstRegisterInd.SwizzleY,
-            reg->DstRegisterInd.SwizzleZ,
-            reg->DstRegisterInd.SwizzleW,
-            reg->DstRegisterInd.Negate,
-            reg->DstRegisterInd.Indirect,
-            reg->DstRegisterInd.Dimension,
-            reg->DstRegisterInd.Index,
+            reg->Indirect.File,
+            reg->Indirect.SwizzleX,
+            reg->Indirect.SwizzleY,
+            reg->Indirect.SwizzleZ,
+            reg->Indirect.SwizzleW,
+            reg->Indirect.Negate,
+            reg->Indirect.Absolute,
+            reg->Indirect.Indirect,
+            reg->Indirect.Dimension,
+            reg->Indirect.Index,
             instruction,
             header );
       }
    }
 
    for( i = 0;  i < full_inst->Instruction.NumSrcRegs; i++ ) {
-      const struct tgsi_full_src_register *reg = &full_inst->FullSrcRegisters[i];
+      const struct tgsi_full_src_register *reg = &full_inst->Src[i];
       struct tgsi_src_register *src_register;
       struct tgsi_token *prev_token;
 
@@ -645,43 +604,21 @@ tgsi_build_full_instruction(
       size++;
 
       *src_register = tgsi_build_src_register(
-         reg->SrcRegister.File,
-         reg->SrcRegister.SwizzleX,
-         reg->SrcRegister.SwizzleY,
-         reg->SrcRegister.SwizzleZ,
-         reg->SrcRegister.SwizzleW,
-         reg->SrcRegister.Negate,
-         reg->SrcRegister.Indirect,
-         reg->SrcRegister.Dimension,
-         reg->SrcRegister.Index,
+         reg->Register.File,
+         reg->Register.SwizzleX,
+         reg->Register.SwizzleY,
+         reg->Register.SwizzleZ,
+         reg->Register.SwizzleW,
+         reg->Register.Negate,
+         reg->Register.Absolute,
+         reg->Register.Indirect,
+         reg->Register.Dimension,
+         reg->Register.Index,
          instruction,
          header );
       prev_token = (struct tgsi_token  *) src_register;
 
-      if( tgsi_compare_src_register_ext_mod(
-            reg->SrcRegisterExtMod,
-            tgsi_default_src_register_ext_mod() ) ) {
-         struct tgsi_src_register_ext_mod *src_register_ext_mod;
-
-         if( maxsize <= size )
-            return 0;
-         src_register_ext_mod =
-            (struct  tgsi_src_register_ext_mod *) &tokens[size];
-         size++;
-
-         *src_register_ext_mod = tgsi_build_src_register_ext_mod(
-            reg->SrcRegisterExtMod.Complement,
-            reg->SrcRegisterExtMod.Bias,
-            reg->SrcRegisterExtMod.Scale2X,
-            reg->SrcRegisterExtMod.Absolute,
-            reg->SrcRegisterExtMod.Negate,
-            prev_token,
-            instruction,
-            header );
-         prev_token = (struct tgsi_token  *) src_register_ext_mod;
-      }
-
-      if( reg->SrcRegister.Indirect ) {
+      if( reg->Register.Indirect ) {
          struct  tgsi_src_register *ind;
 
          if( maxsize <= size )
@@ -690,23 +627,24 @@ tgsi_build_full_instruction(
          size++;
 
          *ind = tgsi_build_src_register(
-            reg->SrcRegisterInd.File,
-            reg->SrcRegisterInd.SwizzleX,
-            reg->SrcRegisterInd.SwizzleY,
-            reg->SrcRegisterInd.SwizzleZ,
-            reg->SrcRegisterInd.SwizzleW,
-            reg->SrcRegisterInd.Negate,
-            reg->SrcRegisterInd.Indirect,
-            reg->SrcRegisterInd.Dimension,
-            reg->SrcRegisterInd.Index,
+            reg->Indirect.File,
+            reg->Indirect.SwizzleX,
+            reg->Indirect.SwizzleY,
+            reg->Indirect.SwizzleZ,
+            reg->Indirect.SwizzleW,
+            reg->Indirect.Negate,
+            reg->Indirect.Absolute,
+            reg->Indirect.Indirect,
+            reg->Indirect.Dimension,
+            reg->Indirect.Index,
             instruction,
             header );
       }
 
-      if( reg->SrcRegister.Dimension ) {
+      if( reg->Register.Dimension ) {
          struct  tgsi_dimension *dim;
 
-         assert( !reg->SrcRegisterDim.Dimension );
+         assert( !reg->Dimension.Dimension );
 
          if( maxsize <= size )
             return 0;
@@ -714,12 +652,12 @@ tgsi_build_full_instruction(
          size++;
 
          *dim = tgsi_build_dimension(
-            reg->SrcRegisterDim.Indirect,
-            reg->SrcRegisterDim.Index,
+            reg->Dimension.Indirect,
+            reg->Dimension.Index,
             instruction,
             header );
 
-         if( reg->SrcRegisterDim.Indirect ) {
+         if( reg->Dimension.Indirect ) {
             struct tgsi_src_register *ind;
 
             if( maxsize <= size )
@@ -728,15 +666,16 @@ tgsi_build_full_instruction(
             size++;
 
             *ind = tgsi_build_src_register(
-               reg->SrcRegisterDimInd.File,
-               reg->SrcRegisterDimInd.SwizzleX,
-               reg->SrcRegisterDimInd.SwizzleY,
-               reg->SrcRegisterDimInd.SwizzleZ,
-               reg->SrcRegisterDimInd.SwizzleW,
-               reg->SrcRegisterDimInd.Negate,
-               reg->SrcRegisterDimInd.Indirect,
-               reg->SrcRegisterDimInd.Dimension,
-               reg->SrcRegisterDimInd.Index,
+               reg->DimIndirect.File,
+               reg->DimIndirect.SwizzleX,
+               reg->DimIndirect.SwizzleY,
+               reg->DimIndirect.SwizzleZ,
+               reg->DimIndirect.SwizzleW,
+               reg->DimIndirect.Negate,
+               reg->DimIndirect.Absolute,
+               reg->DimIndirect.Indirect,
+               reg->DimIndirect.Dimension,
+               reg->DimIndirect.Index,
                instruction,
                header );
          }
@@ -746,147 +685,103 @@ tgsi_build_full_instruction(
    return size;
 }
 
-/** test for inequality of 32-bit values pointed to by a and b */
-static INLINE boolean
-compare32(const void *a, const void *b)
+struct tgsi_instruction_predicate
+tgsi_default_instruction_predicate(void)
 {
-   return *((uint32_t *) a) != *((uint32_t *) b);
+   struct tgsi_instruction_predicate instruction_predicate;
+
+   instruction_predicate.SwizzleX = TGSI_SWIZZLE_X;
+   instruction_predicate.SwizzleY = TGSI_SWIZZLE_Y;
+   instruction_predicate.SwizzleZ = TGSI_SWIZZLE_Z;
+   instruction_predicate.SwizzleW = TGSI_SWIZZLE_W;
+   instruction_predicate.Negate = 0;
+   instruction_predicate.Index = 0;
+   instruction_predicate.Padding = 0;
+
+   return instruction_predicate;
 }
 
-struct tgsi_instruction_ext_label
-tgsi_default_instruction_ext_label( void )
+struct tgsi_instruction_predicate
+tgsi_build_instruction_predicate(int index,
+                                 unsigned negate,
+                                 unsigned swizzleX,
+                                 unsigned swizzleY,
+                                 unsigned swizzleZ,
+                                 unsigned swizzleW,
+                                 struct tgsi_instruction *instruction,
+                                 struct tgsi_header *header)
 {
-   struct tgsi_instruction_ext_label instruction_ext_label;
+   struct tgsi_instruction_predicate instruction_predicate;
+
+   instruction_predicate = tgsi_default_instruction_predicate();
+   instruction_predicate.SwizzleX = swizzleX;
+   instruction_predicate.SwizzleY = swizzleY;
+   instruction_predicate.SwizzleZ = swizzleZ;
+   instruction_predicate.SwizzleW = swizzleW;
+   instruction_predicate.Negate = negate;
+   instruction_predicate.Index = index;
 
-   instruction_ext_label.Type = TGSI_INSTRUCTION_EXT_TYPE_LABEL;
-   instruction_ext_label.Label = 0;
-   instruction_ext_label.Padding = 0;
-   instruction_ext_label.Extended = 0;
+   instruction_grow(instruction, header);
 
-   return instruction_ext_label;
+   return instruction_predicate;
 }
 
-unsigned
-tgsi_compare_instruction_ext_label(
-   struct tgsi_instruction_ext_label a,
-   struct tgsi_instruction_ext_label b )
+struct tgsi_instruction_label
+tgsi_default_instruction_label( void )
 {
-   a.Padding = b.Padding = 0;
-   a.Extended = b.Extended = 0;
-   return compare32(&a, &b);
+   struct tgsi_instruction_label instruction_label;
+
+   instruction_label.Label = 0;
+   instruction_label.Padding = 0;
+
+   return instruction_label;
 }
 
-struct tgsi_instruction_ext_label
-tgsi_build_instruction_ext_label(
+struct tgsi_instruction_label
+tgsi_build_instruction_label(
    unsigned label,
    struct tgsi_token  *prev_token,
    struct tgsi_instruction *instruction,
    struct tgsi_header *header )
 {
-   struct tgsi_instruction_ext_label instruction_ext_label;
+   struct tgsi_instruction_label instruction_label;
 
-   instruction_ext_label = tgsi_default_instruction_ext_label();
-   instruction_ext_label.Label = label;
+   instruction_label = tgsi_default_instruction_label();
+   instruction_label.Label = label;
+   instruction->Label = 1;
 
-   prev_token->Extended = 1;
    instruction_grow( instruction, header );
 
-   return instruction_ext_label;
+   return instruction_label;
 }
 
-struct tgsi_instruction_ext_texture
-tgsi_default_instruction_ext_texture( void )
+struct tgsi_instruction_texture
+tgsi_default_instruction_texture( void )
 {
-   struct tgsi_instruction_ext_texture instruction_ext_texture;
+   struct tgsi_instruction_texture instruction_texture;
 
-   instruction_ext_texture.Type = TGSI_INSTRUCTION_EXT_TYPE_TEXTURE;
-   instruction_ext_texture.Texture = TGSI_TEXTURE_UNKNOWN;
-   instruction_ext_texture.Padding = 0;
-   instruction_ext_texture.Extended = 0;
-
-   return instruction_ext_texture;
-}
+   instruction_texture.Texture = TGSI_TEXTURE_UNKNOWN;
+   instruction_texture.Padding = 0;
 
-unsigned
-tgsi_compare_instruction_ext_texture(
-   struct tgsi_instruction_ext_texture a,
-   struct tgsi_instruction_ext_texture b )
-{
-   a.Padding = b.Padding = 0;
-   a.Extended = b.Extended = 0;
-   return compare32(&a, &b);
+   return instruction_texture;
 }
 
-struct tgsi_instruction_ext_texture
-tgsi_build_instruction_ext_texture(
+struct tgsi_instruction_texture
+tgsi_build_instruction_texture(
    unsigned texture,
    struct tgsi_token *prev_token,
    struct tgsi_instruction *instruction,
    struct tgsi_header *header )
 {
-   struct tgsi_instruction_ext_texture instruction_ext_texture;
+   struct tgsi_instruction_texture instruction_texture;
 
-   instruction_ext_texture = tgsi_default_instruction_ext_texture();
-   instruction_ext_texture.Texture = texture;
+   instruction_texture = tgsi_default_instruction_texture();
+   instruction_texture.Texture = texture;
+   instruction->Texture = 1;
 
-   prev_token->Extended = 1;
    instruction_grow( instruction, header );
 
-   return instruction_ext_texture;
-}
-
-struct tgsi_instruction_ext_predicate
-tgsi_default_instruction_ext_predicate(void)
-{
-   struct tgsi_instruction_ext_predicate instruction_ext_predicate;
-
-   instruction_ext_predicate.Type = TGSI_INSTRUCTION_EXT_TYPE_PREDICATE;
-   instruction_ext_predicate.SwizzleX = TGSI_SWIZZLE_X;
-   instruction_ext_predicate.SwizzleY = TGSI_SWIZZLE_Y;
-   instruction_ext_predicate.SwizzleZ = TGSI_SWIZZLE_Z;
-   instruction_ext_predicate.SwizzleW = TGSI_SWIZZLE_W;
-   instruction_ext_predicate.Negate = 0;
-   instruction_ext_predicate.SrcIndex = 0;
-   instruction_ext_predicate.Padding = 0;
-   instruction_ext_predicate.Extended = 0;
-
-   return instruction_ext_predicate;
-}
-
-unsigned
-tgsi_compare_instruction_ext_predicate(struct tgsi_instruction_ext_predicate a,
-                                       struct tgsi_instruction_ext_predicate b)
-{
-   a.Padding = b.Padding = 0;
-   a.Extended = b.Extended = 0;
-   return compare32(&a, &b);
-}
-
-struct tgsi_instruction_ext_predicate
-tgsi_build_instruction_ext_predicate(unsigned index,
-                                     unsigned negate,
-                                     unsigned swizzleX,
-                                     unsigned swizzleY,
-                                     unsigned swizzleZ,
-                                     unsigned swizzleW,
-                                     struct tgsi_token *prev_token,
-                                     struct tgsi_instruction *instruction,
-                                     struct tgsi_header *header)
-{
-   struct tgsi_instruction_ext_predicate instruction_ext_predicate;
-
-   instruction_ext_predicate = tgsi_default_instruction_ext_predicate();
-   instruction_ext_predicate.SwizzleX = swizzleX;
-   instruction_ext_predicate.SwizzleY = swizzleY;
-   instruction_ext_predicate.SwizzleZ = swizzleZ;
-   instruction_ext_predicate.SwizzleW = swizzleW;
-   instruction_ext_predicate.Negate = negate;
-   instruction_ext_predicate.SrcIndex = index;
-
-   prev_token->Extended = 1;
-   instruction_grow(instruction, header);
-
-   return instruction_ext_predicate;
+   return instruction_texture;
 }
 
 struct tgsi_src_register
@@ -900,10 +795,10 @@ tgsi_default_src_register( void )
    src_register.SwizzleZ = TGSI_SWIZZLE_Z;
    src_register.SwizzleW = TGSI_SWIZZLE_W;
    src_register.Negate = 0;
+   src_register.Absolute = 0;
    src_register.Indirect = 0;
    src_register.Dimension = 0;
    src_register.Index = 0;
-   src_register.Extended = 0;
 
    return src_register;
 }
@@ -916,6 +811,7 @@ tgsi_build_src_register(
    unsigned swizzle_z,
    unsigned swizzle_w,
    unsigned negate,
+   unsigned absolute,
    unsigned indirect,
    unsigned dimension,
    int index,
@@ -939,6 +835,7 @@ tgsi_build_src_register(
    src_register.SwizzleZ = swizzle_z;
    src_register.SwizzleW = swizzle_w;
    src_register.Negate = negate;
+   src_register.Absolute = absolute;
    src_register.Indirect = indirect;
    src_register.Dimension = dimension;
    src_register.Index = index;
@@ -953,75 +850,15 @@ tgsi_default_full_src_register( void )
 {
    struct tgsi_full_src_register full_src_register;
 
-   full_src_register.SrcRegister = tgsi_default_src_register();
-   full_src_register.SrcRegisterExtMod = tgsi_default_src_register_ext_mod();
-   full_src_register.SrcRegisterInd = tgsi_default_src_register();
-   full_src_register.SrcRegisterDim = tgsi_default_dimension();
-   full_src_register.SrcRegisterDimInd = tgsi_default_src_register();
+   full_src_register.Register = tgsi_default_src_register();
+   full_src_register.Indirect = tgsi_default_src_register();
+   full_src_register.Dimension = tgsi_default_dimension();
+   full_src_register.DimIndirect = tgsi_default_src_register();
 
    return full_src_register;
 }
 
 
-struct tgsi_src_register_ext_mod
-tgsi_default_src_register_ext_mod( void )
-{
-   struct tgsi_src_register_ext_mod src_register_ext_mod;
-
-   src_register_ext_mod.Type = TGSI_SRC_REGISTER_EXT_TYPE_MOD;
-   src_register_ext_mod.Complement = 0;
-   src_register_ext_mod.Bias = 0;
-   src_register_ext_mod.Scale2X = 0;
-   src_register_ext_mod.Absolute = 0;
-   src_register_ext_mod.Negate = 0;
-   src_register_ext_mod.Padding = 0;
-   src_register_ext_mod.Extended = 0;
-
-   return src_register_ext_mod;
-}
-
-unsigned
-tgsi_compare_src_register_ext_mod(
-   struct tgsi_src_register_ext_mod a,
-   struct tgsi_src_register_ext_mod b )
-{
-   a.Padding = b.Padding = 0;
-   a.Extended = b.Extended = 0;
-   return compare32(&a, &b);
-}
-
-struct tgsi_src_register_ext_mod
-tgsi_build_src_register_ext_mod(
-   unsigned complement,
-   unsigned bias,
-   unsigned scale_2x,
-   unsigned absolute,
-   unsigned negate,
-   struct tgsi_token *prev_token,
-   struct tgsi_instruction *instruction,
-   struct tgsi_header *header )
-{
-   struct tgsi_src_register_ext_mod src_register_ext_mod;
-
-   assert( complement <= 1 );
-   assert( bias <= 1 );
-   assert( scale_2x <= 1 );
-   assert( absolute <= 1 );
-   assert( negate <= 1 );
-
-   src_register_ext_mod = tgsi_default_src_register_ext_mod();
-   src_register_ext_mod.Complement = complement;
-   src_register_ext_mod.Bias = bias;
-   src_register_ext_mod.Scale2X = scale_2x;
-   src_register_ext_mod.Absolute = absolute;
-   src_register_ext_mod.Negate = negate;
-
-   prev_token->Extended = 1;
-   instruction_grow( instruction, header );
-
-   return src_register_ext_mod;
-}
-
 struct tgsi_dimension
 tgsi_default_dimension( void )
 {
@@ -1031,7 +868,6 @@ tgsi_default_dimension( void )
    dimension.Dimension = 0;
    dimension.Padding = 0;
    dimension.Index = 0;
-   dimension.Extended = 0;
 
    return dimension;
 }
@@ -1065,7 +901,6 @@ tgsi_default_dst_register( void )
    dst_register.Dimension = 0;
    dst_register.Index = 0;
    dst_register.Padding = 0;
-   dst_register.Extended = 0;
 
    return dst_register;
 }
@@ -1101,53 +936,113 @@ tgsi_default_full_dst_register( void )
 {
    struct tgsi_full_dst_register full_dst_register;
 
-   full_dst_register.DstRegister = tgsi_default_dst_register();
-   full_dst_register.DstRegisterInd = tgsi_default_src_register();
-   full_dst_register.DstRegisterExtModulate =
-      tgsi_default_dst_register_ext_modulate();
+   full_dst_register.Register = tgsi_default_dst_register();
+   full_dst_register.Indirect = tgsi_default_src_register();
 
    return full_dst_register;
 }
 
-struct tgsi_dst_register_ext_modulate
-tgsi_default_dst_register_ext_modulate( void )
+struct tgsi_property
+tgsi_default_property( void )
 {
-   struct tgsi_dst_register_ext_modulate dst_register_ext_modulate;
+   struct tgsi_property property;
 
-   dst_register_ext_modulate.Type = TGSI_DST_REGISTER_EXT_TYPE_MODULATE;
-   dst_register_ext_modulate.Modulate = TGSI_MODULATE_1X;
-   dst_register_ext_modulate.Padding = 0;
-   dst_register_ext_modulate.Extended = 0;
+   property.Type = TGSI_TOKEN_TYPE_PROPERTY;
+   property.NrTokens = 1;
+   property.PropertyName = TGSI_PROPERTY_GS_INPUT_PRIM;
+   property.Padding = 0;
 
-   return dst_register_ext_modulate;
+   return property;
 }
 
-unsigned
-tgsi_compare_dst_register_ext_modulate(
-   struct tgsi_dst_register_ext_modulate a,
-   struct tgsi_dst_register_ext_modulate b )
+struct tgsi_property
+tgsi_build_property(unsigned property_name,
+                    struct tgsi_header *header)
 {
-   a.Padding = b.Padding = 0;
-   a.Extended = b.Extended = 0;
-   return compare32(&a, &b);
+   struct tgsi_property property;
+
+   property = tgsi_default_property();
+   property.PropertyName = property_name;
+
+   header_bodysize_grow( header );
+
+   return property;
 }
 
-struct tgsi_dst_register_ext_modulate
-tgsi_build_dst_register_ext_modulate(
-   unsigned modulate,
-   struct tgsi_token *prev_token,
-   struct tgsi_instruction *instruction,
+
+struct tgsi_full_property
+tgsi_default_full_property( void )
+{
+   struct tgsi_full_property  full_property;
+
+   full_property.Property  = tgsi_default_property();
+   memset(full_property.u, 0,
+          sizeof(struct tgsi_property_data) * 8);
+
+   return full_property;
+}
+
+static void
+property_grow(
+   struct tgsi_property *property,
    struct tgsi_header *header )
 {
-   struct tgsi_dst_register_ext_modulate dst_register_ext_modulate;
+   assert( property->NrTokens < 0xFF );
 
-   assert( modulate <= TGSI_MODULATE_EIGHTH );
+   property->NrTokens++;
 
-   dst_register_ext_modulate = tgsi_default_dst_register_ext_modulate();
-   dst_register_ext_modulate.Modulate = modulate;
+   header_bodysize_grow( header );
+}
 
-   prev_token->Extended = 1;
-   instruction_grow( instruction, header );
+struct tgsi_property_data
+tgsi_build_property_data(
+   unsigned value,
+   struct tgsi_property *property,
+   struct tgsi_header *header )
+{
+   struct tgsi_property_data property_data;
+
+   property_data.Data = value;
+
+   property_grow( property, header );
+
+   return property_data;
+}
+
+unsigned
+tgsi_build_full_property(
+   const struct tgsi_full_property *full_prop,
+   struct tgsi_token *tokens,
+   struct tgsi_header *header,
+   unsigned maxsize )
+{
+   unsigned size = 0, i;
+   struct tgsi_property *property;
 
-   return dst_register_ext_modulate;
+   if( maxsize <= size )
+      return 0;
+   property = (struct tgsi_property *) &tokens[size];
+   size++;
+
+   *property = tgsi_build_property(
+      full_prop->Property.PropertyName,
+      header );
+
+   assert( full_prop->Property.NrTokens <= 8 + 1 );
+
+   for( i = 0; i < full_prop->Property.NrTokens - 1; i++ ) {
+      struct tgsi_property_data *data;
+
+      if( maxsize <= size )
+         return  0;
+      data = (struct tgsi_property_data *) &tokens[size];
+      size++;
+
+      *data = tgsi_build_property_data(
+         full_prop->u[i].Data,
+         property,
+         header );
+   }
+
+   return size;
 }
index 669712eb8f92221ca967e45d4280aed1082e65d5..9de2757fe40b9ecc579784a37d3af0f6d2c43c73 100644 (file)
@@ -36,12 +36,6 @@ struct tgsi_token;
 extern "C" {
 #endif
 
-/*
- * version
- */
-
-struct tgsi_version
-tgsi_build_version( void );
 
 /*
  * header
@@ -132,6 +126,34 @@ tgsi_build_full_immediate(
    struct tgsi_header *header,
    unsigned maxsize );
 
+/*
+ * properties
+ */
+
+struct tgsi_property
+tgsi_default_property( void );
+
+struct tgsi_property
+tgsi_build_property(
+   unsigned property_name,
+   struct tgsi_header *header );
+
+struct tgsi_full_property
+tgsi_default_full_property( void );
+
+struct tgsi_property_data
+tgsi_build_property_data(
+   unsigned value,
+   struct tgsi_property *property,
+   struct tgsi_header *header );
+
+unsigned
+tgsi_build_full_property(
+   const struct tgsi_full_property *full_prop,
+   struct tgsi_token *tokens,
+   struct tgsi_header *header,
+   unsigned maxsize );
+
 /*
  * instruction
  */
@@ -143,6 +165,7 @@ struct tgsi_instruction
 tgsi_build_instruction(
    unsigned opcode,
    unsigned saturate,
+   unsigned predicate,
    unsigned num_dst_regs,
    unsigned num_src_regs,
    struct tgsi_header *header );
@@ -157,54 +180,39 @@ tgsi_build_full_instruction(
    struct tgsi_header *header,
    unsigned maxsize );
 
-struct tgsi_instruction_ext_label
-tgsi_default_instruction_ext_label( void );
+struct tgsi_instruction_predicate
+tgsi_default_instruction_predicate(void);
 
-unsigned
-tgsi_compare_instruction_ext_label(
-   struct tgsi_instruction_ext_label a,
-   struct tgsi_instruction_ext_label b );
+struct tgsi_instruction_predicate
+tgsi_build_instruction_predicate(int index,
+                                 unsigned negate,
+                                 unsigned swizzleX,
+                                 unsigned swizzleY,
+                                 unsigned swizzleZ,
+                                 unsigned swizzleW,
+                                 struct tgsi_instruction *instruction,
+                                 struct tgsi_header *header);
+
+struct tgsi_instruction_label
+tgsi_default_instruction_label( void );
 
-struct tgsi_instruction_ext_label
-tgsi_build_instruction_ext_label(
+struct tgsi_instruction_label
+tgsi_build_instruction_label(
    unsigned label,
    struct tgsi_token *prev_token,
    struct tgsi_instruction *instruction,
    struct tgsi_header *header );
 
-struct tgsi_instruction_ext_texture
-tgsi_default_instruction_ext_texture( void );
+struct tgsi_instruction_texture
+tgsi_default_instruction_texture( void );
 
-unsigned
-tgsi_compare_instruction_ext_texture(
-   struct tgsi_instruction_ext_texture a,
-   struct tgsi_instruction_ext_texture b );
-
-struct tgsi_instruction_ext_texture
-tgsi_build_instruction_ext_texture(
+struct tgsi_instruction_texture
+tgsi_build_instruction_texture(
    unsigned texture,
    struct tgsi_token *prev_token,
    struct tgsi_instruction *instruction,
    struct tgsi_header *header );
 
-struct tgsi_instruction_ext_predicate
-tgsi_default_instruction_ext_predicate(void);
-
-unsigned
-tgsi_compare_instruction_ext_predicate(struct tgsi_instruction_ext_predicate a,
-                                       struct tgsi_instruction_ext_predicate b);
-
-struct tgsi_instruction_ext_predicate
-tgsi_build_instruction_ext_predicate(unsigned index,
-                                     unsigned negate,
-                                     unsigned swizzleX,
-                                     unsigned swizzleY,
-                                     unsigned swizzleZ,
-                                     unsigned swizzleW,
-                                     struct tgsi_token *prev_token,
-                                     struct tgsi_instruction *instruction,
-                                     struct tgsi_header *header);
-
 struct tgsi_src_register
 tgsi_default_src_register( void );
 
@@ -216,6 +224,7 @@ tgsi_build_src_register(
    unsigned swizzle_z,
    unsigned swizzle_w,
    unsigned negate,
+   unsigned absolute,
    unsigned indirect,
    unsigned dimension,
    int index,
@@ -225,24 +234,6 @@ tgsi_build_src_register(
 struct tgsi_full_src_register
 tgsi_default_full_src_register( void );
 
-struct tgsi_src_register_ext_mod
-tgsi_default_src_register_ext_mod( void );
-
-unsigned
-tgsi_compare_src_register_ext_mod(
-   struct tgsi_src_register_ext_mod a,
-   struct tgsi_src_register_ext_mod b );
-
-struct tgsi_src_register_ext_mod
-tgsi_build_src_register_ext_mod(
-   unsigned complement,
-   unsigned bias,
-   unsigned scale_2x,
-   unsigned absolute,
-   unsigned negate,
-   struct tgsi_token *prev_token,
-   struct tgsi_instruction *instruction,
-   struct tgsi_header *header );
 
 struct tgsi_dimension
 tgsi_default_dimension( void );
@@ -269,20 +260,6 @@ tgsi_build_dst_register(
 struct tgsi_full_dst_register
 tgsi_default_full_dst_register( void );
 
-struct tgsi_dst_register_ext_modulate
-tgsi_default_dst_register_ext_modulate( void );
-
-unsigned
-tgsi_compare_dst_register_ext_modulate(
-   struct tgsi_dst_register_ext_modulate a,
-   struct tgsi_dst_register_ext_modulate b );
-
-struct tgsi_dst_register_ext_modulate
-tgsi_build_dst_register_ext_modulate(
-   unsigned modulate,
-   struct tgsi_token *prev_token,
-   struct tgsi_instruction *instruction,
-   struct tgsi_header *header );
 
 #if defined __cplusplus
 }
index d16e64f9c59267da46923de8e05dc353ab6ced23..2c65ff16d81aa402a8fe67fbe8864910ee08f3be 100644 (file)
@@ -101,7 +101,8 @@ static const char *file_names[TGSI_FILE_COUNT] =
    "ADDR",
    "IMM",
    "LOOP",
-   "PRED"
+   "PRED",
+   "SV"
 };
 
 static const char *interpolate_names[] =
@@ -120,7 +121,9 @@ static const char *semantic_names[] =
    "PSIZE",
    "GENERIC",
    "NORMAL",
-   "FACE"
+   "FACE",
+   "EDGEFLAG",
+   "PRIM_ID"
 };
 
 static const char *immediate_type_names[] =
@@ -149,26 +152,42 @@ static const char *texture_names[] =
    "SHADOWRECT"
 };
 
+static const char *property_names[] =
+{
+   "GS_INPUT_PRIMITIVE",
+   "GS_OUTPUT_PRIMITIVE",
+   "GS_MAX_OUTPUT_VERTICES"
+};
 
-static const char *modulate_names[TGSI_MODULATE_COUNT] =
+static const char *primitive_names[] =
 {
-   "",
-   "_2X",
-   "_4X",
-   "_8X",
-   "_D2",
-   "_D4",
-   "_D8"
+   "POINTS",
+   "LINES",
+   "LINE_LOOP",
+   "LINE_STRIP",
+   "TRIANGLES",
+   "TRIANGLE_STRIP",
+   "TRIANGLE_FAN",
+   "QUADS",
+   "QUAD_STRIP",
+   "POLYGON"
 };
 
+
 static void
-_dump_register(
+_dump_register_decl(
    struct dump_ctx *ctx,
    uint file,
    int first,
    int last )
 {
    ENM( file, file_names );
+
+   /* all geometry shader inputs are two dimensional */
+   if (file == TGSI_FILE_INPUT &&
+       ctx->iter.processor.Processor == TGSI_PROCESSOR_GEOMETRY)
+      TXT("[]");
+
    CHR( '[' );
    SID( first );
    if (first != last) {
@@ -178,6 +197,52 @@ _dump_register(
    CHR( ']' );
 }
 
+static void
+_dump_register_dst(
+   struct dump_ctx *ctx,
+   uint file,
+   int index)
+{
+   ENM( file, file_names );
+
+   CHR( '[' );
+   SID( index );
+   CHR( ']' );
+}
+
+
+static void
+_dump_register_src(
+   struct dump_ctx *ctx,
+   const struct tgsi_full_src_register *src )
+{
+   if (src->Register.Indirect) {
+      ENM( src->Register.File, file_names );
+      CHR( '[' );
+      ENM( src->Indirect.File, file_names );
+      CHR( '[' );
+      SID( src->Indirect.Index );
+      TXT( "]." );
+      ENM( src->Indirect.SwizzleX, swizzle_names );
+      if (src->Register.Index != 0) {
+         if (src->Register.Index > 0)
+            CHR( '+' );
+         SID( src->Register.Index );
+      }
+      CHR( ']' );
+   } else {
+      ENM( src->Register.File, file_names );
+      CHR( '[' );
+      SID( src->Register.Index );
+      CHR( ']' );
+   }
+   if (src->Register.Dimension) {
+      CHR( '[' );
+      SID( src->Dimension.Index );
+      CHR( ']' );
+   }
+}
+
 static void
 _dump_register_ind(
    struct dump_ctx *ctx,
@@ -232,22 +297,22 @@ iter_declaration(
 
    TXT( "DCL " );
 
-   _dump_register(
+   _dump_register_decl(
       ctx,
       decl->Declaration.File,
-      decl->DeclarationRange.First,
-      decl->DeclarationRange.Last );
+      decl->Range.First,
+      decl->Range.Last );
    _dump_writemask(
       ctx,
       decl->Declaration.UsageMask );
 
    if (decl->Declaration.Semantic) {
       TXT( ", " );
-      ENM( decl->Semantic.SemanticName, semantic_names );
-      if (decl->Semantic.SemanticIndex != 0 ||
-          decl->Semantic.SemanticName == TGSI_SEMANTIC_GENERIC) {
+      ENM( decl->Semantic.Name, semantic_names );
+      if (decl->Semantic.Index != 0 ||
+          decl->Semantic.Name == TGSI_SEMANTIC_GENERIC) {
          CHR( '[' );
-         UID( decl->Semantic.SemanticIndex );
+         UID( decl->Semantic.Index );
          CHR( ']' );
       }
    }
@@ -283,6 +348,50 @@ tgsi_dump_declaration(
    iter_declaration( &ctx.iter, (struct tgsi_full_declaration *)decl );
 }
 
+static boolean
+iter_property(
+   struct tgsi_iterate_context *iter,
+   struct tgsi_full_property *prop )
+{
+   int i;
+   struct dump_ctx *ctx = (struct dump_ctx *)iter;
+
+   assert(Elements(property_names) == TGSI_PROPERTY_COUNT);
+
+   TXT( "PROPERTY " );
+   ENM(prop->Property.PropertyName, property_names);
+
+   if (prop->Property.NrTokens > 1)
+      TXT(" ");
+
+   for (i = 0; i < prop->Property.NrTokens - 1; ++i) {
+      switch (prop->Property.PropertyName) {
+      case TGSI_PROPERTY_GS_INPUT_PRIM:
+      case TGSI_PROPERTY_GS_OUTPUT_PRIM:
+         ENM(prop->u[i].Data, primitive_names);
+         break;
+      default:
+         SID( prop->u[i].Data );
+         break;
+      }
+      if (i < prop->Property.NrTokens - 2)
+         TXT( ", " );
+   }
+   EOL();
+
+   return TRUE;
+}
+
+void tgsi_dump_property(
+   const struct tgsi_full_property *prop )
+{
+   struct dump_ctx ctx;
+
+   ctx.printf = dump_ctx_printf;
+
+   iter_property( &ctx.iter, (struct tgsi_full_property *)prop );
+}
+
 static boolean
 iter_immediate(
    struct tgsi_iterate_context *iter,
@@ -363,99 +472,66 @@ iter_instruction(
    }
 
    for (i = 0; i < inst->Instruction.NumDstRegs; i++) {
-      const struct tgsi_full_dst_register *dst = &inst->FullDstRegisters[i];
+      const struct tgsi_full_dst_register *dst = &inst->Dst[i];
 
       if (!first_reg)
          CHR( ',' );
       CHR( ' ' );
 
-      if (dst->DstRegister.Indirect) {
+      if (dst->Register.Indirect) {
          _dump_register_ind(
             ctx,
-            dst->DstRegister.File,
-            dst->DstRegister.Index,
-            dst->DstRegisterInd.File,
-            dst->DstRegisterInd.Index,
-            dst->DstRegisterInd.SwizzleX );
+            dst->Register.File,
+            dst->Register.Index,
+            dst->Indirect.File,
+            dst->Indirect.Index,
+            dst->Indirect.SwizzleX );
       }
       else {
-         _dump_register(
+         _dump_register_dst(
             ctx,
-            dst->DstRegister.File,
-            dst->DstRegister.Index,
-            dst->DstRegister.Index );
+            dst->Register.File,
+            dst->Register.Index );
       }
-      ENM( dst->DstRegisterExtModulate.Modulate, modulate_names );
-      _dump_writemask( ctx, dst->DstRegister.WriteMask );
+      _dump_writemask( ctx, dst->Register.WriteMask );
 
       first_reg = FALSE;
    }
 
    for (i = 0; i < inst->Instruction.NumSrcRegs; i++) {
-      const struct tgsi_full_src_register *src = &inst->FullSrcRegisters[i];
+      const struct tgsi_full_src_register *src = &inst->Src[i];
 
       if (!first_reg)
          CHR( ',' );
       CHR( ' ' );
 
-      if (src->SrcRegisterExtMod.Negate)
-         TXT( "-(" );
-      if (src->SrcRegisterExtMod.Absolute)
-         CHR( '|' );
-      if (src->SrcRegisterExtMod.Scale2X)
-         TXT( "2*(" );
-      if (src->SrcRegisterExtMod.Bias)
-         CHR( '(' );
-      if (src->SrcRegisterExtMod.Complement)
-         TXT( "1-(" );
-      if (src->SrcRegister.Negate)
+      if (src->Register.Negate)
          CHR( '-' );
+      if (src->Register.Absolute)
+         CHR( '|' );
 
-      if (src->SrcRegister.Indirect) {
-         _dump_register_ind(
-            ctx,
-            src->SrcRegister.File,
-            src->SrcRegister.Index,
-            src->SrcRegisterInd.File,
-            src->SrcRegisterInd.Index,
-            src->SrcRegisterInd.SwizzleX );
-      }
-      else {
-         _dump_register(
-            ctx,
-            src->SrcRegister.File,
-            src->SrcRegister.Index,
-            src->SrcRegister.Index );
-      }
+      _dump_register_src(ctx, src);
 
-      if (src->SrcRegister.SwizzleX != TGSI_SWIZZLE_X ||
-          src->SrcRegister.SwizzleY != TGSI_SWIZZLE_Y ||
-          src->SrcRegister.SwizzleZ != TGSI_SWIZZLE_Z ||
-          src->SrcRegister.SwizzleW != TGSI_SWIZZLE_W) {
+      if (src->Register.SwizzleX != TGSI_SWIZZLE_X ||
+          src->Register.SwizzleY != TGSI_SWIZZLE_Y ||
+          src->Register.SwizzleZ != TGSI_SWIZZLE_Z ||
+          src->Register.SwizzleW != TGSI_SWIZZLE_W) {
          CHR( '.' );
-         ENM( src->SrcRegister.SwizzleX, swizzle_names );
-         ENM( src->SrcRegister.SwizzleY, swizzle_names );
-         ENM( src->SrcRegister.SwizzleZ, swizzle_names );
-         ENM( src->SrcRegister.SwizzleW, swizzle_names );
+         ENM( src->Register.SwizzleX, swizzle_names );
+         ENM( src->Register.SwizzleY, swizzle_names );
+         ENM( src->Register.SwizzleZ, swizzle_names );
+         ENM( src->Register.SwizzleW, swizzle_names );
       }
 
-      if (src->SrcRegisterExtMod.Complement)
-         CHR( ')' );
-      if (src->SrcRegisterExtMod.Bias)
-         TXT( ")-.5" );
-      if (src->SrcRegisterExtMod.Scale2X)
-         CHR( ')' );
-      if (src->SrcRegisterExtMod.Absolute)
+      if (src->Register.Absolute)
          CHR( '|' );
-      if (src->SrcRegisterExtMod.Negate)
-         CHR( ')' );
 
       first_reg = FALSE;
    }
 
-   if (inst->InstructionExtTexture.Texture != TGSI_TEXTURE_UNKNOWN) {
+   if (inst->Instruction.Texture) {
       TXT( ", " );
-      ENM( inst->InstructionExtTexture.Texture, texture_names );
+      ENM( inst->Texture.Texture, texture_names );
    }
 
    switch (inst->Instruction.Opcode) {
@@ -465,7 +541,7 @@ iter_instruction(
    case TGSI_OPCODE_ENDLOOP:
    case TGSI_OPCODE_CAL:
       TXT( " :" );
-      UID( inst->InstructionExtLabel.Label );
+      UID( inst->Label.Label );
       break;
    }
 
@@ -503,9 +579,6 @@ prolog(
 {
    struct dump_ctx *ctx = (struct dump_ctx *) iter;
    ENM( iter->processor.Processor, processor_type_names );
-   UID( iter->version.MajorVersion );
-   CHR( '.' );
-   UID( iter->version.MinorVersion );
    EOL();
    return TRUE;
 }
@@ -521,6 +594,7 @@ tgsi_dump(
    ctx.iter.iterate_instruction = iter_instruction;
    ctx.iter.iterate_declaration = iter_declaration;
    ctx.iter.iterate_immediate = iter_immediate;
+   ctx.iter.iterate_property = iter_property;
    ctx.iter.epilog = NULL;
 
    ctx.instno = 0;
@@ -575,6 +649,7 @@ tgsi_dump_str(
    ctx.base.iter.iterate_instruction = iter_instruction;
    ctx.base.iter.iterate_declaration = iter_declaration;
    ctx.base.iter.iterate_immediate = iter_immediate;
+   ctx.base.iter.iterate_property = iter_property;
    ctx.base.iter.epilog = NULL;
 
    ctx.base.instno = 0;
index ad1e647ec9000b0b2956627e4393d3c38c082a5c..4cd27317b36792b5d60e53c8efca3ee5dc51b054 100644 (file)
@@ -49,6 +49,7 @@ tgsi_dump(
 struct tgsi_full_immediate;
 struct tgsi_full_instruction;
 struct tgsi_full_declaration;
+struct tgsi_full_property;
 
 void
 tgsi_dump_immediate(
@@ -63,6 +64,10 @@ void
 tgsi_dump_declaration(
    const struct tgsi_full_declaration *decl );
 
+void
+tgsi_dump_property(
+   const struct tgsi_full_property *prop );
+
 #if defined __cplusplus
 }
 #endif
index 4648051e29e5a1ea178bf272926f7e756b2c06a0..47fd1dd590eb71b95f77756fdff480b60660ce21 100644 (file)
@@ -113,13 +113,6 @@ static const char *TGSI_SATS[] =
    "SAT_MINUS_PLUS_ONE"
 };
 
-static const char *TGSI_INSTRUCTION_EXTS[] =
-{
-   "",
-   "INSTRUCTION_EXT_TYPE_LABEL",
-   "INSTRUCTION_EXT_TYPE_TEXTURE"
-};
-
 static const char *TGSI_SWIZZLES[] =
 {
    "SWIZZLE_X",
@@ -141,12 +134,6 @@ static const char *TGSI_TEXTURES[] =
    "TEXTURE_SHADOWRECT"
 };
 
-static const char *TGSI_SRC_REGISTER_EXTS[] =
-{
-   "",
-   "SRC_REGISTER_EXT_TYPE_MOD"
-};
-
 static const char *TGSI_WRITEMASKS[] =
 {
    "0",
@@ -167,23 +154,6 @@ static const char *TGSI_WRITEMASKS[] =
    "WRITEMASK_XYZW"
 };
 
-static const char *TGSI_DST_REGISTER_EXTS[] =
-{
-   "",
-   "DST_REGISTER_EXT_TYPE_MODULATE"
-};
-
-static const char *TGSI_MODULATES[] =
-{
-   "MODULATE_1X",
-   "MODULATE_2X",
-   "MODULATE_4X",
-   "MODULATE_8X",
-   "MODULATE_HALF",
-   "MODULATE_QUARTER",
-   "MODULATE_EIGHTH"
-};
-
 static void
 dump_declaration_verbose(
    struct tgsi_full_declaration  *decl,
@@ -216,6 +186,14 @@ dump_declaration_verbose(
       TXT( "\nSemantic   : " );
       UID( decl->Declaration.Semantic );
    }
+   if (deflt || fd->Declaration.Centroid != decl->Declaration.Centroid) {
+      TXT("\nCentroid   : ");
+      UID(decl->Declaration.Centroid);
+   }
+   if (deflt || fd->Declaration.Invariant != decl->Declaration.Invariant) {
+      TXT("\nInvariant  : ");
+      UID(decl->Declaration.Invariant);
+   }
    if( ignored ) {
       TXT( "\nPadding    : " );
       UIX( decl->Declaration.Padding );
@@ -223,16 +201,16 @@ dump_declaration_verbose(
 
    EOL();
    TXT( "\nFirst: " );
-   UID( decl->DeclarationRange.First );
+   UID( decl->Range.First );
    TXT( "\nLast : " );
-   UID( decl->DeclarationRange.Last );
+   UID( decl->Range.Last );
 
    if( decl->Declaration.Semantic ) {
       EOL();
-      TXT( "\nSemanticName : " );
-      ENM( decl->Semantic.SemanticName, TGSI_SEMANTICS );
-      TXT( "\nSemanticIndex: " );
-      UID( decl->Semantic.SemanticIndex );
+      TXT( "\nName : " );
+      ENM( decl->Semantic.Name, TGSI_SEMANTICS );
+      TXT( "\nIndex: " );
+      UID( decl->Semantic.Index );
       if( ignored ) {
          TXT( "\nPadding      : " );
          UIX( decl->Semantic.Padding );
@@ -292,180 +270,122 @@ dump_instruction_verbose(
       TXT( "\nNumSrcRegs : " );
       UID( inst->Instruction.NumSrcRegs );
    }
+   if (deflt || fi->Instruction.Predicate != inst->Instruction.Predicate) {
+      TXT("\nPredicate  : ");
+      UID(inst->Instruction.Predicate);
+   }
+   if (deflt || fi->Instruction.Label != inst->Instruction.Label) {
+      TXT("\nLabel      : ");
+      UID(inst->Instruction.Label);
+   }
+   if (deflt || fi->Instruction.Texture != inst->Instruction.Texture) {
+      TXT("\nTexture    : ");
+      UID(inst->Instruction.Texture);
+   }
    if( ignored ) {
       TXT( "\nPadding    : " );
       UIX( inst->Instruction.Padding );
    }
 
-   if( deflt || tgsi_compare_instruction_ext_label( inst->InstructionExtLabel, fi->InstructionExtLabel ) ) {
+   if (deflt || inst->Instruction.Label) {
       EOL();
-      TXT( "\nType    : " );
-      ENM( inst->InstructionExtLabel.Type, TGSI_INSTRUCTION_EXTS );
-      if( deflt || fi->InstructionExtLabel.Label != inst->InstructionExtLabel.Label ) {
+      if (deflt || fi->Label.Label != inst->Label.Label) {
          TXT( "\nLabel   : " );
-         UID( inst->InstructionExtLabel.Label );
+         UID(inst->Label.Label);
       }
       if( ignored ) {
          TXT( "\nPadding : " );
-         UIX( inst->InstructionExtLabel.Padding );
-         if( deflt || fi->InstructionExtLabel.Extended != inst->InstructionExtLabel.Extended ) {
-            TXT( "\nExtended: " );
-            UID( inst->InstructionExtLabel.Extended );
-         }
+         UIX(inst->Label.Padding);
       }
    }
 
-   if( deflt || tgsi_compare_instruction_ext_texture( inst->InstructionExtTexture, fi->InstructionExtTexture ) ) {
+   if (deflt || inst->Instruction.Texture) {
       EOL();
-      TXT( "\nType    : " );
-      ENM( inst->InstructionExtTexture.Type, TGSI_INSTRUCTION_EXTS );
-      if( deflt || fi->InstructionExtTexture.Texture != inst->InstructionExtTexture.Texture ) {
+      if (deflt || fi->Texture.Texture != inst->Texture.Texture) {
          TXT( "\nTexture : " );
-         ENM( inst->InstructionExtTexture.Texture, TGSI_TEXTURES );
+         ENM(inst->Texture.Texture, TGSI_TEXTURES);
       }
       if( ignored ) {
          TXT( "\nPadding : " );
-         UIX( inst->InstructionExtTexture.Padding );
-         if( deflt || fi->InstructionExtTexture.Extended != inst->InstructionExtTexture.Extended ) {
-            TXT( "\nExtended: " );
-            UID( inst->InstructionExtTexture.Extended );
-         }
+         UIX(inst->Texture.Padding);
       }
    }
 
    for( i = 0; i < inst->Instruction.NumDstRegs; i++ ) {
-      struct tgsi_full_dst_register *dst = &inst->FullDstRegisters[i];
-      struct tgsi_full_dst_register *fd = &fi->FullDstRegisters[i];
+      struct tgsi_full_dst_register *dst = &inst->Dst[i];
+      struct tgsi_full_dst_register *fd = &fi->Dst[i];
 
       EOL();
       TXT( "\nFile     : " );
-      ENM( dst->DstRegister.File, TGSI_FILES );
-      if( deflt || fd->DstRegister.WriteMask != dst->DstRegister.WriteMask ) {
+      ENM( dst->Register.File, TGSI_FILES );
+      if( deflt || fd->Register.WriteMask != dst->Register.WriteMask ) {
          TXT( "\nWriteMask: " );
-         ENM( dst->DstRegister.WriteMask, TGSI_WRITEMASKS );
+         ENM( dst->Register.WriteMask, TGSI_WRITEMASKS );
       }
       if( ignored ) {
-         if( deflt || fd->DstRegister.Indirect != dst->DstRegister.Indirect ) {
+         if( deflt || fd->Register.Indirect != dst->Register.Indirect ) {
             TXT( "\nIndirect : " );
-            UID( dst->DstRegister.Indirect );
+            UID( dst->Register.Indirect );
          }
-         if( deflt || fd->DstRegister.Dimension != dst->DstRegister.Dimension ) {
+         if( deflt || fd->Register.Dimension != dst->Register.Dimension ) {
             TXT( "\nDimension: " );
-            UID( dst->DstRegister.Dimension );
+            UID( dst->Register.Dimension );
          }
       }
-      if( deflt || fd->DstRegister.Index != dst->DstRegister.Index ) {
+      if( deflt || fd->Register.Index != dst->Register.Index ) {
          TXT( "\nIndex    : " );
-         SID( dst->DstRegister.Index );
+         SID( dst->Register.Index );
       }
       if( ignored ) {
          TXT( "\nPadding  : " );
-         UIX( dst->DstRegister.Padding );
-         if( deflt || fd->DstRegister.Extended != dst->DstRegister.Extended ) {
-            TXT( "\nExtended : " );
-            UID( dst->DstRegister.Extended );
-         }
-      }
-
-      if( deflt || tgsi_compare_dst_register_ext_modulate( dst->DstRegisterExtModulate, fd->DstRegisterExtModulate ) ) {
-         EOL();
-         TXT( "\nType    : " );
-         ENM( dst->DstRegisterExtModulate.Type, TGSI_DST_REGISTER_EXTS );
-         if( deflt || fd->DstRegisterExtModulate.Modulate != dst->DstRegisterExtModulate.Modulate ) {
-            TXT( "\nModulate: " );
-            ENM( dst->DstRegisterExtModulate.Modulate, TGSI_MODULATES );
-         }
-         if( ignored ) {
-            TXT( "\nPadding : " );
-            UIX( dst->DstRegisterExtModulate.Padding );
-            if( deflt || fd->DstRegisterExtModulate.Extended != dst->DstRegisterExtModulate.Extended ) {
-               TXT( "\nExtended: " );
-               UID( dst->DstRegisterExtModulate.Extended );
-            }
-         }
+         UIX( dst->Register.Padding );
       }
    }
 
    for( i = 0; i < inst->Instruction.NumSrcRegs; i++ ) {
-      struct tgsi_full_src_register *src = &inst->FullSrcRegisters[i];
-      struct tgsi_full_src_register *fs = &fi->FullSrcRegisters[i];
+      struct tgsi_full_src_register *src = &inst->Src[i];
+      struct tgsi_full_src_register *fs = &fi->Src[i];
 
       EOL();
       TXT( "\nFile     : ");
-      ENM( src->SrcRegister.File, TGSI_FILES );
-      if( deflt || fs->SrcRegister.SwizzleX != src->SrcRegister.SwizzleX ) {
+      ENM( src->Register.File, TGSI_FILES );
+      if( deflt || fs->Register.SwizzleX != src->Register.SwizzleX ) {
          TXT( "\nSwizzleX : " );
-         ENM( src->SrcRegister.SwizzleX, TGSI_SWIZZLES );
+         ENM( src->Register.SwizzleX, TGSI_SWIZZLES );
       }
-      if( deflt || fs->SrcRegister.SwizzleY != src->SrcRegister.SwizzleY ) {
+      if( deflt || fs->Register.SwizzleY != src->Register.SwizzleY ) {
          TXT( "\nSwizzleY : " );
-         ENM( src->SrcRegister.SwizzleY, TGSI_SWIZZLES );
+         ENM( src->Register.SwizzleY, TGSI_SWIZZLES );
       }
-      if( deflt || fs->SrcRegister.SwizzleZ != src->SrcRegister.SwizzleZ ) {
+      if( deflt || fs->Register.SwizzleZ != src->Register.SwizzleZ ) {
          TXT( "\nSwizzleZ : " );
-         ENM( src->SrcRegister.SwizzleZ, TGSI_SWIZZLES );
+         ENM( src->Register.SwizzleZ, TGSI_SWIZZLES );
       }
-      if( deflt || fs->SrcRegister.SwizzleW != src->SrcRegister.SwizzleW ) {
+      if( deflt || fs->Register.SwizzleW != src->Register.SwizzleW ) {
          TXT( "\nSwizzleW : " );
-         ENM( src->SrcRegister.SwizzleW, TGSI_SWIZZLES );
+         ENM( src->Register.SwizzleW, TGSI_SWIZZLES );
+      }
+      if (deflt || fs->Register.Absolute != src->Register.Absolute) {
+         TXT("\nAbsolute : ");
+         UID(src->Register.Absolute);
       }
-      if( deflt || fs->SrcRegister.Negate != src->SrcRegister.Negate ) {
+      if( deflt || fs->Register.Negate != src->Register.Negate ) {
          TXT( "\nNegate   : " );
-         UID( src->SrcRegister.Negate );
+         UID( src->Register.Negate );
       }
       if( ignored ) {
-         if( deflt || fs->SrcRegister.Indirect != src->SrcRegister.Indirect ) {
+         if( deflt || fs->Register.Indirect != src->Register.Indirect ) {
             TXT( "\nIndirect : " );
-            UID( src->SrcRegister.Indirect );
+            UID( src->Register.Indirect );
          }
-         if( deflt || fs->SrcRegister.Dimension != src->SrcRegister.Dimension ) {
+         if( deflt || fs->Register.Dimension != src->Register.Dimension ) {
             TXT( "\nDimension: " );
-            UID( src->SrcRegister.Dimension );
+            UID( src->Register.Dimension );
          }
       }
-      if( deflt || fs->SrcRegister.Index != src->SrcRegister.Index ) {
+      if( deflt || fs->Register.Index != src->Register.Index ) {
          TXT( "\nIndex    : " );
-         SID( src->SrcRegister.Index );
-      }
-      if( ignored ) {
-         if( deflt || fs->SrcRegister.Extended != src->SrcRegister.Extended ) {
-            TXT( "\nExtended : " );
-            UID( src->SrcRegister.Extended );
-         }
-      }
-
-      if( deflt || tgsi_compare_src_register_ext_mod( src->SrcRegisterExtMod, fs->SrcRegisterExtMod ) ) {
-         EOL();
-         TXT( "\nType     : " );
-         ENM( src->SrcRegisterExtMod.Type, TGSI_SRC_REGISTER_EXTS );
-         if( deflt || fs->SrcRegisterExtMod.Complement != src->SrcRegisterExtMod.Complement ) {
-            TXT( "\nComplement: " );
-            UID( src->SrcRegisterExtMod.Complement );
-         }
-         if( deflt || fs->SrcRegisterExtMod.Bias != src->SrcRegisterExtMod.Bias ) {
-            TXT( "\nBias     : " );
-            UID( src->SrcRegisterExtMod.Bias );
-         }
-         if( deflt || fs->SrcRegisterExtMod.Scale2X != src->SrcRegisterExtMod.Scale2X ) {
-            TXT( "\nScale2X   : " );
-            UID( src->SrcRegisterExtMod.Scale2X );
-         }
-         if( deflt || fs->SrcRegisterExtMod.Absolute != src->SrcRegisterExtMod.Absolute ) {
-            TXT( "\nAbsolute  : " );
-            UID( src->SrcRegisterExtMod.Absolute );
-         }
-         if( deflt || fs->SrcRegisterExtMod.Negate != src->SrcRegisterExtMod.Negate ) {
-            TXT( "\nNegate   : " );
-            UID( src->SrcRegisterExtMod.Negate );
-         }
-         if( ignored ) {
-            TXT( "\nPadding   : " );
-            UIX( src->SrcRegisterExtMod.Padding );
-            if( deflt || fs->SrcRegisterExtMod.Extended != src->SrcRegisterExtMod.Extended ) {
-               TXT( "\nExtended  : " );
-               UID( src->SrcRegisterExtMod.Extended );
-            }
-         }
+         SID( src->Register.Index );
       }
    }
 }
@@ -485,12 +405,6 @@ tgsi_dump_c(
 
    TXT( "tgsi-dump begin -----------------" );
 
-   TXT( "\nMajorVersion: " );
-   UID( parse.FullVersion.Version.MajorVersion );
-   TXT( "\nMinorVersion: " );
-   UID( parse.FullVersion.Version.MinorVersion );
-   EOL();
-
    TXT( "\nHeaderSize: " );
    UID( parse.FullHeader.Header.HeaderSize );
    TXT( "\nBodySize  : " );
@@ -510,10 +424,6 @@ tgsi_dump_c(
       if( ignored ) {
          TXT( "\nSize       : " );
          UID( parse.FullToken.Token.NrTokens );
-         if( deflt || parse.FullToken.Token.Extended ) {
-            TXT( "\nExtended   : " );
-            UID( parse.FullToken.Token.Extended );
-         }
       }
 
       switch( parse.FullToken.Token.Type ) {
index b7569e74d4b6986ddf172eccdfbb707cb229f2e8..ba89f2fbc3cfeda7419b0c53a57e600403f1dd30 100644 (file)
@@ -62,9 +62,6 @@
 
 #define FAST_MATH 1
 
-/** for tgsi_full_instruction::Flags */
-#define SOA_DEPENDENCY_FLAG 0x1
-
 #define TILE_TOP_LEFT     0
 #define TILE_TOP_RIGHT    1
 #define TILE_BOTTOM_LEFT  2
 #define TEMP_P0            TGSI_EXEC_TEMP_P0
 
 #define IS_CHANNEL_ENABLED(INST, CHAN)\
-   ((INST).FullDstRegisters[0].DstRegister.WriteMask & (1 << (CHAN)))
+   ((INST).Dst[0].Register.WriteMask & (1 << (CHAN)))
 
 #define IS_CHANNEL_ENABLED2(INST, CHAN)\
-   ((INST).FullDstRegisters[1].DstRegister.WriteMask & (1 << (CHAN)))
+   ((INST).Dst[1].Register.WriteMask & (1 << (CHAN)))
 
 #define FOR_EACH_ENABLED_CHANNEL(INST, CHAN)\
    for (CHAN = 0; CHAN < NUM_CHANNELS; CHAN++)\
@@ -191,7 +188,7 @@ tgsi_check_soa_dependencies(const struct tgsi_full_instruction *inst)
 {
    uint i, chan;
 
-   uint writemask = inst->FullDstRegisters[0].DstRegister.WriteMask;
+   uint writemask = inst->Dst[0].Register.WriteMask;
    if (writemask == TGSI_WRITEMASK_X ||
        writemask == TGSI_WRITEMASK_Y ||
        writemask == TGSI_WRITEMASK_Z ||
@@ -203,15 +200,15 @@ tgsi_check_soa_dependencies(const struct tgsi_full_instruction *inst)
 
    /* loop over src regs */
    for (i = 0; i < inst->Instruction.NumSrcRegs; i++) {
-      if ((inst->FullSrcRegisters[i].SrcRegister.File ==
-           inst->FullDstRegisters[0].DstRegister.File) &&
-          (inst->FullSrcRegisters[i].SrcRegister.Index ==
-           inst->FullDstRegisters[0].DstRegister.Index)) {
+      if ((inst->Src[i].Register.File ==
+           inst->Dst[0].Register.File) &&
+          (inst->Src[i].Register.Index ==
+           inst->Dst[0].Register.Index)) {
          /* loop over dest channels */
          uint channelsWritten = 0x0;
          FOR_EACH_ENABLED_CHANNEL(*inst, chan) {
             /* check if we're reading a channel that's been written */
-            uint swizzle = tgsi_util_get_full_src_register_swizzle(&inst->FullSrcRegisters[i], chan);
+            uint swizzle = tgsi_util_get_full_src_register_swizzle(&inst->Src[i], chan);
             if (channelsWritten & (1 << swizzle)) {
                return TRUE;
             }
@@ -295,6 +292,14 @@ tgsi_exec_machine_bind_shader(
                                    * sizeof(struct tgsi_full_declaration));
             maxDeclarations += 10;
          }
+         if (parse.FullToken.FullDeclaration.Declaration.File == TGSI_FILE_OUTPUT) {
+            unsigned reg;
+            for (reg = parse.FullToken.FullDeclaration.Range.First;
+                 reg <= parse.FullToken.FullDeclaration.Range.Last;
+                 ++reg) {
+               ++mach->NumOutputs;
+            }
+         }
          memcpy(declarations + numDeclarations,
                 &parse.FullToken.FullDeclaration,
                 sizeof(declarations[0]));
@@ -332,20 +337,6 @@ tgsi_exec_machine_bind_shader(
             maxInstructions += 10;
          }
 
-         if (tgsi_check_soa_dependencies(&parse.FullToken.FullInstruction)) {
-            uint opcode = parse.FullToken.FullInstruction.Instruction.Opcode;
-            parse.FullToken.FullInstruction.Flags = SOA_DEPENDENCY_FLAG;
-            /* XXX we only handle SOA dependencies properly for MOV/SWZ
-             * at this time!
-             */
-            if (opcode != TGSI_OPCODE_MOV) {
-               debug_printf("Warning: SOA dependency in instruction"
-                            " is not handled:\n");
-               tgsi_dump_instruction(&parse.FullToken.FullInstruction,
-                                     numInstructions);
-            }
-         }
-
          memcpy(instructions + numInstructions,
                 &parse.FullToken.FullInstruction,
                 sizeof(instructions[0]));
@@ -353,6 +344,9 @@ tgsi_exec_machine_bind_shader(
          numInstructions++;
          break;
 
+      case TGSI_TOKEN_TYPE_PROPERTY:
+         break;
+
       default:
          assert( 0 );
       }
@@ -386,6 +380,8 @@ tgsi_exec_machine_create( void )
    memset(mach, 0, sizeof(*mach));
 
    mach->Addrs = &mach->Temps[TGSI_EXEC_TEMP_ADDR];
+   mach->MaxGeometryShaderOutputs = TGSI_MAX_TOTAL_VERTICES;
+   mach->Predicates = &mach->Temps[TGSI_EXEC_TEMP_P0];
 
    /* Setup constants. */
    for( i = 0; i < 4; i++ ) {
@@ -517,7 +513,7 @@ micro_ddy(
    dst->f[0] =
    dst->f[1] =
    dst->f[2] =
-   dst->f[3] = src->f[TILE_TOP_LEFT] - src->f[TILE_BOTTOM_LEFT];
+   dst->f[3] = src->f[TILE_BOTTOM_LEFT] - src->f[TILE_TOP_LEFT];
 }
 
 static void
@@ -595,6 +591,24 @@ micro_exp2(
    dst->f[2] = util_fast_exp2( src->f[2] );
    dst->f[3] = util_fast_exp2( src->f[3] );
 #else
+
+#if DEBUG
+   /* Inf is okay for this instruction, so clamp it to silence assertions. */
+   uint i;
+   union tgsi_exec_channel clamped;
+
+   for (i = 0; i < 4; i++) {
+      if (src->f[i] > 127.99999f) {
+         clamped.f[i] = 127.99999f;
+      } else if (src->f[i] < -126.99999f) {
+         clamped.f[i] = -126.99999f;
+      } else {
+         clamped.f[i] = src->f[i];
+      }
+   }
+   src = &clamped;
+#endif
+
    dst->f[0] = powf( 2.0f, src->f[0] );
    dst->f[1] = powf( 2.0f, src->f[1] );
    dst->f[2] = powf( 2.0f, src->f[2] );
@@ -1156,6 +1170,7 @@ fetch_src_file_channel(
          break;
 
       case TGSI_FILE_INPUT:
+      case TGSI_FILE_SYSTEM_VALUE:
          chan->u[0] = mach->Inputs[index->i[0]].xyzw[swizzle].u[0];
          chan->u[1] = mach->Inputs[index->i[1]].xyzw[swizzle].u[1];
          chan->u[2] = mach->Inputs[index->i[2]].xyzw[swizzle].u[2];
@@ -1193,10 +1208,10 @@ fetch_src_file_channel(
          assert(index->i[1] < TGSI_EXEC_NUM_PREDS);
          assert(index->i[2] < TGSI_EXEC_NUM_PREDS);
          assert(index->i[3] < TGSI_EXEC_NUM_PREDS);
-         chan->u[0] = mach->Addrs[0].xyzw[swizzle].u[0];
-         chan->u[1] = mach->Addrs[0].xyzw[swizzle].u[1];
-         chan->u[2] = mach->Addrs[0].xyzw[swizzle].u[2];
-         chan->u[3] = mach->Addrs[0].xyzw[swizzle].u[3];
+         chan->u[0] = mach->Predicates[0].xyzw[swizzle].u[0];
+         chan->u[1] = mach->Predicates[0].xyzw[swizzle].u[1];
+         chan->u[2] = mach->Predicates[0].xyzw[swizzle].u[2];
+         chan->u[3] = mach->Predicates[0].xyzw[swizzle].u[3];
          break;
 
       case TGSI_FILE_OUTPUT:
@@ -1231,13 +1246,13 @@ fetch_source(
     *
     *    file[1],
     *    where:
-    *       file = SrcRegister.File
-    *       [1] = SrcRegister.Index
+    *       file = Register.File
+    *       [1] = Register.Index
     */
    index.i[0] =
    index.i[1] =
    index.i[2] =
-   index.i[3] = reg->SrcRegister.Index;
+   index.i[3] = reg->Register.Index;
 
    /* There is an extra source register that indirectly subscripts
     * a register file. The direct index now becomes an offset
@@ -1245,11 +1260,11 @@ fetch_source(
     *
     *    file[ind[2].x+1],
     *    where:
-    *       ind = SrcRegisterInd.File
-    *       [2] = SrcRegisterInd.Index
-    *       .x = SrcRegisterInd.SwizzleX
+    *       ind = Indirect.File
+    *       [2] = Indirect.Index
+    *       .x = Indirect.SwizzleX
     */
-   if (reg->SrcRegister.Indirect) {
+   if (reg->Register.Indirect) {
       union tgsi_exec_channel index2;
       union tgsi_exec_channel indir_index;
       const uint execmask = mach->ExecMask;
@@ -1259,13 +1274,13 @@ fetch_source(
       index2.i[0] =
       index2.i[1] =
       index2.i[2] =
-      index2.i[3] = reg->SrcRegisterInd.Index;
+      index2.i[3] = reg->Indirect.Index;
 
       /* get current value of address register[swizzle] */
-      swizzle = tgsi_util_get_src_register_swizzle( &reg->SrcRegisterInd, CHAN_X );
+      swizzle = tgsi_util_get_src_register_swizzle( &reg->Indirect, CHAN_X );
       fetch_src_file_channel(
          mach,
-         reg->SrcRegisterInd.File,
+         reg->Indirect.File,
          swizzle,
          &index2,
          &indir_index );
@@ -1291,15 +1306,16 @@ fetch_source(
     *
     *    file[1][3] == file[1*sizeof(file[1])+3],
     *    where:
-    *       [3] = SrcRegisterDim.Index
+    *       [3] = Dimension.Index
     */
-   if (reg->SrcRegister.Dimension) {
+   if (reg->Register.Dimension) {
       /* The size of the first-order array depends on the register file type.
        * We need to multiply the index to the first array to get an effective,
        * "flat" index that points to the beginning of the second-order array.
        */
-      switch (reg->SrcRegister.File) {
+      switch (reg->Register.File) {
       case TGSI_FILE_INPUT:
+      case TGSI_FILE_SYSTEM_VALUE:
          index.i[0] *= TGSI_EXEC_MAX_INPUT_ATTRIBS;
          index.i[1] *= TGSI_EXEC_MAX_INPUT_ATTRIBS;
          index.i[2] *= TGSI_EXEC_MAX_INPUT_ATTRIBS;
@@ -1315,10 +1331,10 @@ fetch_source(
          assert( 0 );
       }
 
-      index.i[0] += reg->SrcRegisterDim.Index;
-      index.i[1] += reg->SrcRegisterDim.Index;
-      index.i[2] += reg->SrcRegisterDim.Index;
-      index.i[3] += reg->SrcRegisterDim.Index;
+      index.i[0] += reg->Dimension.Index;
+      index.i[1] += reg->Dimension.Index;
+      index.i[2] += reg->Dimension.Index;
+      index.i[3] += reg->Dimension.Index;
 
       /* Again, the second subscript index can be addressed indirectly
        * identically to the first one.
@@ -1327,11 +1343,11 @@ fetch_source(
        *
        *    file[1][ind[4].y+3],
        *    where:
-       *       ind = SrcRegisterDimInd.File
-       *       [4] = SrcRegisterDimInd.Index
-       *       .y = SrcRegisterDimInd.SwizzleX
+       *       ind = DimIndirect.File
+       *       [4] = DimIndirect.Index
+       *       .y = DimIndirect.SwizzleX
        */
-      if (reg->SrcRegisterDim.Indirect) {
+      if (reg->Dimension.Indirect) {
          union tgsi_exec_channel index2;
          union tgsi_exec_channel indir_index;
          const uint execmask = mach->ExecMask;
@@ -1340,12 +1356,12 @@ fetch_source(
          index2.i[0] =
          index2.i[1] =
          index2.i[2] =
-         index2.i[3] = reg->SrcRegisterDimInd.Index;
+         index2.i[3] = reg->DimIndirect.Index;
 
-         swizzle = tgsi_util_get_src_register_swizzle( &reg->SrcRegisterDimInd, CHAN_X );
+         swizzle = tgsi_util_get_src_register_swizzle( &reg->DimIndirect, CHAN_X );
          fetch_src_file_channel(
             mach,
-            reg->SrcRegisterDimInd.File,
+            reg->DimIndirect.File,
             swizzle,
             &index2,
             &indir_index );
@@ -1365,7 +1381,7 @@ fetch_source(
       }
 
       /* If by any chance there was a need for a 3D array of register
-       * files, we would have to check whether SrcRegisterDim is followed
+       * files, we would have to check whether Dimension is followed
        * by a dimension register and continue the saga.
        */
    }
@@ -1373,7 +1389,7 @@ fetch_source(
    swizzle = tgsi_util_get_full_src_register_swizzle( reg, chan_index );
    fetch_src_file_channel(
       mach,
-      reg->SrcRegister.File,
+      reg->Register.File,
       swizzle,
       &index,
       chan );
@@ -1395,10 +1411,6 @@ fetch_source(
    case TGSI_UTIL_SIGN_KEEP:
       break;
    }
-
-   if (reg->SrcRegisterExtMod.Complement) {
-      micro_sub( chan, &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], chan );
-   }
 }
 
 static void
@@ -1426,11 +1438,11 @@ store_dest(
     *
     *    file[ind[2].x+1],
     *    where:
-    *       ind = DstRegisterInd.File
-    *       [2] = DstRegisterInd.Index
-    *       .x = DstRegisterInd.SwizzleX
+    *       ind = Indirect.File
+    *       [2] = Indirect.Index
+    *       .x = Indirect.SwizzleX
     */
-   if (reg->DstRegister.Indirect) {
+   if (reg->Register.Indirect) {
       union tgsi_exec_channel index;
       union tgsi_exec_channel indir_index;
       uint swizzle;
@@ -1439,15 +1451,15 @@ store_dest(
       index.i[0] =
       index.i[1] =
       index.i[2] =
-      index.i[3] = reg->DstRegisterInd.Index;
+      index.i[3] = reg->Indirect.Index;
 
       /* get current value of address register[swizzle] */
-      swizzle = tgsi_util_get_src_register_swizzle( &reg->DstRegisterInd, CHAN_X );
+      swizzle = tgsi_util_get_src_register_swizzle( &reg->Indirect, CHAN_X );
 
       /* fetch values from the address/indirection register */
       fetch_src_file_channel(
          mach,
-         reg->DstRegisterInd.File,
+         reg->Indirect.File,
          swizzle,
          &index,
          &indir_index );
@@ -1456,32 +1468,48 @@ store_dest(
       offset = (int) indir_index.f[0];
    }
 
-   switch (reg->DstRegister.File) {
+   switch (reg->Register.File) {
    case TGSI_FILE_NULL:
       dst = &null;
       break;
 
    case TGSI_FILE_OUTPUT:
       index = mach->Temps[TEMP_OUTPUT_I].xyzw[TEMP_OUTPUT_C].u[0]
-         + reg->DstRegister.Index;
+         + reg->Register.Index;
       dst = &mach->Outputs[offset + index].xyzw[chan_index];
+#if 0
+      if (TGSI_PROCESSOR_GEOMETRY == mach->Processor) {
+         fprintf(stderr, "STORING OUT[%d] mask(%d), = (", offset + index, execmask);
+         for (i = 0; i < QUAD_SIZE; i++)
+            if (execmask & (1 << i))
+               fprintf(stderr, "%f, ", chan->f[i]);
+         fprintf(stderr, ")\n");
+      }
+#endif
       break;
 
    case TGSI_FILE_TEMPORARY:
-      index = reg->DstRegister.Index;
+      index = reg->Register.Index;
       assert( index < TGSI_EXEC_NUM_TEMPS );
       dst = &mach->Temps[offset + index].xyzw[chan_index];
       break;
 
    case TGSI_FILE_ADDRESS:
-      index = reg->DstRegister.Index;
+      index = reg->Register.Index;
       dst = &mach->Addrs[index].xyzw[chan_index];
       break;
 
+   case TGSI_FILE_LOOP:
+      assert(reg->Register.Index == 0);
+      assert(mach->LoopCounterStackTop > 0);
+      assert(chan_index == CHAN_X);
+      dst = &mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[chan_index];
+      break;
+
    case TGSI_FILE_PREDICATE:
-      index = reg->DstRegister.Index;
+      index = reg->Register.Index;
       assert(index < TGSI_EXEC_NUM_PREDS);
-      dst = &mach->Addrs[index].xyzw[chan_index];
+      dst = &mach->Predicates[index].xyzw[chan_index];
       break;
 
    default:
@@ -1489,6 +1517,47 @@ store_dest(
       return;
    }
 
+   if (inst->Instruction.Predicate) {
+      uint swizzle;
+      union tgsi_exec_channel *pred;
+
+      switch (chan_index) {
+      case CHAN_X:
+         swizzle = inst->Predicate.SwizzleX;
+         break;
+      case CHAN_Y:
+         swizzle = inst->Predicate.SwizzleY;
+         break;
+      case CHAN_Z:
+         swizzle = inst->Predicate.SwizzleZ;
+         break;
+      case CHAN_W:
+         swizzle = inst->Predicate.SwizzleW;
+         break;
+      default:
+         assert(0);
+         return;
+      }
+
+      assert(inst->Predicate.Index == 0);
+
+      pred = &mach->Predicates[inst->Predicate.Index].xyzw[swizzle];
+
+      if (inst->Predicate.Negate) {
+         for (i = 0; i < QUAD_SIZE; i++) {
+            if (pred->u[i]) {
+               execmask &= ~(1 << i);
+            }
+         }
+      } else {
+         for (i = 0; i < QUAD_SIZE; i++) {
+            if (!pred->u[i]) {
+               execmask &= ~(1 << i);
+            }
+         }
+      }
+   }
+
    switch (inst->Instruction.Saturate) {
    case TGSI_SAT_NONE:
       for (i = 0; i < QUAD_SIZE; i++)
@@ -1526,10 +1595,10 @@ store_dest(
 }
 
 #define FETCH(VAL,INDEX,CHAN)\
-    fetch_source (mach, VAL, &inst->FullSrcRegisters[INDEX], CHAN)
+    fetch_source (mach, VAL, &inst->Src[INDEX], CHAN)
 
 #define STORE(VAL,INDEX,CHAN)\
-    store_dest (mach, VAL, &inst->FullDstRegisters[INDEX], inst, CHAN )
+    store_dest (mach, VAL, &inst->Dst[INDEX], inst, CHAN )
 
 
 /**
@@ -1555,7 +1624,7 @@ exec_kil(struct tgsi_exec_machine *mach,
 
       /* unswizzle channel */
       swizzle = tgsi_util_get_full_src_register_swizzle (
-                        &inst->FullSrcRegisters[0],
+                        &inst->Src[0],
                         chan_index);
 
       /* check if the component has not been already tested */
@@ -1587,6 +1656,35 @@ exec_kilp(struct tgsi_exec_machine *mach,
    mach->Temps[TEMP_KILMASK_I].xyzw[TEMP_KILMASK_C].u[0] |= kilmask;
 }
 
+static void
+emit_vertex(struct tgsi_exec_machine *mach)
+{
+   /* FIXME: check for exec mask correctly
+   unsigned i;
+   for (i = 0; i < QUAD_SIZE; ++i) {
+         if ((mach->ExecMask & (1 << i)))
+   */
+   if (mach->ExecMask) {
+      mach->Temps[TEMP_OUTPUT_I].xyzw[TEMP_OUTPUT_C].u[0] += mach->NumOutputs;
+      mach->Primitives[mach->Temps[TEMP_PRIMITIVE_I].xyzw[TEMP_PRIMITIVE_C].u[0]]++;
+   }
+}
+
+static void
+emit_primitive(struct tgsi_exec_machine *mach)
+{
+   unsigned *prim_count = &mach->Temps[TEMP_PRIMITIVE_I].xyzw[TEMP_PRIMITIVE_C].u[0];
+   /* FIXME: check for exec mask correctly
+   unsigned i;
+   for (i = 0; i < QUAD_SIZE; ++i) {
+         if ((mach->ExecMask & (1 << i)))
+   */
+   if (mach->ExecMask) {
+      ++(*prim_count);
+      debug_assert((*prim_count * mach->NumOutputs) < mach->MaxGeometryShaderOutputs);
+      mach->Primitives[*prim_count] = 0;
+   }
+}
 
 /*
  * Fetch a four texture samples using STR texture coordinates.
@@ -1622,14 +1720,14 @@ exec_tex(struct tgsi_exec_machine *mach,
          boolean biasLod,
          boolean projected)
 {
-   const uint unit = inst->FullSrcRegisters[1].SrcRegister.Index;
+   const uint unit = inst->Src[1].Register.Index;
    union tgsi_exec_channel r[4];
    uint chan_index;
    float lodBias;
 
    /*   debug_printf("Sampler %u unit %u\n", sampler, unit); */
 
-   switch (inst->InstructionExtTexture.Texture) {
+   switch (inst->Texture.Texture) {
    case TGSI_TEXTURE_1D:
    case TGSI_TEXTURE_SHADOW1D:
 
@@ -1715,6 +1813,64 @@ exec_tex(struct tgsi_exec_machine *mach,
    }
 }
 
+static void
+exec_txd(struct tgsi_exec_machine *mach,
+         const struct tgsi_full_instruction *inst)
+{
+   const uint unit = inst->Src[3].Register.Index;
+   union tgsi_exec_channel r[4];
+   uint chan_index;
+
+   /*
+    * XXX: This is fake TXD -- the derivatives are not taken into account, yet.
+    */
+
+   switch (inst->Texture.Texture) {
+   case TGSI_TEXTURE_1D:
+   case TGSI_TEXTURE_SHADOW1D:
+
+      FETCH(&r[0], 0, CHAN_X);
+
+      fetch_texel(mach->Samplers[unit],
+                  &r[0], &ZeroVec, &ZeroVec, 0.0f,  /* S, T, P, BIAS */
+                  &r[0], &r[1], &r[2], &r[3]);      /* R, G, B, A */
+      break;
+
+   case TGSI_TEXTURE_2D:
+   case TGSI_TEXTURE_RECT:
+   case TGSI_TEXTURE_SHADOW2D:
+   case TGSI_TEXTURE_SHADOWRECT:
+
+      FETCH(&r[0], 0, CHAN_X);
+      FETCH(&r[1], 0, CHAN_Y);
+      FETCH(&r[2], 0, CHAN_Z);
+
+      fetch_texel(mach->Samplers[unit],
+                  &r[0], &r[1], &r[2], 0.0f,    /* inputs */
+                  &r[0], &r[1], &r[2], &r[3]);  /* outputs */
+      break;
+
+   case TGSI_TEXTURE_3D:
+   case TGSI_TEXTURE_CUBE:
+
+      FETCH(&r[0], 0, CHAN_X);
+      FETCH(&r[1], 0, CHAN_Y);
+      FETCH(&r[2], 0, CHAN_Z);
+
+      fetch_texel(mach->Samplers[unit],
+                  &r[0], &r[1], &r[2], 0.0f,
+                  &r[0], &r[1], &r[2], &r[3]);
+      break;
+
+   default:
+      assert(0);
+   }
+
+   FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+      STORE(&r[chan_index], 0, chan_index);
+   }
+}
+
 
 /**
  * Evaluate a constant-valued coefficient at the position of the
@@ -1784,53 +1940,59 @@ typedef void (* eval_coef_func)(
    unsigned chan );
 
 static void
-exec_declaration(
-   struct tgsi_exec_machine *mach,
-   const struct tgsi_full_declaration *decl )
+exec_declaration(struct tgsi_exec_machine *mach,
+                 const struct tgsi_full_declaration *decl)
 {
-   if( mach->Processor == TGSI_PROCESSOR_FRAGMENT ) {
-      if( decl->Declaration.File == TGSI_FILE_INPUT ) {
-         unsigned first, last, mask;
-         eval_coef_func eval;
+   if (mach->Processor == TGSI_PROCESSOR_FRAGMENT) {
+      if (decl->Declaration.File == TGSI_FILE_INPUT ||
+          decl->Declaration.File == TGSI_FILE_SYSTEM_VALUE) {
+         uint first, last, mask;
 
-         first = decl->DeclarationRange.First;
-         last = decl->DeclarationRange.Last;
+         first = decl->Range.First;
+         last = decl->Range.Last;
          mask = decl->Declaration.UsageMask;
 
-         switch( decl->Declaration.Interpolate ) {
-         case TGSI_INTERPOLATE_CONSTANT:
-            eval = eval_constant_coef;
-            break;
+         if (decl->Semantic.Name == TGSI_SEMANTIC_POSITION) {
+            assert(decl->Semantic.Index == 0);
+            assert(first == last);
+            assert(mask = TGSI_WRITEMASK_XYZW);
 
-         case TGSI_INTERPOLATE_LINEAR:
-            eval = eval_linear_coef;
-            break;
+            mach->Inputs[first] = mach->QuadPos;
+         } else if (decl->Semantic.Name == TGSI_SEMANTIC_FACE) {
+            uint i;
 
-         case TGSI_INTERPOLATE_PERSPECTIVE:
-            eval = eval_perspective_coef;
-            break;
+            assert(decl->Semantic.Index == 0);
+            assert(first == last);
 
-         default:
-            assert( 0 );
-            return;
-         }
-
-         if( mask == TGSI_WRITEMASK_XYZW ) {
-            unsigned i, j;
-
-            for( i = first; i <= last; i++ ) {
-               for( j = 0; j < NUM_CHANNELS; j++ ) {
-                  eval( mach, i, j );
-               }
+            for (i = 0; i < QUAD_SIZE; i++) {
+               mach->Inputs[first].xyzw[0].f[i] = mach->Face;
+            }
+         } else {
+            eval_coef_func eval;
+            uint i, j;
+
+            switch (decl->Declaration.Interpolate) {
+            case TGSI_INTERPOLATE_CONSTANT:
+               eval = eval_constant_coef;
+               break;
+
+            case TGSI_INTERPOLATE_LINEAR:
+               eval = eval_linear_coef;
+               break;
+
+            case TGSI_INTERPOLATE_PERSPECTIVE:
+               eval = eval_perspective_coef;
+               break;
+
+            default:
+               assert(0);
+               return;
             }
-         }
-         else {
-            unsigned i, j;
 
-            for( j = 0; j < NUM_CHANNELS; j++ ) {
-               if( mask & (1 << j) ) {
-                  for( i = first; i <= last; i++ ) {
-                     eval( mach, i, j );
+            for (j = 0; j < NUM_CHANNELS; j++) {
+               if (mask & (1 << j)) {
+                  for (i = first; i <= last; i++) {
+                     eval(mach, i, j);
                   }
                }
             }
@@ -1847,6 +2009,7 @@ exec_instruction(
 {
    uint chan_index;
    union tgsi_exec_channel r[10];
+   union tgsi_exec_channel d[8];
 
    (*pc)++;
 
@@ -1855,42 +2018,27 @@ exec_instruction(
    case TGSI_OPCODE_FLR:
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
-         micro_flr( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+         micro_flr(&d[chan_index], &r[0]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
    case TGSI_OPCODE_MOV:
-      if (inst->Flags & SOA_DEPENDENCY_FLAG) {
-         /* Do all fetches into temp regs, then do all stores to avoid
-          * intermediate/accidental clobbering.  This could be done all the
-          * time for MOV but for other instructions we'll need more temps...
-          */
-         FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-            FETCH( &r[chan_index], 0, chan_index );
-         }
-         FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-            STORE( &r[chan_index], 0, chan_index );
-         }
+      FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
+         FETCH(&d[chan_index], 0, chan_index);
       }
-      else {
-         FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
-            FETCH( &r[0], 0, chan_index );
-            STORE( &r[0], 0, chan_index );
-         }
+      FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
    case TGSI_OPCODE_LIT:
-      if (IS_CHANNEL_ENABLED( *inst, CHAN_X )) {
-         STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_X );
-      }
-
       if (IS_CHANNEL_ENABLED( *inst, CHAN_Y ) || IS_CHANNEL_ENABLED( *inst, CHAN_Z )) {
          FETCH( &r[0], 0, CHAN_X );
          if (IS_CHANNEL_ENABLED( *inst, CHAN_Y )) {
-            micro_max( &r[0], &r[0], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
-            STORE( &r[0], 0, CHAN_Y );
+            micro_max(&d[CHAN_Y], &r[0], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C]);
          }
 
          if (IS_CHANNEL_ENABLED( *inst, CHAN_Z )) {
@@ -1901,11 +2049,19 @@ exec_instruction(
             micro_min( &r[2], &r[2], &mach->Temps[TEMP_128_I].xyzw[TEMP_128_C] );
             micro_max( &r[2], &r[2], &mach->Temps[TEMP_M128_I].xyzw[TEMP_M128_C] );
             micro_pow( &r[1], &r[1], &r[2] );
-            micro_lt( &r[0], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &r[0], &r[1], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
-            STORE( &r[0], 0, CHAN_Z );
+            micro_lt(&d[CHAN_Z], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &r[0], &r[1], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C]);
          }
-      }
 
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_Y)) {
+            STORE(&d[CHAN_Y], 0, CHAN_Y);
+         }
+         if (IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
+            STORE(&d[CHAN_Z], 0, CHAN_Z);
+         }
+      }
+      if (IS_CHANNEL_ENABLED( *inst, CHAN_X )) {
+         STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_X );
+      }
       if (IS_CHANNEL_ENABLED( *inst, CHAN_W )) {
          STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_W );
       }
@@ -1973,14 +2129,13 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_MUL:
-      FOR_EACH_ENABLED_CHANNEL( *inst, chan_index )
-      {
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
          FETCH(&r[0], 0, chan_index);
          FETCH(&r[1], 1, chan_index);
-
-         micro_mul( &r[0], &r[0], &r[1] );
-
-         STORE(&r[0], 0, chan_index);
+         micro_mul(&d[chan_index], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -1988,8 +2143,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_add( &r[0], &r[0], &r[1] );
-         STORE( &r[0], 0, chan_index );
+         micro_add(&d[chan_index], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2045,25 +2202,29 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_DST:
-      if (IS_CHANNEL_ENABLED( *inst, CHAN_X )) {
-         STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_X );
-      }
-
       if (IS_CHANNEL_ENABLED( *inst, CHAN_Y )) {
          FETCH( &r[0], 0, CHAN_Y );
          FETCH( &r[1], 1, CHAN_Y);
-         micro_mul( &r[0], &r[0], &r[1] );
-         STORE( &r[0], 0, CHAN_Y );
+         micro_mul(&d[CHAN_Y], &r[0], &r[1]);
       }
-
       if (IS_CHANNEL_ENABLED( *inst, CHAN_Z )) {
-         FETCH( &r[0], 0, CHAN_Z );
-         STORE( &r[0], 0, CHAN_Z );
+         FETCH(&d[CHAN_Z], 0, CHAN_Z);
       }
-
       if (IS_CHANNEL_ENABLED( *inst, CHAN_W )) {
-         FETCH( &r[0], 1, CHAN_W );
-         STORE( &r[0], 0, CHAN_W );
+         FETCH(&d[CHAN_W], 1, CHAN_W);
+      }
+
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_X)) {
+         STORE(&mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_X);
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_Y)) {
+         STORE(&d[CHAN_Y], 0, CHAN_Y);
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
+         STORE(&d[CHAN_Z], 0, CHAN_Z);
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_W)) {
+         STORE(&d[CHAN_W], 0, CHAN_W);
       }
       break;
 
@@ -2073,9 +2234,10 @@ exec_instruction(
          FETCH(&r[1], 1, chan_index);
 
          /* XXX use micro_min()?? */
-         micro_lt( &r[0], &r[0], &r[1], &r[0], &r[1] );
-
-         STORE(&r[0], 0, chan_index);
+         micro_lt(&d[chan_index], &r[0], &r[1], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2085,9 +2247,10 @@ exec_instruction(
          FETCH(&r[1], 1, chan_index);
 
          /* XXX use micro_max()?? */
-         micro_lt( &r[0], &r[0], &r[1], &r[1], &r[0] );
-
-         STORE(&r[0], 0, chan_index );
+         micro_lt(&d[chan_index], &r[0], &r[1], &r[1], &r[0] );
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2096,8 +2259,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_lt( &r[0], &r[0], &r[1], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
-         STORE( &r[0], 0, chan_index );
+         micro_lt(&d[chan_index], &r[0], &r[1], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2106,8 +2271,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_le( &r[0], &r[1], &r[0], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
-         STORE( &r[0], 0, chan_index );
+         micro_le(&d[chan_index], &r[1], &r[0], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2118,8 +2285,10 @@ exec_instruction(
          FETCH( &r[1], 1, chan_index );
          micro_mul( &r[0], &r[0], &r[1] );
          FETCH( &r[1], 2, chan_index );
-         micro_add( &r[0], &r[0], &r[1] );
-         STORE( &r[0], 0, chan_index );
+         micro_add(&d[chan_index], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2127,10 +2296,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH(&r[0], 0, chan_index);
          FETCH(&r[1], 1, chan_index);
-
-         micro_sub( &r[0], &r[0], &r[1] );
-
-         STORE(&r[0], 0, chan_index);
+         micro_sub(&d[chan_index], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2139,12 +2308,12 @@ exec_instruction(
          FETCH(&r[0], 0, chan_index);
          FETCH(&r[1], 1, chan_index);
          FETCH(&r[2], 2, chan_index);
-
          micro_sub( &r[1], &r[1], &r[2] );
          micro_mul( &r[0], &r[0], &r[1] );
-         micro_add( &r[0], &r[0], &r[2] );
-
-         STORE(&r[0], 0, chan_index);
+         micro_add(&d[chan_index], &r[0], &r[2]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2153,8 +2322,10 @@ exec_instruction(
          FETCH(&r[0], 0, chan_index);
          FETCH(&r[1], 1, chan_index);
          FETCH(&r[2], 2, chan_index);
-         micro_lt(&r[0], &mach->Temps[TEMP_HALF_I].xyzw[TEMP_HALF_C], &r[2], &r[0], &r[1]);
-         STORE(&r[0], 0, chan_index);
+         micro_lt(&d[chan_index], &mach->Temps[TEMP_HALF_I].xyzw[TEMP_HALF_C], &r[2], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2179,8 +2350,10 @@ exec_instruction(
    case TGSI_OPCODE_FRC:
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
-         micro_frc( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+         micro_frc(&d[chan_index], &r[0]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2190,8 +2363,10 @@ exec_instruction(
          FETCH(&r[1], 1, chan_index);
          micro_max(&r[0], &r[0], &r[1]);
          FETCH(&r[1], 2, chan_index);
-         micro_min(&r[0], &r[0], &r[1]);
-         STORE(&r[0], 0, chan_index);
+         micro_min(&d[chan_index], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2199,19 +2374,17 @@ exec_instruction(
    case TGSI_OPCODE_ARR:
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
-         micro_rnd( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+         micro_rnd(&d[chan_index], &r[0]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
    case TGSI_OPCODE_EX2:
       FETCH(&r[0], 0, CHAN_X);
 
-#if FAST_MATH
       micro_exp2( &r[0], &r[0] );
-#else
-      micro_pow( &r[0], &mach->Temps[TEMP_2_I].xyzw[TEMP_2_C], &r[0] );
-#endif
 
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          STORE( &r[0], 0, chan_index );
@@ -2247,11 +2420,7 @@ exec_instruction(
       FETCH(&r[4], 1, CHAN_Y);
 
       micro_mul( &r[5], &r[3], &r[4] );
-      micro_sub( &r[2], &r[2], &r[5] );
-
-      if (IS_CHANNEL_ENABLED( *inst, CHAN_X )) {
-         STORE( &r[2], 0, CHAN_X );
-      }
+      micro_sub(&d[CHAN_X], &r[2], &r[5]);
 
       FETCH(&r[2], 1, CHAN_X);
 
@@ -2260,20 +2429,21 @@ exec_instruction(
       FETCH(&r[5], 0, CHAN_X);
 
       micro_mul( &r[1], &r[1], &r[5] );
-      micro_sub( &r[3], &r[3], &r[1] );
-
-      if (IS_CHANNEL_ENABLED( *inst, CHAN_Y )) {
-         STORE( &r[3], 0, CHAN_Y );
-      }
+      micro_sub(&d[CHAN_Y], &r[3], &r[1]);
 
       micro_mul( &r[5], &r[5], &r[4] );
       micro_mul( &r[0], &r[0], &r[2] );
-      micro_sub( &r[5], &r[5], &r[0] );
+      micro_sub(&d[CHAN_Z], &r[5], &r[0]);
 
-      if (IS_CHANNEL_ENABLED( *inst, CHAN_Z )) {
-         STORE( &r[5], 0, CHAN_Z );
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_X)) {
+         STORE(&d[CHAN_X], 0, CHAN_X);
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_Y)) {
+         STORE(&d[CHAN_Y], 0, CHAN_Y);
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
+         STORE(&d[CHAN_Z], 0, CHAN_Z);
       }
-
       if (IS_CHANNEL_ENABLED( *inst, CHAN_W )) {
          STORE( &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], 0, CHAN_W );
       }
@@ -2282,11 +2452,11 @@ exec_instruction(
     case TGSI_OPCODE_ABS:
        FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
           FETCH(&r[0], 0, chan_index);
-
-          micro_abs( &r[0], &r[0] );
-
-          STORE(&r[0], 0, chan_index);
+          micro_abs(&d[chan_index], &r[0]);
        }
+       FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
+      }
        break;
 
    case TGSI_OPCODE_RCC:
@@ -2338,16 +2508,20 @@ exec_instruction(
    case TGSI_OPCODE_DDX:
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
-         micro_ddx( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+         micro_ddx(&d[chan_index], &r[0]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
    case TGSI_OPCODE_DDY:
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
-         micro_ddy( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+         micro_ddy(&d[chan_index], &r[0]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2428,10 +2602,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_eq( &r[0], &r[0], &r[1],
-                   &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C],
-                   &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
-         STORE( &r[0], 0, chan_index );
+         micro_eq(&d[chan_index], &r[0], &r[1], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2445,8 +2619,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_le( &r[0], &r[0], &r[1], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C] );
-         STORE( &r[0], 0, chan_index );
+         micro_le(&d[chan_index], &r[0], &r[1], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2462,8 +2638,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_le( &r[0], &r[0], &r[1], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C] );
-         STORE( &r[0], 0, chan_index );
+         micro_le(&d[chan_index], &r[0], &r[1], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2471,8 +2649,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_eq( &r[0], &r[0], &r[1], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C] );
-         STORE( &r[0], 0, chan_index );
+         micro_eq(&d[chan_index], &r[0], &r[1], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2502,7 +2682,7 @@ exec_instruction(
       /* src[1] = d[strq]/dx */
       /* src[2] = d[strq]/dy */
       /* src[3] = sampler unit */
-      assert (0);
+      exec_txd(mach, inst);
       break;
 
    case TGSI_OPCODE_TXL:
@@ -2546,13 +2726,8 @@ exec_instruction(
          micro_mul(&r[3], &r[3], &r[1]);
          micro_add(&r[2], &r[2], &r[3]);
          FETCH(&r[3], 0, CHAN_X);
-         micro_add(&r[2], &r[2], &r[3]);
-         if (IS_CHANNEL_ENABLED(*inst, CHAN_X)) {
-            STORE(&r[2], 0, CHAN_X);
-         }
-         if (IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
-            STORE(&r[2], 0, CHAN_Z);
-         }
+         micro_add(&d[CHAN_X], &r[2], &r[3]);
+         
       }
       if (IS_CHANNEL_ENABLED(*inst, CHAN_Y) ||
           IS_CHANNEL_ENABLED(*inst, CHAN_W)) {
@@ -2562,13 +2737,20 @@ exec_instruction(
          micro_mul(&r[3], &r[3], &r[1]);
          micro_add(&r[2], &r[2], &r[3]);
          FETCH(&r[3], 0, CHAN_Y);
-         micro_add(&r[2], &r[2], &r[3]);
-         if (IS_CHANNEL_ENABLED(*inst, CHAN_Y)) {
-            STORE(&r[2], 0, CHAN_Y);
-         }
-         if (IS_CHANNEL_ENABLED(*inst, CHAN_W)) {
-            STORE(&r[2], 0, CHAN_W);
-         }
+         micro_add(&d[CHAN_Y], &r[2], &r[3]);
+         
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_X)) {
+         STORE(&d[CHAN_X], 0, CHAN_X);
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_Y)) {
+         STORE(&d[CHAN_Y], 0, CHAN_Y);
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_Z)) {
+         STORE(&d[CHAN_X], 0, CHAN_Z);
+      }
+      if (IS_CHANNEL_ENABLED(*inst, CHAN_W)) {
+         STORE(&d[CHAN_Y], 0, CHAN_W);
       }
       break;
 
@@ -2611,7 +2793,7 @@ exec_instruction(
          mach->FuncStack[mach->FuncStackTop++] = mach->FuncMask;
 
          /* Finally, jump to the subroutine */
-         *pc = inst->InstructionExtLabel.Label;
+         *pc = inst->Label.Label;
       }
       break;
 
@@ -2653,8 +2835,10 @@ exec_instruction(
    /* TGSI_OPCODE_SGN */
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
-         micro_sgn( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+         micro_sgn(&d[chan_index], &r[0]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2663,10 +2847,10 @@ exec_instruction(
          FETCH(&r[0], 0, chan_index);
          FETCH(&r[1], 1, chan_index);
          FETCH(&r[2], 2, chan_index);
-
-         micro_lt( &r[0], &r[0], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &r[1], &r[2] );
-
-         STORE(&r[0], 0, chan_index);
+         micro_lt(&d[chan_index], &r[0], &mach->Temps[TEMP_0_I].xyzw[TEMP_0_C], &r[1], &r[2]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2841,32 +3025,40 @@ exec_instruction(
    case TGSI_OPCODE_CEIL:
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
-         micro_ceil( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+         micro_ceil(&d[chan_index], &r[0]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
    case TGSI_OPCODE_I2F:
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
-         micro_i2f( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+         micro_i2f(&d[chan_index], &r[0]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
    case TGSI_OPCODE_NOT:
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
-         micro_not( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+         micro_not(&d[chan_index], &r[0]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
    case TGSI_OPCODE_TRUNC:
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
-         micro_trunc( &r[0], &r[0] );
-         STORE( &r[0], 0, chan_index );
+         micro_trunc(&d[chan_index], &r[0]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2874,8 +3066,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_shl( &r[0], &r[0], &r[1] );
-         STORE( &r[0], 0, chan_index );
+         micro_shl(&d[chan_index], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2883,8 +3077,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_ishr( &r[0], &r[0], &r[1] );
-         STORE( &r[0], 0, chan_index );
+         micro_ishr(&d[chan_index], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2892,8 +3088,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_and( &r[0], &r[0], &r[1] );
-         STORE( &r[0], 0, chan_index );
+         micro_and(&d[chan_index], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2901,8 +3099,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_or( &r[0], &r[0], &r[1] );
-         STORE( &r[0], 0, chan_index );
+         micro_or(&d[chan_index], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2914,8 +3114,10 @@ exec_instruction(
       FOR_EACH_ENABLED_CHANNEL( *inst, chan_index ) {
          FETCH( &r[0], 0, chan_index );
          FETCH( &r[1], 1, chan_index );
-         micro_xor( &r[0], &r[0], &r[1] );
-         STORE( &r[0], 0, chan_index );
+         micro_xor(&d[chan_index], &r[0], &r[1]);
+      }
+      FOR_EACH_ENABLED_CHANNEL(*inst, chan_index) {
+         STORE(&d[chan_index], 0, chan_index);
       }
       break;
 
@@ -2932,13 +3134,11 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_EMIT:
-      mach->Temps[TEMP_OUTPUT_I].xyzw[TEMP_OUTPUT_C].u[0] += 16;
-      mach->Primitives[mach->Temps[TEMP_PRIMITIVE_I].xyzw[TEMP_PRIMITIVE_C].u[0]]++;
+      emit_vertex(mach);
       break;
 
    case TGSI_OPCODE_ENDPRIM:
-      mach->Temps[TEMP_PRIMITIVE_I].xyzw[TEMP_PRIMITIVE_C].u[0]++;
-      mach->Primitives[mach->Temps[TEMP_PRIMITIVE_I].xyzw[TEMP_PRIMITIVE_C].u[0]] = 0;
+      emit_primitive(mach);
       break;
 
    case TGSI_OPCODE_BGNFOR:
@@ -2946,8 +3146,23 @@ exec_instruction(
       for (chan_index = 0; chan_index < 3; chan_index++) {
          FETCH( &mach->LoopCounterStack[mach->LoopCounterStackTop].xyzw[chan_index], 0, chan_index );
       }
-      STORE( &mach->LoopCounterStack[mach->LoopCounterStackTop].xyzw[CHAN_Y], 0, CHAN_X );
       ++mach->LoopCounterStackTop;
+      STORE(&mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_X], 0, CHAN_X);
+      /* update LoopMask */
+      if (mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y].f[0] <= 0.0f) {
+         mach->LoopMask &= ~0x1;
+      }
+      if (mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y].f[1] <= 0.0f) {
+         mach->LoopMask &= ~0x2;
+      }
+      if (mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y].f[2] <= 0.0f) {
+         mach->LoopMask &= ~0x4;
+      }
+      if (mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y].f[3] <= 0.0f) {
+         mach->LoopMask &= ~0x8;
+      }
+      /* TODO: if mach->LoopMask == 0, jump to end of loop */
+      UPDATE_EXEC_MASK(mach);
       /* fall-through (for now) */
    case TGSI_OPCODE_BGNLOOP:
       /* push LoopMask and ContMasks */
@@ -2961,28 +3176,28 @@ exec_instruction(
 
    case TGSI_OPCODE_ENDFOR:
       assert(mach->LoopCounterStackTop > 0);
-      micro_sub( &mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_X], 
-                 &mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_X],
-                 &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C] );
+      micro_sub(&mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y], 
+                &mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y],
+                &mach->Temps[TEMP_1_I].xyzw[TEMP_1_C]);
       /* update LoopMask */
-      if( mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_X].f[0] <= 0) {
+      if (mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y].f[0] <= 0.0f) {
          mach->LoopMask &= ~0x1;
       }
-      if( mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_X].f[1] <= 0 ) {
+      if (mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y].f[1] <= 0.0f) {
          mach->LoopMask &= ~0x2;
       }
-      if( mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_X].f[2] <= 0 ) {
+      if (mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y].f[2] <= 0.0f) {
          mach->LoopMask &= ~0x4;
       }
-      if( mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_X].f[3] <= 0 ) {
+      if (mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y].f[3] <= 0.0f) {
          mach->LoopMask &= ~0x8;
       }
-      micro_add( &mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y], 
-                 &mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Y], 
-                 &mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Z]);
+      micro_add(&mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_X], 
+                &mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_X], 
+                &mach->LoopCounterStack[mach->LoopCounterStackTop - 1].xyzw[CHAN_Z]);
       assert(mach->LoopLabelStackTop > 0);
       inst = mach->Instructions + mach->LoopLabelStack[mach->LoopLabelStackTop - 1];
-      STORE( &mach->LoopCounterStack[mach->LoopCounterStackTop].xyzw[CHAN_Y], 0, CHAN_X );
+      STORE(&mach->LoopCounterStack[mach->LoopCounterStackTop].xyzw[CHAN_X], 0, CHAN_X);
       /* Restore ContMask, but don't pop */
       assert(mach->ContStackTop > 0);
       mach->ContMask = mach->ContStack[mach->ContStackTop - 1];
@@ -3049,17 +3264,59 @@ exec_instruction(
       break;
 
    case TGSI_OPCODE_ENDSUB:
-      /* no-op */
+      /*
+       * XXX: This really should be a no-op. We should never reach this opcode.
+       */
+
+      assert(mach->CallStackTop > 0);
+      mach->CallStackTop--;
+
+      mach->CondStackTop = mach->CallStack[mach->CallStackTop].CondStackTop;
+      mach->CondMask = mach->CondStack[mach->CondStackTop];
+
+      mach->LoopStackTop = mach->CallStack[mach->CallStackTop].LoopStackTop;
+      mach->LoopMask = mach->LoopStack[mach->LoopStackTop];
+
+      mach->ContStackTop = mach->CallStack[mach->CallStackTop].ContStackTop;
+      mach->ContMask = mach->ContStack[mach->ContStackTop];
+
+      assert(mach->FuncStackTop > 0);
+      mach->FuncMask = mach->FuncStack[--mach->FuncStackTop];
+
+      *pc = mach->CallStack[mach->CallStackTop].ReturnAddr;
+
+      UPDATE_EXEC_MASK(mach);
       break;
 
    case TGSI_OPCODE_NOP:
       break;
 
+   case TGSI_OPCODE_BREAKC:
+      FETCH(&r[0], 0, CHAN_X);
+      /* update CondMask */
+      if (r[0].u[0] && (mach->ExecMask & 0x1)) {
+         mach->LoopMask &= ~0x1;
+      }
+      if (r[0].u[1] && (mach->ExecMask & 0x2)) {
+         mach->LoopMask &= ~0x2;
+      }
+      if (r[0].u[2] && (mach->ExecMask & 0x4)) {
+         mach->LoopMask &= ~0x4;
+      }
+      if (r[0].u[3] && (mach->ExecMask & 0x8)) {
+         mach->LoopMask &= ~0x8;
+      }
+      /* Todo: if mach->LoopMask == 0, jump to end of loop */
+      UPDATE_EXEC_MASK(mach);
+      break;
+
    default:
       assert( 0 );
    }
 }
 
+#define DEBUG_EXECUTION 0
+
 
 /**
  * Run TGSI interpreter.
@@ -3103,10 +3360,67 @@ tgsi_exec_machine_run( struct tgsi_exec_machine *mach )
       exec_declaration( mach, mach->Declarations+i );
    }
 
-   /* execute instructions, until pc is set to -1 */
-   while (pc != -1) {
-      assert(pc < (int) mach->NumInstructions);
-      exec_instruction( mach, mach->Instructions + pc, &pc );
+   {
+#if DEBUG_EXECUTION
+      struct tgsi_exec_vector temps[TGSI_EXEC_NUM_TEMPS + TGSI_EXEC_NUM_TEMP_EXTRAS];
+      struct tgsi_exec_vector outputs[PIPE_MAX_ATTRIBS];
+      uint inst = 1;
+
+      memcpy(temps, mach->Temps, sizeof(temps));
+      memcpy(outputs, mach->Outputs, sizeof(outputs));
+#endif
+
+      /* execute instructions, until pc is set to -1 */
+      while (pc != -1) {
+
+#if DEBUG_EXECUTION
+         uint i;
+
+         tgsi_dump_instruction(&mach->Instructions[pc], inst++);
+#endif
+
+         assert(pc < (int) mach->NumInstructions);
+         exec_instruction(mach, mach->Instructions + pc, &pc);
+
+#if DEBUG_EXECUTION
+         for (i = 0; i < TGSI_EXEC_NUM_TEMPS + TGSI_EXEC_NUM_TEMP_EXTRAS; i++) {
+            if (memcmp(&temps[i], &mach->Temps[i], sizeof(temps[i]))) {
+               uint j;
+
+               memcpy(&temps[i], &mach->Temps[i], sizeof(temps[i]));
+               debug_printf("TEMP[%2u] = ", i);
+               for (j = 0; j < 4; j++) {
+                  if (j > 0) {
+                     debug_printf("           ");
+                  }
+                  debug_printf("(%6f, %6f, %6f, %6f)\n",
+                               temps[i].xyzw[0].f[j],
+                               temps[i].xyzw[1].f[j],
+                               temps[i].xyzw[2].f[j],
+                               temps[i].xyzw[3].f[j]);
+               }
+            }
+         }
+         for (i = 0; i < PIPE_MAX_ATTRIBS; i++) {
+            if (memcmp(&outputs[i], &mach->Outputs[i], sizeof(outputs[i]))) {
+               uint j;
+
+               memcpy(&outputs[i], &mach->Outputs[i], sizeof(outputs[i]));
+               debug_printf("OUT[%2u] =  ", i);
+               for (j = 0; j < 4; j++) {
+                  if (j > 0) {
+                     debug_printf("           ");
+                  }
+                  debug_printf("{%6f, %6f, %6f, %6f}\n",
+                               outputs[i].xyzw[0].f[j],
+                               outputs[i].xyzw[1].f[j],
+                               outputs[i].xyzw[2].f[j],
+                               outputs[i].xyzw[3].f[j]);
+               }
+            }
+         }
+#endif
+      }
    }
 
 #if 0
@@ -3120,5 +3434,10 @@ tgsi_exec_machine_run( struct tgsi_exec_machine *mach )
    }
 #endif
 
+   assert(mach->CondStackTop == 0);
+   assert(mach->LoopStackTop == 0);
+   assert(mach->ContStackTop == 0);
+   assert(mach->CallStackTop == 0);
+
    return ~mach->Temps[TEMP_KILMASK_I].xyzw[TEMP_KILMASK_C].u[0];
 }
index 471f591dd6e3e4cbe51845a6b46bbeae3af216c5..afaf5c39c48c0f9940a5308c4ce4d69f903ac8b7 100644 (file)
@@ -191,6 +191,14 @@ struct tgsi_exec_labels
  */
 #define TGSI_EXEC_MAX_CONST_BUFFER  4096
 
+/* The maximum number of vertices per primitive */
+#define TGSI_MAX_PRIM_VERTICES 6
+
+/* The maximum number of primitives to be generated */
+#define TGSI_MAX_PRIMITIVES 64
+
+/* The maximum total number of vertices */
+#define TGSI_MAX_TOTAL_VERTICES (TGSI_MAX_PRIM_VERTICES * TGSI_MAX_PRIMITIVES * PIPE_MAX_ATTRIBS)
 
 /** function call/activation record */
 struct tgsi_call_record
@@ -201,7 +209,6 @@ struct tgsi_call_record
    uint ReturnAddr;
 };
 
-
 /**
  * Run-time virtual machine state for executing TGSI shader.
  */
@@ -214,10 +221,11 @@ struct tgsi_exec_machine
 
    float                         Imms[TGSI_EXEC_NUM_IMMEDIATES][4];
 
-   struct tgsi_exec_vector       Inputs[PIPE_MAX_ATTRIBS];
-   struct tgsi_exec_vector       Outputs[PIPE_MAX_ATTRIBS];
+   struct tgsi_exec_vector       Inputs[TGSI_MAX_PRIM_VERTICES * PIPE_MAX_ATTRIBS];
+   struct tgsi_exec_vector       Outputs[TGSI_MAX_TOTAL_VERTICES];
 
    struct tgsi_exec_vector       *Addrs;
+   struct tgsi_exec_vector       *Predicates;
 
    struct tgsi_sampler           **Samplers;
 
@@ -228,10 +236,13 @@ struct tgsi_exec_machine
 
    /* GEOMETRY processor only. */
    unsigned                      *Primitives;
+   unsigned                       NumOutputs;
+   unsigned                       MaxGeometryShaderOutputs;
 
    /* FRAGMENT processor only. */
    const struct tgsi_interp_coef *InterpCoefs;
    struct tgsi_exec_vector       QuadPos;
+   float                         Face;    /**< +1 if front facing, -1 if back facing */
 
    /* Conditional execution masks */
    uint CondMask;  /**< For IF/ELSE/ENDIF */
@@ -252,7 +263,7 @@ struct tgsi_exec_machine
    uint LoopLabelStack[TGSI_EXEC_MAX_LOOP_NESTING];
    int LoopLabelStackTop;
 
-   /** Loop counter stack (x = count, y = current, z = step) */
+   /** Loop counter stack (x = index, y = counter, z = step) */
    struct tgsi_exec_vector LoopCounterStack[TGSI_EXEC_MAX_LOOP_NESTING];
    int LoopCounterStackTop;
    
index d88c2558d81dcfd6541a7d255a8b0515719129dd..0ba5fe48419bdcd0df467824ac4c996506e11535 100644 (file)
@@ -39,7 +39,6 @@ tgsi_iterate_shader(
       return FALSE;
 
    ctx->processor = parse.FullHeader.Processor;
-   ctx->version = parse.FullVersion.Version;
 
    if (ctx->prolog)
       if (!ctx->prolog( ctx ))
@@ -67,6 +66,12 @@ tgsi_iterate_shader(
                goto fail;
          break;
 
+      case TGSI_TOKEN_TYPE_PROPERTY:
+         if (ctx->iterate_property)
+            if (!ctx->iterate_property( ctx,  &parse.FullToken.FullProperty ))
+               goto fail;
+         break;
+
       default:
          assert( 0 );
       }
index ec7b85bf63d667f8833ef2be559c748705223298..8d67f22c4295d16a660f55e692e0cb856a9147e2 100644 (file)
@@ -56,12 +56,16 @@ struct tgsi_iterate_context
       struct tgsi_iterate_context *ctx,
       struct tgsi_full_immediate *imm );
 
+   boolean
+   (* iterate_property)(
+      struct tgsi_iterate_context *ctx,
+      struct tgsi_full_property *prop );
+
    boolean
    (* epilog)(
       struct tgsi_iterate_context *ctx );
 
    struct tgsi_processor processor;
-   struct tgsi_version version;
 };
 
 boolean
index 83f9df1183e5aaab06bc53154705d58d64586dda..fa65ecb99754722924fbaa6a3e22253f086ae8f8 100644 (file)
 #include "util/u_debug.h"
 #include "pipe/p_shader_tokens.h"
 #include "tgsi_parse.h"
-#include "tgsi_build.h"
 #include "util/u_memory.h"
 
-void
-tgsi_full_token_init(
-   union tgsi_full_token *full_token )
-{
-   full_token->Token.Type = TGSI_TOKEN_TYPE_DECLARATION;
-}
-
-void
-tgsi_full_token_free(
-   union tgsi_full_token *full_token )
-{
-}
-
 unsigned
 tgsi_parse_init(
    struct tgsi_parse_context *ctx,
    const struct tgsi_token *tokens )
 {
-   ctx->FullVersion.Version = *(struct tgsi_version *) &tokens[0];
-   if( ctx->FullVersion.Version.MajorVersion > 1 ) {
-      return TGSI_PARSE_ERROR;
-   }
-
-   ctx->FullHeader.Header = *(struct tgsi_header *) &tokens[1];
+   ctx->FullHeader.Header = *(struct tgsi_header *) &tokens[0];
    if( ctx->FullHeader.Header.HeaderSize >= 2 ) {
-      ctx->FullHeader.Processor = *(struct tgsi_processor *) &tokens[2];
+      ctx->FullHeader.Processor = *(struct tgsi_processor *) &tokens[1];
    }
    else {
-      ctx->FullHeader.Processor = tgsi_default_processor();
+      return TGSI_PARSE_ERROR;
    }
 
    ctx->Tokens = tokens;
-   ctx->Position = 1 + ctx->FullHeader.Header.HeaderSize;
-
-   tgsi_full_token_init( &ctx->FullToken );
+   ctx->Position = ctx->FullHeader.Header.HeaderSize;
 
    return TGSI_PARSE_OK;
 }
@@ -74,7 +53,6 @@ void
 tgsi_parse_free(
    struct tgsi_parse_context *ctx )
 {
-   tgsi_full_token_free( &ctx->FullToken );
 }
 
 boolean
@@ -82,7 +60,7 @@ tgsi_parse_end_of_tokens(
    struct tgsi_parse_context *ctx )
 {
    return ctx->Position >=
-      1 + ctx->FullHeader.Header.HeaderSize + ctx->FullHeader.Header.BodySize;
+      ctx->FullHeader.Header.HeaderSize + ctx->FullHeader.Header.BodySize;
 }
 
 
@@ -119,9 +97,6 @@ tgsi_parse_token(
    struct tgsi_token token;
    unsigned i;
 
-   tgsi_full_token_free( &ctx->FullToken );
-   tgsi_full_token_init( &ctx->FullToken );
-
    next_token( ctx, &token );
 
    switch( token.Type ) {
@@ -129,10 +104,10 @@ tgsi_parse_token(
    {
       struct tgsi_full_declaration *decl = &ctx->FullToken.FullDeclaration;
 
-      *decl = tgsi_default_full_declaration();
+      memset(decl, 0, sizeof *decl);
       copy_token(&decl->Declaration, &token);
 
-      next_token( ctx, &decl->DeclarationRange );
+      next_token( ctx, &decl->Range );
 
       if( decl->Declaration.Semantic ) {
          next_token( ctx, &decl->Semantic );
@@ -145,9 +120,8 @@ tgsi_parse_token(
    {
       struct tgsi_full_immediate *imm = &ctx->FullToken.FullImmediate;
 
-      *imm = tgsi_default_full_immediate();
+      memset(imm, 0, sizeof *imm);
       copy_token(&imm->Immediate, &token);
-      assert( !imm->Immediate.Extended );
 
       switch (imm->Immediate.DataType) {
       case TGSI_IMM_FLOAT32:
@@ -169,137 +143,76 @@ tgsi_parse_token(
    case TGSI_TOKEN_TYPE_INSTRUCTION:
    {
       struct tgsi_full_instruction *inst = &ctx->FullToken.FullInstruction;
-      unsigned extended;
 
-      *inst = tgsi_default_full_instruction();
+      memset(inst, 0, sizeof *inst);
       copy_token(&inst->Instruction, &token);
-      extended = inst->Instruction.Extended;
-
-      while( extended ) {
-         struct tgsi_src_register_ext token;
-
-         next_token( ctx, &token );
-
-         switch( token.Type ) {
-         case TGSI_INSTRUCTION_EXT_TYPE_LABEL:
-            copy_token(&inst->InstructionExtLabel, &token);
-            break;
-
-         case TGSI_INSTRUCTION_EXT_TYPE_TEXTURE:
-            copy_token(&inst->InstructionExtTexture, &token);
-            break;
 
-         case TGSI_INSTRUCTION_EXT_TYPE_PREDICATE:
-            copy_token(&inst->InstructionExtPredicate, &token);
-            break;
+      if (inst->Instruction.Predicate) {
+         next_token(ctx, &inst->Predicate);
+      }
 
-         default:
-            assert( 0 );
-         }
+      if (inst->Instruction.Label) {
+         next_token( ctx, &inst->Label);
+      }
 
-         extended = token.Extended;
+      if (inst->Instruction.Texture) {
+         next_token( ctx, &inst->Texture);
       }
 
       assert( inst->Instruction.NumDstRegs <= TGSI_FULL_MAX_DST_REGISTERS );
 
       for(  i = 0; i < inst->Instruction.NumDstRegs; i++ ) {
-         unsigned extended;
 
-         next_token( ctx, &inst->FullDstRegisters[i].DstRegister );
+         next_token( ctx, &inst->Dst[i].Register );
 
          /*
           * No support for indirect or multi-dimensional addressing.
           */
-         assert( !inst->FullDstRegisters[i].DstRegister.Dimension );
-
-         extended = inst->FullDstRegisters[i].DstRegister.Extended;
-
-         while( extended ) {
-            struct tgsi_src_register_ext token;
+         assert( !inst->Dst[i].Register.Dimension );
 
-            next_token( ctx, &token );
-
-            switch( token.Type ) {
-            case TGSI_DST_REGISTER_EXT_TYPE_MODULATE:
-               copy_token(&inst->FullDstRegisters[i].DstRegisterExtModulate,
-                          &token);
-               break;
-
-            default:
-               assert( 0 );
-            }
-
-            extended = token.Extended;
-         }
-
-         if( inst->FullDstRegisters[i].DstRegister.Indirect ) {
-            next_token( ctx, &inst->FullDstRegisters[i].DstRegisterInd );
+         if( inst->Dst[i].Register.Indirect ) {
+            next_token( ctx, &inst->Dst[i].Indirect );
 
             /*
              * No support for indirect or multi-dimensional addressing.
              */
-            assert( !inst->FullDstRegisters[i].DstRegisterInd.Indirect );
-            assert( !inst->FullDstRegisters[i].DstRegisterInd.Dimension );
-            assert( !inst->FullDstRegisters[i].DstRegisterInd.Extended );
+            assert( !inst->Dst[i].Indirect.Dimension );
+            assert( !inst->Dst[i].Indirect.Indirect );
          }
       }
 
       assert( inst->Instruction.NumSrcRegs <= TGSI_FULL_MAX_SRC_REGISTERS );
 
       for( i = 0; i < inst->Instruction.NumSrcRegs; i++ ) {
-         unsigned extended;
-
-         next_token( ctx, &inst->FullSrcRegisters[i].SrcRegister );
-
-         extended = inst->FullSrcRegisters[i].SrcRegister.Extended;
-
-         while( extended ) {
-            struct tgsi_src_register_ext token;
-
-            next_token( ctx, &token );
-
-            switch( token.Type ) {
-            case TGSI_SRC_REGISTER_EXT_TYPE_MOD:
-               copy_token(&inst->FullSrcRegisters[i].SrcRegisterExtMod,
-                          &token);
-               break;
 
-            default:
-               assert( 0 );
-            }
-
-            extended = token.Extended;
-         }
+         next_token( ctx, &inst->Src[i].Register );
 
-         if( inst->FullSrcRegisters[i].SrcRegister.Indirect ) {
-            next_token( ctx, &inst->FullSrcRegisters[i].SrcRegisterInd );
+         if( inst->Src[i].Register.Indirect ) {
+            next_token( ctx, &inst->Src[i].Indirect );
 
             /*
              * No support for indirect or multi-dimensional addressing.
              */
-            assert( !inst->FullSrcRegisters[i].SrcRegisterInd.Indirect );
-            assert( !inst->FullSrcRegisters[i].SrcRegisterInd.Dimension );
-            assert( !inst->FullSrcRegisters[i].SrcRegisterInd.Extended );
+            assert( !inst->Src[i].Indirect.Indirect );
+            assert( !inst->Src[i].Indirect.Dimension );
          }
 
-         if( inst->FullSrcRegisters[i].SrcRegister.Dimension ) {
-            next_token( ctx, &inst->FullSrcRegisters[i].SrcRegisterDim );
+         if( inst->Src[i].Register.Dimension ) {
+            next_token( ctx, &inst->Src[i].Dimension );
 
             /*
              * No support for multi-dimensional addressing.
              */
-            assert( !inst->FullSrcRegisters[i].SrcRegisterDim.Dimension );
-            assert( !inst->FullSrcRegisters[i].SrcRegisterDim.Extended );
+            assert( !inst->Src[i].Dimension.Dimension );
 
-            if( inst->FullSrcRegisters[i].SrcRegisterDim.Indirect ) {
-               next_token( ctx, &inst->FullSrcRegisters[i].SrcRegisterDimInd );
+            if( inst->Src[i].Dimension.Indirect ) {
+               next_token( ctx, &inst->Src[i].DimIndirect );
 
                /*
                * No support for indirect or multi-dimensional addressing.
                */
-               assert( !inst->FullSrcRegisters[i].SrcRegisterInd.Indirect );
-               assert( !inst->FullSrcRegisters[i].SrcRegisterInd.Dimension );
-               assert( !inst->FullSrcRegisters[i].SrcRegisterInd.Extended );
+               assert( !inst->Src[i].Indirect.Indirect );
+               assert( !inst->Src[i].Indirect.Dimension );
             }
          }
       }
@@ -307,6 +220,22 @@ tgsi_parse_token(
       break;
    }
 
+   case TGSI_TOKEN_TYPE_PROPERTY:
+   {
+      struct tgsi_full_property *prop = &ctx->FullToken.FullProperty;
+      uint prop_count;
+
+      memset(prop, 0, sizeof *prop);
+      copy_token(&prop->Property, &token);
+
+      prop_count = prop->Property.NrTokens - 1;
+      for (i = 0; i < prop_count; i++) {
+         next_token(ctx, &prop->u[i]);
+      }
+
+      break;
+   }
+
    default:
       assert( 0 );
    }
@@ -319,8 +248,7 @@ tgsi_num_tokens(const struct tgsi_token *tokens)
    struct tgsi_parse_context ctx;
    if (tgsi_parse_init(&ctx, tokens) == TGSI_PARSE_OK) {
       unsigned len = (ctx.FullHeader.Header.HeaderSize +
-                      ctx.FullHeader.Header.BodySize +
-                      1);
+                      ctx.FullHeader.Header.BodySize);
       return len;
    }
    return 0;
index 76f1676d85d1e27f9b87ae2635e0f305aafac7c4..439a57269b7ef4e85565b08113d88b08cdb455b9 100644 (file)
 extern "C" {
 #endif
 
-struct tgsi_full_version
-{
-   struct tgsi_version  Version;
-};
-
 struct tgsi_full_header
 {
    struct tgsi_header      Header;
@@ -47,24 +42,22 @@ struct tgsi_full_header
 
 struct tgsi_full_dst_register
 {
-   struct tgsi_dst_register               DstRegister;
-   struct tgsi_src_register               DstRegisterInd;
-   struct tgsi_dst_register_ext_modulate  DstRegisterExtModulate;
+   struct tgsi_dst_register               Register;
+   struct tgsi_src_register               Indirect;
 };
 
 struct tgsi_full_src_register
 {
-   struct tgsi_src_register         SrcRegister;
-   struct tgsi_src_register_ext_mod SrcRegisterExtMod;
-   struct tgsi_src_register         SrcRegisterInd;
-   struct tgsi_dimension            SrcRegisterDim;
-   struct tgsi_src_register         SrcRegisterDimInd;
+   struct tgsi_src_register         Register;
+   struct tgsi_src_register         Indirect;
+   struct tgsi_dimension            Dimension;
+   struct tgsi_src_register         DimIndirect;
 };
 
 struct tgsi_full_declaration
 {
    struct tgsi_declaration Declaration;
-   struct tgsi_declaration_range DeclarationRange;
+   struct tgsi_declaration_range Range;
    struct tgsi_declaration_semantic Semantic;
 };
 
@@ -74,18 +67,23 @@ struct tgsi_full_immediate
    union tgsi_immediate_data u[4];
 };
 
+struct tgsi_full_property
+{
+   struct tgsi_property   Property;
+   struct tgsi_property_data u[8];
+};
+
 #define TGSI_FULL_MAX_DST_REGISTERS 2
 #define TGSI_FULL_MAX_SRC_REGISTERS 4 /* TXD has 4 */
 
 struct tgsi_full_instruction
 {
    struct tgsi_instruction             Instruction;
-   struct tgsi_instruction_ext_label   InstructionExtLabel;
-   struct tgsi_instruction_ext_texture InstructionExtTexture;
-   struct tgsi_instruction_ext_predicate InstructionExtPredicate;
-   struct tgsi_full_dst_register       FullDstRegisters[TGSI_FULL_MAX_DST_REGISTERS];
-   struct tgsi_full_src_register       FullSrcRegisters[TGSI_FULL_MAX_SRC_REGISTERS];
-   uint Flags;  /**< user-defined usage */
+   struct tgsi_instruction_predicate   Predicate;
+   struct tgsi_instruction_label       Label;
+   struct tgsi_instruction_texture     Texture;
+   struct tgsi_full_dst_register       Dst[TGSI_FULL_MAX_DST_REGISTERS];
+   struct tgsi_full_src_register       Src[TGSI_FULL_MAX_SRC_REGISTERS];
 };
 
 union tgsi_full_token
@@ -94,21 +92,13 @@ union tgsi_full_token
    struct tgsi_full_declaration  FullDeclaration;
    struct tgsi_full_immediate    FullImmediate;
    struct tgsi_full_instruction  FullInstruction;
+   struct tgsi_full_property     FullProperty;
 };
 
-void
-tgsi_full_token_init(
-   union tgsi_full_token *full_token );
-
-void
-tgsi_full_token_free(
-   union tgsi_full_token *full_token );
-
 struct tgsi_parse_context
 {
    const struct tgsi_token    *Tokens;
    unsigned                   Position;
-   struct tgsi_full_version   FullVersion;
    struct tgsi_full_header    FullHeader;
    union tgsi_full_token      FullToken;
 };
index 617fd7f6be1f23580e9d9c503c9f791550de94d6..138d2d095bb037577148113ccc8a43ee958c5a10 100644 (file)
@@ -60,7 +60,7 @@ const float ppc_builtin_constants[] ALIGN16_ATTRIB = {
    for (CHAN = 0; CHAN < NUM_CHANNELS; CHAN++)
 
 #define IS_DST0_CHANNEL_ENABLED( INST, CHAN )\
-   ((INST).FullDstRegisters[0].DstRegister.WriteMask & (1 << (CHAN)))
+   ((INST).Dst[0].Register.WriteMask & (1 << (CHAN)))
 
 #define IF_IS_DST0_CHANNEL_ENABLED( INST, CHAN )\
    if (IS_DST0_CHANNEL_ENABLED( INST, CHAN ))
@@ -156,8 +156,8 @@ init_gen_context(struct gen_context *gen, struct ppc_function *func)
 static boolean
 is_ppc_vec_temporary(const struct tgsi_full_src_register *reg)
 {
-   return (reg->SrcRegister.File == TGSI_FILE_TEMPORARY &&
-           reg->SrcRegister.Index < MAX_PPC_TEMPS);
+   return (reg->Register.File == TGSI_FILE_TEMPORARY &&
+           reg->Register.Index < MAX_PPC_TEMPS);
 }
 
 
@@ -167,8 +167,8 @@ is_ppc_vec_temporary(const struct tgsi_full_src_register *reg)
 static boolean
 is_ppc_vec_temporary_dst(const struct tgsi_full_dst_register *reg)
 {
-   return (reg->DstRegister.File == TGSI_FILE_TEMPORARY &&
-           reg->DstRegister.Index < MAX_PPC_TEMPS);
+   return (reg->Register.File == TGSI_FILE_TEMPORARY &&
+           reg->Register.Index < MAX_PPC_TEMPS);
 }
 
 
@@ -291,10 +291,11 @@ emit_fetch(struct gen_context *gen,
    case TGSI_SWIZZLE_Y:
    case TGSI_SWIZZLE_Z:
    case TGSI_SWIZZLE_W:
-      switch (reg->SrcRegister.File) {
+      switch (reg->Register.File) {
       case TGSI_FILE_INPUT:
+      case TGSI_FILE_SYSTEM_VALUE:
          {
-            int offset = (reg->SrcRegister.Index * 4 + swizzle) * 16;
+            int offset = (reg->Register.Index * 4 + swizzle) * 16;
             int offset_reg = emit_li_offset(gen, offset);
             dst_vec = ppc_allocate_vec_register(gen->f);
             ppc_lvx(gen->f, dst_vec, gen->inputs_reg, offset_reg);
@@ -303,11 +304,11 @@ emit_fetch(struct gen_context *gen,
       case TGSI_FILE_TEMPORARY:
          if (is_ppc_vec_temporary(reg)) {
             /* use PPC vec register */
-            dst_vec = gen->temps_map[reg->SrcRegister.Index][swizzle];
+            dst_vec = gen->temps_map[reg->Register.Index][swizzle];
          }
          else {
             /* use memory-based temp register "file" */
-            int offset = (reg->SrcRegister.Index * 4 + swizzle) * 16;
+            int offset = (reg->Register.Index * 4 + swizzle) * 16;
             int offset_reg = emit_li_offset(gen, offset);
             dst_vec = ppc_allocate_vec_register(gen->f);
             ppc_lvx(gen->f, dst_vec, gen->temps_reg, offset_reg);
@@ -315,7 +316,7 @@ emit_fetch(struct gen_context *gen,
          break;
       case TGSI_FILE_IMMEDIATE:
          {
-            int offset = (reg->SrcRegister.Index * 4 + swizzle) * 4;
+            int offset = (reg->Register.Index * 4 + swizzle) * 4;
             int offset_reg = emit_li_offset(gen, offset);
             dst_vec = ppc_allocate_vec_register(gen->f);
             /* Load 4-byte word into vector register.
@@ -331,7 +332,7 @@ emit_fetch(struct gen_context *gen,
          break;
       case TGSI_FILE_CONSTANT:
          {
-            int offset = (reg->SrcRegister.Index * 4 + swizzle) * 4;
+            int offset = (reg->Register.Index * 4 + swizzle) * 4;
             int offset_reg = emit_li_offset(gen, offset);
             dst_vec = ppc_allocate_vec_register(gen->f);
             /* Load 4-byte word into vector register.
@@ -404,9 +405,9 @@ equal_src_locs(const struct tgsi_full_src_register *a, uint chan_a,
 {
    int swz_a, swz_b;
    int sign_a, sign_b;
-   if (a->SrcRegister.File != b->SrcRegister.File)
+   if (a->Register.File != b->Register.File)
       return FALSE;
-   if (a->SrcRegister.Index != b->SrcRegister.Index)
+   if (a->Register.Index != b->Register.Index)
       return FALSE;
    swz_a = tgsi_util_get_full_src_register_swizzle(a, chan_a);
    swz_b = tgsi_util_get_full_src_register_swizzle(b, chan_b);
@@ -431,7 +432,7 @@ get_src_vec(struct gen_context *gen,
             struct tgsi_full_instruction *inst, int src_reg, uint chan)
 {
    const const struct tgsi_full_src_register *src = 
-      &inst->FullSrcRegisters[src_reg];
+      &inst->Src[src_reg];
    int vec;
    uint i;
 
@@ -482,10 +483,10 @@ get_dst_vec(struct gen_context *gen,
             const struct tgsi_full_instruction *inst,
             unsigned chan_index)
 {
-   const struct tgsi_full_dst_register *reg = &inst->FullDstRegisters[0];
+   const struct tgsi_full_dst_register *reg = &inst->Dst[0];
 
    if (is_ppc_vec_temporary_dst(reg)) {
-      int vec = gen->temps_map[reg->DstRegister.Index][chan_index];
+      int vec = gen->temps_map[reg->Register.Index][chan_index];
       return vec;
    }
    else {
@@ -505,12 +506,12 @@ emit_store(struct gen_context *gen,
            unsigned chan_index,
            boolean free_vec)
 {
-   const struct tgsi_full_dst_register *reg = &inst->FullDstRegisters[0];
+   const struct tgsi_full_dst_register *reg = &inst->Dst[0];
 
-   switch (reg->DstRegister.File) {
+   switch (reg->Register.File) {
    case TGSI_FILE_OUTPUT:
       {
-         int offset = (reg->DstRegister.Index * 4 + chan_index) * 16;
+         int offset = (reg->Register.Index * 4 + chan_index) * 16;
          int offset_reg = emit_li_offset(gen, offset);
          ppc_stvx(gen->f, src_vec, gen->outputs_reg, offset_reg);
       }
@@ -518,14 +519,14 @@ emit_store(struct gen_context *gen,
    case TGSI_FILE_TEMPORARY:
       if (is_ppc_vec_temporary_dst(reg)) {
          if (!free_vec) {
-            int dst_vec = gen->temps_map[reg->DstRegister.Index][chan_index];
+            int dst_vec = gen->temps_map[reg->Register.Index][chan_index];
             if (dst_vec != src_vec)
                ppc_vmove(gen->f, dst_vec, src_vec);
          }
          free_vec = FALSE;
       }
       else {
-         int offset = (reg->DstRegister.Index * 4 + chan_index) * 16;
+         int offset = (reg->Register.Index * 4 + chan_index) * 16;
          int offset_reg = emit_li_offset(gen, offset);
          ppc_stvx(gen->f, src_vec, gen->temps_reg, offset_reg);
       }
@@ -535,7 +536,7 @@ emit_store(struct gen_context *gen,
       emit_addrs(
          func,
          xmm,
-         reg->DstRegister.Index,
+         reg->Register.Index,
          chan_index );
       break;
 #endif
@@ -1173,13 +1174,14 @@ emit_declaration(
    struct ppc_function *func,
    struct tgsi_full_declaration *decl )
 {
-   if( decl->Declaration.File == TGSI_FILE_INPUT ) {
+   if( decl->Declaration.File == TGSI_FILE_INPUT ||
+       decl->Declaration.File == TGSI_FILE_SYSTEM_VALUE ) {
 #if 0
       unsigned first, last, mask;
       unsigned i, j;
 
-      first = decl->DeclarationRange.First;
-      last = decl->DeclarationRange.Last;
+      first = decl->Range.First;
+      last = decl->Range.Last;
       mask = decl->Declaration.UsageMask;
 
       for( i = first; i <= last; i++ ) {
@@ -1339,6 +1341,9 @@ tgsi_emit_ppc(const struct tgsi_token *tokens,
          }
          break;
 
+      case TGSI_TOKEN_TYPE_PROPERTY:
+         break;
+
       default:
         ok = 0;
          assert( 0 );
index 36e27ea52f485e59421e3e02abd322f39058ebe2..16b8ec60518458759d01a6e19586f7adf9d90b9f 100644 (file)
  **************************************************************************/
 
 #include "util/u_debug.h"
+#include "util/u_memory.h"
+#include "util/u_prim.h"
+#include "cso_cache/cso_hash.h"
 #include "tgsi_sanity.h"
 #include "tgsi_info.h"
 #include "tgsi_iterate.h"
 
-typedef uint reg_flag;
-
-#define BITS_IN_REG_FLAG (sizeof( reg_flag ) * 8)
-
-#define MAX_REGISTERS 256
-#define MAX_REG_FLAGS ((MAX_REGISTERS + BITS_IN_REG_FLAG - 1) / BITS_IN_REG_FLAG)
+typedef struct {
+   uint file : 28;
+   /* max 2 dimensions */
+   uint dimensions : 4;
+   uint indices[2];
+} scan_register;
 
 struct sanity_check_ctx
 {
    struct tgsi_iterate_context iter;
+   struct cso_hash *regs_decl;
+   struct cso_hash *regs_used;
+   struct cso_hash *regs_ind_used;
 
-   reg_flag regs_decl[TGSI_FILE_COUNT][MAX_REG_FLAGS];
-   reg_flag regs_used[TGSI_FILE_COUNT][MAX_REG_FLAGS];
-   boolean regs_ind_used[TGSI_FILE_COUNT];
    uint num_imms;
    uint num_instructions;
    uint index_of_END;
 
    uint errors;
    uint warnings;
+   uint implied_array_size;
 };
 
+static INLINE unsigned
+scan_register_key(const scan_register *reg)
+{
+   unsigned key = reg->file;
+   key |= (reg->indices[0] << 4);
+   key |= (reg->indices[1] << 18);
+
+   return key;
+}
+
+static void
+fill_scan_register1d(scan_register *reg,
+                     uint file, uint index)
+{
+   reg->file = file;
+   reg->dimensions = 1;
+   reg->indices[0] = index;
+   reg->indices[1] = 0;
+}
+
+static void
+fill_scan_register2d(scan_register *reg,
+                     uint file, uint index1, uint index2)
+{
+   reg->file = file;
+   reg->dimensions = 2;
+   reg->indices[0] = index1;
+   reg->indices[1] = index2;
+}
+
+static void
+scan_register_dst(scan_register *reg,
+                  struct tgsi_full_dst_register *dst)
+{
+   fill_scan_register1d(reg,
+                        dst->Register.File,
+                        dst->Register.Index);
+}
+
+static void
+scan_register_src(scan_register *reg,
+                  struct tgsi_full_src_register *src)
+{
+   if (src->Register.Dimension) {
+      /*FIXME: right now we don't support indirect
+       * multidimensional addressing */
+      debug_assert(!src->Dimension.Indirect);
+      fill_scan_register2d(reg,
+                           src->Register.File,
+                           src->Register.Index,
+                           src->Dimension.Index);
+   } else {
+      fill_scan_register1d(reg,
+                           src->Register.File,
+                           src->Register.Index);
+   }
+}
+
+static scan_register *
+create_scan_register_src(struct tgsi_full_src_register *src)
+{
+   scan_register *reg = MALLOC(sizeof(scan_register));
+   scan_register_src(reg, src);
+
+   return reg;
+}
+
+static scan_register *
+create_scan_register_dst(struct tgsi_full_dst_register *dst)
+{
+   scan_register *reg = MALLOC(sizeof(scan_register));
+   scan_register_dst(reg, dst);
+
+   return reg;
+}
+
 static void
 report_error(
    struct sanity_check_ctx *ctx,
@@ -99,12 +179,12 @@ check_file_name(
 static boolean
 is_register_declared(
    struct sanity_check_ctx *ctx,
-   uint file,
-   int index )
+   const scan_register *reg)
 {
-   assert( index >= 0 && index < MAX_REGISTERS );
-
-   return (ctx->regs_decl[file][index / BITS_IN_REG_FLAG] & (1 << (index % BITS_IN_REG_FLAG))) ? TRUE : FALSE;
+   void *data = cso_hash_find_data_from_template(
+      ctx->regs_decl, scan_register_key(reg),
+      (void*)reg, sizeof(scan_register));
+   return  data ? TRUE : FALSE;
 }
 
 static boolean
@@ -112,23 +192,37 @@ is_any_register_declared(
    struct sanity_check_ctx *ctx,
    uint file )
 {
-   uint i;
+   struct cso_hash_iter iter =
+      cso_hash_first_node(ctx->regs_decl);
 
-   for (i = 0; i < MAX_REG_FLAGS; i++)
-      if (ctx->regs_decl[file][i])
+   while (cso_hash_iter_is_null(iter)) {
+      scan_register *reg = (scan_register *)cso_hash_iter_data(iter);
+      if (reg->file == file)
          return TRUE;
+      iter = cso_hash_iter_next(iter);
+   }
+
    return FALSE;
 }
 
 static boolean
 is_register_used(
    struct sanity_check_ctx *ctx,
-   uint file,
-   int index )
+   scan_register *reg)
 {
-   assert( index < MAX_REGISTERS );
+   void *data = cso_hash_find_data_from_template(
+      ctx->regs_used, scan_register_key(reg),
+      reg, sizeof(scan_register));
+   return  data ? TRUE : FALSE;
+}
 
-   return (ctx->regs_used[file][index / BITS_IN_REG_FLAG] & (1 << (index % BITS_IN_REG_FLAG))) ? TRUE : FALSE;
+
+static boolean
+is_ind_register_used(
+   struct sanity_check_ctx *ctx,
+   scan_register *reg)
+{
+   return cso_hash_contains(ctx->regs_ind_used, reg->file);
 }
 
 static const char *file_names[TGSI_FILE_COUNT] =
@@ -148,31 +242,40 @@ static const char *file_names[TGSI_FILE_COUNT] =
 static boolean
 check_register_usage(
    struct sanity_check_ctx *ctx,
-   uint file,
-   int index,
+   scan_register *reg,
    const char *name,
    boolean indirect_access )
 {
-   if (!check_file_name( ctx, file ))
+   if (!check_file_name( ctx, reg->file )) {
+      free(reg);
       return FALSE;
+   }
 
    if (indirect_access) {
       /* Note that 'index' is an offset relative to the value of the
-       * address register.  No range checking done here.
-       */
-      if (!is_any_register_declared( ctx, file ))
-         report_error( ctx, "%s: Undeclared %s register", file_names[file], name );
-      ctx->regs_ind_used[file] = TRUE;
+       * address register.  No range checking done here.*/
+      reg->indices[0] = 0;
+      reg->indices[1] = 0;
+      if (!is_any_register_declared( ctx, reg->file ))
+         report_error( ctx, "%s: Undeclared %s register", file_names[reg->file], name );
+      if (!is_ind_register_used(ctx, reg))
+         cso_hash_insert(ctx->regs_ind_used, reg->file, reg);
+      else
+         free(reg);
    }
    else {
-      if (index < 0 || index >= MAX_REGISTERS) {
-         report_error( ctx, "%s[%d]: Invalid %s index", file_names[file], index, name );
-         return FALSE;
-      }
-
-      if (!is_register_declared( ctx, file, index ))
-         report_error( ctx, "%s[%d]: Undeclared %s register", file_names[file], index, name );
-      ctx->regs_used[file][index / BITS_IN_REG_FLAG] |= (1 << (index % BITS_IN_REG_FLAG));
+      if (!is_register_declared( ctx, reg )) {
+         if (reg->dimensions == 2)
+            report_error( ctx, "%s[%d][%d]: Undeclared %s register", file_names[reg->file],
+                          reg->indices[0], reg->indices[1], name );
+         else
+            report_error( ctx, "%s[%d]: Undeclared %s register", file_names[reg->file],
+                          reg->indices[0], name );
+         }
+      if (!is_register_used( ctx, reg ))
+         cso_hash_insert(ctx->regs_used, scan_register_key(reg), reg);
+      else
+         free(reg);
    }
    return TRUE;
 }
@@ -210,33 +313,33 @@ iter_instruction(
     * Mark the registers as used.
     */
    for (i = 0; i < inst->Instruction.NumDstRegs; i++) {
+      scan_register *reg = create_scan_register_dst(&inst->Dst[i]);
       check_register_usage(
          ctx,
-         inst->FullDstRegisters[i].DstRegister.File,
-         inst->FullDstRegisters[i].DstRegister.Index,
+         reg,
          "destination",
          FALSE );
    }
    for (i = 0; i < inst->Instruction.NumSrcRegs; i++) {
+      scan_register *reg = create_scan_register_src(&inst->Src[i]);
       check_register_usage(
          ctx,
-         inst->FullSrcRegisters[i].SrcRegister.File,
-         inst->FullSrcRegisters[i].SrcRegister.Index,
+         reg,
          "source",
-         (boolean)inst->FullSrcRegisters[i].SrcRegister.Indirect );
-      if (inst->FullSrcRegisters[i].SrcRegister.Indirect) {
-         uint file;
-         int index;
+         (boolean)inst->Src[i].Register.Indirect );
+      if (inst->Src[i].Register.Indirect) {
+         scan_register *ind_reg = MALLOC(sizeof(scan_register));
 
-         file = inst->FullSrcRegisters[i].SrcRegisterInd.File;
-         index = inst->FullSrcRegisters[i].SrcRegisterInd.Index;
+         fill_scan_register1d(ind_reg,
+                              inst->Src[i].Indirect.File,
+                              inst->Src[i].Indirect.Index);
          check_register_usage(
             ctx,
-            file,
-            index,
+            reg,
             "indirect",
             FALSE );
-         if (!(file == TGSI_FILE_ADDRESS || file == TGSI_FILE_LOOP) || index != 0) {
+         if (!(reg->file == TGSI_FILE_ADDRESS || reg->file == TGSI_FILE_LOOP) ||
+             reg->indices[0] != 0) {
             report_warning(ctx, "Indirect register neither ADDR[0] nor LOOP[0]");
          }
       }
@@ -245,8 +348,8 @@ iter_instruction(
    switch (inst->Instruction.Opcode) {
    case TGSI_OPCODE_BGNFOR:
    case TGSI_OPCODE_ENDFOR:
-      if (inst->FullDstRegisters[0].DstRegister.File != TGSI_FILE_LOOP ||
-          inst->FullDstRegisters[0].DstRegister.Index != 0) {
+      if (inst->Dst[0].Register.File != TGSI_FILE_LOOP ||
+          inst->Dst[0].Register.Index != 0) {
          report_error(ctx, "Destination register must be LOOP[0]");
       }
       break;
@@ -254,8 +357,8 @@ iter_instruction(
 
    switch (inst->Instruction.Opcode) {
    case TGSI_OPCODE_BGNFOR:
-      if (inst->FullSrcRegisters[0].SrcRegister.File != TGSI_FILE_CONSTANT &&
-          inst->FullSrcRegisters[0].SrcRegister.File != TGSI_FILE_IMMEDIATE) {
+      if (inst->Src[0].Register.File != TGSI_FILE_CONSTANT &&
+          inst->Src[0].Register.File != TGSI_FILE_IMMEDIATE) {
          report_error(ctx, "Source register file must be either CONST or IMM");
       }
       break;
@@ -266,6 +369,19 @@ iter_instruction(
    return TRUE;
 }
 
+static void
+check_and_declare(struct sanity_check_ctx *ctx,
+                  scan_register *reg)
+{
+   if (is_register_declared( ctx, reg))
+      report_error( ctx, "%s[%u]: The same register declared more than once",
+                    file_names[reg->file], reg->indices[0] );
+   cso_hash_insert(ctx->regs_decl,
+                   scan_register_key(reg),
+                   reg);
+}
+
+
 static boolean
 iter_declaration(
    struct tgsi_iterate_context *iter,
@@ -286,10 +402,22 @@ iter_declaration(
    file = decl->Declaration.File;
    if (!check_file_name( ctx, file ))
       return TRUE;
-   for (i = decl->DeclarationRange.First; i <= decl->DeclarationRange.Last; i++) {
-      if (is_register_declared( ctx, file, i ))
-         report_error( ctx, "%s[%u]: The same register declared more than once", file_names[file], i );
-      ctx->regs_decl[file][i / BITS_IN_REG_FLAG] |= (1 << (i % BITS_IN_REG_FLAG));
+   for (i = decl->Range.First; i <= decl->Range.Last; i++) {
+      /* declared TGSI_FILE_INPUT's for geometry processor
+       * have an implied second dimension */
+      if (file == TGSI_FILE_INPUT &&
+          ctx->iter.processor.Processor == TGSI_PROCESSOR_GEOMETRY) {
+         uint vert;
+         for (vert = 0; vert < ctx->implied_array_size; ++vert) {
+            scan_register *reg = MALLOC(sizeof(scan_register));
+            fill_scan_register2d(reg, file, vert, i);
+            check_and_declare(ctx, reg);
+         }
+      } else {
+         scan_register *reg = MALLOC(sizeof(scan_register));
+         fill_scan_register1d(reg, file, i);
+         check_and_declare(ctx, reg);
+      }
    }
 
    return TRUE;
@@ -301,8 +429,7 @@ iter_immediate(
    struct tgsi_full_immediate *imm )
 {
    struct sanity_check_ctx *ctx = (struct sanity_check_ctx *) iter;
-
-   assert( ctx->num_imms < MAX_REGISTERS );
+   scan_register *reg;
 
    /* No immediates allowed after the first instruction.
     */
@@ -311,7 +438,9 @@ iter_immediate(
 
    /* Mark the register as declared.
     */
-   ctx->regs_decl[TGSI_FILE_IMMEDIATE][ctx->num_imms / BITS_IN_REG_FLAG] |= (1 << (ctx->num_imms % BITS_IN_REG_FLAG));
+   reg = MALLOC(sizeof(scan_register));
+   fill_scan_register1d(reg, TGSI_FILE_IMMEDIATE, ctx->num_imms);
+   cso_hash_insert(ctx->regs_decl, scan_register_key(reg), reg);
    ctx->num_imms++;
 
    /* Check data type validity.
@@ -324,12 +453,26 @@ iter_immediate(
    return TRUE;
 }
 
+
+static boolean
+iter_property(
+   struct tgsi_iterate_context *iter,
+   struct tgsi_full_property *prop )
+{
+   struct sanity_check_ctx *ctx = (struct sanity_check_ctx *) iter;
+
+   if (iter->processor.Processor == TGSI_PROCESSOR_GEOMETRY &&
+       prop->Property.PropertyName == TGSI_PROPERTY_GS_INPUT_PRIM) {
+      ctx->implied_array_size = u_vertices_per_prim(prop->u[0].Data);
+   }
+   return TRUE;
+}
+
 static boolean
 epilog(
    struct tgsi_iterate_context *iter )
 {
    struct sanity_check_ctx *ctx = (struct sanity_check_ctx *) iter;
-   uint file;
 
    /* There must be an END instruction somewhere.
     */
@@ -339,13 +482,17 @@ epilog(
 
    /* Check if all declared registers were used.
     */
-   for (file = TGSI_FILE_NULL; file < TGSI_FILE_COUNT; file++) {
-      uint i;
-
-      for (i = 0; i < MAX_REGISTERS; i++) {
-         if (is_register_declared( ctx, file, i ) && !is_register_used( ctx, file, i ) && !ctx->regs_ind_used[file]) {
-            report_warning( ctx, "%s[%u]: Register never used", file_names[file], i );
+   {
+      struct cso_hash_iter iter =
+         cso_hash_first_node(ctx->regs_decl);
+
+      while (cso_hash_iter_is_null(iter)) {
+         scan_register *reg = (scan_register *)cso_hash_iter_data(iter);
+         if (!is_register_used(ctx, reg) && !is_ind_register_used(ctx, reg)) {
+            report_warning( ctx, "%s[%u]: Register never used",
+                            file_names[reg->file], reg->indices[0] );
          }
+         iter = cso_hash_iter_next(iter);
       }
    }
 
@@ -357,6 +504,18 @@ epilog(
    return TRUE;
 }
 
+static void
+regs_hash_destroy(struct cso_hash *hash)
+{
+   struct cso_hash_iter iter = cso_hash_first_node(hash);
+   while (!cso_hash_iter_is_null(iter)) {
+      scan_register *reg = (scan_register *)cso_hash_iter_data(iter);
+      iter = cso_hash_erase(hash, iter);
+      free(reg);
+   }
+   cso_hash_delete(hash);
+}
+
 boolean
 tgsi_sanity_check(
    const struct tgsi_token *tokens )
@@ -367,20 +526,26 @@ tgsi_sanity_check(
    ctx.iter.iterate_instruction = iter_instruction;
    ctx.iter.iterate_declaration = iter_declaration;
    ctx.iter.iterate_immediate = iter_immediate;
+   ctx.iter.iterate_property = iter_property;
    ctx.iter.epilog = epilog;
 
-   memset( ctx.regs_decl, 0, sizeof( ctx.regs_decl ) );
-   memset( ctx.regs_used, 0, sizeof( ctx.regs_used ) );
-   memset( ctx.regs_ind_used, 0, sizeof( ctx.regs_ind_used ) );
+   ctx.regs_decl = cso_hash_create();
+   ctx.regs_used = cso_hash_create();
+   ctx.regs_ind_used = cso_hash_create();
+
    ctx.num_imms = 0;
    ctx.num_instructions = 0;
    ctx.index_of_END = ~0;
 
    ctx.errors = 0;
    ctx.warnings = 0;
+   ctx.implied_array_size = 0;
 
    if (!tgsi_iterate_shader( tokens, &ctx.iter ))
       return FALSE;
 
+   regs_hash_destroy(ctx.regs_decl);
+   regs_hash_destroy(ctx.regs_used);
+   regs_hash_destroy(ctx.regs_ind_used);
    return ctx.errors == 0;
 }
index f9c16f1b6cb78968ed029ad4862bfe9af6775057..0f48b0dc3a113ff53ad3e5121bd5f5d91f0dd81b 100644 (file)
@@ -35,7 +35,6 @@
 
 
 #include "util/u_math.h"
-#include "tgsi/tgsi_build.h"
 #include "tgsi/tgsi_parse.h"
 #include "tgsi/tgsi_scan.h"
 
@@ -97,14 +96,15 @@ tgsi_scan_shader(const struct tgsi_token *tokens,
                uint i;
                for (i = 0; i < fullinst->Instruction.NumSrcRegs; i++) {
                   const struct tgsi_full_src_register *src =
-                     &fullinst->FullSrcRegisters[i];
-                  if (src->SrcRegister.File == TGSI_FILE_INPUT) {
-                     const int ind = src->SrcRegister.Index;
+                     &fullinst->Src[i];
+                  if (src->Register.File == TGSI_FILE_INPUT ||
+                      src->Register.File == TGSI_FILE_SYSTEM_VALUE) {
+                     const int ind = src->Register.Index;
                      if (info->input_semantic_name[ind] == TGSI_SEMANTIC_FOG) {
-                        if (src->SrcRegister.SwizzleX == TGSI_SWIZZLE_X) {
+                        if (src->Register.SwizzleX == TGSI_SWIZZLE_X) {
                            info->uses_fogcoord = TRUE;
                         }
-                        else if (src->SrcRegister.SwizzleX == TGSI_SWIZZLE_Y) {
+                        else if (src->Register.SwizzleX == TGSI_SWIZZLE_Y) {
                            info->uses_frontfacing = TRUE;
                         }
                      }
@@ -120,8 +120,8 @@ tgsi_scan_shader(const struct tgsi_token *tokens,
                = &parse.FullToken.FullDeclaration;
             const uint file = fulldecl->Declaration.File;
             uint reg;
-            for (reg = fulldecl->DeclarationRange.First;
-                 reg <= fulldecl->DeclarationRange.Last;
+            for (reg = fulldecl->Range.First;
+                 reg <= fulldecl->Range.Last;
                  reg++) {
 
                /* only first 32 regs will appear in this bitfield */
@@ -129,25 +129,29 @@ tgsi_scan_shader(const struct tgsi_token *tokens,
                info->file_count[file]++;
                info->file_max[file] = MAX2(info->file_max[file], (int)reg);
 
-               if (file == TGSI_FILE_INPUT) {
-                  info->input_semantic_name[reg] = (ubyte)fulldecl->Semantic.SemanticName;
-                  info->input_semantic_index[reg] = (ubyte)fulldecl->Semantic.SemanticIndex;
+               if (file == TGSI_FILE_INPUT || file == TGSI_FILE_SYSTEM_VALUE) {
+                  info->input_semantic_name[reg] = (ubyte)fulldecl->Semantic.Name;
+                  info->input_semantic_index[reg] = (ubyte)fulldecl->Semantic.Index;
                   info->input_interpolate[reg] = (ubyte)fulldecl->Declaration.Interpolate;
                   info->num_inputs++;
                }
                else if (file == TGSI_FILE_OUTPUT) {
-                  info->output_semantic_name[reg] = (ubyte)fulldecl->Semantic.SemanticName;
-                  info->output_semantic_index[reg] = (ubyte)fulldecl->Semantic.SemanticIndex;
+                  info->output_semantic_name[reg] = (ubyte)fulldecl->Semantic.Name;
+                  info->output_semantic_index[reg] = (ubyte)fulldecl->Semantic.Index;
                   info->num_outputs++;
-               }
 
-               /* special case */
-               if (procType == TGSI_PROCESSOR_FRAGMENT &&
-                   file == TGSI_FILE_OUTPUT &&
-                   fulldecl->Semantic.SemanticName == TGSI_SEMANTIC_POSITION) {
-                  info->writes_z = TRUE;
+                  /* extra info for special outputs */
+                  if (procType == TGSI_PROCESSOR_FRAGMENT &&
+                      fulldecl->Semantic.Name == TGSI_SEMANTIC_POSITION) {
+                     info->writes_z = TRUE;
+                  }
+                  if (procType == TGSI_PROCESSOR_VERTEX &&
+                      fulldecl->Semantic.Name == TGSI_SEMANTIC_EDGEFLAG) {
+                     info->writes_edgeflag = TRUE;
+                  }
                }
-            }
+
+             }
          }
          break;
 
@@ -161,6 +165,19 @@ tgsi_scan_shader(const struct tgsi_token *tokens,
             info->file_max[file] = MAX2(info->file_max[file], (int)reg);
          }
          break;
+      case TGSI_TOKEN_TYPE_PROPERTY:
+      {
+         const struct tgsi_full_property *fullprop
+            = &parse.FullToken.FullProperty;
+
+         info->properties[info->num_properties].name =
+            fullprop->Property.PropertyName;
+         memcpy(info->properties[info->num_properties].data,
+                fullprop->u, 8 * sizeof(unsigned));;
+
+         ++info->num_properties;
+      }
+      break;
 
       default:
          assert( 0 );
@@ -206,29 +223,26 @@ tgsi_is_passthrough_shader(const struct tgsi_token *tokens)
             struct tgsi_full_instruction *fullinst =
                &parse.FullToken.FullInstruction;
             const struct tgsi_full_src_register *src =
-               &fullinst->FullSrcRegisters[0];
+               &fullinst->Src[0];
             const struct tgsi_full_dst_register *dst =
-               &fullinst->FullDstRegisters[0];
+               &fullinst->Dst[0];
 
             /* Do a whole bunch of checks for a simple move */
             if (fullinst->Instruction.Opcode != TGSI_OPCODE_MOV ||
-                src->SrcRegister.File != TGSI_FILE_INPUT ||
-                dst->DstRegister.File != TGSI_FILE_OUTPUT ||
-                src->SrcRegister.Index != dst->DstRegister.Index ||
-
-                src->SrcRegister.Negate ||
-                src->SrcRegisterExtMod.Negate ||
-                src->SrcRegisterExtMod.Absolute ||
-                src->SrcRegisterExtMod.Scale2X ||
-                src->SrcRegisterExtMod.Bias ||
-                src->SrcRegisterExtMod.Complement ||
-
-                src->SrcRegister.SwizzleX != TGSI_SWIZZLE_X ||
-                src->SrcRegister.SwizzleY != TGSI_SWIZZLE_Y ||
-                src->SrcRegister.SwizzleZ != TGSI_SWIZZLE_Z ||
-                src->SrcRegister.SwizzleW != TGSI_SWIZZLE_W ||
-
-                dst->DstRegister.WriteMask != TGSI_WRITEMASK_XYZW)
+                src->Register.File != TGSI_FILE_INPUT ||
+                src->Register.File != TGSI_FILE_SYSTEM_VALUE ||
+                dst->Register.File != TGSI_FILE_OUTPUT ||
+                src->Register.Index != dst->Register.Index ||
+
+                src->Register.Negate ||
+                src->Register.Absolute ||
+
+                src->Register.SwizzleX != TGSI_SWIZZLE_X ||
+                src->Register.SwizzleY != TGSI_SWIZZLE_Y ||
+                src->Register.SwizzleZ != TGSI_SWIZZLE_Z ||
+                src->Register.SwizzleW != TGSI_SWIZZLE_W ||
+
+                dst->Register.WriteMask != TGSI_WRITEMASK_XYZW)
             {
                tgsi_parse_free(&parse);
                return FALSE;
@@ -240,6 +254,8 @@ tgsi_is_passthrough_shader(const struct tgsi_token *tokens)
          /* fall-through */
       case TGSI_TOKEN_TYPE_IMMEDIATE:
          /* fall-through */
+      case TGSI_TOKEN_TYPE_PROPERTY:
+         /* fall-through */
       default:
          ; /* no-op */
       }
index 8a7ee0c7e4f1f1e2e69f159586f45d0a8cd28018..dae5376c24aae63dbebcab6f9d6d2e36adc75ca7 100644 (file)
@@ -33,7 +33,6 @@
 #include "pipe/p_state.h"
 #include "pipe/p_shader_tokens.h"
 
-
 /**
  * Shader summary info
  */
@@ -58,11 +57,17 @@ struct tgsi_shader_info
    uint opcode_count[TGSI_OPCODE_LAST];  /**< opcode histogram */
 
    boolean writes_z;  /**< does fragment shader write Z value? */
+   boolean writes_edgeflag; /**< vertex shader outputs edgeflag */
    boolean uses_kill;  /**< KIL or KILP instruction used? */
    boolean uses_fogcoord; /**< fragment shader uses fog coord? */
    boolean uses_frontfacing; /**< fragment shader uses front/back-face flag? */
-};
 
+   struct {
+      unsigned name;
+      unsigned data[8];
+   } properties[TGSI_PROPERTY_COUNT];
+   uint num_properties;
+};
 
 extern void
 tgsi_scan_shader(const struct tgsi_token *tokens,
index a96fc94c7ad99c4dfff043f86f4e53cc0e81fce0..d63c75dafb37aae87982a6ca48ccbc5805cdf84b 100644 (file)
@@ -58,7 +58,7 @@
    for (CHAN = 0; CHAN < NUM_CHANNELS; CHAN++)
 
 #define IS_DST0_CHANNEL_ENABLED( INST, CHAN )\
-   ((INST).FullDstRegisters[0].DstRegister.WriteMask & (1 << (CHAN)))
+   ((INST).Dst[0].Register.WriteMask & (1 << (CHAN)))
 
 #define IF_IS_DST0_CHANNEL_ENABLED( INST, CHAN )\
    if (IS_DST0_CHANNEL_ENABLED( INST, CHAN ))
@@ -1267,31 +1267,32 @@ emit_fetch(
    case TGSI_SWIZZLE_Y:
    case TGSI_SWIZZLE_Z:
    case TGSI_SWIZZLE_W:
-      switch (reg->SrcRegister.File) {
+      switch (reg->Register.File) {
       case TGSI_FILE_CONSTANT:
          emit_const(
             func,
             xmm,
-            reg->SrcRegister.Index,
+            reg->Register.Index,
             swizzle,
-            reg->SrcRegister.Indirect,
-            reg->SrcRegisterInd.File,
-            reg->SrcRegisterInd.Index );
+            reg->Register.Indirect,
+            reg->Indirect.File,
+            reg->Indirect.Index );
          break;
 
       case TGSI_FILE_IMMEDIATE:
          emit_immediate(
             func,
             xmm,
-            reg->SrcRegister.Index,
+            reg->Register.Index,
             swizzle );
          break;
 
       case TGSI_FILE_INPUT:
+      case TGSI_FILE_SYSTEM_VALUE:
          emit_inputf(
             func,
             xmm,
-            reg->SrcRegister.Index,
+            reg->Register.Index,
             swizzle );
          break;
 
@@ -1299,7 +1300,7 @@ emit_fetch(
          emit_tempf(
             func,
             xmm,
-            reg->SrcRegister.Index,
+            reg->Register.Index,
             swizzle );
          break;
 
@@ -1331,7 +1332,7 @@ emit_fetch(
 }
 
 #define FETCH( FUNC, INST, XMM, INDEX, CHAN )\
-   emit_fetch( FUNC, XMM, &(INST).FullSrcRegisters[INDEX], CHAN )
+   emit_fetch( FUNC, XMM, &(INST).Src[INDEX], CHAN )
 
 /**
  * Register store.
@@ -1371,12 +1372,12 @@ emit_store(
    }
 
 
-   switch( reg->DstRegister.File ) {
+   switch( reg->Register.File ) {
    case TGSI_FILE_OUTPUT:
       emit_output(
          func,
          xmm,
-         reg->DstRegister.Index,
+         reg->Register.Index,
          chan_index );
       break;
 
@@ -1384,7 +1385,7 @@ emit_store(
       emit_temps(
          func,
          xmm,
-         reg->DstRegister.Index,
+         reg->Register.Index,
          chan_index );
       break;
 
@@ -1392,7 +1393,7 @@ emit_store(
       emit_addrs(
          func,
          xmm,
-         reg->DstRegister.Index,
+         reg->Register.Index,
          chan_index );
       break;
 
@@ -1402,7 +1403,7 @@ emit_store(
 }
 
 #define STORE( FUNC, INST, XMM, INDEX, CHAN )\
-   emit_store( FUNC, XMM, &(INST).FullDstRegisters[INDEX], &(INST), CHAN )
+   emit_store( FUNC, XMM, &(INST).Dst[INDEX], &(INST), CHAN )
 
 
 static void PIPE_CDECL
@@ -1459,12 +1460,13 @@ emit_tex( struct x86_function *func,
           boolean lodbias,
           boolean projected)
 {
-   const uint unit = inst->FullSrcRegisters[1].SrcRegister.Index;
+   const uint unit = inst->Src[1].Register.Index;
    struct x86_reg args[2];
    unsigned count;
    unsigned i;
 
-   switch (inst->InstructionExtTexture.Texture) {
+   assert(inst->Instruction.Texture);
+   switch (inst->Texture.Texture) {
    case TGSI_TEXTURE_1D:
       count = 1;
       break;
@@ -1719,15 +1721,15 @@ indirect_temp_reference(const struct tgsi_full_instruction *inst)
 {
    uint i;
    for (i = 0; i < inst->Instruction.NumSrcRegs; i++) {
-      const struct tgsi_full_src_register *reg = &inst->FullSrcRegisters[i];
-      if (reg->SrcRegister.File == TGSI_FILE_TEMPORARY &&
-          reg->SrcRegister.Indirect)
+      const struct tgsi_full_src_register *reg = &inst->Src[i];
+      if (reg->Register.File == TGSI_FILE_TEMPORARY &&
+          reg->Register.Indirect)
          return TRUE;
    }
    for (i = 0; i < inst->Instruction.NumDstRegs; i++) {
-      const struct tgsi_full_dst_register *reg = &inst->FullDstRegisters[i];
-      if (reg->DstRegister.File == TGSI_FILE_TEMPORARY &&
-          reg->DstRegister.Indirect)
+      const struct tgsi_full_dst_register *reg = &inst->Dst[i];
+      if (reg->Register.File == TGSI_FILE_TEMPORARY &&
+          reg->Register.Indirect)
          return TRUE;
    }
    return FALSE;
@@ -2243,7 +2245,7 @@ emit_instruction(
 
    case TGSI_OPCODE_KIL:
       /* conditional kill */
-      emit_kil( func, &inst->FullSrcRegisters[0] );
+      emit_kil( func, &inst->Src[0] );
       break;
 
    case TGSI_OPCODE_PK2H:
@@ -2632,12 +2634,13 @@ emit_declaration(
    struct x86_function *func,
    struct tgsi_full_declaration *decl )
 {
-   if( decl->Declaration.File == TGSI_FILE_INPUT ) {
+   if( decl->Declaration.File == TGSI_FILE_INPUT ||
+       decl->Declaration.File == TGSI_FILE_SYSTEM_VALUE ) {
       unsigned first, last, mask;
       unsigned i, j;
 
-      first = decl->DeclarationRange.First;
-      last = decl->DeclarationRange.Last;
+      first = decl->Range.First;
+      last = decl->Range.Last;
       mask = decl->Declaration.UsageMask;
 
       for( i = first; i <= last; i++ ) {
@@ -2951,6 +2954,9 @@ tgsi_emit_sse2(
             num_immediates++;
          }
          break;
+      case TGSI_TOKEN_TYPE_PROPERTY:
+         /* we just ignore them for now */
+         break;
 
       default:
         ok = 0;
index d2b03ffb2fc72b0a9fcd0924419f5a25bffdf51f..2e3f9a90e3aed5a3fa0aaf0e522abf884c4ba142 100644 (file)
@@ -27,6 +27,9 @@
 
 #include "util/u_debug.h"
 #include "util/u_memory.h"
+#include "util/u_prim.h"
+#include "pipe/p_defines.h"
+#include "pipe/p_inlines.h"
 #include "tgsi_text.h"
 #include "tgsi_build.h"
 #include "tgsi_info.h"
@@ -59,6 +62,23 @@ static boolean uprcase( char c )
    return c;
 }
 
+/*
+ * Ignore case of str1 and assume str1 is already uppercase.
+ * Return TRUE iff str1 and str2 are equal.
+ */
+static int
+streq_nocase_uprcase(const char *str1,
+                     const char *str2)
+{
+   while (*str1 && *str2) {
+      if (*str1 != uprcase(*str2))
+         return FALSE;
+      str1++;
+      str2++;
+   }
+   return TRUE;
+}
+
 static boolean str_match_no_case( const char **pcur, const char *str )
 {
    const char *cur = *pcur;
@@ -110,6 +130,20 @@ static boolean parse_uint( const char **pcur, uint *val )
    return FALSE;
 }
 
+static boolean parse_identifier( const char **pcur, char *ret )
+{
+   const char *cur = *pcur;
+   int i = 0;
+   if (is_alpha_underscore( cur )) {
+      ret[i++] = *cur++;
+      while (is_alpha_underscore( cur ))
+         ret[i++] = *cur++;
+      *pcur = cur;
+      return TRUE;
+   }
+   return FALSE;
+}
+
 /* Parse floating point.
  */
 static boolean parse_float( const char **pcur, float *val )
@@ -163,38 +197,49 @@ struct translate_ctx
    struct tgsi_token *tokens_cur;
    struct tgsi_token *tokens_end;
    struct tgsi_header *header;
+   unsigned processor : 4;
+   int implied_array_size : 5;
 };
 
 static void report_error( struct translate_ctx *ctx, const char *msg )
 {
-   debug_printf( "\nError: %s", msg );
+   int line = 1;
+   int column = 1;
+   const char *itr = ctx->text;
+
+   while (itr != ctx->cur) {
+      if (*itr == '\n') {
+         column = 1;
+         ++line;
+      }
+      ++column;
+      ++itr;
+   }
+
+   debug_printf( "\nTGSI asm error: %s [%d : %d] \n", msg, line, column );
 }
 
 /* Parse shader header.
  * Return TRUE for one of the following headers.
- *    FRAG1.1
- *    GEOM1.1
- *    VERT1.1
+ *    FRAG
+ *    GEOM
+ *    VERT
  */
 static boolean parse_header( struct translate_ctx *ctx )
 {
    uint processor;
 
-   if (str_match_no_case( &ctx->cur, "FRAG1.1" ))
+   if (str_match_no_case( &ctx->cur, "FRAG" ))
       processor = TGSI_PROCESSOR_FRAGMENT;
-   else if (str_match_no_case( &ctx->cur, "VERT1.1" ))
+   else if (str_match_no_case( &ctx->cur, "VERT" ))
       processor = TGSI_PROCESSOR_VERTEX;
-   else if (str_match_no_case( &ctx->cur, "GEOM1.1" ))
+   else if (str_match_no_case( &ctx->cur, "GEOM" ))
       processor = TGSI_PROCESSOR_GEOMETRY;
    else {
       report_error( ctx, "Unknown header" );
       return FALSE;
    }
 
-   if (ctx->tokens_cur >= ctx->tokens_end)
-      return FALSE;
-   *(struct tgsi_version *) ctx->tokens_cur++ = tgsi_build_version();
-
    if (ctx->tokens_cur >= ctx->tokens_end)
       return FALSE;
    ctx->header = (struct tgsi_header *) ctx->tokens_cur++;
@@ -203,6 +248,7 @@ static boolean parse_header( struct translate_ctx *ctx )
    if (ctx->tokens_cur >= ctx->tokens_end)
       return FALSE;
    *(struct tgsi_processor *) ctx->tokens_cur++ = tgsi_build_processor( processor, ctx->header );
+   ctx->processor = processor;
 
    return TRUE;
 }
@@ -233,7 +279,8 @@ static const char *file_names[TGSI_FILE_COUNT] =
    "ADDR",
    "IMM",
    "LOOP",
-   "PRED"
+   "PRED",
+   "SV"
 };
 
 static boolean
@@ -298,92 +345,36 @@ parse_opt_writemask(
    return TRUE;
 }
 
-/* <register_file_bracket> ::= <file> `['
- */
 static boolean
-parse_register_file_bracket(
-   struct translate_ctx *ctx,
-   uint *file )
-{
-   if (!parse_file( &ctx->cur, file )) {
-      report_error( ctx, "Unknown register file" );
-      return FALSE;
-   }
-   eat_opt_white( &ctx->cur );
-   if (*ctx->cur != '[') {
-      report_error( ctx, "Expected `['" );
-      return FALSE;
-   }
-   ctx->cur++;
-   return TRUE;
-}
+parse_register_dst( struct translate_ctx *ctx,
+                    uint *file,
+                    int *index );
 
-/* <register_file_bracket_index> ::= <register_file_bracket> <uint>
- */
-static boolean
-parse_register_file_bracket_index(
-   struct translate_ctx *ctx,
-   uint *file,
-   int *index )
-{
-   uint uindex;
+struct parsed_src_bracket {
+   int index;
 
-   if (!parse_register_file_bracket( ctx, file ))
-      return FALSE;
-   eat_opt_white( &ctx->cur );
-   if (!parse_uint( &ctx->cur, &uindex )) {
-      report_error( ctx, "Expected literal unsigned integer" );
-      return FALSE;
-   }
-   *index = (int) uindex;
-   return TRUE;
-}
+   uint ind_file;
+   int ind_index;
+   uint ind_comp;
+};
 
-/* Parse destination register operand.
- *    <register_dst> ::= <register_file_bracket_index> `]'
- */
-static boolean
-parse_register_dst(
-   struct translate_ctx *ctx,
-   uint *file,
-   int *index )
-{
-   if (!parse_register_file_bracket_index( ctx, file, index ))
-      return FALSE;
-   eat_opt_white( &ctx->cur );
-   if (*ctx->cur != ']') {
-      report_error( ctx, "Expected `]'" );
-      return FALSE;
-   }
-   ctx->cur++;
-   return TRUE;
-}
 
-/* Parse source register operand.
- *    <register_src> ::= <register_file_bracket_index> `]' |
- *                       <register_file_bracket> <register_dst> [`.' (`x' | `y' | `z' | `w')] `]' |
- *                       <register_file_bracket> <register_dst> [`.' (`x' | `y' | `z' | `w')] `+' <uint> `]' |
- *                       <register_file_bracket> <register_dst> [`.' (`x' | `y' | `z' | `w')] `-' <uint> `]'
- */
 static boolean
-parse_register_src(
+parse_register_src_bracket(
    struct translate_ctx *ctx,
-   uint *file,
-   int *index,
-   uint *ind_file,
-   int *ind_index,
-   uint *ind_comp)
+   struct parsed_src_bracket *brackets)
 {
    const char *cur;
    uint uindex;
 
-   *ind_comp = TGSI_SWIZZLE_X;
-   if (!parse_register_file_bracket( ctx, file ))
-      return FALSE;
+   memset(brackets, 0, sizeof(struct parsed_src_bracket));
+
    eat_opt_white( &ctx->cur );
+
    cur = ctx->cur;
-   if (parse_file( &cur, ind_file )) {
-      if (!parse_register_dst( ctx, ind_file, ind_index ))
+   if (parse_file( &cur, &brackets->ind_file )) {
+      if (!parse_register_dst( ctx, &brackets->ind_file,
+                               &brackets->ind_index ))
          return FALSE;
       eat_opt_white( &ctx->cur );
 
@@ -393,16 +384,16 @@ parse_register_src(
 
          switch (uprcase(*ctx->cur)) {
          case 'X':
-            *ind_comp = TGSI_SWIZZLE_X;
+            brackets->ind_comp = TGSI_SWIZZLE_X;
             break;
          case 'Y':
-            *ind_comp = TGSI_SWIZZLE_Y;
+            brackets->ind_comp = TGSI_SWIZZLE_Y;
             break;
          case 'Z':
-            *ind_comp = TGSI_SWIZZLE_Z;
+            brackets->ind_comp = TGSI_SWIZZLE_Z;
             break;
          case 'W':
-            *ind_comp = TGSI_SWIZZLE_W;
+            brackets->ind_comp = TGSI_SWIZZLE_W;
             break;
          default:
             report_error(ctx, "Expected indirect register swizzle component `x', `y', `z' or `w'");
@@ -423,12 +414,12 @@ parse_register_src(
             return FALSE;
          }
          if (negate)
-            *index = -(int) uindex;
+            brackets->index = -(int) uindex;
          else
-            *index = (int) uindex;
+            brackets->index = (int) uindex;
       }
       else {
-         *index = 0;
+         brackets->index = 0;
       }
    }
    else {
@@ -436,9 +427,9 @@ parse_register_src(
          report_error( ctx, "Expected literal unsigned integer" );
          return FALSE;
       }
-      *index = (int) uindex;
-      *ind_file = TGSI_FILE_NULL;
-      *ind_index = 0;
+      brackets->index = (int) uindex;
+      brackets->ind_file = TGSI_FILE_NULL;
+      brackets->ind_index = 0;
    }
    eat_opt_white( &ctx->cur );
    if (*ctx->cur != ']') {
@@ -449,20 +440,123 @@ parse_register_src(
    return TRUE;
 }
 
-/* Parse register declaration.
- *    <register_dcl> ::= <register_file_bracket_index> `]' |
- *                       <register_file_bracket_index> `..' <index> `]'
+static boolean
+parse_opt_register_src_bracket(
+   struct translate_ctx *ctx,
+   struct parsed_src_bracket *brackets,
+   int *parsed_brackets)
+{
+   const char *cur = ctx->cur;
+
+   *parsed_brackets = 0;
+
+   eat_opt_white( &cur );
+   if (cur[0] == '[') {
+      ++cur;
+      ctx->cur = cur;
+
+      if (!parse_register_src_bracket(ctx, brackets))
+         return FALSE;
+
+      *parsed_brackets = 1;
+   }
+
+   return TRUE;
+}
+
+/* <register_file_bracket> ::= <file> `['
  */
 static boolean
-parse_register_dcl(
+parse_register_file_bracket(
+   struct translate_ctx *ctx,
+   uint *file )
+{
+   if (!parse_file( &ctx->cur, file )) {
+      report_error( ctx, "Unknown register file" );
+      return FALSE;
+   }
+   eat_opt_white( &ctx->cur );
+   if (*ctx->cur != '[') {
+      report_error( ctx, "Expected `['" );
+      return FALSE;
+   }
+   ctx->cur++;
+   return TRUE;
+}
+
+/* <register_file_bracket_index> ::= <register_file_bracket> <uint>
+ */
+static boolean
+parse_register_file_bracket_index(
+   struct translate_ctx *ctx,
+   uint *file,
+   int *index )
+{
+   uint uindex;
+
+   if (!parse_register_file_bracket( ctx, file ))
+      return FALSE;
+   eat_opt_white( &ctx->cur );
+   if (!parse_uint( &ctx->cur, &uindex )) {
+      report_error( ctx, "Expected literal unsigned integer" );
+      return FALSE;
+   }
+   *index = (int) uindex;
+   return TRUE;
+}
+
+/* Parse source register operand.
+ *    <register_src> ::= <register_file_bracket_index> `]' |
+ *                       <register_file_bracket> <register_dst> [`.' (`x' | `y' | `z' | `w')] `]' |
+ *                       <register_file_bracket> <register_dst> [`.' (`x' | `y' | `z' | `w')] `+' <uint> `]' |
+ *                       <register_file_bracket> <register_dst> [`.' (`x' | `y' | `z' | `w')] `-' <uint> `]'
+ */
+static boolean
+parse_register_src(
    struct translate_ctx *ctx,
    uint *file,
-   int *first,
-   int *last )
+   struct parsed_src_bracket *brackets)
 {
-   if (!parse_register_file_bracket_index( ctx, file, first ))
+
+   brackets->ind_comp = TGSI_SWIZZLE_X;
+   if (!parse_register_file_bracket( ctx, file ))
       return FALSE;
+   if (!parse_register_src_bracket( ctx, brackets ))
+       return FALSE;
+
+   return TRUE;
+}
+
+struct parsed_dcl_bracket {
+   uint first;
+   uint last;
+};
+
+static boolean
+parse_register_dcl_bracket(
+   struct translate_ctx *ctx,
+   struct parsed_dcl_bracket *bracket)
+{
+   uint uindex;
+   memset(bracket, 0, sizeof(struct parsed_dcl_bracket));
+
    eat_opt_white( &ctx->cur );
+
+   if (!parse_uint( &ctx->cur, &uindex )) {
+      /* it can be an empty bracket [] which means its range
+       * is from 0 to some implied size */
+      if (ctx->cur[0] == ']' && ctx->implied_array_size != 0) {
+         bracket->first = 0;
+         bracket->last = ctx->implied_array_size - 1;
+         goto cleanup;
+      }
+      report_error( ctx, "Expected literal unsigned integer" );
+      return FALSE;
+   }
+   bracket->first = (int) uindex;
+
+   eat_opt_white( &ctx->cur );
+
    if (ctx->cur[0] == '.' && ctx->cur[1] == '.') {
       uint uindex;
 
@@ -472,12 +566,14 @@ parse_register_dcl(
          report_error( ctx, "Expected literal integer" );
          return FALSE;
       }
-      *last = (int) uindex;
+      bracket->last = (int) uindex;
       eat_opt_white( &ctx->cur );
    }
    else {
-      *last = *first;
+      bracket->last = bracket->first;
    }
+
+cleanup:
    if (*ctx->cur != ']') {
       report_error( ctx, "Expected `]' or `..'" );
       return FALSE;
@@ -486,16 +582,70 @@ parse_register_dcl(
    return TRUE;
 }
 
-static const char *modulate_names[TGSI_MODULATE_COUNT] =
+/* Parse register declaration.
+ *    <register_dcl> ::= <register_file_bracket_index> `]' |
+ *                       <register_file_bracket_index> `..' <index> `]'
+ */
+static boolean
+parse_register_dcl(
+   struct translate_ctx *ctx,
+   uint *file,
+   struct parsed_dcl_bracket *brackets,
+   int *num_brackets)
 {
-   "_1X",
-   "_2X",
-   "_4X",
-   "_8X",
-   "_D2",
-   "_D4",
-   "_D8"
-};
+   const char *cur;
+
+   *num_brackets = 0;
+
+   if (!parse_register_file_bracket( ctx, file ))
+      return FALSE;
+   if (!parse_register_dcl_bracket( ctx, &brackets[0] ))
+      return FALSE;
+
+   *num_brackets = 1;
+
+   cur = ctx->cur;
+   eat_opt_white( &cur );
+
+   if (cur[0] == '[') {
+      ++cur;
+      ctx->cur = cur;
+      if (!parse_register_dcl_bracket( ctx, &brackets[1] ))
+         return FALSE;
+      /* for geometry shader we don't really care about
+       * the first brackets it's always the size of the
+       * input primitive. so we want to declare just
+       * the index relevant to the semantics which is in
+       * the second bracket */
+      if (ctx->processor == TGSI_PROCESSOR_GEOMETRY) {
+         brackets[0] = brackets[1];
+      }
+      *num_brackets = 2;
+   }
+
+   return TRUE;
+}
+
+
+/* Parse destination register operand.
+ *    <register_dst> ::= <register_file_bracket_index> `]'
+ */
+static boolean
+parse_register_dst(
+   struct translate_ctx *ctx,
+   uint *file,
+   int *index )
+{
+   if (!parse_register_file_bracket_index( ctx, file, index ))
+      return FALSE;
+   eat_opt_white( &ctx->cur );
+   if (*ctx->cur != ']') {
+      report_error( ctx, "Expected `]'" );
+      return FALSE;
+   }
+   ctx->cur++;
+   return TRUE;
+}
 
 static boolean
 parse_dst_operand(
@@ -512,26 +662,13 @@ parse_dst_operand(
 
    cur = ctx->cur;
    eat_opt_white( &cur );
-   if (*cur == '_') {
-      uint i;
-
-      for (i = 0; i < TGSI_MODULATE_COUNT; i++) {
-         if (str_match_no_case( &cur, modulate_names[i] )) {
-            if (!is_digit_alpha_underscore( cur )) {
-               dst->DstRegisterExtModulate.Modulate = i;
-               ctx->cur = cur;
-               break;
-            }
-         }
-      }
-   }
 
    if (!parse_opt_writemask( ctx, &writemask ))
       return FALSE;
 
-   dst->DstRegister.File = file;
-   dst->DstRegister.Index = index;
-   dst->DstRegister.WriteMask = writemask;
+   dst->Register.File = file;
+   dst->Register.Index = index;
+   dst->Register.WriteMask = writemask;
    return TRUE;
 }
 
@@ -577,162 +714,59 @@ parse_src_operand(
    struct translate_ctx *ctx,
    struct tgsi_full_src_register *src )
 {
-   const char *cur;
-   float value;
    uint file;
-   int index;
-   uint ind_file;
-   int ind_index;
-   uint ind_comp;
    uint swizzle[4];
-   boolean parsed_ext_negate_paren = FALSE;
    boolean parsed_swizzle;
+   struct parsed_src_bracket bracket[2];
+   int parsed_opt_brackets;
 
    if (*ctx->cur == '-') {
-      cur = ctx->cur;
-      cur++;
-      eat_opt_white( &cur );
-      if (*cur == '(') {
-         cur++;
-         src->SrcRegisterExtMod.Negate = 1;
-         eat_opt_white( &cur );
-         ctx->cur = cur;
-         parsed_ext_negate_paren = TRUE;
-      }
-      else if (*cur == '|') {
-         cur++;
-         src->SrcRegisterExtMod.Negate = 1;
-         src->SrcRegisterExtMod.Absolute = 1;
-         eat_opt_white(&cur);
-         ctx->cur = cur;
-      }
-   }
-   else if (*ctx->cur == '|') {
       ctx->cur++;
       eat_opt_white( &ctx->cur );
-      src->SrcRegisterExtMod.Absolute = 1;
+      src->Register.Negate = 1;
    }
 
-   if (*ctx->cur == '-') {
+   if (*ctx->cur == '|') {
       ctx->cur++;
       eat_opt_white( &ctx->cur );
-      src->SrcRegister.Negate = 1;
+      src->Register.Absolute = 1;
    }
 
-   cur = ctx->cur;
-   if (parse_float( &cur, &value )) {
-      if (value == 2.0f) {
-         eat_opt_white( &cur );
-         if (*cur != '*') {
-            report_error( ctx, "Expected `*'" );
-            return FALSE;
-         }
-         cur++;
-         if (*cur != '(') {
-            report_error( ctx, "Expected `('" );
-            return FALSE;
-         }
-         cur++;
-         src->SrcRegisterExtMod.Scale2X = 1;
-         eat_opt_white( &cur );
-         ctx->cur = cur;
-      }
-   }
+   if (!parse_register_src(ctx, &file, &bracket[0]))
+      return FALSE;
+   if (!parse_opt_register_src_bracket(ctx, &bracket[1], &parsed_opt_brackets))
+      return FALSE;
 
-   if (*ctx->cur == '(') {
-      ctx->cur++;
-      eat_opt_white( &ctx->cur );
-      src->SrcRegisterExtMod.Bias = 1;
+   src->Register.File = file;
+   src->Register.Index = bracket[0].index;
+   if (bracket[0].ind_file != TGSI_FILE_NULL) {
+      src->Register.Indirect = 1;
+      src->Indirect.File = bracket[0].ind_file;
+      src->Indirect.Index = bracket[0].ind_index;
+      src->Indirect.SwizzleX = bracket[0].ind_comp;
+      src->Indirect.SwizzleY = bracket[0].ind_comp;
+      src->Indirect.SwizzleZ = bracket[0].ind_comp;
+      src->Indirect.SwizzleW = bracket[0].ind_comp;
    }
-
-   cur = ctx->cur;
-   if (parse_float( &cur, &value )) {
-      if (value == 1.0f) {
-         eat_opt_white( &cur );
-         if (*cur != '-') {
-            report_error( ctx, "Expected `-'" );
-            return FALSE;
-         }
-         cur++;
-         if (*cur != '(') {
-            report_error( ctx, "Expected `('" );
-            return FALSE;
-         }
-         cur++;
-         src->SrcRegisterExtMod.Complement = 1;
-         eat_opt_white( &cur );
-         ctx->cur = cur;
-      }
-   }
-
-   if (!parse_register_src(ctx, &file, &index, &ind_file, &ind_index, &ind_comp))
-      return FALSE;
-   src->SrcRegister.File = file;
-   src->SrcRegister.Index = index;
-   if (ind_file != TGSI_FILE_NULL) {
-      src->SrcRegister.Indirect = 1;
-      src->SrcRegisterInd.File = ind_file;
-      src->SrcRegisterInd.Index = ind_index;
-      src->SrcRegisterInd.SwizzleX = ind_comp;
-      src->SrcRegisterInd.SwizzleY = ind_comp;
-      src->SrcRegisterInd.SwizzleZ = ind_comp;
-      src->SrcRegisterInd.SwizzleW = ind_comp;
+   if (parsed_opt_brackets) {
+      src->Register.Dimension = 1;
+      src->Dimension.Indirect = 0;
+      src->Dimension.Dimension = 0;
+      src->Dimension.Index = bracket[1].index;
    }
 
    /* Parse optional swizzle.
     */
    if (parse_optional_swizzle( ctx, swizzle, &parsed_swizzle )) {
       if (parsed_swizzle) {
-         src->SrcRegister.SwizzleX = swizzle[0];
-         src->SrcRegister.SwizzleY = swizzle[1];
-         src->SrcRegister.SwizzleZ = swizzle[2];
-         src->SrcRegister.SwizzleW = swizzle[3];
-      }
-   }
-
-   if (src->SrcRegisterExtMod.Complement) {
-      eat_opt_white( &ctx->cur );
-      if (*ctx->cur != ')') {
-         report_error( ctx, "Expected `)'" );
-         return FALSE;
-      }
-      ctx->cur++;
-   }
-
-   if (src->SrcRegisterExtMod.Bias) {
-      eat_opt_white( &ctx->cur );
-      if (*ctx->cur != ')') {
-         report_error( ctx, "Expected `)'" );
-         return FALSE;
-      }
-      ctx->cur++;
-      eat_opt_white( &ctx->cur );
-      if (*ctx->cur != '-') {
-         report_error( ctx, "Expected `-'" );
-         return FALSE;
-      }
-      ctx->cur++;
-      eat_opt_white( &ctx->cur );
-      if (!parse_float( &ctx->cur, &value )) {
-         report_error( ctx, "Expected literal floating point" );
-         return FALSE;
-      }
-      if (value != 0.5f) {
-         report_error( ctx, "Expected 0.5" );
-         return FALSE;
+         src->Register.SwizzleX = swizzle[0];
+         src->Register.SwizzleY = swizzle[1];
+         src->Register.SwizzleZ = swizzle[2];
+         src->Register.SwizzleW = swizzle[3];
       }
    }
 
-   if (src->SrcRegisterExtMod.Scale2X) {
-      eat_opt_white( &ctx->cur );
-      if (*ctx->cur != ')') {
-         report_error( ctx, "Expected `)'" );
-         return FALSE;
-      }
-      ctx->cur++;
-   }
-
-   if (src->SrcRegisterExtMod.Absolute) {
+   if (src->Register.Absolute) {
       eat_opt_white( &ctx->cur );
       if (*ctx->cur != '|') {
          report_error( ctx, "Expected `|'" );
@@ -741,14 +775,6 @@ parse_src_operand(
       ctx->cur++;
    }
 
-   if (parsed_ext_negate_paren) {
-      eat_opt_white( &ctx->cur );
-      if (*ctx->cur != ')') {
-         report_error( ctx, "Expected `)'" );
-         return FALSE;
-      }
-      ctx->cur++;
-   }
 
    return TRUE;
 }
@@ -840,11 +866,11 @@ parse_instruction(
       }
 
       if (i < info->num_dst) {
-         if (!parse_dst_operand( ctx, &inst.FullDstRegisters[i] ))
+         if (!parse_dst_operand( ctx, &inst.Dst[i] ))
             return FALSE;
       }
       else if (i < info->num_dst + info->num_src) {
-         if (!parse_src_operand( ctx, &inst.FullSrcRegisters[i - info->num_dst] ))
+         if (!parse_src_operand( ctx, &inst.Src[i - info->num_dst] ))
             return FALSE;
       }
       else {
@@ -853,7 +879,8 @@ parse_instruction(
          for (j = 0; j < TGSI_TEXTURE_COUNT; j++) {
             if (str_match_no_case( &ctx->cur, texture_names[j] )) {
                if (!is_digit_alpha_underscore( ctx->cur )) {
-                  inst.InstructionExtTexture.Texture = j;
+                  inst.Instruction.Texture = 1;
+                  inst.Texture.Texture = j;
                   break;
                }
             }
@@ -879,7 +906,8 @@ parse_instruction(
          report_error( ctx, "Expected a label" );
          return FALSE;
       }
-      inst.InstructionExtLabel.Label = target;
+      inst.Instruction.Label = 1;
+      inst.Label.Label = target;
    }
 
    advance = tgsi_build_full_instruction(
@@ -903,7 +931,8 @@ static const char *semantic_names[TGSI_SEMANTIC_COUNT] =
    "PSIZE",
    "GENERIC",
    "NORMAL",
-   "FACE"
+   "FACE",
+   "PRIM_ID"
 };
 
 static const char *interpolate_names[TGSI_INTERPOLATE_COUNT] =
@@ -917,8 +946,8 @@ static boolean parse_declaration( struct translate_ctx *ctx )
 {
    struct tgsi_full_declaration decl;
    uint file;
-   int first;
-   int last;
+   struct parsed_dcl_bracket brackets[2];
+   int num_brackets;
    uint writemask;
    const char *cur;
    uint advance;
@@ -930,7 +959,7 @@ static boolean parse_declaration( struct translate_ctx *ctx )
       report_error( ctx, "Syntax error" );
       return FALSE;
    }
-   if (!parse_register_dcl( ctx, &file, &first, &last ))
+   if (!parse_register_dcl( ctx, &file, brackets, &num_brackets))
       return FALSE;
    if (!parse_opt_writemask( ctx, &writemask ))
       return FALSE;
@@ -938,8 +967,8 @@ static boolean parse_declaration( struct translate_ctx *ctx )
    decl = tgsi_default_full_declaration();
    decl.Declaration.File = file;
    decl.Declaration.UsageMask = writemask;
-   decl.DeclarationRange.First = first;
-   decl.DeclarationRange.Last = last;
+   decl.Range.First = brackets[0].first;
+   decl.Range.Last = brackets[0].last;
 
    cur = ctx->cur;
    eat_opt_white( &cur );
@@ -970,13 +999,13 @@ static boolean parse_declaration( struct translate_ctx *ctx )
                }
                cur2++;
 
-               decl.Semantic.SemanticIndex = index;
+               decl.Semantic.Index = index;
 
                cur = cur2;
             }
 
             decl.Declaration.Semantic = 1;
-            decl.Semantic.SemanticName = i;
+            decl.Semantic.Name = i;
 
             ctx->cur = cur;
             break;
@@ -1082,6 +1111,110 @@ static boolean parse_immediate( struct translate_ctx *ctx )
    return TRUE;
 }
 
+static const char *property_names[] =
+{
+   "GS_INPUT_PRIMITIVE",
+   "GS_OUTPUT_PRIMITIVE",
+   "GS_MAX_OUTPUT_VERTICES"
+};
+
+static const char *primitive_names[] =
+{
+   "POINTS",
+   "LINES",
+   "LINE_LOOP",
+   "LINE_STRIP",
+   "TRIANGLES",
+   "TRIANGLE_STRIP",
+   "TRIANGLE_FAN",
+   "QUADS",
+   "QUAD_STRIP",
+   "POLYGON"
+};
+
+static boolean
+parse_primitive( const char **pcur, uint *primitive )
+{
+   uint i;
+
+   for (i = 0; i < PIPE_PRIM_MAX; i++) {
+      const char *cur = *pcur;
+
+      if (str_match_no_case( &cur, primitive_names[i])) {
+         *primitive = i;
+         *pcur = cur;
+         return TRUE;
+      }
+   }
+   return FALSE;
+}
+
+
+static boolean parse_property( struct translate_ctx *ctx )
+{
+   struct tgsi_full_property prop;
+   uint property_name;
+   uint values[8];
+   uint advance;
+   char id[64];
+
+   if (!eat_white( &ctx->cur )) {
+      report_error( ctx, "Syntax error" );
+      return FALSE;
+   }
+   if (!parse_identifier( &ctx->cur, id )) {
+      report_error( ctx, "Syntax error" );
+      return FALSE;
+   }
+   for (property_name = 0; property_name < TGSI_PROPERTY_COUNT;
+        ++property_name) {
+      if (streq_nocase_uprcase(property_names[property_name], id)) {
+         break;
+      }
+   }
+   if (property_name >= TGSI_PROPERTY_COUNT) {
+      debug_printf( "\nError: Unknown property : '%s'", id );
+      return FALSE;
+   }
+
+   eat_opt_white( &ctx->cur );
+   switch(property_name) {
+   case TGSI_PROPERTY_GS_INPUT_PRIM:
+   case TGSI_PROPERTY_GS_OUTPUT_PRIM:
+      if (!parse_primitive(&ctx->cur, &values[0] )) {
+         report_error( ctx, "Unknown primitive name as property!" );
+         return FALSE;
+      }
+      if (property_name == TGSI_PROPERTY_GS_INPUT_PRIM &&
+          ctx->processor == TGSI_PROCESSOR_GEOMETRY) {
+         ctx->implied_array_size = u_vertices_per_prim(values[0]);
+      }
+      break;
+   default:
+      if (!parse_uint(&ctx->cur, &values[0] )) {
+         report_error( ctx, "Expected unsigned integer as property!" );
+         return FALSE;
+      }
+   }
+
+   prop = tgsi_default_full_property();
+   prop.Property.PropertyName = property_name;
+   prop.Property.NrTokens += 1;
+   prop.u[0].Data = values[0];
+
+   advance = tgsi_build_full_property(
+      &prop,
+      ctx->tokens_cur,
+      ctx->header,
+      (uint) (ctx->tokens_end - ctx->tokens_cur) );
+   if (advance == 0)
+      return FALSE;
+   ctx->tokens_cur += advance;
+
+   return TRUE;
+}
+
+
 static boolean translate( struct translate_ctx *ctx )
 {
    eat_opt_white( &ctx->cur );
@@ -1090,7 +1223,6 @@ static boolean translate( struct translate_ctx *ctx )
 
    while (*ctx->cur != '\0') {
       uint label_val = 0;
-
       if (!eat_white( &ctx->cur )) {
          report_error( ctx, "Syntax error" );
          return FALSE;
@@ -1098,7 +1230,6 @@ static boolean translate( struct translate_ctx *ctx )
 
       if (*ctx->cur == '\0')
          break;
-
       if (parse_label( ctx, &label_val )) {
          if (!parse_instruction( ctx, TRUE ))
             return FALSE;
@@ -1111,6 +1242,10 @@ static boolean translate( struct translate_ctx *ctx )
          if (!parse_immediate( ctx ))
             return FALSE;
       }
+      else if (str_match_no_case( &ctx->cur, "PROPERTY" )) {
+         if (!parse_property( ctx ))
+            return FALSE;
+      }
       else if (!parse_instruction( ctx, FALSE )) {
          return FALSE;
       }
index bc9c18fd4a725df628b7b64f57a60c405cca832b..ae875f29abfb23334a42fa5c55d42d6461909455 100644 (file)
@@ -79,6 +79,19 @@ emit_immediate(struct tgsi_transform_context *ctx,
 }
 
 
+static void
+emit_property(struct tgsi_transform_context *ctx,
+              const struct tgsi_full_property *prop)
+{
+   uint ti = ctx->ti;
+
+   ti += tgsi_build_full_property(prop,
+                                  ctx->tokens_out + ti,
+                                  ctx->header,
+                                  ctx->max_tokens_out - ti);
+   ctx->ti = ti;
+}
+
 
 /**
  * Apply user-defined transformations to the input shader to produce
@@ -110,6 +123,7 @@ tgsi_transform_shader(const struct tgsi_token *tokens_in,
    ctx->emit_instruction = emit_instruction;
    ctx->emit_declaration = emit_declaration;
    ctx->emit_immediate = emit_immediate;
+   ctx->emit_property = emit_property;
    ctx->tokens_out = tokens_out;
    ctx->max_tokens_out = max_tokens_out;
 
@@ -130,15 +144,13 @@ tgsi_transform_shader(const struct tgsi_token *tokens_in,
    /**
     **  Setup output shader
     **/
-   *(struct tgsi_version *) &tokens_out[0] = tgsi_build_version();
-
-   ctx->header = (struct tgsi_header *) (tokens_out + 1);
+   ctx->header = (struct tgsi_header *)tokens_out;
    *ctx->header = tgsi_build_header();
 
-   processor = (struct tgsi_processor *) (tokens_out + 2);
+   processor = (struct tgsi_processor *) (tokens_out + 1);
    *processor = tgsi_build_processor( procType, ctx->header );
 
-   ctx->ti = 3;
+   ctx->ti = 2;
 
 
    /**
@@ -184,6 +196,17 @@ tgsi_transform_shader(const struct tgsi_token *tokens_in,
                ctx->emit_immediate(ctx, fullimm);
          }
          break;
+      case TGSI_TOKEN_TYPE_PROPERTY:
+         {
+            struct tgsi_full_property *fullprop
+               = &parse.FullToken.FullProperty;
+
+            if (ctx->transform_property)
+               ctx->transform_property(ctx, fullprop);
+            else
+               ctx->emit_property(ctx, fullprop);
+         }
+         break;
 
       default:
          assert( 0 );
@@ -215,7 +238,7 @@ tgsi_transform_foo( struct tgsi_token *tokens_out,
                     uint max_tokens_out )
 {
    const char *text = 
-      "FRAG1.1\n"
+      "FRAG\n"
       "DCL IN[0], COLOR, CONSTANT\n"
       "DCL OUT[0], COLOR\n"
       "  0: MOV OUT[0], IN[0]\n"
index a121adbaef4056fa296fc0f7591dadb72479a176..818478e277a91e8ed55105453578987550965e7f 100644 (file)
@@ -53,6 +53,8 @@ struct tgsi_transform_context
 
    void (*transform_immediate)(struct tgsi_transform_context *ctx,
                                struct tgsi_full_immediate *imm);
+   void (*transform_property)(struct tgsi_transform_context *ctx,
+                              struct tgsi_full_property *prop);
 
    /**
     * Called at end of input program to allow caller to append extra
@@ -73,6 +75,8 @@ struct tgsi_transform_context
                             const struct tgsi_full_declaration *decl);
    void (*emit_immediate)(struct tgsi_transform_context *ctx,
                           const struct tgsi_full_immediate *imm);
+   void (*emit_property)(struct tgsi_transform_context *ctx,
+                         const struct tgsi_full_property *prop);
 
    struct tgsi_header *header;
    uint max_tokens_out;
index 3f752e9352f6a0ab99472d10bdfd370ba82cea52..1e730e534276be4d20d50a48063a3ef0a9f0f029 100644 (file)
@@ -37,7 +37,6 @@
 #include "util/u_math.h"
 
 union tgsi_any_token {
-   struct tgsi_version version;
    struct tgsi_header header;
    struct tgsi_processor processor;
    struct tgsi_token token;
@@ -47,14 +46,12 @@ union tgsi_any_token {
    struct tgsi_immediate imm;
    union  tgsi_immediate_data imm_data;
    struct tgsi_instruction insn;
-   struct tgsi_instruction_ext_label insn_ext_label;
-   struct tgsi_instruction_ext_texture insn_ext_texture;
-   struct tgsi_instruction_ext_predicate insn_ext_predicate;
+   struct tgsi_instruction_predicate insn_predicate;
+   struct tgsi_instruction_label insn_label;
+   struct tgsi_instruction_texture insn_texture;
    struct tgsi_src_register src;
-   struct tgsi_src_register_ext_mod src_ext_mod;
    struct tgsi_dimension dim;
    struct tgsi_dst_register dst;
-   struct tgsi_dst_register_ext_modulate dst_ext_mod;
    unsigned value;
 };
 
@@ -72,6 +69,7 @@ struct ureg_tokens {
 #define UREG_MAX_IMMEDIATE 32
 #define UREG_MAX_TEMP 256
 #define UREG_MAX_ADDR 2
+#define UREG_MAX_LOOP 1
 #define UREG_MAX_PRED 1
 
 #define DOMAIN_DECL 0
@@ -91,6 +89,11 @@ struct ureg_program
 
    unsigned vs_inputs[UREG_MAX_INPUT/32];
 
+   struct {
+      unsigned index;
+   } gs_input[UREG_MAX_INPUT];
+   unsigned nr_gs_inputs;
+
    struct {
       unsigned semantic_name;
       unsigned semantic_index;
@@ -117,6 +120,7 @@ struct ureg_program
 
    unsigned nr_addrs;
    unsigned nr_preds;
+   unsigned nr_loops;
    unsigned nr_instructions;
 
    struct ureg_tokens domain[2];
@@ -279,6 +283,22 @@ ureg_DECL_vs_input( struct ureg_program *ureg,
 }
 
 
+struct ureg_src
+ureg_DECL_gs_input(struct ureg_program *ureg,
+                   unsigned index)
+{
+   if (ureg->nr_gs_inputs < UREG_MAX_INPUT) {
+      ureg->gs_input[ureg->nr_gs_inputs].index = index;
+      ureg->nr_gs_inputs++;
+   } else {
+      set_bad(ureg);
+   }
+
+   /* XXX: Add suport for true 2D input registers. */
+   return ureg_src_register(TGSI_FILE_INPUT, index);
+}
+
+
 struct ureg_dst 
 ureg_DECL_output( struct ureg_program *ureg,
                   unsigned name,
@@ -350,6 +370,7 @@ struct ureg_src ureg_DECL_constant(struct ureg_program *ureg,
       i = ureg->nr_constant_ranges++;
       ureg->constant_range[i].first = index;
       ureg->constant_range[i].last = index;
+      goto out;
    }
 
    /* Collapse all ranges down to one:
@@ -417,6 +438,19 @@ struct ureg_dst ureg_DECL_address( struct ureg_program *ureg )
    return ureg_dst_register( TGSI_FILE_ADDRESS, 0 );
 }
 
+/* Allocate a new loop register.
+ */
+struct ureg_dst
+ureg_DECL_loop(struct ureg_program *ureg)
+{
+   if (ureg->nr_loops < UREG_MAX_LOOP) {
+      return ureg_dst_register(TGSI_FILE_LOOP, ureg->nr_loops++);
+   }
+
+   assert(0);
+   return ureg_dst_register(TGSI_FILE_LOOP, 0);
+}
+
 /* Allocate a new predicate register.
  */
 struct ureg_dst
@@ -541,9 +575,7 @@ void
 ureg_emit_src( struct ureg_program *ureg,
                struct ureg_src src )
 {
-   unsigned size = (1 + 
-                    (src.Absolute ? 1 : 0) +
-                    (src.Indirect ? 1 : 0));
+   unsigned size = 1 + (src.Indirect ? 1 : 0);
 
    union tgsi_any_token *out = get_tokens( ureg, DOMAIN_INSN, size );
    unsigned n = 0;
@@ -560,17 +592,8 @@ ureg_emit_src( struct ureg_program *ureg,
    out[n].src.SwizzleW = src.SwizzleW;
    out[n].src.Index = src.Index;
    out[n].src.Negate = src.Negate;
+   out[0].src.Absolute = src.Absolute;
    n++;
-   
-   if (src.Absolute) {
-      out[0].src.Extended = 1;
-      out[0].src.Negate = 0;
-      out[n].value = 0;
-      out[n].src_ext_mod.Type = TGSI_SRC_REGISTER_EXT_TYPE_MOD;
-      out[n].src_ext_mod.Absolute = 1;
-      out[n].src_ext_mod.Negate = src.Negate;
-      n++;
-   }
 
    if (src.Indirect) {
       out[0].src.Indirect = 1;
@@ -661,35 +684,27 @@ ureg_emit_insn(struct ureg_program *ureg,
    validate( opcode, num_dst, num_src );
    
    out = get_tokens( ureg, DOMAIN_INSN, count );
-   out[0].value = 0;
-   out[0].insn.Type = TGSI_TOKEN_TYPE_INSTRUCTION;
-   out[0].insn.NrTokens = 0;
+   out[0].insn = tgsi_default_instruction();
    out[0].insn.Opcode = opcode;
    out[0].insn.Saturate = saturate;
    out[0].insn.NumDstRegs = num_dst;
    out[0].insn.NumSrcRegs = num_src;
-   out[0].insn.Padding = 0;
 
    result.insn_token = ureg->domain[DOMAIN_INSN].count - count;
+   result.extended_token = result.insn_token;
 
    if (predicate) {
-      out[0].insn.Extended = 1;
-      out[1].insn_ext_predicate = tgsi_default_instruction_ext_predicate();
-      out[1].insn_ext_predicate.Negate = pred_negate;
-      out[1].insn_ext_predicate.SwizzleX = pred_swizzle_x;
-      out[1].insn_ext_predicate.SwizzleY = pred_swizzle_y;
-      out[1].insn_ext_predicate.SwizzleZ = pred_swizzle_z;
-      out[1].insn_ext_predicate.SwizzleW = pred_swizzle_w;
-
-      result.extended_token = result.insn_token + 1;
-   } else {
-      out[0].insn.Extended = 0;
-
-      result.extended_token = result.insn_token;
+      out[0].insn.Predicate = 1;
+      out[1].insn_predicate = tgsi_default_instruction_predicate();
+      out[1].insn_predicate.Negate = pred_negate;
+      out[1].insn_predicate.SwizzleX = pred_swizzle_x;
+      out[1].insn_predicate.SwizzleY = pred_swizzle_y;
+      out[1].insn_predicate.SwizzleZ = pred_swizzle_z;
+      out[1].insn_predicate.SwizzleW = pred_swizzle_w;
    }
 
    ureg->nr_instructions++;
-   
+
    return result;
 }
 
@@ -705,13 +720,11 @@ ureg_emit_label(struct ureg_program *ureg,
       return;
 
    out = get_tokens( ureg, DOMAIN_INSN, 1 );
-   insn = retrieve_token( ureg, DOMAIN_INSN, extended_token );
+   out[0].value = 0;
 
-   insn->token.Extended = 1;
+   insn = retrieve_token( ureg, DOMAIN_INSN, extended_token );
+   insn->insn.Label = 1;
 
-   out[0].value = 0;
-   out[0].insn_ext_label.Type = TGSI_INSTRUCTION_EXT_TYPE_LABEL;
-   
    *label_token = ureg->domain[DOMAIN_INSN].count - 1;
 }
 
@@ -734,8 +747,7 @@ ureg_fixup_label(struct ureg_program *ureg,
 {
    union tgsi_any_token *out = retrieve_token( ureg, DOMAIN_INSN, label_token );
 
-   assert(out->insn_ext_label.Type == TGSI_INSTRUCTION_EXT_TYPE_LABEL);
-   out->insn_ext_label.Label = instruction_number;
+   out->insn_label.Label = instruction_number;
 }
 
 
@@ -749,11 +761,10 @@ ureg_emit_texture(struct ureg_program *ureg,
    out = get_tokens( ureg, DOMAIN_INSN, 1 );
    insn = retrieve_token( ureg, DOMAIN_INSN, extended_token );
 
-   insn->token.Extended = 1;
+   insn->insn.Texture = 1;
 
    out[0].value = 0;
-   out[0].insn_ext_texture.Type = TGSI_INSTRUCTION_EXT_TYPE_TEXTURE;
-   out[0].insn_ext_texture.Texture = target;
+   out[0].insn_texture.Texture = target;
 }
 
 
@@ -918,8 +929,8 @@ static void emit_decl( struct ureg_program *ureg,
       out[1].decl_range.Last = index;
 
    out[2].value = 0;
-   out[2].decl_semantic.SemanticName = semantic_name;
-   out[2].decl_semantic.SemanticIndex = semantic_index;
+   out[2].decl_semantic.Name = semantic_name;
+   out[2].decl_semantic.Index = semantic_index;
 
 }
 
@@ -954,7 +965,6 @@ static void emit_immediate( struct ureg_program *ureg,
    out[0].imm.NrTokens = 5;
    out[0].imm.DataType = TGSI_IMM_FLOAT32;
    out[0].imm.Padding = 0;
-   out[0].imm.Extended = 0;
 
    out[1].imm_data.Float = v[0];
    out[2].imm_data.Float = v[1];
@@ -975,8 +985,7 @@ static void emit_decls( struct ureg_program *ureg )
             emit_decl_range( ureg, TGSI_FILE_INPUT, i, 1 );
          }
       }
-   }
-   else {
+   } else if (ureg->processor == TGSI_PROCESSOR_FRAGMENT) {
       for (i = 0; i < ureg->nr_fs_inputs; i++) {
          emit_decl( ureg, 
                     TGSI_FILE_INPUT, 
@@ -985,6 +994,13 @@ static void emit_decls( struct ureg_program *ureg )
                     ureg->fs_input[i].semantic_index,
                     ureg->fs_input[i].interp );
       }
+   } else {
+      for (i = 0; i < ureg->nr_gs_inputs; i++) {
+         emit_decl_range(ureg, 
+                         TGSI_FILE_INPUT, 
+                         ureg->gs_input[i].index,
+                         1);
+      }
    }
 
    for (i = 0; i < ureg->nr_outputs; i++) {
@@ -1023,6 +1039,13 @@ static void emit_decls( struct ureg_program *ureg )
                        0, ureg->nr_addrs );
    }
 
+   if (ureg->nr_loops) {
+      emit_decl_range(ureg,
+                      TGSI_FILE_LOOP,
+                      0,
+                      ureg->nr_loops);
+   }
+
    if (ureg->nr_preds) {
       emit_decl_range(ureg,
                       TGSI_FILE_PREDICATE,
@@ -1055,26 +1078,22 @@ static void copy_instructions( struct ureg_program *ureg )
 static void
 fixup_header_size(struct ureg_program *ureg)
 {
-   union tgsi_any_token *out = retrieve_token( ureg, DOMAIN_DECL, 1 );
+   union tgsi_any_token *out = retrieve_token( ureg, DOMAIN_DECL, 0 );
 
-   out->header.BodySize = ureg->domain[DOMAIN_DECL].count - 3;
+   out->header.BodySize = ureg->domain[DOMAIN_DECL].count - 2;
 }
 
 
 static void
 emit_header( struct ureg_program *ureg )
 {
-   union tgsi_any_token *out = get_tokens( ureg, DOMAIN_DECL, 3 );
-
-   out[0].version.MajorVersion = 1;
-   out[0].version.MinorVersion = 1;
-   out[0].version.Padding = 0;
+   union tgsi_any_token *out = get_tokens( ureg, DOMAIN_DECL, 2 );
 
-   out[1].header.HeaderSize = 2;
-   out[1].header.BodySize = 0;
+   out[0].header.HeaderSize = 2;
+   out[0].header.BodySize = 0;
 
-   out[2].processor.Processor = ureg->processor;
-   out[2].processor.Padding = 0;
+   out[1].processor.Processor = ureg->processor;
+   out[1].processor.Padding = 0;
 }
 
 
index dae429119470ca3003a973bb2a140ea74f80f623..7e3e7bcf1d36aea5f04e90987c44b42214d67489 100644 (file)
@@ -133,6 +133,10 @@ struct ureg_src
 ureg_DECL_vs_input( struct ureg_program *,
                     unsigned index );
 
+struct ureg_src
+ureg_DECL_gs_input(struct ureg_program *,
+                   unsigned index);
+
 struct ureg_dst
 ureg_DECL_output( struct ureg_program *,
                   unsigned semantic_name,
@@ -157,6 +161,9 @@ ureg_release_temporary( struct ureg_program *ureg,
 struct ureg_dst
 ureg_DECL_address( struct ureg_program * );
 
+struct ureg_dst
+ureg_DECL_loop( struct ureg_program * );
+
 struct ureg_dst
 ureg_DECL_predicate(struct ureg_program *);
 
index 4dee1be9e8cb2f89976708c33fab09cd8136fb7a..f4ca9e21ed95cc7686cd34e2f00465ea00525991 100644 (file)
@@ -76,7 +76,7 @@ tgsi_util_get_full_src_register_swizzle(
    unsigned component )
 {
    return tgsi_util_get_src_register_swizzle(
-      &reg->SrcRegister,
+      &reg->Register,
       component );
 }
 
@@ -111,10 +111,10 @@ tgsi_util_get_full_src_register_sign_mode(
 {
    unsigned sign_mode;
 
-   if( reg->SrcRegisterExtMod.Absolute ) {
+   if( reg->Register.Absolute ) {
       /* Consider only the post-abs negation. */
 
-      if( reg->SrcRegisterExtMod.Negate ) {
+      if( reg->Register.Negate ) {
          sign_mode = TGSI_UTIL_SIGN_SET;
       }
       else {
@@ -122,17 +122,7 @@ tgsi_util_get_full_src_register_sign_mode(
       }
    }
    else {
-      /* Accumulate the three negations. */
-
-      unsigned negate;
-
-      negate = reg->SrcRegister.Negate;
-
-      if( reg->SrcRegisterExtMod.Negate ) {
-         negate = !negate;
-      }
-
-      if( negate ) {
+      if( reg->Register.Negate ) {
          sign_mode = TGSI_UTIL_SIGN_TOGGLE;
       }
       else {
@@ -151,27 +141,23 @@ tgsi_util_set_full_src_register_sign_mode(
    switch (sign_mode)
    {
    case TGSI_UTIL_SIGN_CLEAR:
-      reg->SrcRegister.Negate = 0;
-      reg->SrcRegisterExtMod.Absolute = 1;
-      reg->SrcRegisterExtMod.Negate = 0;
+      reg->Register.Negate = 0;
+      reg->Register.Absolute = 1;
       break;
 
    case TGSI_UTIL_SIGN_SET:
-      reg->SrcRegister.Negate = 0;
-      reg->SrcRegisterExtMod.Absolute = 1;
-      reg->SrcRegisterExtMod.Negate = 1;
+      reg->Register.Absolute = 1;
+      reg->Register.Negate = 1;
       break;
 
    case TGSI_UTIL_SIGN_TOGGLE:
-      reg->SrcRegister.Negate = 1;
-      reg->SrcRegisterExtMod.Absolute = 0;
-      reg->SrcRegisterExtMod.Negate = 0;
+      reg->Register.Negate = 1;
+      reg->Register.Absolute = 0;
       break;
 
    case TGSI_UTIL_SIGN_KEEP:
-      reg->SrcRegister.Negate = 0;
-      reg->SrcRegisterExtMod.Absolute = 0;
-      reg->SrcRegisterExtMod.Negate = 0;
+      reg->Register.Negate = 0;
+      reg->Register.Absolute = 0;
       break;
 
    default:
index 1d8bb55bbd68d4bd9bdae78f69bd2ed268007fca..3ed90fd1b70dd615ad424643a0271a6c3a1c77aa 100644 (file)
@@ -9,8 +9,10 @@ C_SOURCES = \
        u_debug_symbol.c \
        u_debug_stack.c \
        u_blit.c \
+       u_blitter.c \
        u_cache.c \
        u_cpu_detect.c \
+       u_dl.c \
        u_draw_quad.c \
        u_format.c \
        u_format_access.c \
@@ -30,6 +32,7 @@ C_SOURCES = \
        u_stream_stdc.c \
        u_stream_wd.c \
        u_surface.c \
+       u_texture.c \
        u_tile.c \
        u_time.c \
        u_timed_winsys.c \
index 8d99106d0b852b252fb62110cadf46437fc8bd0e..2a546d19dc0d796698b11d12036be3f83f6b1e57 100644 (file)
@@ -23,6 +23,7 @@ util = env.ConvenienceLibrary(
        source = [
                'u_bitmask.c',
                'u_blit.c',
+               'u_blitter.c',
                'u_cache.c',
                'u_cpu_detect.c',
                'u_debug.c',
@@ -30,6 +31,7 @@ util = env.ConvenienceLibrary(
                'u_debug_memory.c',
                'u_debug_stack.c',
                'u_debug_symbol.c',
+               'u_dl.c',
                'u_draw_quad.c',
                'u_format.c',
                'u_format_access.c',
@@ -48,6 +50,7 @@ util = env.ConvenienceLibrary(
                'u_stream_stdc.c',
                'u_stream_wd.c',
                'u_surface.c',
+               'u_texture.c',
                'u_tile.c',
                'u_time.c',
                'u_timed_winsys.c',
index 5038642599531044a819d1f993b7066b442998b8..3f74e2aa8b803a45a9637255287fdbc260045e5f 100644 (file)
@@ -42,6 +42,7 @@
 
 #include "util/u_blit.h"
 #include "util/u_draw_quad.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 #include "util/u_simple_shaders.h"
@@ -126,7 +127,8 @@ util_create_blit(struct pipe_context *pipe, struct cso_context *cso)
    }
 
    /* fragment shader */
-   ctx->fs[TGSI_WRITEMASK_XYZW] = util_make_fragment_tex_shader(pipe);
+   ctx->fs[TGSI_WRITEMASK_XYZW] =
+      util_make_fragment_tex_shader(pipe, TGSI_TEXTURE_2D);
    ctx->vbuf = NULL;
 
    /* init vertex data that doesn't change */
@@ -354,10 +356,9 @@ util_blit_pixels_writemask(struct blit_state *ctx,
       texTemp.target = PIPE_TEXTURE_2D;
       texTemp.format = src->format;
       texTemp.last_level = 0;
-      texTemp.width[0] = srcW;
-      texTemp.height[0] = srcH;
-      texTemp.depth[0] = 1;
-      pf_get_block(src->format, &texTemp.block);
+      texTemp.width0 = srcW;
+      texTemp.height0 = srcH;
+      texTemp.depth0 = 1;
 
       tex = screen->texture_create(screen, &texTemp);
       if (!tex)
@@ -389,10 +390,10 @@ util_blit_pixels_writemask(struct blit_state *ctx,
    }
    else {
       pipe_texture_reference(&tex, src->texture);
-      s0 = srcX0 / (float)tex->width[0];
-      s1 = srcX1 / (float)tex->width[0];
-      t0 = srcY0 / (float)tex->height[0];
-      t1 = srcY1 / (float)tex->height[0];
+      s0 = srcX0 / (float)tex->width0;
+      s1 = srcX1 / (float)tex->width0;
+      t0 = srcY0 / (float)tex->height0;
+      t1 = srcY1 / (float)tex->height0;
    }
 
 
@@ -421,7 +422,9 @@ util_blit_pixels_writemask(struct blit_state *ctx,
    cso_set_sampler_textures(ctx->cso, 1, &tex);
 
    if (ctx->fs[writemask] == NULL)
-      ctx->fs[writemask] = util_make_fragment_tex_shader_writemask(pipe, writemask);
+      ctx->fs[writemask] =
+         util_make_fragment_tex_shader_writemask(pipe, TGSI_TEXTURE_2D,
+                                                 writemask);
 
    /* shaders */
    cso_set_fragment_shader_handle(ctx->cso, ctx->fs[writemask]);
@@ -518,13 +521,13 @@ util_blit_pixels_tex(struct blit_state *ctx,
    assert(filter == PIPE_TEX_MIPFILTER_NEAREST ||
           filter == PIPE_TEX_MIPFILTER_LINEAR);
 
-   assert(tex->width[0] != 0);
-   assert(tex->height[0] != 0);
+   assert(tex->width0 != 0);
+   assert(tex->height0 != 0);
 
-   s0 = srcX0 / (float)tex->width[0];
-   s1 = srcX1 / (float)tex->width[0];
-   t0 = srcY0 / (float)tex->height[0];
-   t1 = srcY1 / (float)tex->height[0];
+   s0 = srcX0 / (float)tex->width0;
+   s1 = srcX1 / (float)tex->width0;
+   t0 = srcY0 / (float)tex->height0;
+   t1 = srcY1 / (float)tex->height0;
 
    assert(ctx->pipe->screen->is_format_supported(ctx->pipe->screen, dst->format,
                                                  PIPE_TEXTURE_2D,
diff --git a/src/gallium/auxiliary/util/u_blitter.c b/src/gallium/auxiliary/util/u_blitter.c
new file mode 100644 (file)
index 0000000..1f794d3
--- /dev/null
@@ -0,0 +1,719 @@
+/**************************************************************************
+ *
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * Blitter utility to facilitate acceleration of the clear, surface_copy,
+ * and surface_fill functions.
+ *
+ * @author Marek Olšák
+ */
+
+#include "pipe/p_context.h"
+#include "pipe/p_defines.h"
+#include "pipe/p_inlines.h"
+#include "pipe/p_shader_tokens.h"
+#include "pipe/p_state.h"
+
+#include "util/u_format.h"
+#include "util/u_memory.h"
+#include "util/u_math.h"
+#include "util/u_blitter.h"
+#include "util/u_draw_quad.h"
+#include "util/u_pack_color.h"
+#include "util/u_rect.h"
+#include "util/u_simple_shaders.h"
+#include "util/u_texture.h"
+
+struct blitter_context_priv
+{
+   struct blitter_context blitter;
+
+   struct pipe_context *pipe; /**< pipe context */
+   struct pipe_buffer *vbuf;  /**< quad */
+
+   float vertices[4][2][4];   /**< {pos, color} or {pos, texcoord} */
+
+   /* Templates for various state objects. */
+   struct pipe_depth_stencil_alpha_state template_dsa;
+   struct pipe_sampler_state template_sampler_state;
+
+   /* Constant state objects. */
+   /* Vertex shaders. */
+   void *vs_col; /**< Vertex shader which passes {pos, color} to the output */
+   void *vs_tex; /**<Vertex shader which passes {pos, texcoord} to the output.*/
+
+   /* Fragment shaders. */
+   /* FS which outputs a color to multiple color buffers. */
+   void *fs_col[PIPE_MAX_COLOR_BUFS];
+
+   /* FS which outputs a color from a texture,
+      where the index is PIPE_TEXTURE_* to be sampled. */
+   void *fs_texfetch_col[PIPE_MAX_TEXTURE_TYPES];
+
+   /* FS which outputs a depth from a texture,
+      where the index is PIPE_TEXTURE_* to be sampled. */
+   void *fs_texfetch_depth[PIPE_MAX_TEXTURE_TYPES];
+
+   /* Blend state. */
+   void *blend_write_color;   /**< blend state with writemask of RGBA */
+   void *blend_keep_color;    /**< blend state with writemask of 0 */
+
+   /* Depth stencil alpha state. */
+   void *dsa_write_depth_stencil[0xff]; /**< indices are stencil clear values */
+   void *dsa_write_depth_keep_stencil;
+   void *dsa_keep_depth_stencil;
+
+   /* Sampler state for clamping to a miplevel. */
+   void *sampler_state[PIPE_MAX_TEXTURE_LEVELS];
+
+   /* Rasterizer state. */
+   void *rs_state;
+};
+
+struct blitter_context *util_blitter_create(struct pipe_context *pipe)
+{
+   struct blitter_context_priv *ctx;
+   struct pipe_blend_state blend;
+   struct pipe_depth_stencil_alpha_state *dsa;
+   struct pipe_rasterizer_state rs_state;
+   struct pipe_sampler_state *sampler_state;
+   unsigned i;
+
+   ctx = CALLOC_STRUCT(blitter_context_priv);
+   if (!ctx)
+      return NULL;
+
+   ctx->pipe = pipe;
+
+   /* init state objects for them to be considered invalid */
+   ctx->blitter.saved_fb_state.nr_cbufs = ~0;
+   ctx->blitter.saved_num_textures = ~0;
+   ctx->blitter.saved_num_sampler_states = ~0;
+
+   /* blend state objects */
+   memset(&blend, 0, sizeof(blend));
+   ctx->blend_keep_color = pipe->create_blend_state(pipe, &blend);
+
+   blend.colormask = PIPE_MASK_RGBA;
+   ctx->blend_write_color = pipe->create_blend_state(pipe, &blend);
+
+   /* depth stencil alpha state objects */
+   dsa = &ctx->template_dsa;
+   ctx->dsa_keep_depth_stencil =
+      pipe->create_depth_stencil_alpha_state(pipe, dsa);
+
+   dsa->depth.enabled = 1;
+   dsa->depth.writemask = 1;
+   dsa->depth.func = PIPE_FUNC_ALWAYS;
+   ctx->dsa_write_depth_keep_stencil =
+      pipe->create_depth_stencil_alpha_state(pipe, dsa);
+
+   dsa->stencil[0].enabled = 1;
+   dsa->stencil[0].func = PIPE_FUNC_ALWAYS;
+   dsa->stencil[0].fail_op = PIPE_STENCIL_OP_REPLACE;
+   dsa->stencil[0].zpass_op = PIPE_STENCIL_OP_REPLACE;
+   dsa->stencil[0].zfail_op = PIPE_STENCIL_OP_REPLACE;
+   dsa->stencil[0].valuemask = 0xff;
+   dsa->stencil[0].writemask = 0xff;
+   /* The DSA state objects which write depth and stencil are created
+    * on-demand. */
+
+   /* sampler state */
+   sampler_state = &ctx->template_sampler_state;
+   sampler_state->wrap_s = PIPE_TEX_WRAP_CLAMP_TO_EDGE;
+   sampler_state->wrap_t = PIPE_TEX_WRAP_CLAMP_TO_EDGE;
+   sampler_state->wrap_r = PIPE_TEX_WRAP_CLAMP_TO_EDGE;
+   /* The sampler state objects which sample from a specified mipmap level
+    * are created on-demand. */
+
+   /* rasterizer state */
+   memset(&rs_state, 0, sizeof(rs_state));
+   rs_state.front_winding = PIPE_WINDING_CW;
+   rs_state.cull_mode = PIPE_WINDING_NONE;
+   rs_state.bypass_vs_clip_and_viewport = 1;
+   rs_state.gl_rasterization_rules = 1;
+   ctx->rs_state = pipe->create_rasterizer_state(pipe, &rs_state);
+
+   /* fragment shaders are created on-demand */
+
+   /* vertex shaders */
+   {
+      const uint semantic_names[] = { TGSI_SEMANTIC_POSITION,
+                                      TGSI_SEMANTIC_COLOR };
+      const uint semantic_indices[] = { 0, 0 };
+      ctx->vs_col =
+         util_make_vertex_passthrough_shader(pipe, 2, semantic_names,
+                                             semantic_indices);
+   }
+   {
+      const uint semantic_names[] = { TGSI_SEMANTIC_POSITION,
+                                      TGSI_SEMANTIC_GENERIC };
+      const uint semantic_indices[] = { 0, 0 };
+      ctx->vs_tex =
+         util_make_vertex_passthrough_shader(pipe, 2, semantic_names,
+                                             semantic_indices);
+   }
+
+   /* set invariant vertex coordinates */
+   for (i = 0; i < 4; i++)
+      ctx->vertices[i][0][3] = 1; /*v.w*/
+
+   /* create the vertex buffer */
+   ctx->vbuf = pipe_buffer_create(ctx->pipe->screen,
+                                  32,
+                                  PIPE_BUFFER_USAGE_VERTEX,
+                                  sizeof(ctx->vertices));
+
+   return &ctx->blitter;
+}
+
+void util_blitter_destroy(struct blitter_context *blitter)
+{
+   struct blitter_context_priv *ctx = (struct blitter_context_priv*)blitter;
+   struct pipe_context *pipe = ctx->pipe;
+   int i;
+
+   pipe->delete_blend_state(pipe, ctx->blend_write_color);
+   pipe->delete_blend_state(pipe, ctx->blend_keep_color);
+   pipe->delete_depth_stencil_alpha_state(pipe, ctx->dsa_keep_depth_stencil);
+   pipe->delete_depth_stencil_alpha_state(pipe,
+                                          ctx->dsa_write_depth_keep_stencil);
+
+   for (i = 0; i < 0xff; i++)
+      if (ctx->dsa_write_depth_stencil[i])
+         pipe->delete_depth_stencil_alpha_state(pipe,
+            ctx->dsa_write_depth_stencil[i]);
+
+   pipe->delete_rasterizer_state(pipe, ctx->rs_state);
+   pipe->delete_vs_state(pipe, ctx->vs_col);
+   pipe->delete_vs_state(pipe, ctx->vs_tex);
+
+   for (i = 0; i < PIPE_MAX_TEXTURE_TYPES; i++) {
+      if (ctx->fs_texfetch_col[i])
+         pipe->delete_fs_state(pipe, ctx->fs_texfetch_col[i]);
+      if (ctx->fs_texfetch_depth[i])
+         pipe->delete_fs_state(pipe, ctx->fs_texfetch_depth[i]);
+   }
+
+   for (i = 0; i < PIPE_MAX_COLOR_BUFS && ctx->fs_col[i]; i++)
+      if (ctx->fs_col[i])
+         pipe->delete_fs_state(pipe, ctx->fs_col[i]);
+
+   for (i = 0; i < PIPE_MAX_TEXTURE_LEVELS; i++)
+      if (ctx->sampler_state[i])
+         pipe->delete_sampler_state(pipe, ctx->sampler_state[i]);
+
+   pipe_buffer_reference(&ctx->vbuf, NULL);
+   FREE(ctx);
+}
+
+static void blitter_check_saved_CSOs(struct blitter_context_priv *ctx)
+{
+   /* make sure these CSOs have been saved */
+   assert(ctx->blitter.saved_blend_state &&
+          ctx->blitter.saved_dsa_state &&
+          ctx->blitter.saved_rs_state &&
+          ctx->blitter.saved_fs &&
+          ctx->blitter.saved_vs);
+}
+
+static void blitter_restore_CSOs(struct blitter_context_priv *ctx)
+{
+   struct pipe_context *pipe = ctx->pipe;
+
+   /* restore the state objects which are always required to be saved */
+   pipe->bind_blend_state(pipe, ctx->blitter.saved_blend_state);
+   pipe->bind_depth_stencil_alpha_state(pipe, ctx->blitter.saved_dsa_state);
+   pipe->bind_rasterizer_state(pipe, ctx->blitter.saved_rs_state);
+   pipe->bind_fs_state(pipe, ctx->blitter.saved_fs);
+   pipe->bind_vs_state(pipe, ctx->blitter.saved_vs);
+
+   ctx->blitter.saved_blend_state = 0;
+   ctx->blitter.saved_dsa_state = 0;
+   ctx->blitter.saved_rs_state = 0;
+   ctx->blitter.saved_fs = 0;
+   ctx->blitter.saved_vs = 0;
+
+   /* restore the state objects which are required to be saved before copy/fill
+    */
+   if (ctx->blitter.saved_fb_state.nr_cbufs != ~0) {
+      pipe->set_framebuffer_state(pipe, &ctx->blitter.saved_fb_state);
+      ctx->blitter.saved_fb_state.nr_cbufs = ~0;
+   }
+
+   if (ctx->blitter.saved_num_sampler_states != ~0) {
+      pipe->bind_fragment_sampler_states(pipe,
+                                         ctx->blitter.saved_num_sampler_states,
+                                         ctx->blitter.saved_sampler_states);
+      ctx->blitter.saved_num_sampler_states = ~0;
+   }
+
+   if (ctx->blitter.saved_num_textures != ~0) {
+      pipe->set_fragment_sampler_textures(pipe,
+                                          ctx->blitter.saved_num_textures,
+                                          ctx->blitter.saved_textures);
+      ctx->blitter.saved_num_textures = ~0;
+   }
+}
+
+static void blitter_set_rectangle(struct blitter_context_priv *ctx,
+                                  unsigned x1, unsigned y1,
+                                  unsigned x2, unsigned y2,
+                                  float depth)
+{
+   int i;
+
+   /* set vertex positions */
+   ctx->vertices[0][0][0] = x1; /*v0.x*/
+   ctx->vertices[0][0][1] = y1; /*v0.y*/
+
+   ctx->vertices[1][0][0] = x2; /*v1.x*/
+   ctx->vertices[1][0][1] = y1; /*v1.y*/
+
+   ctx->vertices[2][0][0] = x2; /*v2.x*/
+   ctx->vertices[2][0][1] = y2; /*v2.y*/
+
+   ctx->vertices[3][0][0] = x1; /*v3.x*/
+   ctx->vertices[3][0][1] = y2; /*v3.y*/
+
+   for (i = 0; i < 4; i++)
+      ctx->vertices[i][0][2] = depth; /*z*/
+}
+
+static void blitter_set_clear_color(struct blitter_context_priv *ctx,
+                                    const float *rgba)
+{
+   int i;
+
+   for (i = 0; i < 4; i++) {
+      ctx->vertices[i][1][0] = rgba[0];
+      ctx->vertices[i][1][1] = rgba[1];
+      ctx->vertices[i][1][2] = rgba[2];
+      ctx->vertices[i][1][3] = rgba[3];
+   }
+}
+
+static void blitter_set_texcoords_2d(struct blitter_context_priv *ctx,
+                                     struct pipe_surface *surf,
+                                     unsigned x1, unsigned y1,
+                                     unsigned x2, unsigned y2)
+{
+   int i;
+   float s1 = x1 / (float)surf->width;
+   float t1 = y1 / (float)surf->height;
+   float s2 = x2 / (float)surf->width;
+   float t2 = y2 / (float)surf->height;
+
+   ctx->vertices[0][1][0] = s1; /*t0.s*/
+   ctx->vertices[0][1][1] = t1; /*t0.t*/
+
+   ctx->vertices[1][1][0] = s2; /*t1.s*/
+   ctx->vertices[1][1][1] = t1; /*t1.t*/
+
+   ctx->vertices[2][1][0] = s2; /*t2.s*/
+   ctx->vertices[2][1][1] = t2; /*t2.t*/
+
+   ctx->vertices[3][1][0] = s1; /*t3.s*/
+   ctx->vertices[3][1][1] = t2; /*t3.t*/
+
+   for (i = 0; i < 4; i++) {
+      ctx->vertices[i][1][2] = 0; /*r*/
+      ctx->vertices[i][1][3] = 1; /*q*/
+   }
+}
+
+static void blitter_set_texcoords_3d(struct blitter_context_priv *ctx,
+                                     struct pipe_surface *surf,
+                                     unsigned x1, unsigned y1,
+                                     unsigned x2, unsigned y2)
+{
+   int i;
+   float depth = u_minify(surf->texture->depth0, surf->level);
+   float r = surf->zslice / depth;
+
+   blitter_set_texcoords_2d(ctx, surf, x1, y1, x2, y2);
+
+   for (i = 0; i < 4; i++)
+      ctx->vertices[i][1][2] = r; /*r*/
+}
+
+static void blitter_set_texcoords_cube(struct blitter_context_priv *ctx,
+                                       struct pipe_surface *surf,
+                                       unsigned x1, unsigned y1,
+                                       unsigned x2, unsigned y2)
+{
+   int i;
+   float s1 = x1 / (float)surf->width;
+   float t1 = y1 / (float)surf->height;
+   float s2 = x2 / (float)surf->width;
+   float t2 = y2 / (float)surf->height;
+   const float st[4][2] = {
+      {s1, t1}, {s2, t1}, {s2, t2}, {s1, t2}
+   };
+
+   util_map_texcoords2d_onto_cubemap(surf->face,
+                                     /* pointer, stride in floats */
+                                     &st[0][0], 2,
+                                     &ctx->vertices[0][1][0], 8);
+
+   for (i = 0; i < 4; i++)
+      ctx->vertices[i][1][3] = 1; /*q*/
+}
+
+static void blitter_draw_quad(struct blitter_context_priv *ctx)
+{
+   struct pipe_context *pipe = ctx->pipe;
+
+   /* write vertices and draw them */
+   pipe_buffer_write(pipe->screen, ctx->vbuf,
+                     0, sizeof(ctx->vertices), ctx->vertices);
+
+   util_draw_vertex_buffer(pipe, ctx->vbuf, 0, PIPE_PRIM_TRIANGLE_FAN,
+                           4,  /* verts */
+                           2); /* attribs/vert */
+}
+
+static INLINE
+void *blitter_get_state_write_depth_stencil(
+               struct blitter_context_priv *ctx,
+               unsigned stencil)
+{
+   struct pipe_context *pipe = ctx->pipe;
+
+   stencil &= 0xff;
+
+   /* Create the DSA state on-demand. */
+   if (!ctx->dsa_write_depth_stencil[stencil]) {
+      ctx->template_dsa.stencil[0].ref_value = stencil;
+
+      ctx->dsa_write_depth_stencil[stencil] =
+         pipe->create_depth_stencil_alpha_state(pipe, &ctx->template_dsa);
+   }
+
+   return ctx->dsa_write_depth_stencil[stencil];
+}
+
+static INLINE
+void **blitter_get_sampler_state(struct blitter_context_priv *ctx,
+                                 int miplevel)
+{
+   struct pipe_context *pipe = ctx->pipe;
+   struct pipe_sampler_state *sampler_state = &ctx->template_sampler_state;
+
+   assert(miplevel < PIPE_MAX_TEXTURE_LEVELS);
+
+   /* Create the sampler state on-demand. */
+   if (!ctx->sampler_state[miplevel]) {
+      sampler_state->lod_bias = miplevel;
+      sampler_state->min_lod = miplevel;
+      sampler_state->max_lod = miplevel;
+
+      ctx->sampler_state[miplevel] = pipe->create_sampler_state(pipe,
+                                                                sampler_state);
+   }
+
+   /* Return void** so that it can be passed to bind_fragment_sampler_states
+    * directly. */
+   return &ctx->sampler_state[miplevel];
+}
+
+static INLINE
+void *blitter_get_fs_col(struct blitter_context_priv *ctx, unsigned num_cbufs)
+{
+   struct pipe_context *pipe = ctx->pipe;
+   unsigned index = num_cbufs ? num_cbufs - 1 : 0;
+
+   assert(num_cbufs <= PIPE_MAX_COLOR_BUFS);
+
+   if (!ctx->fs_col[index])
+      ctx->fs_col[index] =
+         util_make_fragment_clonecolor_shader(pipe, num_cbufs);
+
+   return ctx->fs_col[index];
+}
+
+static INLINE
+void *blitter_get_fs_texfetch_col(struct blitter_context_priv *ctx,
+                                  unsigned tex_target)
+{
+   struct pipe_context *pipe = ctx->pipe;
+
+   assert(tex_target < PIPE_MAX_TEXTURE_TYPES);
+
+   /* Create the fragment shader on-demand. */
+   if (!ctx->fs_texfetch_col[tex_target]) {
+      switch (tex_target) {
+         case PIPE_TEXTURE_1D:
+            ctx->fs_texfetch_col[PIPE_TEXTURE_1D] =
+               util_make_fragment_tex_shader(pipe, TGSI_TEXTURE_1D);
+            break;
+         case PIPE_TEXTURE_2D:
+            ctx->fs_texfetch_col[PIPE_TEXTURE_2D] =
+               util_make_fragment_tex_shader(pipe, TGSI_TEXTURE_2D);
+            break;
+         case PIPE_TEXTURE_3D:
+            ctx->fs_texfetch_col[PIPE_TEXTURE_3D] =
+               util_make_fragment_tex_shader(pipe, TGSI_TEXTURE_3D);
+            break;
+         case PIPE_TEXTURE_CUBE:
+            ctx->fs_texfetch_col[PIPE_TEXTURE_CUBE] =
+               util_make_fragment_tex_shader(pipe, TGSI_TEXTURE_CUBE);
+            break;
+         default:;
+      }
+   }
+
+   return ctx->fs_texfetch_col[tex_target];
+}
+
+static INLINE
+void *blitter_get_fs_texfetch_depth(struct blitter_context_priv *ctx,
+                                    unsigned tex_target)
+{
+   struct pipe_context *pipe = ctx->pipe;
+
+   assert(tex_target < PIPE_MAX_TEXTURE_TYPES);
+
+   /* Create the fragment shader on-demand. */
+   if (!ctx->fs_texfetch_depth[tex_target]) {
+      switch (tex_target) {
+         case PIPE_TEXTURE_1D:
+            ctx->fs_texfetch_depth[PIPE_TEXTURE_1D] =
+               util_make_fragment_tex_shader_writedepth(pipe, TGSI_TEXTURE_1D);
+            break;
+         case PIPE_TEXTURE_2D:
+            ctx->fs_texfetch_depth[PIPE_TEXTURE_2D] =
+               util_make_fragment_tex_shader_writedepth(pipe, TGSI_TEXTURE_2D);
+            break;
+         case PIPE_TEXTURE_3D:
+            ctx->fs_texfetch_depth[PIPE_TEXTURE_3D] =
+               util_make_fragment_tex_shader_writedepth(pipe, TGSI_TEXTURE_3D);
+            break;
+         case PIPE_TEXTURE_CUBE:
+            ctx->fs_texfetch_depth[PIPE_TEXTURE_CUBE] =
+               util_make_fragment_tex_shader_writedepth(pipe,TGSI_TEXTURE_CUBE);
+            break;
+         default:;
+      }
+   }
+
+   return ctx->fs_texfetch_depth[tex_target];
+}
+
+void util_blitter_clear(struct blitter_context *blitter,
+                        unsigned width, unsigned height,
+                        unsigned num_cbufs,
+                        unsigned clear_buffers,
+                        const float *rgba,
+                        double depth, unsigned stencil)
+{
+   struct blitter_context_priv *ctx = (struct blitter_context_priv*)blitter;
+   struct pipe_context *pipe = ctx->pipe;
+
+   assert(num_cbufs <= PIPE_MAX_COLOR_BUFS);
+
+   blitter_check_saved_CSOs(ctx);
+
+   /* bind CSOs */
+   if (clear_buffers & PIPE_CLEAR_COLOR)
+      pipe->bind_blend_state(pipe, ctx->blend_write_color);
+   else
+      pipe->bind_blend_state(pipe, ctx->blend_keep_color);
+
+   if (clear_buffers & PIPE_CLEAR_DEPTHSTENCIL)
+      pipe->bind_depth_stencil_alpha_state(pipe,
+         blitter_get_state_write_depth_stencil(ctx, stencil));
+   else
+      pipe->bind_depth_stencil_alpha_state(pipe, ctx->dsa_keep_depth_stencil);
+
+   pipe->bind_rasterizer_state(pipe, ctx->rs_state);
+   pipe->bind_fs_state(pipe, blitter_get_fs_col(ctx, num_cbufs));
+   pipe->bind_vs_state(pipe, ctx->vs_col);
+
+   blitter_set_clear_color(ctx, rgba);
+   blitter_set_rectangle(ctx, 0, 0, width, height, depth);
+   blitter_draw_quad(ctx);
+   blitter_restore_CSOs(ctx);
+}
+
+void util_blitter_copy(struct blitter_context *blitter,
+                       struct pipe_surface *dst,
+                       unsigned dstx, unsigned dsty,
+                       struct pipe_surface *src,
+                       unsigned srcx, unsigned srcy,
+                       unsigned width, unsigned height,
+                       boolean ignore_stencil)
+{
+   struct blitter_context_priv *ctx = (struct blitter_context_priv*)blitter;
+   struct pipe_context *pipe = ctx->pipe;
+   struct pipe_screen *screen = pipe->screen;
+   struct pipe_framebuffer_state fb_state;
+   boolean is_stencil, is_depth;
+   unsigned dst_tex_usage;
+
+   /* give up if textures are not set */
+   assert(dst->texture && src->texture);
+   if (!dst->texture || !src->texture)
+      return;
+
+   is_depth = util_format_get_component_bits(src->format, UTIL_FORMAT_COLORSPACE_ZS, 0) != 0;
+   is_stencil = util_format_get_component_bits(src->format, UTIL_FORMAT_COLORSPACE_ZS, 1) != 0;
+   dst_tex_usage = is_depth || is_stencil ? PIPE_TEXTURE_USAGE_DEPTH_STENCIL :
+                                            PIPE_TEXTURE_USAGE_RENDER_TARGET;
+
+   /* check if we can sample from and render to the surfaces */
+   /* (assuming copying a stencil buffer is not possible) */
+   if ((!ignore_stencil && is_stencil) ||
+       !screen->is_format_supported(screen, dst->format, dst->texture->target,
+                                    dst_tex_usage, 0) ||
+       !screen->is_format_supported(screen, src->format, src->texture->target,
+                                    PIPE_TEXTURE_USAGE_SAMPLER, 0)) {
+      util_surface_copy(pipe, FALSE, dst, dstx, dsty, src, srcx, srcy,
+                        width, height);
+      return;
+   }
+
+   /* check whether the states are properly saved */
+   blitter_check_saved_CSOs(ctx);
+   assert(blitter->saved_fb_state.nr_cbufs != ~0);
+   assert(blitter->saved_num_textures != ~0);
+   assert(blitter->saved_num_sampler_states != ~0);
+   assert(src->texture->target < PIPE_MAX_TEXTURE_TYPES);
+
+   /* bind CSOs */
+   fb_state.width = dst->width;
+   fb_state.height = dst->height;
+
+   if (is_depth) {
+      pipe->bind_blend_state(pipe, ctx->blend_keep_color);
+      pipe->bind_depth_stencil_alpha_state(pipe,
+                                           ctx->dsa_write_depth_keep_stencil);
+      pipe->bind_fs_state(pipe,
+         blitter_get_fs_texfetch_depth(ctx, src->texture->target));
+
+      fb_state.nr_cbufs = 0;
+      fb_state.zsbuf = dst;
+   } else {
+      pipe->bind_blend_state(pipe, ctx->blend_write_color);
+      pipe->bind_depth_stencil_alpha_state(pipe, ctx->dsa_keep_depth_stencil);
+      pipe->bind_fs_state(pipe,
+         blitter_get_fs_texfetch_col(ctx, src->texture->target));
+
+      fb_state.nr_cbufs = 1;
+      fb_state.cbufs[0] = dst;
+      fb_state.zsbuf = 0;
+   }
+
+   pipe->bind_rasterizer_state(pipe, ctx->rs_state);
+   pipe->bind_vs_state(pipe, ctx->vs_tex);
+   pipe->bind_fragment_sampler_states(pipe, 1,
+      blitter_get_sampler_state(ctx, src->level));
+   pipe->set_fragment_sampler_textures(pipe, 1, &src->texture);
+   pipe->set_framebuffer_state(pipe, &fb_state);
+
+   /* set texture coordinates */
+   switch (src->texture->target) {
+      case PIPE_TEXTURE_1D:
+      case PIPE_TEXTURE_2D:
+         blitter_set_texcoords_2d(ctx, src, srcx, srcy,
+                                  srcx+width, srcy+height);
+         break;
+      case PIPE_TEXTURE_3D:
+         blitter_set_texcoords_3d(ctx, src, srcx, srcy,
+                                  srcx+width, srcy+height);
+         break;
+      case PIPE_TEXTURE_CUBE:
+         blitter_set_texcoords_cube(ctx, src, srcx, srcy,
+                                    srcx+width, srcy+height);
+         break;
+      default:
+         assert(0);
+   }
+
+   blitter_set_rectangle(ctx, dstx, dsty, dstx+width, dsty+height, 0);
+   blitter_draw_quad(ctx);
+   blitter_restore_CSOs(ctx);
+}
+
+void util_blitter_fill(struct blitter_context *blitter,
+                       struct pipe_surface *dst,
+                       unsigned dstx, unsigned dsty,
+                       unsigned width, unsigned height,
+                       unsigned value)
+{
+   struct blitter_context_priv *ctx = (struct blitter_context_priv*)blitter;
+   struct pipe_context *pipe = ctx->pipe;
+   struct pipe_screen *screen = pipe->screen;
+   struct pipe_framebuffer_state fb_state;
+   float rgba[4];
+   ubyte ub_rgba[4] = {0};
+   union util_color color;
+   int i;
+
+   assert(dst->texture);
+   if (!dst->texture)
+      return;
+
+   /* check if we can render to the surface */
+   if (util_format_is_depth_or_stencil(dst->format) || /* unlikely, but you never know */
+       !screen->is_format_supported(screen, dst->format, dst->texture->target,
+                                    PIPE_TEXTURE_USAGE_RENDER_TARGET, 0)) {
+      util_surface_fill(pipe, dst, dstx, dsty, width, height, value);
+      return;
+   }
+
+   /* unpack the color */
+   color.ui = value;
+   util_unpack_color_ub(dst->format, &color,
+                        ub_rgba, ub_rgba+1, ub_rgba+2, ub_rgba+3);
+   for (i = 0; i < 4; i++)
+      rgba[i] = ubyte_to_float(ub_rgba[i]);
+
+   /* check the saved state */
+   blitter_check_saved_CSOs(ctx);
+   assert(blitter->saved_fb_state.nr_cbufs != ~0);
+
+   /* bind CSOs */
+   pipe->bind_blend_state(pipe, ctx->blend_write_color);
+   pipe->bind_depth_stencil_alpha_state(pipe, ctx->dsa_keep_depth_stencil);
+   pipe->bind_rasterizer_state(pipe, ctx->rs_state);
+   pipe->bind_fs_state(pipe, blitter_get_fs_col(ctx, 1));
+   pipe->bind_vs_state(pipe, ctx->vs_col);
+
+   /* set a framebuffer state */
+   fb_state.width = dst->width;
+   fb_state.height = dst->height;
+   fb_state.nr_cbufs = 1;
+   fb_state.cbufs[0] = dst;
+   fb_state.zsbuf = 0;
+   pipe->set_framebuffer_state(pipe, &fb_state);
+
+   blitter_set_clear_color(ctx, rgba);
+   blitter_set_rectangle(ctx, 0, 0, width, height, 0);
+   blitter_draw_quad(ctx);
+   blitter_restore_CSOs(ctx);
+}
diff --git a/src/gallium/auxiliary/util/u_blitter.h b/src/gallium/auxiliary/util/u_blitter.h
new file mode 100644 (file)
index 0000000..3da5a6c
--- /dev/null
@@ -0,0 +1,230 @@
+/**************************************************************************
+ *
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+#ifndef U_BLITTER_H
+#define U_BLITTER_H
+
+#include "util/u_memory.h"
+
+#include "pipe/p_state.h"
+
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+struct pipe_context;
+
+struct blitter_context
+{
+   /* Private members, really. */
+   void *saved_blend_state;   /**< blend state */
+   void *saved_dsa_state;     /**< depth stencil alpha state */
+   void *saved_rs_state;      /**< rasterizer state */
+   void *saved_fs, *saved_vs; /**< fragment shader, vertex shader */
+
+   struct pipe_framebuffer_state saved_fb_state;  /**< framebuffer state */
+
+   int saved_num_sampler_states;
+   void *saved_sampler_states[32];
+
+   int saved_num_textures;
+   struct pipe_texture *saved_textures[32]; /* is 32 enough? */
+};
+
+/**
+ * Create a blitter context.
+ */
+struct blitter_context *util_blitter_create(struct pipe_context *pipe);
+
+/**
+ * Destroy a blitter context.
+ */
+void util_blitter_destroy(struct blitter_context *blitter);
+
+/*
+ * These CSOs must be saved before any of the following functions is called:
+ * - blend state
+ * - depth stencil alpha state
+ * - rasterizer state
+ * - vertex shader
+ * - fragment shader
+ */
+
+/**
+ * Clear a specified set of currently bound buffers to specified values.
+ */
+void util_blitter_clear(struct blitter_context *blitter,
+                        unsigned width, unsigned height,
+                        unsigned num_cbufs,
+                        unsigned clear_buffers,
+                        const float *rgba,
+                        double depth, unsigned stencil);
+
+/**
+ * Copy a block of pixels from one surface to another.
+ *
+ * You can copy from any color format to any other color format provided
+ * the former can be sampled and the latter can be rendered to. Otherwise,
+ * a software fallback path is taken and both surfaces must be of the same
+ * format.
+ *
+ * The same holds for depth-stencil formats with the exception that stencil
+ * cannot be copied unless you set ignore_stencil to FALSE. In that case,
+ * a software fallback path is taken and both surfaces must be of the same
+ * format.
+ *
+ * Use pipe_screen->is_format_supported to know your options.
+ *
+ * These states must be saved in the blitter in addition to the state objects
+ * already required to be saved:
+ * - framebuffer state
+ * - fragment sampler states
+ * - fragment sampler textures
+ */
+void util_blitter_copy(struct blitter_context *blitter,
+                       struct pipe_surface *dst,
+                       unsigned dstx, unsigned dsty,
+                       struct pipe_surface *src,
+                       unsigned srcx, unsigned srcy,
+                       unsigned width, unsigned height,
+                       boolean ignore_stencil);
+
+/**
+ * Fill a region of a surface with a constant value.
+ *
+ * If the surface cannot be rendered to or it's a depth-stencil format,
+ * a software fallback path is taken.
+ *
+ * These states must be saved in the blitter in addition to the state objects
+ * already required to be saved:
+ * - framebuffer state
+ */
+void util_blitter_fill(struct blitter_context *blitter,
+                       struct pipe_surface *dst,
+                       unsigned dstx, unsigned dsty,
+                       unsigned width, unsigned height,
+                       unsigned value);
+
+/**
+ * Copy all pixels from one surface to another.
+ *
+ * The rules are the same as in util_blitter_copy with the addition that
+ * surfaces must have the same size.
+ */
+static INLINE
+void util_blitter_copy_surface(struct blitter_context *blitter,
+                               struct pipe_surface *dst,
+                               struct pipe_surface *src,
+                               boolean ignore_stencil)
+{
+   assert(dst->width == src->width && dst->height == src->height);
+
+   util_blitter_copy(blitter, dst, 0, 0, src, 0, 0, src->width, src->height,
+                     ignore_stencil);
+}
+
+
+/* The functions below should be used to save currently bound constant state
+ * objects inside a driver. The objects are automatically restored at the end
+ * of the util_blitter_{clear, fill, copy, copy_surface} functions and then
+ * forgotten.
+ *
+ * CSOs not listed here are not affected by util_blitter. */
+
+static INLINE
+void util_blitter_save_blend(struct blitter_context *blitter,
+                             void *state)
+{
+   blitter->saved_blend_state = state;
+}
+
+static INLINE
+void util_blitter_save_depth_stencil_alpha(struct blitter_context *blitter,
+                                           void *state)
+{
+   blitter->saved_dsa_state = state;
+}
+
+static INLINE
+void util_blitter_save_rasterizer(struct blitter_context *blitter,
+                                  void *state)
+{
+   blitter->saved_rs_state = state;
+}
+
+static INLINE
+void util_blitter_save_fragment_shader(struct blitter_context *blitter,
+                                       void *fs)
+{
+   blitter->saved_fs = fs;
+}
+
+static INLINE
+void util_blitter_save_vertex_shader(struct blitter_context *blitter,
+                                     void *vs)
+{
+   blitter->saved_vs = vs;
+}
+
+static INLINE
+void util_blitter_save_framebuffer(struct blitter_context *blitter,
+                                   struct pipe_framebuffer_state *state)
+{
+   blitter->saved_fb_state = *state;
+}
+
+static INLINE
+void util_blitter_save_fragment_sampler_states(
+                  struct blitter_context *blitter,
+                  int num_sampler_states,
+                  void **sampler_states)
+{
+   assert(num_sampler_states <= Elements(blitter->saved_sampler_states));
+
+   blitter->saved_num_sampler_states = num_sampler_states;
+   memcpy(blitter->saved_sampler_states, sampler_states,
+          num_sampler_states * sizeof(void *));
+}
+
+static INLINE
+void util_blitter_save_fragment_sampler_textures(
+                  struct blitter_context *blitter,
+                  int num_textures,
+                  struct pipe_texture **textures)
+{
+   assert(num_textures <= Elements(blitter->saved_textures));
+
+   blitter->saved_num_textures = num_textures;
+   memcpy(blitter->saved_textures, textures,
+          num_textures * sizeof(struct pipe_texture *));
+}
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif
index 1e65a035aed103fee2b886dfe732b2909c6b4ed1..2c32db61756035aae298daf777c062acd2f31982 100644 (file)
@@ -46,13 +46,13 @@ util_clear(struct pipe_context *pipe,
 {
    if (buffers & PIPE_CLEAR_COLOR) {
       struct pipe_surface *ps = framebuffer->cbufs[0];
-      unsigned color;
+      union util_color uc;
 
-      util_pack_color(rgba, ps->format, &color);
+      util_pack_color(rgba, ps->format, &uc);
       if (pipe->surface_fill) {
-         pipe->surface_fill(pipe, ps, 0, 0, ps->width, ps->height, color);
+         pipe->surface_fill(pipe, ps, 0, 0, ps->width, ps->height, uc.ui);
       } else {
-         util_surface_fill(pipe, ps, 0, 0, ps->width, ps->height, color);
+         util_surface_fill(pipe, ps, 0, 0, ps->width, ps->height, uc.ui);
       }
    }
 
index 96d400c839b166669b3702657e5d01116dd56792..4e01123fff17101831a3c5b5a5926e4cd97d5bcd 100644 (file)
 #include "pipe/p_format.h" 
 #include "pipe/p_state.h" 
 #include "pipe/p_inlines.h" 
+#include "util/u_format.h"
 #include "util/u_memory.h" 
 #include "util/u_string.h" 
 #include "util/u_stream.h" 
 #include "util/u_math.h" 
 #include "util/u_tile.h" 
+#include "util/u_prim.h" 
 
 
 #ifdef PIPE_SUBSYSTEM_WINDOWS_DISPLAY
@@ -600,6 +602,32 @@ const char *pf_name( enum pipe_format format )
 }
 
 
+
+static const struct debug_named_value pipe_prim_names[] = {
+#ifdef DEBUG
+   DEBUG_NAMED_VALUE(PIPE_PRIM_POINTS),
+   DEBUG_NAMED_VALUE(PIPE_PRIM_LINES),
+   DEBUG_NAMED_VALUE(PIPE_PRIM_LINE_LOOP),
+   DEBUG_NAMED_VALUE(PIPE_PRIM_LINE_STRIP),
+   DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLES),
+   DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLE_STRIP),
+   DEBUG_NAMED_VALUE(PIPE_PRIM_TRIANGLE_FAN),
+   DEBUG_NAMED_VALUE(PIPE_PRIM_QUADS),
+   DEBUG_NAMED_VALUE(PIPE_PRIM_QUAD_STRIP),
+   DEBUG_NAMED_VALUE(PIPE_PRIM_POLYGON),
+#endif
+   DEBUG_NAMED_VALUE_END
+};
+
+
+const char *u_prim_name( unsigned prim )
+{
+   return debug_dump_enum(pipe_prim_names, prim);
+}
+
+
+
+
 #ifdef DEBUG
 void debug_dump_image(const char *prefix,
                       unsigned format, unsigned cpp,
@@ -669,10 +697,10 @@ void debug_dump_surface(const char *prefix,
       goto error;
    
    debug_dump_image(prefix, 
-                    transfer->format,
-                    transfer->block.size
-                    transfer->nblocksx,
-                    transfer->nblocksy,
+                    texture->format,
+                    util_format_get_blocksize(texture->format)
+                    util_format_get_nblocksx(texture->format, transfer->width),
+                    util_format_get_nblocksy(texture->format, transfer->height),
                     transfer->stride,
                     data);
    
diff --git a/src/gallium/auxiliary/util/u_dl.c b/src/gallium/auxiliary/util/u_dl.c
new file mode 100644 (file)
index 0000000..b42b429
--- /dev/null
@@ -0,0 +1,79 @@
+/**************************************************************************
+ *
+ * Copyright 2009 VMware, Inc.
+ * Copyright 1999-2008  Brian Paul
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ **************************************************************************/
+
+
+#include "pipe/p_config.h"
+
+#if defined(PIPE_OS_UNIX)
+#include <dlfcn.h>
+#endif
+#if defined(PIPE_OS_WINDOWS)
+#include <windows.h>
+#endif
+
+#include "u_dl.h"
+
+
+struct util_dl_library *
+util_dl_open(const char *filename)
+{
+#if defined(PIPE_OS_UNIX)
+   return (struct util_dl_library *)dlopen(filename, RTLD_LAZY | RTLD_GLOBAL);
+#elif defined(PIPE_OS_WINDOWS)
+   return (struct util_dl_library *)LoadLibraryA(filename);
+#else
+   return NULL;
+#endif
+}
+
+
+util_dl_proc
+util_dl_get_proc_address(struct util_dl_library *library,
+                         const char *procname)
+{
+#if defined(PIPE_OS_UNIX)
+   return (util_dl_proc)dlsym((void *)library, procname);
+#elif defined(PIPE_OS_WINDOWS)
+   return (util_dl_proc)GetProcAddress((HMODULE)library, procname);
+#else
+   return (util_dl_proc)NULL;
+#endif
+}
+
+
+void
+util_dl_close(struct util_dl_library *library)
+{
+#if defined(PIPE_OS_UNIX)
+   dlclose((void *)library);
+#elif defined(PIPE_OS_WINDOWS)
+   FreeLibrary((HMODULE)library);
+#else
+   (void)library;
+#endif
+}
diff --git a/src/gallium/auxiliary/util/u_dl.h b/src/gallium/auxiliary/util/u_dl.h
new file mode 100644 (file)
index 0000000..85296c5
--- /dev/null
@@ -0,0 +1,73 @@
+/**************************************************************************
+ *
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ **************************************************************************/
+
+
+#ifndef U_DL_H_
+#define U_DL_H_
+
+
+#include "pipe/p_config.h"
+
+
+#if defined(PIPE_OS_WINDOWS)
+#  define UTIL_DL_EXT ".dll"
+#elif defined(PIPE_OS_APPLE)
+#  define UTIL_DL_EXT ".dylib"
+#else
+#  define UTIL_DL_EXT ".so"
+#endif
+
+
+struct util_dl_library;
+
+
+typedef void (*util_dl_proc)(void);
+
+
+/**
+ * Open a library dynamically.
+ */
+struct util_dl_library *
+util_dl_open(const char *filename);
+
+
+/**
+ * Lookup a function in a library.
+ */
+util_dl_proc
+util_dl_get_proc_address(struct util_dl_library *library,
+                         const char *procname);
+
+
+/**
+ * Close a library.
+ */
+void
+util_dl_close(struct util_dl_library *library);
+
+
+#endif /* U_DL_H_ */
index 98ea13b60b50e2ada5c30c1498073252da130c3e..e0724a1a8be909a09f8742d2e40c58e386f9adb3 100644 (file)
 const struct util_format_description *
 util_format_description(enum pipe_format format)
 {
-   const struct util_format_description *desc = util_format_description_table;
+   const struct util_format_description *desc;
 
-   while(TRUE) {
-      if(desc->format == format)
-         return desc;
+   if (format >= PIPE_FORMAT_COUNT) {
+      return NULL;
+   }
 
-      if(desc->format == PIPE_FORMAT_NONE)
-         return NULL;
+   desc = &util_format_description_table[format];
+   assert(desc->format == format);
 
-      ++desc;
-   };
+   return desc;
 }
index f1bf94f17ddc6f15487d3b5139898979c8f7544f..866b18ff16002d3c32ab8abd0c6355eee9612429 100644 (file)
@@ -11,6 +11,8 @@ PIPE_FORMAT_A8_UNORM              , arith , 1, 1, un8 ,     ,     ,     , 000x,
 PIPE_FORMAT_I8_UNORM              , arith , 1, 1, un8 ,     ,     ,     , xxxx, rgb
 PIPE_FORMAT_A8L8_UNORM            , arith , 1, 1, un8 , un8 ,     ,     , xxxy, rgb
 PIPE_FORMAT_L16_UNORM             , arith , 1, 1, un16,     ,     ,     , xxx1, rgb
+PIPE_FORMAT_YCBCR                 , yuv   , 2, 1, x32 ,     ,     ,     , xyz1, yuv
+PIPE_FORMAT_YCBCR_REV             , yuv   , 2, 1, x32 ,     ,     ,     , xyz1, yuv
 PIPE_FORMAT_Z16_UNORM             , array , 1, 1, un16,     ,     ,     , x___, zs 
 PIPE_FORMAT_Z32_UNORM             , array , 1, 1, un32,     ,     ,     , x___, zs 
 PIPE_FORMAT_Z32_FLOAT             , array , 1, 1, f32 ,     ,     ,     , x___, zs 
@@ -97,3 +99,11 @@ PIPE_FORMAT_B8G8R8A8_SRGB         , arith , 1, 1, u8  , u8  , u8  , u8  , zyxw,
 PIPE_FORMAT_B8G8R8X8_SRGB         , arith , 1, 1, u8  , u8  , u8  , u8  , zyx1, srgb 
 PIPE_FORMAT_X8UB8UG8SR8S_NORM     , arith , 1, 1, sn8 , sn8 , un8 , x8  , 1zyx, rgb
 PIPE_FORMAT_B6UG5SR5S_NORM        , arith , 1, 1, sn5 , sn5 , un6 ,     , xyz1, rgb
+PIPE_FORMAT_DXT1_RGB              , dxt   , 4, 4, x64 ,     ,     ,     , xyz1, rgb
+PIPE_FORMAT_DXT1_RGBA             , dxt   , 4, 4, x64 ,     ,     ,     , xyzw, rgb
+PIPE_FORMAT_DXT3_RGBA             , dxt   , 4, 4, x128,     ,     ,     , xyzw, rgb
+PIPE_FORMAT_DXT5_RGBA             , dxt   , 4, 4, x128,     ,     ,     , xyzw, rgb
+PIPE_FORMAT_DXT1_SRGB             , dxt   , 4, 4, x64 ,     ,     ,     , xyz1, srgb
+PIPE_FORMAT_DXT1_SRGBA            , dxt   , 4, 4, x64 ,     ,     ,     , xyzw, srgb
+PIPE_FORMAT_DXT3_SRGBA            , dxt   , 4, 4, x128,     ,     ,     , xyzw, srgb
+PIPE_FORMAT_DXT5_SRGBA            , dxt   , 4, 4, x128,     ,     ,     , xyzw, srgb
index 7b5b7fcda5bea8018339af61aed20e7399db1b0b..a558923b2eda6e1e5837379dcaf134753400f8c6 100644 (file)
 
 #include "pipe/p_format.h"
 
+#ifdef __cplusplus
+extern "C" {
+#endif
 
+
+/**
+ * Describe how to best pack/unpack pixels into/from the prescribed format.
+ *
+ * These are used for automatic code generation of pixel packing and unpacking
+ * routines (in compile time, e.g., u_format_access.py, or in runtime, like
+ * llvmpipe does).
+ *
+ * Thumb rule is: if you're not code generating pixel packing/unpacking then
+ * these are irrelevant for you.
+ *
+ * Note that this can be deduced from other values in util_format_description
+ * structure. This is by design, to make code generation of pixel
+ * packing/unpacking/sampling routines simple and efficient.
+ *
+ * XXX: This should be renamed to something like util_format_pack.
+ */
 enum util_format_layout {
+   /**
+    * Single scalar component.
+    */
    UTIL_FORMAT_LAYOUT_SCALAR = 0,
+
+   /**
+    * One or more components of mixed integer formats, arithmetically encoded
+    * in a word up to 32bits.
+    */
    UTIL_FORMAT_LAYOUT_ARITH = 1,
+
+   /**
+    * One or more components, no mixed formats, each with equal power of two
+    * number of bytes.
+    */
    UTIL_FORMAT_LAYOUT_ARRAY = 2,
+
+   /**
+    * XXX: Not used yet. These might go away and be replaced by a single entry,
+    * for formats where multiple pixels have to be
+    * read in order to determine a single pixel value (i.e., block.width > 1
+    * || block.height > 1)
+    */
    UTIL_FORMAT_LAYOUT_YUV = 3,
    UTIL_FORMAT_LAYOUT_DXT = 4
 };
@@ -50,7 +90,7 @@ struct util_format_block
    /** Block height in pixels */
    unsigned height;
 
-   /** Block size in bytes */
+   /** Block size in bits */
    unsigned bits;
 };
 
@@ -79,7 +119,7 @@ enum util_format_colorspace {
    UTIL_FORMAT_COLORSPACE_RGB = 0,
    UTIL_FORMAT_COLORSPACE_SRGB = 1,
    UTIL_FORMAT_COLORSPACE_YUV = 2,
-   UTIL_FORMAT_COLORSPACE_ZS = 3,
+   UTIL_FORMAT_COLORSPACE_ZS = 3
 };
 
 
@@ -111,6 +151,250 @@ const struct util_format_description *
 util_format_description(enum pipe_format format);
 
 
+/*
+ * Format query functions.
+ */
+
+static INLINE boolean 
+util_format_is_compressed(enum pipe_format format)
+{
+   const struct util_format_description *desc = util_format_description(format);
+
+   assert(format);
+   if (!format) {
+      return FALSE;
+   }
+
+   return desc->layout == UTIL_FORMAT_LAYOUT_DXT ? TRUE : FALSE;
+}
+
+static INLINE boolean 
+util_format_is_depth_or_stencil(enum pipe_format format)
+{
+   const struct util_format_description *desc = util_format_description(format);
+
+   assert(format);
+   if (!format) {
+      return FALSE;
+   }
+
+   return desc->colorspace == UTIL_FORMAT_COLORSPACE_ZS ? TRUE : FALSE;
+}
+
+static INLINE boolean 
+util_format_is_depth_and_stencil(enum pipe_format format)
+{
+   const struct util_format_description *desc = util_format_description(format);
+
+   assert(format);
+   if (!format) {
+      return FALSE;
+   }
+
+   if (desc->colorspace != UTIL_FORMAT_COLORSPACE_ZS) {
+      return FALSE;
+   }
+
+   return (desc->swizzle[0] != UTIL_FORMAT_SWIZZLE_NONE &&
+           desc->swizzle[1] != UTIL_FORMAT_SWIZZLE_NONE) ? TRUE : FALSE;
+}
+
+
+/**
+ * Return total bits needed for the pixel format per block.
+ */
+static INLINE uint
+util_format_get_blocksizebits(enum pipe_format format)
+{
+   const struct util_format_description *desc = util_format_description(format);
+
+   assert(format);
+   if (!format) {
+      return 0;
+   }
+
+   return desc->block.bits;
+}
+
+/**
+ * Return bytes per block (not pixel) for the given format.
+ */
+static INLINE uint
+util_format_get_blocksize(enum pipe_format format)
+{
+   uint bits = util_format_get_blocksizebits(format);
+
+   assert(bits % 8 == 0);
+
+   return bits / 8;
+}
+
+static INLINE uint
+util_format_get_blockwidth(enum pipe_format format)
+{
+   const struct util_format_description *desc = util_format_description(format);
+
+   assert(format);
+   if (!format) {
+      return 1;
+   }
+
+   switch (desc->layout) {
+   case UTIL_FORMAT_LAYOUT_YUV:
+      return 2;
+   case UTIL_FORMAT_LAYOUT_DXT:
+      return 4;
+   default:
+      return 1;
+   }
+}
+
+static INLINE uint
+util_format_get_blockheight(enum pipe_format format)
+{
+   const struct util_format_description *desc = util_format_description(format);
+
+   assert(format);
+   if (!format) {
+      return 1;
+   }
+
+   switch (desc->layout) {
+   case UTIL_FORMAT_LAYOUT_DXT:
+      return 4;
+   default:
+      return 1;
+   }
+}
+
+static INLINE unsigned
+util_format_get_nblocksx(enum pipe_format format,
+                         unsigned x)
+{
+   unsigned blockwidth = util_format_get_blockwidth(format);
+   return (x + blockwidth - 1) / blockwidth;
+}
+
+static INLINE unsigned
+util_format_get_nblocksy(enum pipe_format format,
+                         unsigned y)
+{
+   unsigned blockheight = util_format_get_blockheight(format);
+   return (y + blockheight - 1) / blockheight;
+}
+
+static INLINE unsigned
+util_format_get_nblocks(enum pipe_format format,
+                        unsigned width,
+                        unsigned height)
+{
+   return util_format_get_nblocksx(format, width) * util_format_get_nblocksy(format, height);
+}
+
+static INLINE size_t
+util_format_get_stride(enum pipe_format format,
+                       unsigned width)
+{
+   return util_format_get_nblocksx(format, width) * util_format_get_blocksize(format);
+}
+
+static INLINE size_t
+util_format_get_2d_size(enum pipe_format format,
+                        size_t stride,
+                        unsigned height)
+{
+   return util_format_get_nblocksy(format, height) * stride;
+}
+
+static INLINE uint
+util_format_get_component_bits(enum pipe_format format,
+                               enum util_format_colorspace colorspace,
+                               uint component)
+{
+   const struct util_format_description *desc = util_format_description(format);
+   enum util_format_colorspace desc_colorspace;
+
+   assert(format);
+   if (!format) {
+      return 0;
+   }
+
+   assert(component < 4);
+
+   /* Treat RGB and SRGB as equivalent. */
+   if (colorspace == UTIL_FORMAT_COLORSPACE_SRGB) {
+      colorspace = UTIL_FORMAT_COLORSPACE_RGB;
+   }
+   if (desc->colorspace == UTIL_FORMAT_COLORSPACE_SRGB) {
+      desc_colorspace = UTIL_FORMAT_COLORSPACE_RGB;
+   } else {
+      desc_colorspace = desc->colorspace;
+   }
+
+   if (desc_colorspace != colorspace) {
+      return 0;
+   }
+
+   switch (desc->swizzle[component]) {
+   case UTIL_FORMAT_SWIZZLE_X:
+      return desc->channel[0].size;
+   case UTIL_FORMAT_SWIZZLE_Y:
+      return desc->channel[1].size;
+   case UTIL_FORMAT_SWIZZLE_Z:
+      return desc->channel[2].size;
+   case UTIL_FORMAT_SWIZZLE_W:
+      return desc->channel[3].size;
+   default:
+      return 0;
+   }
+}
+
+static INLINE boolean
+util_format_has_alpha(enum pipe_format format)
+{
+   const struct util_format_description *desc = util_format_description(format);
+
+   assert(format);
+   if (!format) {
+      return FALSE;
+   }
+
+   switch (desc->layout) {
+   case UTIL_FORMAT_LAYOUT_SCALAR:
+   case UTIL_FORMAT_LAYOUT_ARITH:
+   case UTIL_FORMAT_LAYOUT_ARRAY:
+      /* FIXME: pf_get_component_bits( PIPE_FORMAT_A8L8_UNORM, PIPE_FORMAT_COMP_A ) should not return 0 right? */
+      if (format == PIPE_FORMAT_A8_UNORM ||
+          format == PIPE_FORMAT_A8L8_UNORM ||
+          format == PIPE_FORMAT_A8L8_SRGB) {
+         return TRUE;
+      }
+      return util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 3) != 0;
+   case UTIL_FORMAT_LAYOUT_YUV:
+      return FALSE;
+   case UTIL_FORMAT_LAYOUT_DXT:
+      switch (format) {
+      case PIPE_FORMAT_DXT1_RGBA:
+      case PIPE_FORMAT_DXT3_RGBA:
+      case PIPE_FORMAT_DXT5_RGBA:
+      case PIPE_FORMAT_DXT1_SRGBA:
+      case PIPE_FORMAT_DXT3_SRGBA:
+      case PIPE_FORMAT_DXT5_SRGBA:
+         return TRUE;
+      default:
+         return FALSE;
+      }
+   default:
+      assert(0);
+      return FALSE;
+   }
+}
+
+
+/*
+ * Format access functions.
+ */
+
 void
 util_format_read_4f(enum pipe_format format,
                     float *dst, unsigned dst_stride, 
@@ -135,4 +419,8 @@ util_format_write_4ub(enum pipe_format format,
                       void *dst, unsigned dst_stride, 
                       unsigned x, unsigned y, unsigned w, unsigned h);
 
+#ifdef __cplusplus
+} // extern "C" {
+#endif
+
 #endif /* ! U_FORMAT_H */
index eeb1a9657fd05ce5fdcce2a8f162dee9f290e31b..0b05ddb9312c700911b3073177fb6b1412bbbdd8 100644 (file)
@@ -325,14 +325,14 @@ def generate_format_read(format, dst_type, dst_native_type, dst_suffix):
             elif swizzle == SWIZZLE_0:
                 value = '0'
             elif swizzle == SWIZZLE_1:
-                value = '1'
+                value = get_one(dst_type)
             else:
                 assert False
         elif format.colorspace == 'zs':
             if i < 3:
                 value = 'z'
             else:
-                value = '1'
+                value = get_one(dst_type)
         else:
             assert False
         print '         *dst_pixel++ = %s; /* %s */' % (value, 'rgba'[i])
index 8834568e8ee607022381eef0bd5a85c6388d3e77..571cab55dc8ffa30361a8b355e2d5b04fa461553 100755 (executable)
@@ -44,11 +44,10 @@ def colorspace_map(colorspace):
 
 
 colorspace_channels_map = {
-    'rgb': 'rgba',
-    'rgba': 'rgba',
-    'zs': 'zs',
-    'yuv': ['y1', 'y2', 'u', 'v'],
-    'dxt': []
+    'rgb': ['r', 'g', 'b', 'a'],
+    'srgb': ['sr', 'sg', 'sb', 'a'],
+    'zs': ['z', 's'],
+    'yuv': ['y', 'u', 'v'],
 }
 
 
@@ -90,11 +89,20 @@ def write_format_table(formats):
     print 'const struct util_format_description'
     print 'util_format_description_table[] = '
     print "{"
+    print "   {"
+    print "      PIPE_FORMAT_NONE,"
+    print "      \"PIPE_FORMAT_NONE\","
+    print "      {0, 0, 0},"
+    print "      0,"
+    print "      {{0, 0, 0}, {0, 0, 0}, {0, 0, 0}, {0, 0, 0}},"
+    print "      {0, 0, 0, 0},"
+    print "      0"
+    print "   },"
     for format in formats:
         print "   {"
         print "      %s," % (format.name,)
         print "      \"%s\"," % (format.name,)
-        print "      {%u, %u, %u}, /* block */" % (format.block_width, format.block_height, format.block_size())
+        print "      {%u, %u, %u},\t/* block */" % (format.block_width, format.block_height, format.block_size())
         print "      %s," % (layout_map(format.layout),)
         print "      {"
         for i in range(4):
@@ -103,7 +111,7 @@ def write_format_table(formats):
                 sep = ","
             else:
                 sep = ""
-            print "         {%s, %s, %u}%s /* %s */" % (kind_map[type.kind], bool_map(type.norm), type.size, sep, "xyzw"[i])
+            print "         {%s, %s, %u}%s\t/* %s */" % (kind_map[type.kind], bool_map(type.norm), type.size, sep, "xyzw"[i])
         print "      },"
         print "      {"
         for i in range(4):
@@ -113,22 +121,13 @@ def write_format_table(formats):
             else:
                 sep = ""
             try:
-                comment = layout_channels_map[format.layout][i]
-            except:
+                comment = colorspace_channels_map[format.colorspace][i]
+            except (KeyError, IndexError):
                 comment = 'ignored'
-            print "         %s%s /* %s */" % (swizzle_map[swizzle], sep, comment)
+            print "         %s%s\t/* %s */" % (swizzle_map[swizzle], sep, comment)
         print "      },"
         print "      %s," % (colorspace_map(format.colorspace),)
         print "   },"
-    print "   {"
-    print "      PIPE_FORMAT_NONE,"
-    print "      \"PIPE_FORMAT_NONE\","
-    print "      {0, 0, 0},"
-    print "      0,"
-    print "      {{0, 0, 0}, {0, 0, 0}, {0, 0, 0}, {0, 0, 0}},"
-    print "      {0, 0, 0, 0},"
-    print "      0"
-    print "   },"
     print "};"
 
 
index aa823aa218bde6f95f6f0bc683cec9fae5955c1e..76023794dcdaecdacb0cea2a509c17bf0b037105 100644 (file)
 #include "pipe/p_shader_tokens.h"
 #include "pipe/p_state.h"
 
+#include "util/u_format.h"
 #include "util/u_memory.h"
 #include "util/u_draw_quad.h"
 #include "util/u_gen_mipmap.h"
 #include "util/u_simple_shaders.h"
+#include "util/u_math.h"
+#include "util/u_texture.h"
 
 #include "cso_cache/cso_context.h"
 
@@ -60,7 +63,7 @@ struct gen_mipmap_state
    struct pipe_sampler_state sampler;
 
    void *vs;
-   void *fs;
+   void *fs2d, *fsCube;
 
    struct pipe_buffer *vbuf;  /**< quad vertices */
    unsigned vbuf_slot;
@@ -995,7 +998,7 @@ reduce_2d(enum pipe_format pformat,
 {
    enum dtype datatype;
    uint comps;
-   const int bpt = pf_get_size(pformat);
+   const int bpt = util_format_get_blocksize(pformat);
    const ubyte *srcA, *srcB;
    ubyte *dst;
    int row;
@@ -1034,7 +1037,7 @@ reduce_3d(enum pipe_format pformat,
           int dstWidth, int dstHeight, int dstDepth,
           int dstRowStride, ubyte *dstPtr)
 {
-   const int bpt = pf_get_size(pformat);
+   const int bpt = util_format_get_blocksize(pformat);
    const int border = 0;
    int img, row;
    int bytesPerSrcImage, bytesPerDstImage;
@@ -1125,12 +1128,12 @@ make_1d_mipmap(struct gen_mipmap_state *ctx,
       
       srcTrans = screen->get_tex_transfer(screen, pt, face, srcLevel, zslice,
                                           PIPE_TRANSFER_READ, 0, 0,
-                                          pt->width[srcLevel],
-                                          pt->height[srcLevel]);
+                                          u_minify(pt->width0, srcLevel),
+                                          u_minify(pt->height0, srcLevel));
       dstTrans = screen->get_tex_transfer(screen, pt, face, dstLevel, zslice,
                                           PIPE_TRANSFER_WRITE, 0, 0,
-                                          pt->width[dstLevel],
-                                          pt->height[dstLevel]);
+                                          u_minify(pt->width0, dstLevel),
+                                          u_minify(pt->height0, dstLevel));
 
       srcMap = (ubyte *) screen->transfer_map(screen, srcTrans);
       dstMap = (ubyte *) screen->transfer_map(screen, dstTrans);
@@ -1158,8 +1161,8 @@ make_2d_mipmap(struct gen_mipmap_state *ctx,
    const uint zslice = 0;
    uint dstLevel;
    
-   assert(pt->block.width == 1);
-   assert(pt->block.height == 1);
+   assert(util_format_get_blockwidth(pt->format) == 1);
+   assert(util_format_get_blockheight(pt->format) == 1);
 
    for (dstLevel = baseLevel + 1; dstLevel <= lastLevel; dstLevel++) {
       const uint srcLevel = dstLevel - 1;
@@ -1168,12 +1171,12 @@ make_2d_mipmap(struct gen_mipmap_state *ctx,
       
       srcTrans = screen->get_tex_transfer(screen, pt, face, srcLevel, zslice,
                                           PIPE_TRANSFER_READ, 0, 0,
-                                          pt->width[srcLevel],
-                                          pt->height[srcLevel]);
+                                          u_minify(pt->width0, srcLevel),
+                                          u_minify(pt->height0, srcLevel));
       dstTrans = screen->get_tex_transfer(screen, pt, face, dstLevel, zslice,
                                           PIPE_TRANSFER_WRITE, 0, 0,
-                                          pt->width[dstLevel],
-                                          pt->height[dstLevel]);
+                                          u_minify(pt->width0, dstLevel),
+                                          u_minify(pt->height0, dstLevel));
 
       srcMap = (ubyte *) screen->transfer_map(screen, srcTrans);
       dstMap = (ubyte *) screen->transfer_map(screen, dstTrans);
@@ -1203,8 +1206,8 @@ make_3d_mipmap(struct gen_mipmap_state *ctx,
    struct pipe_screen *screen = pipe->screen;
    uint dstLevel, zslice = 0;
 
-   assert(pt->block.width == 1);
-   assert(pt->block.height == 1);
+   assert(util_format_get_blockwidth(pt->format) == 1);
+   assert(util_format_get_blockheight(pt->format) == 1);
 
    for (dstLevel = baseLevel + 1; dstLevel <= lastLevel; dstLevel++) {
       const uint srcLevel = dstLevel - 1;
@@ -1213,12 +1216,12 @@ make_3d_mipmap(struct gen_mipmap_state *ctx,
       
       srcTrans = screen->get_tex_transfer(screen, pt, face, srcLevel, zslice,
                                           PIPE_TRANSFER_READ, 0, 0,
-                                          pt->width[srcLevel],
-                                          pt->height[srcLevel]);
+                                          u_minify(pt->width0, srcLevel),
+                                          u_minify(pt->height0, srcLevel));
       dstTrans = screen->get_tex_transfer(screen, pt, face, dstLevel, zslice,
                                           PIPE_TRANSFER_WRITE, 0, 0,
-                                          pt->width[dstLevel],
-                                          pt->height[dstLevel]);
+                                          u_minify(pt->width0, dstLevel),
+                                          u_minify(pt->height0, dstLevel));
 
       srcMap = (ubyte *) screen->transfer_map(screen, srcTrans);
       dstMap = (ubyte *) screen->transfer_map(screen, dstTrans);
@@ -1316,7 +1319,8 @@ util_create_gen_mipmap(struct pipe_context *pipe,
    }
 
    /* fragment shader */
-   ctx->fs = util_make_fragment_tex_shader(pipe);
+   ctx->fs2d = util_make_fragment_tex_shader(pipe, TGSI_TEXTURE_2D);
+   ctx->fsCube = util_make_fragment_tex_shader(pipe, TGSI_TEXTURE_CUBE);
 
    /* vertex data that doesn't change */
    for (i = 0; i < 4; i++) {
@@ -1382,59 +1386,9 @@ set_vertex_data(struct gen_mipmap_state *ctx,
       static const float st[4][2] = {
          {0.0f, 0.0f}, {1.0f, 0.0f}, {1.0f, 1.0f}, {0.0f, 1.0f}
       };
-      float rx, ry, rz;
-      uint i;
-
-      /* loop over quad verts */
-      for (i = 0; i < 4; i++) {
-         /* Compute sc = +/-scale and tc = +/-scale.
-          * Not +/-1 to avoid cube face selection ambiguity near the edges,
-          * though that can still sometimes happen with this scale factor...
-          */
-         const float scale = 0.9999f;
-         const float sc = (2.0f * st[i][0] - 1.0f) * scale;
-         const float tc = (2.0f * st[i][1] - 1.0f) * scale;
-
-         switch (face) {
-         case PIPE_TEX_FACE_POS_X:
-            rx = 1.0f;
-            ry = -tc;
-            rz = -sc;
-            break;
-         case PIPE_TEX_FACE_NEG_X:
-            rx = -1.0f;
-            ry = -tc;
-            rz = sc;
-            break;
-         case PIPE_TEX_FACE_POS_Y:
-            rx = sc;
-            ry = 1.0f;
-            rz = tc;
-            break;
-         case PIPE_TEX_FACE_NEG_Y:
-            rx = sc;
-            ry = -1.0f;
-            rz = -tc;
-            break;
-         case PIPE_TEX_FACE_POS_Z:
-            rx = sc;
-            ry = -tc;
-            rz = 1.0f;
-            break;
-         case PIPE_TEX_FACE_NEG_Z:
-            rx = -sc;
-            ry = -tc;
-            rz = -1.0f;
-            break;
-         default:
-            rx = ry = rz = 0.0f;
-            assert(0);
-         }
 
-         ctx->vertices[i][1][0] = rx; /*s*/
-         ctx->vertices[i][1][1] = ry; /*t*/
-         ctx->vertices[i][1][2] = rz; /*r*/
-      }
+      util_map_texcoords2d_onto_cubemap(face, &st[0][0], 2,
+                                        &ctx->vertices[0][1][0], 8);
    }
    else {
       /* 1D/2D */
@@ -1474,7 +1428,8 @@ util_destroy_gen_mipmap(struct gen_mipmap_state *ctx)
    struct pipe_context *pipe = ctx->pipe;
 
    pipe->delete_vs_state(pipe, ctx->vs);
-   pipe->delete_fs_state(pipe, ctx->fs);
+   pipe->delete_fs_state(pipe, ctx->fs2d);
+   pipe->delete_fs_state(pipe, ctx->fsCube);
 
    pipe_buffer_reference(&ctx->vbuf, NULL);
 
@@ -1512,6 +1467,7 @@ util_gen_mipmap(struct gen_mipmap_state *ctx,
    struct pipe_context *pipe = ctx->pipe;
    struct pipe_screen *screen = pipe->screen;
    struct pipe_framebuffer_state fb;
+   void *fs = (pt->target == PIPE_TEXTURE_CUBE) ? ctx->fsCube : ctx->fs2d;
    uint dstLevel;
    uint zslice = 0;
    uint offset;
@@ -1549,7 +1505,7 @@ util_gen_mipmap(struct gen_mipmap_state *ctx,
    cso_set_depth_stencil_alpha(ctx->cso, &ctx->depthstencil);
    cso_set_rasterizer(ctx->cso, &ctx->rasterizer);
 
-   cso_set_fragment_shader_handle(ctx->cso, ctx->fs);
+   cso_set_fragment_shader_handle(ctx->cso, fs);
    cso_set_vertex_shader_handle(ctx->cso, ctx->vs);
 
    /* init framebuffer state */
@@ -1575,8 +1531,8 @@ util_gen_mipmap(struct gen_mipmap_state *ctx,
        * Setup framebuffer / dest surface
        */
       fb.cbufs[0] = surf;
-      fb.width = pt->width[dstLevel];
-      fb.height = pt->height[dstLevel];
+      fb.width = u_minify(pt->width0, dstLevel);
+      fb.height = u_minify(pt->height0, dstLevel);
       cso_set_framebuffer(ctx->cso, &fb);
 
       /*
@@ -1597,8 +1553,8 @@ util_gen_mipmap(struct gen_mipmap_state *ctx,
       offset = set_vertex_data(ctx,
                                pt->target,
                                face,
-                               (float) pt->width[dstLevel],
-                               (float) pt->height[dstLevel]);
+                               (float) u_minify(pt->width0, dstLevel),
+                               (float) u_minify(pt->height0, dstLevel));
 
       util_draw_vertex_buffer(ctx->pipe, 
                               ctx->vbuf,
index a1dce3f5cf495c4ffada6496e1f72dd3f4805f9f..f1aef2167713f592ff458edf3e45727415c881e2 100644 (file)
@@ -82,7 +82,7 @@ void pipe_linear_from_tile(struct pipe_tile_info *t, const void *src_ptr,
 
 void
 pipe_linear_fill_info(struct pipe_tile_info *t,
-                     const struct pipe_format_block *block,
+                     const struct u_linear_format_block *block,
                      unsigned tile_width, unsigned tile_height,
                      unsigned tiles_x, unsigned tiles_y)
 {
index b74308ffa3d17679aa14f4be68c8dd7f68b12a0c..42c40b2aa75570b29cccb37fba5ae910df0e2cb8 100644 (file)
 
 #include "pipe/p_format.h"
 
+struct u_linear_format_block
+{
+   /** Block size in bytes */
+   unsigned size;
+   
+   /** Block width in pixels */
+   unsigned width;
+   
+   /** Block height in pixels */
+   unsigned height;
+};
+
+
 struct pipe_tile_info
 {
    unsigned size;
@@ -49,10 +62,10 @@ struct pipe_tile_info
    unsigned rows;
 
    /* Describe the tile in pixels */
-   struct pipe_format_block tile;
+   struct u_linear_format_block tile;
 
    /* Describe each block within the tile */
-   struct pipe_format_block block;
+   struct u_linear_format_block block;
 };
 
 void pipe_linear_to_tile(size_t src_stride, const void *src_ptr,
@@ -71,7 +84,7 @@ void pipe_linear_from_tile(struct pipe_tile_info *t, const void *src_ptr,
  * @tiles_y number of tiles in y axis
  */
 void pipe_linear_fill_info(struct pipe_tile_info *t,
-                          const struct pipe_format_block *block,
+                          const struct u_linear_format_block *block,
                           unsigned tile_width, unsigned tile_height,
                           unsigned tiles_x, unsigned tiles_y);
 
index 75b075f160d922c14c0277a57dd9c5b7b60ec680..81aeb83cbb5f149951492ce0f9cdc27ce3b8f200 100644 (file)
@@ -490,6 +490,47 @@ util_next_power_of_two(unsigned x)
 }
 
 
+/**
+ * Return number of bits set in n.
+ */
+static INLINE unsigned
+util_bitcount(unsigned n)
+{
+#if defined(PIPE_CC_GCC)
+   return __builtin_popcount(n);
+#else
+   /* K&R classic bitcount.
+    *
+    * For each iteration, clear the LSB from the bitfield.
+    * Requires only one iteration per set bit, instead of
+    * one iteration per bit less than highest set bit.
+    */
+   unsigned bits = 0;
+   for (bits; n; bits++) {
+      n &= n - 1;
+   }
+   return bits;
+#endif
+}
+
+
+/**
+ * Reverse byte order of a 32 bit word.
+ */
+static INLINE uint32_t
+util_bswap32(uint32_t n)
+{
+#if defined(PIPE_CC_GCC) && (PIPE_CC_GCC_VERSION >= 403)
+   return __builtin_bswap32(n);
+#else
+   return (n >> 24) |
+          ((n >> 8) & 0x0000ff00) |
+          ((n << 8) & 0x00ff0000) |
+          (n << 24);
+#endif
+}
+
+
 /**
  * Clamp X to [MIN, MAX].
  * This is a macro to allow float, int, uint, etc. types.
@@ -499,6 +540,9 @@ util_next_power_of_two(unsigned x)
 #define MIN2( A, B )   ( (A)<(B) ? (A) : (B) )
 #define MAX2( A, B )   ( (A)>(B) ? (A) : (B) )
 
+#define MIN3( A, B, C ) MIN2( MIN2( A, B ), C )
+#define MAX3( A, B, C ) MAX2( MAX2( A, B ), C )
+
 
 static INLINE int
 align(int value, int alignment)
@@ -507,9 +551,9 @@ align(int value, int alignment)
 }
 
 static INLINE unsigned
-minify(unsigned value)
+u_minify(unsigned value, unsigned levels)
 {
-    return MAX2(1, value >> 1);
+    return MAX2(1, value >> levels);
 }
 
 #ifndef COPY_4V
@@ -539,6 +583,19 @@ do {                                     \
 #endif
 
 
+static INLINE uint32_t util_unsigned_fixed(float value, unsigned frac_bits)
+{
+   value *= (1<<frac_bits);
+   return value < 0 ? 0 : value;
+}
+
+static INLINE int32_t util_signed_fixed(float value, unsigned frac_bits)
+{
+   return value * (1<<frac_bits);
+}
+
+
+
 #ifdef __cplusplus
 }
 #endif
index 9dacc6d83dbaad81c1adf53c8420d1570b0db6c2..43eb0153ee7936d3c58273b9cbe516d39e6b3086 100644 (file)
 
 #include "pipe/p_compiler.h"
 #include "pipe/p_format.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 
 
+
+union util_color {
+   ubyte ub;
+   ushort us;
+   uint ui;
+   float f[4];
+};
+
 /**
  * Pack ubyte R,G,B,A into dest pixel.
  */
 static INLINE void
 util_pack_color_ub(ubyte r, ubyte g, ubyte b, ubyte a,
-                   enum pipe_format format, void *dest)
+                   enum pipe_format format, union util_color *uc)
 {
    switch (format) {
    case PIPE_FORMAT_R8G8B8A8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (r << 24) | (g << 16) | (b << 8) | a;
+         uc->ui = (r << 24) | (g << 16) | (b << 8) | a;
       }
       return;
    case PIPE_FORMAT_R8G8B8X8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (r << 24) | (g << 16) | (b << 8) | 0xff;
+         uc->ui = (r << 24) | (g << 16) | (b << 8) | 0xff;
       }
       return;
    case PIPE_FORMAT_A8R8G8B8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (a << 24) | (r << 16) | (g << 8) | b;
+         uc->ui = (a << 24) | (r << 16) | (g << 8) | b;
       }
       return;
    case PIPE_FORMAT_X8R8G8B8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (0xff << 24) | (r << 16) | (g << 8) | b;
+         uc->ui = (0xff << 24) | (r << 16) | (g << 8) | b;
       }
       return;
    case PIPE_FORMAT_B8G8R8A8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (b << 24) | (g << 16) | (r << 8) | a;
+         uc->ui = (b << 24) | (g << 16) | (r << 8) | a;
       }
       return;
    case PIPE_FORMAT_B8G8R8X8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (b << 24) | (g << 16) | (r << 8) | 0xff;
+         uc->ui = (b << 24) | (g << 16) | (r << 8) | 0xff;
       }
       return;
    case PIPE_FORMAT_R5G6B5_UNORM:
       {
-         ushort *d = (ushort *) dest;
-         *d = ((r & 0xf8) << 8) | ((g & 0xfc) << 3) | (b >> 3);
+         uc->us = ((r & 0xf8) << 8) | ((g & 0xfc) << 3) | (b >> 3);
       }
       return;
    case PIPE_FORMAT_A1R5G5B5_UNORM:
       {
-         ushort *d = (ushort *) dest;
-         *d = ((a & 0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);
+         uc->us = ((a & 0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);
       }
       return;
    case PIPE_FORMAT_A4R4G4B4_UNORM:
       {
-         ushort *d = (ushort *) dest;
-         *d = ((a & 0xf0) << 8) | ((r & 0xf0) << 4) | ((g & 0xf0) << 0) | (b >> 4);
+         uc->us = ((a & 0xf0) << 8) | ((r & 0xf0) << 4) | ((g & 0xf0) << 0) | (b >> 4);
       }
       return;
    case PIPE_FORMAT_A8_UNORM:
       {
-         ubyte *d = (ubyte *) dest;
-         *d = a;
+         uc->ub = a;
       }
       return;
    case PIPE_FORMAT_L8_UNORM:
    case PIPE_FORMAT_I8_UNORM:
       {
-         ubyte *d = (ubyte *) dest;
-         *d = r;
+         uc->ub = a;
       }
       return;
    case PIPE_FORMAT_R32G32B32A32_FLOAT:
       {
-         float *d = (float *) dest;
-         d[0] = (float)r / 255.0f;
-         d[1] = (float)g / 255.0f;
-         d[2] = (float)b / 255.0f;
-         d[3] = (float)a / 255.0f;
+         uc->f[0] = (float)r / 255.0f;
+         uc->f[1] = (float)g / 255.0f;
+         uc->f[2] = (float)b / 255.0f;
+         uc->f[3] = (float)a / 255.0f;
       }
       return;
    case PIPE_FORMAT_R32G32B32_FLOAT:
       {
-         float *d = (float *) dest;
-         d[0] = (float)r / 255.0f;
-         d[1] = (float)g / 255.0f;
-         d[2] = (float)b / 255.0f;
+         uc->f[0] = (float)r / 255.0f;
+         uc->f[1] = (float)g / 255.0f;
+         uc->f[2] = (float)b / 255.0f;
       }
       return;
 
-   /* XXX lots more cases to add */
+   /* Handle other cases with a generic function.
+    */
    default:
-      debug_print_format("gallium: unhandled format in util_pack_color_ub()", format);
-      assert(0);
+      {
+         ubyte src[4];
+
+         src[0] = r;
+         src[1] = g;
+         src[2] = b;
+         src[3] = a;
+         util_format_write_4ub(format, src, 0, uc, 0, 0, 0, 1, 1);
+      }
    }
 }
  
@@ -145,13 +149,13 @@ util_pack_color_ub(ubyte r, ubyte g, ubyte b, ubyte a,
  * Unpack RGBA from a packed pixel, returning values as ubytes in [0,255].
  */
 static INLINE void
-util_unpack_color_ub(enum pipe_format format, const void *src,
+util_unpack_color_ub(enum pipe_format format, union util_color *uc,
                      ubyte *r, ubyte *g, ubyte *b, ubyte *a)
 {
    switch (format) {
    case PIPE_FORMAT_R8G8B8A8_UNORM:
       {
-         uint p = ((const uint *) src)[0];
+         uint p = uc->ui;
          *r = (ubyte) ((p >> 24) & 0xff);
          *g = (ubyte) ((p >> 16) & 0xff);
          *b = (ubyte) ((p >>  8) & 0xff);
@@ -160,7 +164,7 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
       return;
    case PIPE_FORMAT_R8G8B8X8_UNORM:
       {
-         uint p = ((const uint *) src)[0];
+         uint p = uc->ui;
          *r = (ubyte) ((p >> 24) & 0xff);
          *g = (ubyte) ((p >> 16) & 0xff);
          *b = (ubyte) ((p >>  8) & 0xff);
@@ -169,7 +173,7 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
       return;
    case PIPE_FORMAT_A8R8G8B8_UNORM:
       {
-         uint p = ((const uint *) src)[0];
+         uint p = uc->ui;
          *r = (ubyte) ((p >> 16) & 0xff);
          *g = (ubyte) ((p >>  8) & 0xff);
          *b = (ubyte) ((p >>  0) & 0xff);
@@ -178,7 +182,7 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
       return;
    case PIPE_FORMAT_X8R8G8B8_UNORM:
       {
-         uint p = ((const uint *) src)[0];
+         uint p = uc->ui;
          *r = (ubyte) ((p >> 16) & 0xff);
          *g = (ubyte) ((p >>  8) & 0xff);
          *b = (ubyte) ((p >>  0) & 0xff);
@@ -187,7 +191,7 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
       return;
    case PIPE_FORMAT_B8G8R8A8_UNORM:
       {
-         uint p = ((const uint *) src)[0];
+         uint p = uc->ui;
          *r = (ubyte) ((p >>  8) & 0xff);
          *g = (ubyte) ((p >> 16) & 0xff);
          *b = (ubyte) ((p >> 24) & 0xff);
@@ -196,7 +200,7 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
       return;
    case PIPE_FORMAT_B8G8R8X8_UNORM:
       {
-         uint p = ((const uint *) src)[0];
+         uint p = uc->ui;
          *r = (ubyte) ((p >>  8) & 0xff);
          *g = (ubyte) ((p >> 16) & 0xff);
          *b = (ubyte) ((p >> 24) & 0xff);
@@ -205,7 +209,7 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
       return;
    case PIPE_FORMAT_R5G6B5_UNORM:
       {
-         ushort p = ((const ushort *) src)[0];
+         ushort p = uc->us;
          *r = (ubyte) (((p >> 8) & 0xf8) | ((p >> 13) & 0x7));
          *g = (ubyte) (((p >> 3) & 0xfc) | ((p >>  9) & 0x3));
          *b = (ubyte) (((p << 3) & 0xf8) | ((p >>  2) & 0x7));
@@ -214,7 +218,7 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
       return;
    case PIPE_FORMAT_A1R5G5B5_UNORM:
       {
-         ushort p = ((const ushort *) src)[0];
+         ushort p = uc->us;
          *r = (ubyte) (((p >>  7) & 0xf8) | ((p >> 12) & 0x7));
          *g = (ubyte) (((p >>  2) & 0xf8) | ((p >>  7) & 0x7));
          *b = (ubyte) (((p <<  3) & 0xf8) | ((p >>  2) & 0x7));
@@ -223,7 +227,7 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
       return;
    case PIPE_FORMAT_A4R4G4B4_UNORM:
       {
-         ushort p = ((const ushort *) src)[0];
+         ushort p = uc->us;
          *r = (ubyte) (((p >> 4) & 0xf0) | ((p >>  8) & 0xf));
          *g = (ubyte) (((p >> 0) & 0xf0) | ((p >>  4) & 0xf));
          *b = (ubyte) (((p << 4) & 0xf0) | ((p >>  0) & 0xf));
@@ -232,27 +236,27 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
       return;
    case PIPE_FORMAT_A8_UNORM:
       {
-         ubyte p = ((const ubyte *) src)[0];
+         ubyte p = uc->ub;
          *r = *g = *b = (ubyte) 0xff;
          *a = p;
       }
       return;
    case PIPE_FORMAT_L8_UNORM:
       {
-         ubyte p = ((const ubyte *) src)[0];
+         ubyte p = uc->ub;
          *r = *g = *b = p;
          *a = (ubyte) 0xff;
       }
       return;
    case PIPE_FORMAT_I8_UNORM:
       {
-         ubyte p = ((const ubyte *) src)[0];
+         ubyte p = uc->ub;
          *r = *g = *b = *a = p;
       }
       return;
    case PIPE_FORMAT_R32G32B32A32_FLOAT:
       {
-         const float *p = (const float *) src;
+         const float *p = &uc->f[0];
          *r = float_to_ubyte(p[0]);
          *g = float_to_ubyte(p[1]);
          *b = float_to_ubyte(p[2]);
@@ -261,7 +265,7 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
       return;
    case PIPE_FORMAT_R32G32B32_FLOAT:
       {
-         const float *p = (const float *) src;
+         const float *p = &uc->f[0];
          *r = float_to_ubyte(p[0]);
          *g = float_to_ubyte(p[1]);
          *b = float_to_ubyte(p[2]);
@@ -271,7 +275,7 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
 
    case PIPE_FORMAT_R32G32_FLOAT:
       {
-         const float *p = (const float *) src;
+         const float *p = &uc->f[0];
          *r = float_to_ubyte(p[0]);
          *g = float_to_ubyte(p[1]);
          *b = *a = (ubyte) 0xff;
@@ -280,34 +284,40 @@ util_unpack_color_ub(enum pipe_format format, const void *src,
 
    case PIPE_FORMAT_R32_FLOAT:
       {
-         const float *p = (const float *) src;
+         const float *p = &uc->f[0];
          *r = float_to_ubyte(p[0]);
          *g = *b = *a = (ubyte) 0xff;
       }
       return;
 
-   /* XXX lots more cases to add */
+   /* Handle other cases with a generic function.
+    */
    default:
-      debug_print_format("gallium: unhandled format in util_unpack_color_ub()",
-                         format);
-      assert(0);
+      {
+         ubyte dst[4];
+
+         util_format_read_4ub(format, dst, 0, uc, 0, 0, 0, 1, 1);
+         *r = dst[0];
+         *g = dst[1];
+         *b = dst[2];
+         *a = dst[3];
+      }
    }
 }
 
 
 /**
  * Note rgba outside [0,1] will be clamped for int pixel formats.
  */
 static INLINE void
-util_pack_color(const float rgba[4], enum pipe_format format, void *dest)
+util_pack_color(const float rgba[4], enum pipe_format format, union util_color *uc)
 {
    ubyte r = 0;
    ubyte g = 0;
    ubyte b = 0;
    ubyte a = 0;
 
-   if (pf_size_x(format) <= 8) {
+   if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0) <= 8) {
       /* format uses 8-bit components or less */
       r = float_to_ubyte(rgba[0]);
       g = float_to_ubyte(rgba[1]);
@@ -318,92 +328,80 @@ util_pack_color(const float rgba[4], enum pipe_format format, void *dest)
    switch (format) {
    case PIPE_FORMAT_R8G8B8A8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (r << 24) | (g << 16) | (b << 8) | a;
+         uc->ui = (r << 24) | (g << 16) | (b << 8) | a;
       }
       return;
    case PIPE_FORMAT_R8G8B8X8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (r << 24) | (g << 16) | (b << 8) | 0xff;
+         uc->ui = (r << 24) | (g << 16) | (b << 8) | 0xff;
       }
       return;
    case PIPE_FORMAT_A8R8G8B8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (a << 24) | (r << 16) | (g << 8) | b;
+         uc->ui = (a << 24) | (r << 16) | (g << 8) | b;
       }
       return;
    case PIPE_FORMAT_X8R8G8B8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (0xff << 24) | (r << 16) | (g << 8) | b;
+         uc->ui = (0xff << 24) | (r << 16) | (g << 8) | b;
       }
       return;
    case PIPE_FORMAT_B8G8R8A8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (b << 24) | (g << 16) | (r << 8) | a;
+         uc->ui = (b << 24) | (g << 16) | (r << 8) | a;
       }
       return;
    case PIPE_FORMAT_B8G8R8X8_UNORM:
       {
-         uint *d = (uint *) dest;
-         *d = (b << 24) | (g << 16) | (r << 8) | 0xff;
+         uc->ui = (b << 24) | (g << 16) | (r << 8) | 0xff;
       }
       return;
    case PIPE_FORMAT_R5G6B5_UNORM:
       {
-         ushort *d = (ushort *) dest;
-         *d = ((r & 0xf8) << 8) | ((g & 0xfc) << 3) | (b >> 3);
+         uc->us = ((r & 0xf8) << 8) | ((g & 0xfc) << 3) | (b >> 3);
       }
       return;
    case PIPE_FORMAT_A1R5G5B5_UNORM:
       {
-         ushort *d = (ushort *) dest;
-         *d = ((a & 0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);
+         uc->us = ((a & 0x80) << 8) | ((r & 0xf8) << 7) | ((g & 0xf8) << 2) | (b >> 3);
       }
       return;
    case PIPE_FORMAT_A4R4G4B4_UNORM:
       {
-         ushort *d = (ushort *) dest;
-         *d = ((a & 0xf0) << 8) | ((r & 0xf0) << 4) | ((g & 0xf0) << 0) | (b >> 4);
+         uc->ub = ((a & 0xf0) << 8) | ((r & 0xf0) << 4) | ((g & 0xf0) << 0) | (b >> 4);
       }
       return;
    case PIPE_FORMAT_A8_UNORM:
       {
-         ubyte *d = (ubyte *) dest;
-         *d = a;
+         uc->ub = a;
       }
       return;
    case PIPE_FORMAT_L8_UNORM:
    case PIPE_FORMAT_I8_UNORM:
       {
-         ubyte *d = (ubyte *) dest;
-         *d = r;
+         uc->ub = r;
       }
       return;
    case PIPE_FORMAT_R32G32B32A32_FLOAT:
       {
-         float *d = (float *) dest;
-         d[0] = rgba[0];
-         d[1] = rgba[1];
-         d[2] = rgba[2];
-         d[3] = rgba[3];
+         uc->f[0] = rgba[0];
+         uc->f[1] = rgba[1];
+         uc->f[2] = rgba[2];
+         uc->f[3] = rgba[3];
       }
       return;
    case PIPE_FORMAT_R32G32B32_FLOAT:
       {
-         float *d = (float *) dest;
-         d[0] = rgba[0];
-         d[1] = rgba[1];
-         d[2] = rgba[2];
+         uc->f[0] = rgba[0];
+         uc->f[1] = rgba[1];
+         uc->f[2] = rgba[2];
       }
       return;
-   /* XXX lots more cases to add */
+
+   /* Handle other cases with a generic function.
+    */
    default:
-      debug_print_format("gallium: unhandled format in util_pack_color()", format);
-      assert(0);
+      util_format_write_4f(format, rgba, 0, uc, 0, 0, 0, 1, 1);
    }
 }
  
index a9b533eea700458bf512607aa8b5fd9f29c4ce7f..10a874f3416d1c8957bb1e18b68c99752ed8b760 100644 (file)
@@ -135,4 +135,39 @@ static INLINE unsigned u_reduced_prim( unsigned pipe_prim )
    }
 }
 
+static INLINE unsigned
+u_vertices_per_prim(int primitive)
+{
+   switch(primitive) {
+   case PIPE_PRIM_POINTS:
+      return 1;
+   case PIPE_PRIM_LINES:
+   case PIPE_PRIM_LINE_LOOP:
+   case PIPE_PRIM_LINE_STRIP:
+      return 2;
+   case PIPE_PRIM_TRIANGLES:
+   case PIPE_PRIM_TRIANGLE_STRIP:
+   case PIPE_PRIM_TRIANGLE_FAN:
+      return 3;
+   case PIPE_PRIM_LINES_ADJACENCY:
+   case PIPE_PRIM_LINE_STRIP_ADJACENCY:
+      return 4;
+   case PIPE_PRIM_TRIANGLES_ADJACENCY:
+   case PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY:
+      return 6;
+
+   /* following primitives should never be used
+    * with geometry shaders abd their size is
+    * undefined */
+   case PIPE_PRIM_POLYGON:
+   case PIPE_PRIM_QUADS:
+   case PIPE_PRIM_QUAD_STRIP:
+   default:
+      debug_printf("Unrecognized geometry shader primitive");
+      return 3;
+   }
+}
+
+const char *u_prim_name( unsigned pipe_prim );
+
 #endif
index 9866b6fc8a0e626eb3ff1cfc17777ea94b2962f0..298fbacecbae7c09411cf6b32e9bd55ab856cb2f 100644 (file)
@@ -34,6 +34,7 @@
 #include "pipe/p_format.h"
 #include "pipe/p_context.h"
 #include "pipe/p_screen.h"
+#include "util/u_format.h"
 #include "util/u_rect.h"
 
 
@@ -44,7 +45,7 @@
  */
 void
 util_copy_rect(ubyte * dst,
-               const struct pipe_format_block *block,
+               enum pipe_format format,
                unsigned dst_stride,
                unsigned dst_x,
                unsigned dst_y,
@@ -57,27 +58,30 @@ util_copy_rect(ubyte * dst,
 {
    unsigned i;
    int src_stride_pos = src_stride < 0 ? -src_stride : src_stride;
+   int blocksize = util_format_get_blocksize(format);
+   int blockwidth = util_format_get_blockwidth(format);
+   int blockheight = util_format_get_blockheight(format);
 
-   assert(block->size > 0);
-   assert(block->width > 0);
-   assert(block->height > 0);
+   assert(blocksize > 0);
+   assert(blockwidth > 0);
+   assert(blockheight > 0);
    assert(src_x >= 0);
    assert(src_y >= 0);
    assert(dst_x >= 0);
    assert(dst_y >= 0);
 
-   dst_x /= block->width;
-   dst_y /= block->height;
-   width = (width + block->width - 1)/block->width;
-   height = (height + block->height - 1)/block->height;
-   src_x /= block->width;
-   src_y /= block->height;
+   dst_x /= blockwidth;
+   dst_y /= blockheight;
+   width = (width + blockwidth - 1)/blockwidth;
+   height = (height + blockheight - 1)/blockheight;
+   src_x /= blockwidth;
+   src_y /= blockheight;
    
-   dst += dst_x * block->size;
-   src += src_x * block->size;
+   dst += dst_x * blocksize;
+   src += src_x * blocksize;
    dst += dst_y * dst_stride;
    src += src_y * src_stride_pos;
-   width *= block->size;
+   width *= blocksize;
 
    if (width == dst_stride && width == src_stride)
       memcpy(dst, src, height * width);
@@ -92,7 +96,7 @@ util_copy_rect(ubyte * dst,
 
 void
 util_fill_rect(ubyte * dst,
-               const struct pipe_format_block *block,
+               enum pipe_format format,
                unsigned dst_stride,
                unsigned dst_x,
                unsigned dst_y,
@@ -102,23 +106,26 @@ util_fill_rect(ubyte * dst,
 {
    unsigned i, j;
    unsigned width_size;
+   int blocksize = util_format_get_blocksize(format);
+   int blockwidth = util_format_get_blockwidth(format);
+   int blockheight = util_format_get_blockheight(format);
 
-   assert(block->size > 0);
-   assert(block->width > 0);
-   assert(block->height > 0);
+   assert(blocksize > 0);
+   assert(blockwidth > 0);
+   assert(blockheight > 0);
    assert(dst_x >= 0);
    assert(dst_y >= 0);
 
-   dst_x /= block->width;
-   dst_y /= block->height;
-   width = (width + block->width - 1)/block->width;
-   height = (height + block->height - 1)/block->height;
+   dst_x /= blockwidth;
+   dst_y /= blockheight;
+   width = (width + blockwidth - 1)/blockwidth;
+   height = (height + blockheight - 1)/blockheight;
    
-   dst += dst_x * block->size;
+   dst += dst_x * blocksize;
    dst += dst_y * dst_stride;
-   width_size = width * block->size;
+   width_size = width * blocksize;
    
-   switch (block->size) {
+   switch (blocksize) {
    case 1:
       if(dst_stride == width_size)
         memset(dst, (ubyte) value, height * width_size);
@@ -172,10 +179,15 @@ util_surface_copy(struct pipe_context *pipe,
    struct pipe_transfer *src_trans, *dst_trans;
    void *dst_map;
    const void *src_map;
+   enum pipe_format src_format, dst_format;
 
    assert(src->texture && dst->texture);
    if (!src->texture || !dst->texture)
       return;
+
+   src_format = src->texture->format;
+   dst_format = dst->texture->format;
+
    src_trans = screen->get_tex_transfer(screen,
                                         src->texture,
                                         src->face,
@@ -192,9 +204,9 @@ util_surface_copy(struct pipe_context *pipe,
                                         PIPE_TRANSFER_WRITE,
                                         dst_x, dst_y, w, h);
 
-   assert(dst_trans->block.size == src_trans->block.size);
-   assert(dst_trans->block.width == src_trans->block.width);
-   assert(dst_trans->block.height == src_trans->block.height);
+   assert(util_format_get_blocksize(dst_format) == util_format_get_blocksize(src_format));
+   assert(util_format_get_blockwidth(dst_format) == util_format_get_blockwidth(src_format));
+   assert(util_format_get_blockheight(dst_format) == util_format_get_blockheight(src_format));
 
    src_map = pipe->screen->transfer_map(screen, src_trans);
    dst_map = pipe->screen->transfer_map(screen, dst_trans);
@@ -205,7 +217,7 @@ util_surface_copy(struct pipe_context *pipe,
    if (src_map && dst_map) {
       /* If do_flip, invert src_y position and pass negative src stride */
       util_copy_rect(dst_map,
-                     &dst_trans->block,
+                     dst_format,
                      dst_trans->stride,
                      0, 0,
                      w, h,
@@ -259,11 +271,11 @@ util_surface_fill(struct pipe_context *pipe,
    if (dst_map) {
       assert(dst_trans->stride > 0);
 
-      switch (dst_trans->block.size) {
+      switch (util_format_get_blocksize(dst_trans->texture->format)) {
       case 1:
       case 2:
       case 4:
-         util_fill_rect(dst_map, &dst_trans->block, dst_trans->stride,
+         util_fill_rect(dst_map, dst_trans->texture->format, dst_trans->stride,
                         0, 0, width, height, value);
          break;
       case 8:
index daa50834d367bc39d57a0db346e8b0ecb82ffb8c..5e444ffae210452fc441d6e79405b52e8c9fd4c0 100644 (file)
@@ -42,13 +42,13 @@ struct pipe_surface;
 
 
 extern void
-util_copy_rect(ubyte * dst, const struct pipe_format_block *block,
+util_copy_rect(ubyte * dst, enum pipe_format format,
                unsigned dst_stride, unsigned dst_x, unsigned dst_y,
                unsigned width, unsigned height, const ubyte * src,
                int src_stride, unsigned src_x, int src_y);
 
 extern void
-util_fill_rect(ubyte * dst, const struct pipe_format_block *block,
+util_fill_rect(ubyte * dst, enum pipe_format format,
                unsigned dst_stride, unsigned dst_x, unsigned dst_y,
                unsigned width, unsigned height, uint32_t value);
 
index 1c8b157d91fcd1ac9d21b00fa69b3d9c01fd4e53..8172ead0201e836548178033f762dabbfa2ef680 100644 (file)
@@ -2,6 +2,7 @@
  *
  * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas.
  * All Rights Reserved.
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the
@@ -30,6 +31,7 @@
  * Simple vertex/fragment shader generators.
  *  
  * @author Brian Paul
+           Marek Olšák
  */
 
 
@@ -87,6 +89,7 @@ util_make_vertex_passthrough_shader(struct pipe_context *pipe,
  */
 void *
 util_make_fragment_tex_shader_writemask(struct pipe_context *pipe,
+                                        unsigned tex_target,
                                         unsigned writemask )
 {
    struct ureg_program *ureg;
@@ -116,30 +119,82 @@ util_make_fragment_tex_shader_writemask(struct pipe_context *pipe,
 
    ureg_TEX( ureg, 
              ureg_writemask(out, writemask),
-             TGSI_TEXTURE_2D, tex, sampler );
+             tex_target, tex, sampler );
    ureg_END( ureg );
 
    return ureg_create_shader_and_destroy( ureg, pipe );
 }
 
 void *
-util_make_fragment_tex_shader(struct pipe_context *pipe )
+util_make_fragment_tex_shader(struct pipe_context *pipe, unsigned tex_target )
 {
    return util_make_fragment_tex_shader_writemask( pipe,
+                                                   tex_target,
                                                    TGSI_WRITEMASK_XYZW );
 }
 
+/**
+ * Make a simple fragment texture shader which reads an X component from
+ * a texture and writes it as depth.
+ */
+void *
+util_make_fragment_tex_shader_writedepth(struct pipe_context *pipe,
+                                         unsigned tex_target)
+{
+   struct ureg_program *ureg;
+   struct ureg_src sampler;
+   struct ureg_src tex;
+   struct ureg_dst out, depth;
+   struct ureg_src imm;
 
+   ureg = ureg_create( TGSI_PROCESSOR_FRAGMENT );
+   if (ureg == NULL)
+      return NULL;
+
+   sampler = ureg_DECL_sampler( ureg, 0 );
+
+   tex = ureg_DECL_fs_input( ureg,
+                             TGSI_SEMANTIC_GENERIC, 0,
+                             TGSI_INTERPOLATE_PERSPECTIVE );
+
+   out = ureg_DECL_output( ureg,
+                           TGSI_SEMANTIC_COLOR,
+                           0 );
+
+   depth = ureg_DECL_output( ureg,
+                             TGSI_SEMANTIC_POSITION,
+                             0 );
+
+   imm = ureg_imm4f( ureg, 0, 0, 0, 1 );
+
+   ureg_MOV( ureg, out, imm );
+
+   ureg_TEX( ureg,
+             ureg_writemask(depth, TGSI_WRITEMASK_Z),
+             tex_target, tex, sampler );
+   ureg_END( ureg );
+
+   return ureg_create_shader_and_destroy( ureg, pipe );
+}
 
 /**
  * Make simple fragment color pass-through shader.
  */
 void *
 util_make_fragment_passthrough_shader(struct pipe_context *pipe)
+{
+   return util_make_fragment_clonecolor_shader(pipe, 1);
+}
+
+void *
+util_make_fragment_clonecolor_shader(struct pipe_context *pipe, int num_cbufs)
 {
    struct ureg_program *ureg;
    struct ureg_src src;
-   struct ureg_dst dst;
+   struct ureg_dst dst[8];
+   int i;
+
+   assert(num_cbufs <= 8);
 
    ureg = ureg_create( TGSI_PROCESSOR_FRAGMENT );
    if (ureg == NULL)
@@ -148,12 +203,13 @@ util_make_fragment_passthrough_shader(struct pipe_context *pipe)
    src = ureg_DECL_fs_input( ureg, TGSI_SEMANTIC_COLOR, 0, 
                              TGSI_INTERPOLATE_PERSPECTIVE );
 
-   dst = ureg_DECL_output( ureg, TGSI_SEMANTIC_COLOR, 0 );
+   for (i = 0; i < num_cbufs; i++)
+      dst[i] = ureg_DECL_output( ureg, TGSI_SEMANTIC_COLOR, i );
+
+   for (i = 0; i < num_cbufs; i++)
+      ureg_MOV( ureg, dst[i], src );
 
-   ureg_MOV( ureg, dst, src );
    ureg_END( ureg );
 
    return ureg_create_shader_and_destroy( ureg, pipe );
 }
-
-
index d2e80d6eb4d1dcaf9c63ec09bd36288fc8d3200e..6e760942e25242d48fa9587d9e52626071a43378 100644 (file)
@@ -51,16 +51,25 @@ util_make_vertex_passthrough_shader(struct pipe_context *pipe,
 
 extern void *
 util_make_fragment_tex_shader_writemask(struct pipe_context *pipe, 
-                                        unsigned writemask );
+                                        unsigned tex_target,
+                                        unsigned writemask);
 
 extern void *
-util_make_fragment_tex_shader(struct pipe_context *pipe);
+util_make_fragment_tex_shader(struct pipe_context *pipe, unsigned tex_target);
+
+
+extern void *
+util_make_fragment_tex_shader_writedepth(struct pipe_context *pipe,
+                                         unsigned tex_target);
 
 
 extern void *
 util_make_fragment_passthrough_shader(struct pipe_context *pipe);
 
 
+extern void *
+util_make_fragment_clonecolor_shader(struct pipe_context *pipe, int num_cbufs);
+
 #ifdef __cplusplus
 }
 #endif
index 85e443204e312a2e1d6270f12230220c7ccc53c9..35c497820434e5bbb506853101e844c341c8502a 100644 (file)
@@ -36,6 +36,7 @@
 #include "pipe/p_state.h"
 #include "pipe/p_defines.h"
 
+#include "util/u_format.h"
 #include "util/u_surface.h"
 
 
@@ -79,10 +80,9 @@ util_create_rgba_surface(struct pipe_screen *screen,
    templ.target = target;
    templ.format = format;
    templ.last_level = 0;
-   templ.width[0] = width;
-   templ.height[0] = height;
-   templ.depth[0] = 1;
-   pf_get_block(format, &templ.block);
+   templ.width0 = width;
+   templ.height0 = height;
+   templ.depth0 = 1;
    templ.tex_usage = usage;
 
    *textureOut = screen->texture_create(screen, &templ);
diff --git a/src/gallium/auxiliary/util/u_texture.c b/src/gallium/auxiliary/util/u_texture.c
new file mode 100644 (file)
index 0000000..cd477ab
--- /dev/null
@@ -0,0 +1,102 @@
+/**************************************************************************
+ *
+ * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * Copyright 2008 VMware, Inc.  All rights reserved.
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * Texture mapping utility functions.
+ *
+ * @author Brian Paul
+ *         Marek Olšák
+ */
+
+#include "pipe/p_defines.h"
+
+#include "util/u_texture.h"
+
+void util_map_texcoords2d_onto_cubemap(unsigned face,
+                                       const float *in_st, unsigned in_stride,
+                                       float *out_str, unsigned out_stride)
+{
+   int i;
+   float rx, ry, rz;
+
+   /* loop over quad verts */
+   for (i = 0; i < 4; i++) {
+      /* Compute sc = +/-scale and tc = +/-scale.
+       * Not +/-1 to avoid cube face selection ambiguity near the edges,
+       * though that can still sometimes happen with this scale factor...
+       */
+      const float scale = 0.9999f;
+      const float sc = (2 * in_st[0] - 1) * scale;
+      const float tc = (2 * in_st[1] - 1) * scale;
+
+      switch (face) {
+         case PIPE_TEX_FACE_POS_X:
+            rx = 1;
+            ry = -tc;
+            rz = -sc;
+            break;
+         case PIPE_TEX_FACE_NEG_X:
+            rx = -1;
+            ry = -tc;
+            rz = sc;
+            break;
+         case PIPE_TEX_FACE_POS_Y:
+            rx = sc;
+            ry = 1;
+            rz = tc;
+            break;
+         case PIPE_TEX_FACE_NEG_Y:
+            rx = sc;
+            ry = -1;
+            rz = -tc;
+            break;
+         case PIPE_TEX_FACE_POS_Z:
+            rx = sc;
+            ry = -tc;
+            rz = 1;
+            break;
+         case PIPE_TEX_FACE_NEG_Z:
+            rx = -sc;
+            ry = -tc;
+            rz = -1;
+            break;
+         default:
+            rx = ry = rz = 0;
+            assert(0);
+      }
+
+      out_str[0] = rx; /*s*/
+      out_str[1] = ry; /*t*/
+      out_str[2] = rz; /*r*/
+
+      in_st += in_stride;
+      out_str += out_stride;
+   }
+}
diff --git a/src/gallium/auxiliary/util/u_texture.h b/src/gallium/auxiliary/util/u_texture.h
new file mode 100644 (file)
index 0000000..93b2f1e
--- /dev/null
@@ -0,0 +1,54 @@
+/**************************************************************************
+ *
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+#ifndef U_TEXTURE_H
+#define U_TEXTURE_H
+
+#ifdef __cplusplus
+extern "C" {
+#endif
+
+/**
+ * Convert 2D texture coordinates of 4 vertices into cubemap coordinates
+ * in the given face.
+ * Coordinates must be in the range [0,1].
+ *
+ * \param face          Cubemap face.
+ * \param in_st         4 pairs of 2D texture coordinates to convert.
+ * \param in_stride     Stride of in_st in floats.
+ * \param out_str       STR cubemap texture coordinates to compute.
+ * \param out_stride    Stride of out_str in floats.
+ */
+void util_map_texcoords2d_onto_cubemap(unsigned face,
+                                       const float *in_st, unsigned in_stride,
+                                       float *out_str, unsigned out_stride);
+
+
+#ifdef __cplusplus
+}
+#endif
+
+#endif
index 8a22f584bee766962c09a1d01fd642670d3a8337..5b8dd1abb949b68145b1c3337205496afa270078 100644 (file)
@@ -34,6 +34,7 @@
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
 
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 #include "util/u_rect.h"
@@ -52,7 +53,7 @@ pipe_get_tile_raw(struct pipe_transfer *pt,
    const void *src;
 
    if (dst_stride == 0)
-      dst_stride = pf_get_nblocksx(&pt->block, w) * pt->block.size;
+      dst_stride = util_format_get_stride(pt->texture->format, w);
 
    if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
@@ -62,7 +63,7 @@ pipe_get_tile_raw(struct pipe_transfer *pt,
    if(!src)
       return;
 
-   util_copy_rect(dst, &pt->block, dst_stride, 0, 0, w, h, src, pt->stride, x, y);
+   util_copy_rect(dst, pt->texture->format, dst_stride, 0, 0, w, h, src, pt->stride, x, y);
 
    screen->transfer_unmap(screen, pt);
 }
@@ -78,9 +79,10 @@ pipe_put_tile_raw(struct pipe_transfer *pt,
 {
    struct pipe_screen *screen = pt->texture->screen;
    void *dst;
+   enum pipe_format format = pt->texture->format;
 
    if (src_stride == 0)
-      src_stride = pf_get_nblocksx(&pt->block, w) * pt->block.size;
+      src_stride = util_format_get_stride(format, w);
 
    if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
@@ -90,7 +92,7 @@ pipe_put_tile_raw(struct pipe_transfer *pt,
    if(!dst)
       return;
 
-   util_copy_rect(dst, &pt->block, pt->stride, x, y, w, h, src, src_stride, 0, 0);
+   util_copy_rect(dst, format, pt->stride, x, y, w, h, src, src_stride, 0, 0);
 
    screen->transfer_unmap(screen, pt);
 }
@@ -246,6 +248,53 @@ b8g8r8a8_put_tile_rgba(unsigned *dst,
 }
 
 
+/*** PIPE_FORMAT_R8G8B8A8_UNORM ***/
+
+static void
+r8g8b8a8_get_tile_rgba(const unsigned *src,
+                       unsigned w, unsigned h,
+                       float *p,
+                       unsigned dst_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         const unsigned pixel = *src++;
+         pRow[0] = ubyte_to_float((pixel >> 24) & 0xff);
+         pRow[1] = ubyte_to_float((pixel >> 16) & 0xff);
+         pRow[2] = ubyte_to_float((pixel >>  8) & 0xff);
+         pRow[3] = ubyte_to_float((pixel >>  0) & 0xff);
+      }
+      p += dst_stride;
+   }
+}
+
+
+static void
+r8g8b8a8_put_tile_rgba(unsigned *dst,
+                       unsigned w, unsigned h,
+                       const float *p,
+                       unsigned src_stride)
+{
+   unsigned i, j;
+
+   for (i = 0; i < h; i++) {
+      const float *pRow = p;
+      for (j = 0; j < w; j++, pRow += 4) {
+         unsigned r, g, b, a;
+         r = float_to_ubyte(pRow[0]);
+         g = float_to_ubyte(pRow[1]);
+         b = float_to_ubyte(pRow[2]);
+         a = float_to_ubyte(pRow[3]);
+         *dst++ = (r << 24) | (g << 16) | (b << 8) | a;
+      }
+      p += src_stride;
+   }
+}
+
+
 /*** PIPE_FORMAT_A1R5G5B5_UNORM ***/
 
 static void
@@ -1143,6 +1192,9 @@ pipe_tile_raw_to_rgba(enum pipe_format format,
    case PIPE_FORMAT_B8G8R8A8_UNORM:
       b8g8r8a8_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
       break;
+   case PIPE_FORMAT_R8G8B8A8_UNORM:
+      r8g8b8a8_get_tile_rgba((unsigned *) src, w, h, dst, dst_stride);
+      break;
    case PIPE_FORMAT_A1R5G5B5_UNORM:
       a1r5g5b5_get_tile_rgba((ushort *) src, w, h, dst, dst_stride);
       break;
@@ -1219,21 +1271,22 @@ pipe_get_tile_rgba(struct pipe_transfer *pt,
 {
    unsigned dst_stride = w * 4;
    void *packed;
+   enum pipe_format format = pt->texture->format;
 
    if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
 
-   packed = MALLOC(pf_get_nblocks(&pt->block, w, h) * pt->block.size);
+   packed = MALLOC(util_format_get_nblocks(format, w, h) * util_format_get_blocksize(format));
 
    if (!packed)
       return;
 
-   if(pt->format == PIPE_FORMAT_YCBCR || pt->format == PIPE_FORMAT_YCBCR_REV)
+   if(format == PIPE_FORMAT_YCBCR || format == PIPE_FORMAT_YCBCR_REV)
       assert((x & 1) == 0);
 
    pipe_get_tile_raw(pt, x, y, w, h, packed, 0);
 
-   pipe_tile_raw_to_rgba(pt->format, packed, w, h, p, dst_stride);
+   pipe_tile_raw_to_rgba(format, packed, w, h, p, dst_stride);
 
    FREE(packed);
 }
@@ -1246,16 +1299,17 @@ pipe_put_tile_rgba(struct pipe_transfer *pt,
 {
    unsigned src_stride = w * 4;
    void *packed;
+   enum pipe_format format = pt->texture->format;
 
    if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
 
-   packed = MALLOC(pf_get_nblocks(&pt->block, w, h) * pt->block.size);
+   packed = MALLOC(util_format_get_nblocks(format, w, h) * util_format_get_blocksize(format));
 
    if (!packed)
       return;
 
-   switch (pt->format) {
+   switch (format) {
    case PIPE_FORMAT_A8R8G8B8_UNORM:
       a8r8g8b8_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);
       break;
@@ -1265,6 +1319,9 @@ pipe_put_tile_rgba(struct pipe_transfer *pt,
    case PIPE_FORMAT_B8G8R8A8_UNORM:
       b8g8r8a8_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);
       break;
+   case PIPE_FORMAT_R8G8B8A8_UNORM:
+      r8g8b8a8_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);
+      break;
    case PIPE_FORMAT_A1R5G5B5_UNORM:
       a1r5g5b5_put_tile_rgba((ushort *) packed, w, h, p, src_stride);
       break;
@@ -1274,9 +1331,6 @@ pipe_put_tile_rgba(struct pipe_transfer *pt,
    case PIPE_FORMAT_R8G8B8_UNORM:
       r8g8b8_put_tile_rgba((ubyte *) packed, w, h, p, src_stride);
       break;
-   case PIPE_FORMAT_R8G8B8A8_UNORM:
-      assert(0);
-      break;
    case PIPE_FORMAT_A4R4G4B4_UNORM:
       a4r4g4b4_put_tile_rgba((ushort *) packed, w, h, p, src_stride);
       break;
@@ -1322,7 +1376,7 @@ pipe_put_tile_rgba(struct pipe_transfer *pt,
       /*z24s8_put_tile_rgba((unsigned *) packed, w, h, p, src_stride);*/
       break;
    default:
-      debug_printf("%s: unsupported format %s\n", __FUNCTION__, pf_name(pt->format));
+      debug_printf("%s: unsupported format %s\n", __FUNCTION__, pf_name(format));
    }
 
    pipe_put_tile_raw(pt, x, y, w, h, packed, 0);
@@ -1344,6 +1398,7 @@ pipe_get_tile_z(struct pipe_transfer *pt,
    ubyte *map;
    uint *pDest = z;
    uint i, j;
+   enum pipe_format format = pt->texture->format;
 
    if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
@@ -1354,7 +1409,7 @@ pipe_get_tile_z(struct pipe_transfer *pt,
       return;
    }
 
-   switch (pt->format) {
+   switch (format) {
    case PIPE_FORMAT_Z32_UNORM:
       {
          const uint *ptrc
@@ -1428,6 +1483,7 @@ pipe_put_tile_z(struct pipe_transfer *pt,
    const uint *ptrc = zSrc;
    ubyte *map;
    uint i, j;
+   enum pipe_format format = pt->texture->format;
 
    if (pipe_clip_tile(x, y, &w, &h, pt))
       return;
@@ -1438,7 +1494,7 @@ pipe_put_tile_z(struct pipe_transfer *pt,
       return;
    }
 
-   switch (pt->format) {
+   switch (format) {
    case PIPE_FORMAT_Z32_UNORM:
       {
          uint *pDest = (uint *) (map + y * pt->stride + x*4);
index 745b5834af648b571c0da27fda50707283b84584..e158bed9d04291910e373463525f55b26413734f 100644 (file)
@@ -32,6 +32,8 @@
 #ifndef U_UPLOAD_MGR_H
 #define U_UPLOAD_MGR_H
 
+#include "pipe/p_defines.h"
+
 struct pipe_screen;
 struct pipe_buffer;
 struct u_upload_mgr;
index cda6dc134a05530b667af978405ac433cdee782b..fc2a1c59a6b13bb28a1b777d90d898d26b567598 100644 (file)
@@ -95,12 +95,11 @@ create_vert_shader(struct vl_compositor *c)
    assert(c);
 
    tokens = (struct tgsi_token*)MALLOC(max_tokens * sizeof(struct tgsi_token));
-   *(struct tgsi_version*)&tokens[0] = tgsi_build_version();
-   header = (struct tgsi_header*)&tokens[1];
+   header = (struct tgsi_header*)&tokens[0];
    *header = tgsi_build_header();
-   *(struct tgsi_processor*)&tokens[2] = tgsi_build_processor(TGSI_PROCESSOR_VERTEX, header);
+   *(struct tgsi_processor*)&tokens[1] = tgsi_build_processor(TGSI_PROCESSOR_VERTEX, header);
 
-   ti = 3;
+   ti = 2;
 
    /*
     * decl i0             ; Vertex pos
@@ -172,12 +171,11 @@ create_frag_shader(struct vl_compositor *c)
    assert(c);
 
    tokens = (struct tgsi_token*)MALLOC(max_tokens * sizeof(struct tgsi_token));
-   *(struct tgsi_version*)&tokens[0] = tgsi_build_version();
-   header = (struct tgsi_header*)&tokens[1];
+   header = (struct tgsi_header*)&tokens[0];
    *header = tgsi_build_header();
-   *(struct tgsi_processor*)&tokens[2] = tgsi_build_processor(TGSI_PROCESSOR_FRAGMENT, header);
+   *(struct tgsi_processor*)&tokens[1] = tgsi_build_processor(TGSI_PROCESSOR_FRAGMENT, header);
 
-   ti = 3;
+   ti = 2;
 
    /* decl i0             ; Texcoords for s0 */
    decl = vl_decl_interpolated_input(TGSI_SEMANTIC_GENERIC, 1, 0, 0, TGSI_INTERPOLATE_LINEAR);
@@ -213,7 +211,7 @@ create_frag_shader(struct vl_compositor *c)
     */
    for (i = 0; i < 4; ++i) {
       inst = vl_inst3(TGSI_OPCODE_DP4, TGSI_FILE_OUTPUT, 0, TGSI_FILE_TEMPORARY, 0, TGSI_FILE_CONSTANT, i);
-      inst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_X << i;
+      inst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_X << i;
       ti += tgsi_build_full_instruction(&inst, &tokens[ti], header, max_tokens - ti);
    }
 
@@ -455,8 +453,8 @@ void vl_compositor_render(struct vl_compositor          *compositor,
    assert(dst_area);
    assert(picture_type == PIPE_MPEG12_PICTURE_TYPE_FRAME);
 
-   compositor->fb_state.width = dst_surface->width[0];
-   compositor->fb_state.height = dst_surface->height[0];
+   compositor->fb_state.width = dst_surface->width0;
+   compositor->fb_state.height = dst_surface->height0;
    compositor->fb_state.cbufs[0] = compositor->pipe->screen->get_tex_surface
    (
       compositor->pipe->screen,
@@ -479,8 +477,8 @@ void vl_compositor_render(struct vl_compositor          *compositor,
    compositor->pipe->set_framebuffer_state(compositor->pipe, &compositor->fb_state);
    compositor->pipe->set_viewport_state(compositor->pipe, &compositor->viewport);
    compositor->pipe->set_scissor_state(compositor->pipe, &compositor->scissor);
-   compositor->pipe->bind_sampler_states(compositor->pipe, 1, &compositor->sampler);
-   compositor->pipe->set_sampler_textures(compositor->pipe, 1, &src_surface);
+   compositor->pipe->bind_fragment_sampler_states(compositor->pipe, 1, &compositor->sampler);
+   compositor->pipe->set_fragment_sampler_textures(compositor->pipe, 1, &src_surface);
    compositor->pipe->bind_vs_state(compositor->pipe, compositor->vertex_shader);
    compositor->pipe->bind_fs_state(compositor->pipe, compositor->fragment_shader);
    compositor->pipe->set_vertex_buffers(compositor->pipe, 2, compositor->vertex_bufs);
@@ -504,12 +502,12 @@ void vl_compositor_render(struct vl_compositor          *compositor,
    vs_consts->dst_trans.z = 0;
    vs_consts->dst_trans.w = 0;
 
-   vs_consts->src_scale.x = src_area->w / (float)src_surface->width[0];
-   vs_consts->src_scale.y = src_area->h / (float)src_surface->height[0];
+   vs_consts->src_scale.x = src_area->w / (float)src_surface->width0;
+   vs_consts->src_scale.y = src_area->h / (float)src_surface->height0;
    vs_consts->src_scale.z = 1;
    vs_consts->src_scale.w = 1;
-   vs_consts->src_trans.x = src_area->x / (float)src_surface->width[0];
-   vs_consts->src_trans.y = src_area->y / (float)src_surface->height[0];
+   vs_consts->src_trans.x = src_area->x / (float)src_surface->width0;
+   vs_consts->src_trans.y = src_area->y / (float)src_surface->height0;
    vs_consts->src_trans.z = 0;
    vs_consts->src_trans.w = 0;
 
index c4ba69817f9df99f12aeff3a6a9b0197c3405c9b..ab196c21f8782fa82db79f1ae18a6e6dc7f4a511 100644 (file)
@@ -29,6 +29,7 @@
 #include <assert.h>
 #include <pipe/p_context.h>
 #include <pipe/p_inlines.h>
+#include <util/u_format.h>
 #include <util/u_math.h>
 #include <util/u_memory.h>
 #include <tgsi/tgsi_parse.h>
@@ -115,12 +116,11 @@ create_intra_vert_shader(struct vl_mpeg12_mc_renderer *r)
    assert(r);
 
    tokens = (struct tgsi_token *) malloc(max_tokens * sizeof(struct tgsi_token));
-   *(struct tgsi_version *) &tokens[0] = tgsi_build_version();
-   header = (struct tgsi_header *) &tokens[1];
+   header = (struct tgsi_header *) &tokens[0];
    *header = tgsi_build_header();
-   *(struct tgsi_processor *) &tokens[2] = tgsi_build_processor(TGSI_PROCESSOR_VERTEX, header);
+   *(struct tgsi_processor *) &tokens[1] = tgsi_build_processor(TGSI_PROCESSOR_VERTEX, header);
 
-   ti = 3;
+   ti = 2;
 
    /*
     * decl i0              ; Vertex pos
@@ -185,12 +185,11 @@ create_intra_frag_shader(struct vl_mpeg12_mc_renderer *r)
    assert(r);
 
    tokens = (struct tgsi_token *) malloc(max_tokens * sizeof(struct tgsi_token));
-   *(struct tgsi_version *) &tokens[0] = tgsi_build_version();
-   header = (struct tgsi_header *) &tokens[1];
+   header = (struct tgsi_header *) &tokens[0];
    *header = tgsi_build_header();
-   *(struct tgsi_processor *) &tokens[2] = tgsi_build_processor(TGSI_PROCESSOR_FRAGMENT, header);
+   *(struct tgsi_processor *) &tokens[1] = tgsi_build_processor(TGSI_PROCESSOR_FRAGMENT, header);
 
-   ti = 3;
+   ti = 2;
 
    /*
     * decl i0                      ; Luma texcoords
@@ -237,10 +236,10 @@ create_intra_frag_shader(struct vl_mpeg12_mc_renderer *r)
       ti += tgsi_build_full_instruction(&inst, &tokens[ti], header, max_tokens - ti);
 
       inst = vl_inst2(TGSI_OPCODE_MOV, TGSI_FILE_TEMPORARY, 0, TGSI_FILE_TEMPORARY, 1);
-      inst.FullSrcRegisters[0].SrcRegister.SwizzleX = TGSI_SWIZZLE_X;
-      inst.FullSrcRegisters[0].SrcRegister.SwizzleY = TGSI_SWIZZLE_X;
-      inst.FullSrcRegisters[0].SrcRegister.SwizzleZ = TGSI_SWIZZLE_X;
-      inst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_X << i;
+      inst.Src[0].Register.SwizzleX = TGSI_SWIZZLE_X;
+      inst.Src[0].Register.SwizzleY = TGSI_SWIZZLE_X;
+      inst.Src[0].Register.SwizzleZ = TGSI_SWIZZLE_X;
+      inst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_X << i;
       ti += tgsi_build_full_instruction(&inst, &tokens[ti], header, max_tokens - ti);
    }
 
@@ -278,12 +277,11 @@ create_frame_pred_vert_shader(struct vl_mpeg12_mc_renderer *r)
    assert(r);
 
    tokens = (struct tgsi_token *) malloc(max_tokens * sizeof(struct tgsi_token));
-   *(struct tgsi_version *) &tokens[0] = tgsi_build_version();
-   header = (struct tgsi_header *) &tokens[1];
+   header = (struct tgsi_header *) &tokens[0];
    *header = tgsi_build_header();
-   *(struct tgsi_processor *) &tokens[2] = tgsi_build_processor(TGSI_PROCESSOR_VERTEX, header);
+   *(struct tgsi_processor *) &tokens[1] = tgsi_build_processor(TGSI_PROCESSOR_VERTEX, header);
 
-   ti = 3;
+   ti = 2;
 
    /*
     * decl i0              ; Vertex pos
@@ -361,12 +359,11 @@ create_frame_pred_frag_shader(struct vl_mpeg12_mc_renderer *r)
    assert(r);
 
    tokens = (struct tgsi_token *) malloc(max_tokens * sizeof(struct tgsi_token));
-   *(struct tgsi_version *) &tokens[0] = tgsi_build_version();
-   header = (struct tgsi_header *) &tokens[1];
+   header = (struct tgsi_header *) &tokens[0];
    *header = tgsi_build_header();
-   *(struct tgsi_processor *) &tokens[2] = tgsi_build_processor(TGSI_PROCESSOR_FRAGMENT, header);
+   *(struct tgsi_processor *) &tokens[1] = tgsi_build_processor(TGSI_PROCESSOR_FRAGMENT, header);
 
-   ti = 3;
+   ti = 2;
 
    /*
     * decl i0                      ; Luma texcoords
@@ -415,10 +412,10 @@ create_frame_pred_frag_shader(struct vl_mpeg12_mc_renderer *r)
       ti += tgsi_build_full_instruction(&inst, &tokens[ti], header, max_tokens - ti);
 
       inst = vl_inst2(TGSI_OPCODE_MOV, TGSI_FILE_TEMPORARY, 0, TGSI_FILE_TEMPORARY, 1);
-      inst.FullSrcRegisters[0].SrcRegister.SwizzleX = TGSI_SWIZZLE_X;
-      inst.FullSrcRegisters[0].SrcRegister.SwizzleY = TGSI_SWIZZLE_X;
-      inst.FullSrcRegisters[0].SrcRegister.SwizzleZ = TGSI_SWIZZLE_X;
-      inst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_X << i;
+      inst.Src[0].Register.SwizzleX = TGSI_SWIZZLE_X;
+      inst.Src[0].Register.SwizzleY = TGSI_SWIZZLE_X;
+      inst.Src[0].Register.SwizzleZ = TGSI_SWIZZLE_X;
+      inst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_X << i;
       ti += tgsi_build_full_instruction(&inst, &tokens[ti], header, max_tokens - ti);
    }
 
@@ -470,12 +467,11 @@ create_frame_bi_pred_vert_shader(struct vl_mpeg12_mc_renderer *r)
    assert(r);
 
    tokens = (struct tgsi_token *) malloc(max_tokens * sizeof(struct tgsi_token));
-   *(struct tgsi_version *) &tokens[0] = tgsi_build_version();
-   header = (struct tgsi_header *) &tokens[1];
+   header = (struct tgsi_header *) &tokens[0];
    *header = tgsi_build_header();
-   *(struct tgsi_processor *) &tokens[2] = tgsi_build_processor(TGSI_PROCESSOR_VERTEX, header);
+   *(struct tgsi_processor *) &tokens[1] = tgsi_build_processor(TGSI_PROCESSOR_VERTEX, header);
 
-   ti = 3;
+   ti = 2;
 
    /*
     * decl i0              ; Vertex pos
@@ -561,12 +557,11 @@ create_frame_bi_pred_frag_shader(struct vl_mpeg12_mc_renderer *r)
    assert(r);
 
    tokens = (struct tgsi_token *) malloc(max_tokens * sizeof(struct tgsi_token));
-   *(struct tgsi_version *) &tokens[0] = tgsi_build_version();
-   header = (struct tgsi_header *) &tokens[1];
+   header = (struct tgsi_header *) &tokens[0];
    *header = tgsi_build_header();
-   *(struct tgsi_processor *) &tokens[2] = tgsi_build_processor(TGSI_PROCESSOR_FRAGMENT, header);
+   *(struct tgsi_processor *) &tokens[1] = tgsi_build_processor(TGSI_PROCESSOR_FRAGMENT, header);
 
-   ti = 3;
+   ti = 2;
 
    /*
     * decl i0                      ; Luma texcoords
@@ -620,10 +615,10 @@ create_frame_bi_pred_frag_shader(struct vl_mpeg12_mc_renderer *r)
       ti += tgsi_build_full_instruction(&inst, &tokens[ti], header, max_tokens - ti);
 
       inst = vl_inst2(TGSI_OPCODE_MOV, TGSI_FILE_TEMPORARY, 0, TGSI_FILE_TEMPORARY, 1);
-      inst.FullSrcRegisters[0].SrcRegister.SwizzleX = TGSI_SWIZZLE_X;
-      inst.FullSrcRegisters[0].SrcRegister.SwizzleY = TGSI_SWIZZLE_X;
-      inst.FullSrcRegisters[0].SrcRegister.SwizzleZ = TGSI_SWIZZLE_X;
-      inst.FullDstRegisters[0].DstRegister.WriteMask = TGSI_WRITEMASK_X << i;
+      inst.Src[0].Register.SwizzleX = TGSI_SWIZZLE_X;
+      inst.Src[0].Register.SwizzleY = TGSI_SWIZZLE_X;
+      inst.Src[0].Register.SwizzleZ = TGSI_SWIZZLE_X;
+      inst.Dst[0].Register.WriteMask = TGSI_WRITEMASK_X << i;
       ti += tgsi_build_full_instruction(&inst, &tokens[ti], header, max_tokens - ti);
    }
 
@@ -642,10 +637,10 @@ create_frame_bi_pred_frag_shader(struct vl_mpeg12_mc_renderer *r)
 
    /* lerp t1, c1.x, t1, t2        ; Blend past and future texels */
    inst = vl_inst4(TGSI_OPCODE_LRP, TGSI_FILE_TEMPORARY, 1, TGSI_FILE_CONSTANT, 1, TGSI_FILE_TEMPORARY, 1, TGSI_FILE_TEMPORARY, 2);
-   inst.FullSrcRegisters[0].SrcRegister.SwizzleX = TGSI_SWIZZLE_X;
-   inst.FullSrcRegisters[0].SrcRegister.SwizzleY = TGSI_SWIZZLE_X;
-   inst.FullSrcRegisters[0].SrcRegister.SwizzleZ = TGSI_SWIZZLE_X;
-   inst.FullSrcRegisters[0].SrcRegister.SwizzleW = TGSI_SWIZZLE_X;
+   inst.Src[0].Register.SwizzleX = TGSI_SWIZZLE_X;
+   inst.Src[0].Register.SwizzleY = TGSI_SWIZZLE_X;
+   inst.Src[0].Register.SwizzleZ = TGSI_SWIZZLE_X;
+   inst.Src[0].Register.SwizzleW = TGSI_SWIZZLE_X;
    ti += tgsi_build_full_instruction(&inst, &tokens[ti], header, max_tokens - ti);
 
    /* add o0, t0, t1               ; Add past/future ref and differential to form final output */
@@ -681,7 +676,7 @@ xfer_buffers_map(struct vl_mpeg12_mc_renderer *r)
       (
          r->pipe->screen, r->textures.all[i],
          0, 0, 0, PIPE_TRANSFER_WRITE, 0, 0,
-         r->textures.all[i]->width[0], r->textures.all[i]->height[0]
+         r->textures.all[i]->width0, r->textures.all[i]->height0
       );
 
       r->texels[i] = r->pipe->screen->transfer_map(r->pipe->screen, r->tex_transfer[i]);
@@ -835,26 +830,25 @@ init_buffers(struct vl_mpeg12_mc_renderer *r)
    /* TODO: Accomodate HW that can't do this and also for cases when this isn't precise enough */
    template.format = PIPE_FORMAT_R16_SNORM;
    template.last_level = 0;
-   template.width[0] = r->pot_buffers ?
+   template.width0 = r->pot_buffers ?
       util_next_power_of_two(r->picture_width) : r->picture_width;
-   template.height[0] = r->pot_buffers ?
+   template.height0 = r->pot_buffers ?
       util_next_power_of_two(r->picture_height) : r->picture_height;
-   template.depth[0] = 1;
-   pf_get_block(template.format, &template.block);
+   template.depth0 = 1;
    template.tex_usage = PIPE_TEXTURE_USAGE_SAMPLER | PIPE_TEXTURE_USAGE_DYNAMIC;
 
    r->textures.individual.y = r->pipe->screen->texture_create(r->pipe->screen, &template);
 
    if (r->chroma_format == PIPE_VIDEO_CHROMA_FORMAT_420) {
-      template.width[0] = r->pot_buffers ?
+      template.width0 = r->pot_buffers ?
          util_next_power_of_two(r->picture_width / 2) :
          r->picture_width / 2;
-      template.height[0] = r->pot_buffers ?
+      template.height0 = r->pot_buffers ?
          util_next_power_of_two(r->picture_height / 2) :
          r->picture_height / 2;
    }
    else if (r->chroma_format == PIPE_VIDEO_CHROMA_FORMAT_422)
-      template.height[0] = r->pot_buffers ?
+      template.height0 = r->pot_buffers ?
          util_next_power_of_two(r->picture_height / 2) :
          r->picture_height / 2;
 
@@ -1283,8 +1277,8 @@ flush(struct vl_mpeg12_mc_renderer *r)
       PIPE_BUFFER_USAGE_CPU_WRITE | PIPE_BUFFER_USAGE_DISCARD
    );
 
-   vs_consts->denorm.x = r->surface->width[0];
-   vs_consts->denorm.y = r->surface->height[0];
+   vs_consts->denorm.x = r->surface->width0;
+   vs_consts->denorm.y = r->surface->height0;
 
    pipe_buffer_unmap(r->pipe->screen, r->vs_const_buf.buffer);
 
@@ -1296,8 +1290,8 @@ flush(struct vl_mpeg12_mc_renderer *r)
    if (num_macroblocks[MACROBLOCK_TYPE_INTRA] > 0) {
       r->pipe->set_vertex_buffers(r->pipe, 1, r->vertex_bufs.all);
       r->pipe->set_vertex_elements(r->pipe, 4, r->vertex_elems);
-      r->pipe->set_sampler_textures(r->pipe, 3, r->textures.all);
-      r->pipe->bind_sampler_states(r->pipe, 3, r->samplers.all);
+      r->pipe->set_fragment_sampler_textures(r->pipe, 3, r->textures.all);
+      r->pipe->bind_fragment_sampler_states(r->pipe, 3, r->samplers.all);
       r->pipe->bind_vs_state(r->pipe, r->i_vs);
       r->pipe->bind_fs_state(r->pipe, r->i_fs);
 
@@ -1310,8 +1304,8 @@ flush(struct vl_mpeg12_mc_renderer *r)
       r->pipe->set_vertex_buffers(r->pipe, 2, r->vertex_bufs.all);
       r->pipe->set_vertex_elements(r->pipe, 6, r->vertex_elems);
       r->textures.individual.ref[0] = r->past;
-      r->pipe->set_sampler_textures(r->pipe, 4, r->textures.all);
-      r->pipe->bind_sampler_states(r->pipe, 4, r->samplers.all);
+      r->pipe->set_fragment_sampler_textures(r->pipe, 4, r->textures.all);
+      r->pipe->bind_fragment_sampler_states(r->pipe, 4, r->samplers.all);
       r->pipe->bind_vs_state(r->pipe, r->p_vs[0]);
       r->pipe->bind_fs_state(r->pipe, r->p_fs[0]);
 
@@ -1324,8 +1318,8 @@ flush(struct vl_mpeg12_mc_renderer *r)
       r->pipe->set_vertex_buffers(r->pipe, 2, r->vertex_bufs.all);
       r->pipe->set_vertex_elements(r->pipe, 6, r->vertex_elems);
       r->textures.individual.ref[0] = r->past;
-      r->pipe->set_sampler_textures(r->pipe, 4, r->textures.all);
-      r->pipe->bind_sampler_states(r->pipe, 4, r->samplers.all);
+      r->pipe->set_fragment_sampler_textures(r->pipe, 4, r->textures.all);
+      r->pipe->bind_fragment_sampler_states(r->pipe, 4, r->samplers.all);
       r->pipe->bind_vs_state(r->pipe, r->p_vs[1]);
       r->pipe->bind_fs_state(r->pipe, r->p_fs[1]);
 
@@ -1338,8 +1332,8 @@ flush(struct vl_mpeg12_mc_renderer *r)
       r->pipe->set_vertex_buffers(r->pipe, 2, r->vertex_bufs.all);
       r->pipe->set_vertex_elements(r->pipe, 6, r->vertex_elems);
       r->textures.individual.ref[0] = r->future;
-      r->pipe->set_sampler_textures(r->pipe, 4, r->textures.all);
-      r->pipe->bind_sampler_states(r->pipe, 4, r->samplers.all);
+      r->pipe->set_fragment_sampler_textures(r->pipe, 4, r->textures.all);
+      r->pipe->bind_fragment_sampler_states(r->pipe, 4, r->samplers.all);
       r->pipe->bind_vs_state(r->pipe, r->p_vs[0]);
       r->pipe->bind_fs_state(r->pipe, r->p_fs[0]);
 
@@ -1352,8 +1346,8 @@ flush(struct vl_mpeg12_mc_renderer *r)
       r->pipe->set_vertex_buffers(r->pipe, 2, r->vertex_bufs.all);
       r->pipe->set_vertex_elements(r->pipe, 6, r->vertex_elems);
       r->textures.individual.ref[0] = r->future;
-      r->pipe->set_sampler_textures(r->pipe, 4, r->textures.all);
-      r->pipe->bind_sampler_states(r->pipe, 4, r->samplers.all);
+      r->pipe->set_fragment_sampler_textures(r->pipe, 4, r->textures.all);
+      r->pipe->bind_fragment_sampler_states(r->pipe, 4, r->samplers.all);
       r->pipe->bind_vs_state(r->pipe, r->p_vs[1]);
       r->pipe->bind_fs_state(r->pipe, r->p_fs[1]);
 
@@ -1367,8 +1361,8 @@ flush(struct vl_mpeg12_mc_renderer *r)
       r->pipe->set_vertex_elements(r->pipe, 8, r->vertex_elems);
       r->textures.individual.ref[0] = r->past;
       r->textures.individual.ref[1] = r->future;
-      r->pipe->set_sampler_textures(r->pipe, 5, r->textures.all);
-      r->pipe->bind_sampler_states(r->pipe, 5, r->samplers.all);
+      r->pipe->set_fragment_sampler_textures(r->pipe, 5, r->textures.all);
+      r->pipe->bind_fragment_sampler_states(r->pipe, 5, r->samplers.all);
       r->pipe->bind_vs_state(r->pipe, r->b_vs[0]);
       r->pipe->bind_fs_state(r->pipe, r->b_fs[0]);
 
@@ -1382,8 +1376,8 @@ flush(struct vl_mpeg12_mc_renderer *r)
       r->pipe->set_vertex_elements(r->pipe, 8, r->vertex_elems);
       r->textures.individual.ref[0] = r->past;
       r->textures.individual.ref[1] = r->future;
-      r->pipe->set_sampler_textures(r->pipe, 5, r->textures.all);
-      r->pipe->bind_sampler_states(r->pipe, 5, r->samplers.all);
+      r->pipe->set_fragment_sampler_textures(r->pipe, 5, r->textures.all);
+      r->pipe->bind_fragment_sampler_states(r->pipe, 5, r->samplers.all);
       r->pipe->bind_vs_state(r->pipe, r->b_vs[1]);
       r->pipe->bind_fs_state(r->pipe, r->b_fs[1]);
 
@@ -1450,7 +1444,7 @@ grab_blocks(struct vl_mpeg12_mc_renderer *r, unsigned mbx, unsigned mby,
    assert(r);
    assert(blocks);
 
-   tex_pitch = r->tex_transfer[0]->stride / r->tex_transfer[0]->block.size;
+   tex_pitch = r->tex_transfer[0]->stride / util_format_get_blocksize(r->tex_transfer[0]->texture->format);
    texels = r->texels[0] + mbpy * tex_pitch + mbpx;
 
    for (y = 0; y < 2; ++y) {
@@ -1489,7 +1483,7 @@ grab_blocks(struct vl_mpeg12_mc_renderer *r, unsigned mbx, unsigned mby,
    mbpy /= 2;
 
    for (tb = 0; tb < 2; ++tb) {
-      tex_pitch = r->tex_transfer[tb + 1]->stride / r->tex_transfer[tb + 1]->block.size;
+      tex_pitch = r->tex_transfer[tb + 1]->stride / util_format_get_blocksize(r->tex_transfer[tb + 1]->texture->format);
       texels = r->texels[tb + 1] + mbpy * tex_pitch + mbpx;
 
       if ((cbp >> (1 - tb)) & 1) {
@@ -1633,8 +1627,8 @@ vl_mpeg12_mc_renderer_render_macroblocks(struct vl_mpeg12_mc_renderer
       renderer->past = past;
       renderer->future = future;
       renderer->fence = fence;
-      renderer->surface_tex_inv_size.x = 1.0f / surface->width[0];
-      renderer->surface_tex_inv_size.y = 1.0f / surface->height[0];
+      renderer->surface_tex_inv_size.x = 1.0f / surface->width0;
+      renderer->surface_tex_inv_size.y = 1.0f / surface->height0;
    }
 
    while (num_macroblocks) {
index faa20a903cdd27a60845631e1fa029f0488f1b57..d011ef97bd63815b1ef67026a0ef61cc39f90726 100644 (file)
@@ -36,10 +36,10 @@ struct tgsi_full_declaration vl_decl_input(unsigned int name, unsigned int index
 
    decl.Declaration.File = TGSI_FILE_INPUT;
    decl.Declaration.Semantic = 1;
-   decl.Semantic.SemanticName = name;
-   decl.Semantic.SemanticIndex = index;
-   decl.DeclarationRange.First = first;
-   decl.DeclarationRange.Last = last;
+   decl.Semantic.Name = name;
+   decl.Semantic.Index = index;
+   decl.Range.First = first;
+   decl.Range.Last = last;
 
    return decl;
 }
@@ -64,11 +64,11 @@ struct tgsi_full_declaration vl_decl_interpolated_input
 
    decl.Declaration.File = TGSI_FILE_INPUT;
    decl.Declaration.Semantic = 1;
-   decl.Semantic.SemanticName = name;
-   decl.Semantic.SemanticIndex = index;
+   decl.Semantic.Name = name;
+   decl.Semantic.Index = index;
    decl.Declaration.Interpolate = interpolation;;
-   decl.DeclarationRange.First = first;
-   decl.DeclarationRange.Last = last;
+   decl.Range.First = first;
+   decl.Range.Last = last;
 
    return decl;
 }
@@ -79,10 +79,10 @@ struct tgsi_full_declaration vl_decl_constants(unsigned int name, unsigned int i
 
    decl.Declaration.File = TGSI_FILE_CONSTANT;
    decl.Declaration.Semantic = 1;
-   decl.Semantic.SemanticName = name;
-   decl.Semantic.SemanticIndex = index;
-   decl.DeclarationRange.First = first;
-   decl.DeclarationRange.Last = last;
+   decl.Semantic.Name = name;
+   decl.Semantic.Index = index;
+   decl.Range.First = first;
+   decl.Range.Last = last;
 
    return decl;
 }
@@ -93,10 +93,10 @@ struct tgsi_full_declaration vl_decl_output(unsigned int name, unsigned int inde
 
    decl.Declaration.File = TGSI_FILE_OUTPUT;
    decl.Declaration.Semantic = 1;
-   decl.Semantic.SemanticName = name;
-   decl.Semantic.SemanticIndex = index;
-   decl.DeclarationRange.First = first;
-   decl.DeclarationRange.Last = last;
+   decl.Semantic.Name = name;
+   decl.Semantic.Index = index;
+   decl.Range.First = first;
+   decl.Range.Last = last;
 
    return decl;
 }
@@ -107,8 +107,8 @@ struct tgsi_full_declaration vl_decl_temps(unsigned int first, unsigned int last
 
    decl = tgsi_default_full_declaration();
    decl.Declaration.File = TGSI_FILE_TEMPORARY;
-   decl.DeclarationRange.First = first;
-   decl.DeclarationRange.Last = last;
+   decl.Range.First = first;
+   decl.Range.Last = last;
 
    return decl;
 }
@@ -119,8 +119,8 @@ struct tgsi_full_declaration vl_decl_samplers(unsigned int first, unsigned int l
 
    decl = tgsi_default_full_declaration();
    decl.Declaration.File = TGSI_FILE_SAMPLER;
-   decl.DeclarationRange.First = first;
-   decl.DeclarationRange.Last = last;
+   decl.Range.First = first;
+   decl.Range.Last = last;
 
    return decl;
 }
@@ -138,11 +138,11 @@ struct tgsi_full_instruction vl_inst2
 
    inst.Instruction.Opcode = opcode;
    inst.Instruction.NumDstRegs = 1;
-   inst.FullDstRegisters[0].DstRegister.File = dst_file;
-   inst.FullDstRegisters[0].DstRegister.Index = dst_index;
+   inst.Dst[0].Register.File = dst_file;
+   inst.Dst[0].Register.Index = dst_index;
    inst.Instruction.NumSrcRegs = 1;
-   inst.FullSrcRegisters[0].SrcRegister.File = src_file;
-   inst.FullSrcRegisters[0].SrcRegister.Index = src_index;
+   inst.Src[0].Register.File = src_file;
+   inst.Src[0].Register.Index = src_index;
 
    return inst;
 }
@@ -162,13 +162,13 @@ struct tgsi_full_instruction vl_inst3
 
    inst.Instruction.Opcode = opcode;
    inst.Instruction.NumDstRegs = 1;
-   inst.FullDstRegisters[0].DstRegister.File = dst_file;
-   inst.FullDstRegisters[0].DstRegister.Index = dst_index;
+   inst.Dst[0].Register.File = dst_file;
+   inst.Dst[0].Register.Index = dst_index;
    inst.Instruction.NumSrcRegs = 2;
-   inst.FullSrcRegisters[0].SrcRegister.File = src1_file;
-   inst.FullSrcRegisters[0].SrcRegister.Index = src1_index;
-   inst.FullSrcRegisters[1].SrcRegister.File = src2_file;
-   inst.FullSrcRegisters[1].SrcRegister.Index = src2_index;
+   inst.Src[0].Register.File = src1_file;
+   inst.Src[0].Register.Index = src1_index;
+   inst.Src[1].Register.File = src2_file;
+   inst.Src[1].Register.Index = src2_index;
 
    return inst;
 }
@@ -188,14 +188,15 @@ struct tgsi_full_instruction vl_tex
 
    inst.Instruction.Opcode = TGSI_OPCODE_TEX;
    inst.Instruction.NumDstRegs = 1;
-   inst.FullDstRegisters[0].DstRegister.File = dst_file;
-   inst.FullDstRegisters[0].DstRegister.Index = dst_index;
+   inst.Dst[0].Register.File = dst_file;
+   inst.Dst[0].Register.Index = dst_index;
    inst.Instruction.NumSrcRegs = 2;
-   inst.InstructionExtTexture.Texture = tex;
-   inst.FullSrcRegisters[0].SrcRegister.File = src1_file;
-   inst.FullSrcRegisters[0].SrcRegister.Index = src1_index;
-   inst.FullSrcRegisters[1].SrcRegister.File = src2_file;
-   inst.FullSrcRegisters[1].SrcRegister.Index = src2_index;
+   inst.Instruction.Texture = 1;
+   inst.Texture.Texture = tex;
+   inst.Src[0].Register.File = src1_file;
+   inst.Src[0].Register.Index = src1_index;
+   inst.Src[1].Register.File = src2_file;
+   inst.Src[1].Register.Index = src2_index;
 
    return inst;
 }
@@ -217,15 +218,15 @@ struct tgsi_full_instruction vl_inst4
 
    inst.Instruction.Opcode = opcode;
    inst.Instruction.NumDstRegs = 1;
-   inst.FullDstRegisters[0].DstRegister.File = dst_file;
-   inst.FullDstRegisters[0].DstRegister.Index = dst_index;
+   inst.Dst[0].Register.File = dst_file;
+   inst.Dst[0].Register.Index = dst_index;
    inst.Instruction.NumSrcRegs = 3;
-   inst.FullSrcRegisters[0].SrcRegister.File = src1_file;
-   inst.FullSrcRegisters[0].SrcRegister.Index = src1_index;
-   inst.FullSrcRegisters[1].SrcRegister.File = src2_file;
-   inst.FullSrcRegisters[1].SrcRegister.Index = src2_index;
-   inst.FullSrcRegisters[2].SrcRegister.File = src3_file;
-   inst.FullSrcRegisters[2].SrcRegister.Index = src3_index;
+   inst.Src[0].Register.File = src1_file;
+   inst.Src[0].Register.Index = src1_index;
+   inst.Src[1].Register.File = src2_file;
+   inst.Src[1].Register.Index = src2_index;
+   inst.Src[2].Register.File = src3_file;
+   inst.Src[2].Register.Index = src3_index;
 
    return inst;
 }
index 79ad687ea94d03381e4dc1e529d344c38477ea36..3a3f968a492f8530854c4502d9258c7d55a4160a 100644 (file)
@@ -59,9 +59,9 @@ cell_clear(struct pipe_context *pipe, unsigned buffers, const float *rgba,
 
    if (buffers & PIPE_CLEAR_COLOR) {
       uint surfIndex = 0;
-      uint clearValue;
+      union util_color uc;
 
-      util_pack_color(rgba, cell->framebuffer.cbufs[0]->format, &clearValue);
+      util_pack_color(rgba, cell->framebuffer.cbufs[0]->format, &uc);
 
       /* Build a CLEAR command and place it in the current batch buffer */
       STATIC_ASSERT(sizeof(struct cell_command_clear_surface) % 16 == 0);
@@ -70,7 +70,7 @@ cell_clear(struct pipe_context *pipe, unsigned buffers, const float *rgba,
          cell_batch_alloc16(cell, sizeof(*clr));
       clr->opcode[0] = CELL_CMD_CLEAR_SURFACE;
       clr->surface = surfIndex;
-      clr->value = clearValue;
+      clr->value = uc.ui;
    }
 
    if (buffers & PIPE_CLEAR_DEPTHSTENCIL) {
index 644496db40cd8bc6a94682e051e4cf7a9e850f67..01bea0f8ccee3fe72a3216f0b39bd79499188ce0 100644 (file)
@@ -59,7 +59,7 @@ cell_map_constant_buffers(struct cell_context *sp)
       }
    }
 
-   draw_set_mapped_constant_buffer(sp->draw,
+   draw_set_mapped_constant_buffer(sp->draw, PIPE_SHADER_VERTEX,
                                    sp->mapped_constants[PIPE_SHADER_VERTEX],
                                    sp->constants[PIPE_SHADER_VERTEX].buffer->size);
 }
@@ -171,21 +171,11 @@ cell_draw_arrays(struct pipe_context *pipe, unsigned mode,
 }
 
 
-static void
-cell_set_edgeflags(struct pipe_context *pipe, const unsigned *edgeflags)
-{
-   struct cell_context *cell = cell_context(pipe);
-   draw_set_edgeflags(cell->draw, edgeflags);
-}
-
-
-
 void
 cell_init_draw_functions(struct cell_context *cell)
 {
    cell->pipe.draw_arrays = cell_draw_arrays;
    cell->pipe.draw_elements = cell_draw_elements;
    cell->pipe.draw_range_elements = cell_draw_range_elements;
-   cell->pipe.set_edgeflags = cell_set_edgeflags;
 }
 
index 19e3ab0844030b3c39e101443fe3982ced8157ed..1d8a11a4ac97175a2627cd66dff1f38691e75857 100644 (file)
@@ -237,8 +237,8 @@ is_register_src(struct codegen *gen, int channel,
    if (swizzle > TGSI_SWIZZLE_W || sign_op != TGSI_UTIL_SIGN_KEEP) {
       return FALSE;
    }
-   if (src->SrcRegister.File == TGSI_FILE_TEMPORARY ||
-       src->SrcRegister.File == TGSI_FILE_IMMEDIATE) {
+   if (src->Register.File == TGSI_FILE_TEMPORARY ||
+       src->Register.File == TGSI_FILE_IMMEDIATE) {
       return TRUE;
    }
    return FALSE;
@@ -249,7 +249,7 @@ static boolean
 is_memory_dst(struct codegen *gen, int channel,
               const struct tgsi_full_dst_register *dst)
 {
-   if (dst->DstRegister.File == TGSI_FILE_OUTPUT) {
+   if (dst->Register.File == TGSI_FILE_OUTPUT) {
       return TRUE;
    }
    else {
@@ -279,15 +279,15 @@ get_src_reg(struct codegen *gen,
    assert(swizzle <= TGSI_SWIZZLE_W);
 
    {
-      int index = src->SrcRegister.Index;
+      int index = src->Register.Index;
 
       assert(swizzle < 4);
 
-      if (src->SrcRegister.Indirect) {
+      if (src->Register.Indirect) {
          /* XXX unfinished */
       }
 
-      switch (src->SrcRegister.File) {
+      switch (src->Register.File) {
       case TGSI_FILE_TEMPORARY:
          reg = gen->temp_regs[index][swizzle];
          break;
@@ -374,12 +374,12 @@ get_dst_reg(struct codegen *gen,
 {
    int reg = -1;
 
-   switch (dest->DstRegister.File) {
+   switch (dest->Register.File) {
    case TGSI_FILE_TEMPORARY:
       if (gen->if_nesting > 0 || gen->loop_nesting > 0)
          reg = get_itemp(gen);
       else
-         reg = gen->temp_regs[dest->DstRegister.Index][channel];
+         reg = gen->temp_regs[dest->Register.Index][channel];
       break;
    case TGSI_FILE_OUTPUT:
       reg = get_itemp(gen);
@@ -419,10 +419,10 @@ store_dest_reg(struct codegen *gen,
    }
 #endif
 
-   switch (dest->DstRegister.File) {
+   switch (dest->Register.File) {
    case TGSI_FILE_TEMPORARY:
       if (gen->if_nesting > 0 || gen->loop_nesting > 0) {
-         int d_reg = gen->temp_regs[dest->DstRegister.Index][channel];
+         int d_reg = gen->temp_regs[dest->Register.Index][channel];
          int exec_reg = get_exec_mask_reg(gen);
          /* Mix d with new value according to exec mask:
           * d[i] = mask_reg[i] ? value_reg : d_reg
@@ -437,7 +437,7 @@ store_dest_reg(struct codegen *gen,
    case TGSI_FILE_OUTPUT:
       {
          /* offset is measured in quadwords, not bytes */
-         int offset = dest->DstRegister.Index * 4 + channel;
+         int offset = dest->Register.Index * 4 + channel;
          if (gen->if_nesting > 0 || gen->loop_nesting > 0) {
             int exec_reg = get_exec_mask_reg(gen);
             int curval_reg = get_itemp(gen);
@@ -544,7 +544,7 @@ emit_epilogue(struct codegen *gen)
 
 #define FOR_EACH_ENABLED_CHANNEL(inst, ch) \
    for (ch = 0; ch < 4; ch++) \
-      if (inst->FullDstRegisters[0].DstRegister.WriteMask & (1 << ch))
+      if (inst->Dst[0].Register.WriteMask & (1 << ch))
 
 
 static boolean
@@ -552,7 +552,7 @@ emit_ARL(struct codegen *gen, const struct tgsi_full_instruction *inst)
 {
    int ch = 0, src_reg, addr_reg;
 
-   src_reg = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
+   src_reg = get_src_reg(gen, ch, &inst->Src[0]);
    addr_reg = get_address_reg(gen);
 
    /* convert float to int */
@@ -570,19 +570,19 @@ emit_MOV(struct codegen *gen, const struct tgsi_full_instruction *inst)
    int ch, src_reg[4], dst_reg[4];
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      src_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      dst_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      src_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      dst_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
    }
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      if (is_register_src(gen, ch, &inst->FullSrcRegisters[0]) &&
-          is_memory_dst(gen, ch, &inst->FullDstRegisters[0])) {
+      if (is_register_src(gen, ch, &inst->Src[0]) &&
+          is_memory_dst(gen, ch, &inst->Dst[0])) {
          /* special-case: register to memory store */
-         store_dest_reg(gen, src_reg[ch], ch, &inst->FullDstRegisters[0]);
+         store_dest_reg(gen, src_reg[ch], ch, &inst->Dst[0]);
       }
       else {
          spe_move(gen->f, dst_reg[ch], src_reg[ch]);
-         store_dest_reg(gen, dst_reg[ch], ch, &inst->FullDstRegisters[0]);
+         store_dest_reg(gen, dst_reg[ch], ch, &inst->Dst[0]);
       }
    }
 
@@ -601,9 +601,9 @@ emit_binop(struct codegen *gen, const struct tgsi_full_instruction *inst)
 
    /* Loop over Red/Green/Blue/Alpha channels, fetch src operands */
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s1_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      s2_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[1]);
-      d_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      s1_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      s2_reg[ch] = get_src_reg(gen, ch, &inst->Src[1]);
+      d_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
    }
 
    /* Loop over Red/Green/Blue/Alpha channels, do the op, store results */
@@ -626,7 +626,7 @@ emit_binop(struct codegen *gen, const struct tgsi_full_instruction *inst)
 
    /* Store the result (a no-op for TGSI_FILE_TEMPORARY dests) */
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_reg[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg[ch], ch, &inst->Dst[0]);
    }
 
    /* Free any intermediate temps we allocated */
@@ -645,16 +645,16 @@ emit_MAD(struct codegen *gen, const struct tgsi_full_instruction *inst)
    int ch, s1_reg[4], s2_reg[4], s3_reg[4], d_reg[4];
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s1_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      s2_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[1]);
-      s3_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[2]);
-      d_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      s1_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      s2_reg[ch] = get_src_reg(gen, ch, &inst->Src[1]);
+      s3_reg[ch] = get_src_reg(gen, ch, &inst->Src[2]);
+      d_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
    }
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
       spe_fma(gen->f, d_reg[ch], s1_reg[ch], s2_reg[ch], s3_reg[ch]);
    }
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_reg[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg[ch], ch, &inst->Dst[0]);
    }
    free_itemps(gen);
    return TRUE;
@@ -671,10 +671,10 @@ emit_LRP(struct codegen *gen, const struct tgsi_full_instruction *inst)
 
    /* setup/get src/dst/temp regs */
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s1_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      s2_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[1]);
-      s3_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[2]);
-      d_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      s1_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      s2_reg[ch] = get_src_reg(gen, ch, &inst->Src[1]);
+      s3_reg[ch] = get_src_reg(gen, ch, &inst->Src[2]);
+      d_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
       tmp_reg[ch] = get_itemp(gen);
    }
 
@@ -687,7 +687,7 @@ emit_LRP(struct codegen *gen, const struct tgsi_full_instruction *inst)
       spe_fma(gen->f, d_reg[ch], tmp_reg[ch], s1_reg[ch], s3_reg[ch]);
    }
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_reg[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg[ch], ch, &inst->Dst[0]);
    }
    free_itemps(gen);
    return TRUE;
@@ -704,8 +704,8 @@ emit_RCP_RSQ(struct codegen *gen, const struct tgsi_full_instruction *inst)
    int ch, s1_reg[4], d_reg[4], tmp_reg[4];
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s1_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      d_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      s1_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      d_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
       tmp_reg[ch] = get_itemp(gen);
    }
 
@@ -726,7 +726,7 @@ emit_RCP_RSQ(struct codegen *gen, const struct tgsi_full_instruction *inst)
    }
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_reg[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg[ch], ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -747,8 +747,8 @@ emit_ABS(struct codegen *gen, const struct tgsi_full_instruction *inst)
    spe_load_uint(gen->f, bit31mask_reg, (1 << 31));
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s1_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      d_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      s1_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      d_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
    }
 
    /* d = sign bit cleared in s1 */
@@ -757,7 +757,7 @@ emit_ABS(struct codegen *gen, const struct tgsi_full_instruction *inst)
    }
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_reg[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg[ch], ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -775,12 +775,12 @@ emit_DP3(struct codegen *gen, const struct tgsi_full_instruction *inst)
    int s2x_reg, s2y_reg, s2z_reg;
    int t0_reg = get_itemp(gen), t1_reg = get_itemp(gen);
 
-   s1x_reg = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[0]);
-   s2x_reg = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[1]);
-   s1y_reg = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[0]);
-   s2y_reg = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[1]);
-   s1z_reg = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[0]);
-   s2z_reg = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[1]);
+   s1x_reg = get_src_reg(gen, CHAN_X, &inst->Src[0]);
+   s2x_reg = get_src_reg(gen, CHAN_X, &inst->Src[1]);
+   s1y_reg = get_src_reg(gen, CHAN_Y, &inst->Src[0]);
+   s2y_reg = get_src_reg(gen, CHAN_Y, &inst->Src[1]);
+   s1z_reg = get_src_reg(gen, CHAN_Z, &inst->Src[0]);
+   s2z_reg = get_src_reg(gen, CHAN_Z, &inst->Src[1]);
 
    /* t0 = x0 * x1 */
    spe_fm(gen->f, t0_reg, s1x_reg, s2x_reg);
@@ -795,9 +795,9 @@ emit_DP3(struct codegen *gen, const struct tgsi_full_instruction *inst)
    spe_fa(gen->f, t0_reg, t0_reg, t1_reg);
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      int d_reg = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      int d_reg = get_dst_reg(gen, ch, &inst->Dst[0]);
       spe_move(gen->f, d_reg, t0_reg);
-      store_dest_reg(gen, d_reg, ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg, ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -815,14 +815,14 @@ emit_DP4(struct codegen *gen, const struct tgsi_full_instruction *inst)
    int s1x_reg, s1y_reg, s1z_reg, s1w_reg;
    int t0_reg = get_itemp(gen), t1_reg = get_itemp(gen);
 
-   s0x_reg = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[0]);
-   s1x_reg = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[1]);
-   s0y_reg = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[0]);
-   s1y_reg = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[1]);
-   s0z_reg = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[0]);
-   s1z_reg = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[1]);
-   s0w_reg = get_src_reg(gen, CHAN_W, &inst->FullSrcRegisters[0]);
-   s1w_reg = get_src_reg(gen, CHAN_W, &inst->FullSrcRegisters[1]);
+   s0x_reg = get_src_reg(gen, CHAN_X, &inst->Src[0]);
+   s1x_reg = get_src_reg(gen, CHAN_X, &inst->Src[1]);
+   s0y_reg = get_src_reg(gen, CHAN_Y, &inst->Src[0]);
+   s1y_reg = get_src_reg(gen, CHAN_Y, &inst->Src[1]);
+   s0z_reg = get_src_reg(gen, CHAN_Z, &inst->Src[0]);
+   s1z_reg = get_src_reg(gen, CHAN_Z, &inst->Src[1]);
+   s0w_reg = get_src_reg(gen, CHAN_W, &inst->Src[0]);
+   s1w_reg = get_src_reg(gen, CHAN_W, &inst->Src[1]);
 
    /* t0 = x0 * x1 */
    spe_fm(gen->f, t0_reg, s0x_reg, s1x_reg);
@@ -840,9 +840,9 @@ emit_DP4(struct codegen *gen, const struct tgsi_full_instruction *inst)
    spe_fa(gen->f, t0_reg, t0_reg, t1_reg);
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      int d_reg = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      int d_reg = get_dst_reg(gen, ch, &inst->Dst[0]);
       spe_move(gen->f, d_reg, t0_reg);
-      store_dest_reg(gen, d_reg, ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg, ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -857,31 +857,31 @@ emit_DPH(struct codegen *gen, const struct tgsi_full_instruction *inst)
 {
    /* XXX rewrite this function to look more like DP3/DP4 */
    int ch;
-   int s1_reg = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[0]);
-   int s2_reg = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[1]);
+   int s1_reg = get_src_reg(gen, CHAN_X, &inst->Src[0]);
+   int s2_reg = get_src_reg(gen, CHAN_X, &inst->Src[1]);
    int tmp_reg = get_itemp(gen);
 
    /* t = x0 * x1 */
    spe_fm(gen->f, tmp_reg, s1_reg, s2_reg);
 
-   s1_reg = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[0]);
-   s2_reg = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[1]);
+   s1_reg = get_src_reg(gen, CHAN_Y, &inst->Src[0]);
+   s2_reg = get_src_reg(gen, CHAN_Y, &inst->Src[1]);
    /* t = y0 * y1 + t */
    spe_fma(gen->f, tmp_reg, s1_reg, s2_reg, tmp_reg);
 
-   s1_reg = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[0]);
-   s2_reg = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[1]);
+   s1_reg = get_src_reg(gen, CHAN_Z, &inst->Src[0]);
+   s2_reg = get_src_reg(gen, CHAN_Z, &inst->Src[1]);
    /* t = z0 * z1 + t */
    spe_fma(gen->f, tmp_reg, s1_reg, s2_reg, tmp_reg);
 
-   s2_reg = get_src_reg(gen, CHAN_W, &inst->FullSrcRegisters[1]);
+   s2_reg = get_src_reg(gen, CHAN_W, &inst->Src[1]);
    /* t = w1 + t */
    spe_fa(gen->f, tmp_reg, s2_reg, tmp_reg);
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      int d_reg = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      int d_reg = get_dst_reg(gen, ch, &inst->Dst[0]);
       spe_move(gen->f, d_reg, tmp_reg);
-      store_dest_reg(gen, tmp_reg, ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, tmp_reg, ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -898,9 +898,9 @@ emit_NRM3(struct codegen *gen, const struct tgsi_full_instruction *inst)
    int src_reg[3];
    int t0_reg = get_itemp(gen), t1_reg = get_itemp(gen);
 
-   src_reg[0] = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[0]);
-   src_reg[1] = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[0]);
-   src_reg[2] = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[0]);
+   src_reg[0] = get_src_reg(gen, CHAN_X, &inst->Src[0]);
+   src_reg[1] = get_src_reg(gen, CHAN_Y, &inst->Src[0]);
+   src_reg[2] = get_src_reg(gen, CHAN_Z, &inst->Src[0]);
 
    /* t0 = x * x */
    spe_fm(gen->f, t0_reg, src_reg[0], src_reg[0]);
@@ -919,10 +919,10 @@ emit_NRM3(struct codegen *gen, const struct tgsi_full_instruction *inst)
    spe_fi(gen->f, t1_reg, t0_reg, t1_reg);
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      int d_reg = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      int d_reg = get_dst_reg(gen, ch, &inst->Dst[0]);
       /* dst = src[ch] * t1 */
       spe_fm(gen->f, d_reg, src_reg[ch], t1_reg);
-      store_dest_reg(gen, d_reg, ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg, ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -936,48 +936,48 @@ emit_NRM3(struct codegen *gen, const struct tgsi_full_instruction *inst)
 static boolean
 emit_XPD(struct codegen *gen, const struct tgsi_full_instruction *inst)
 {
-   int s1_reg = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[0]);
-   int s2_reg = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[1]);
+   int s1_reg = get_src_reg(gen, CHAN_Z, &inst->Src[0]);
+   int s2_reg = get_src_reg(gen, CHAN_Y, &inst->Src[1]);
    int tmp_reg = get_itemp(gen);
 
    /* t = z0 * y1 */
    spe_fm(gen->f, tmp_reg, s1_reg, s2_reg);
 
-   s1_reg = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[0]);
-   s2_reg = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[1]);
+   s1_reg = get_src_reg(gen, CHAN_Y, &inst->Src[0]);
+   s2_reg = get_src_reg(gen, CHAN_Z, &inst->Src[1]);
    /* t = y0 * z1 - t */
    spe_fms(gen->f, tmp_reg, s1_reg, s2_reg, tmp_reg);
 
-   if (inst->FullDstRegisters[0].DstRegister.WriteMask & (1 << CHAN_X)) {
-      store_dest_reg(gen, tmp_reg, CHAN_X, &inst->FullDstRegisters[0]);
+   if (inst->Dst[0].Register.WriteMask & (1 << CHAN_X)) {
+      store_dest_reg(gen, tmp_reg, CHAN_X, &inst->Dst[0]);
    }
 
-   s1_reg = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[0]);
-   s2_reg = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[1]);
+   s1_reg = get_src_reg(gen, CHAN_X, &inst->Src[0]);
+   s2_reg = get_src_reg(gen, CHAN_Z, &inst->Src[1]);
    /* t = x0 * z1 */
    spe_fm(gen->f, tmp_reg, s1_reg, s2_reg);
 
-   s1_reg = get_src_reg(gen, CHAN_Z, &inst->FullSrcRegisters[0]);
-   s2_reg = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[1]);
+   s1_reg = get_src_reg(gen, CHAN_Z, &inst->Src[0]);
+   s2_reg = get_src_reg(gen, CHAN_X, &inst->Src[1]);
    /* t = z0 * x1 - t */
    spe_fms(gen->f, tmp_reg, s1_reg, s2_reg, tmp_reg);
 
-   if (inst->FullDstRegisters[0].DstRegister.WriteMask & (1 << CHAN_Y)) {
-      store_dest_reg(gen, tmp_reg, CHAN_Y, &inst->FullDstRegisters[0]);
+   if (inst->Dst[0].Register.WriteMask & (1 << CHAN_Y)) {
+      store_dest_reg(gen, tmp_reg, CHAN_Y, &inst->Dst[0]);
    }
 
-   s1_reg = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[0]);
-   s2_reg = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[1]);
+   s1_reg = get_src_reg(gen, CHAN_Y, &inst->Src[0]);
+   s2_reg = get_src_reg(gen, CHAN_X, &inst->Src[1]);
    /* t = y0 * x1 */
    spe_fm(gen->f, tmp_reg, s1_reg, s2_reg);
 
-   s1_reg = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[0]);
-   s2_reg = get_src_reg(gen, CHAN_Y, &inst->FullSrcRegisters[1]);
+   s1_reg = get_src_reg(gen, CHAN_X, &inst->Src[0]);
+   s2_reg = get_src_reg(gen, CHAN_Y, &inst->Src[1]);
    /* t = x0 * y1 - t */
    spe_fms(gen->f, tmp_reg, s1_reg, s2_reg, tmp_reg);
 
-   if (inst->FullDstRegisters[0].DstRegister.WriteMask & (1 << CHAN_Z)) {
-      store_dest_reg(gen, tmp_reg, CHAN_Z, &inst->FullDstRegisters[0]);
+   if (inst->Dst[0].Register.WriteMask & (1 << CHAN_Z)) {
+      store_dest_reg(gen, tmp_reg, CHAN_Z, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -995,14 +995,14 @@ static boolean
 emit_inequality(struct codegen *gen, const struct tgsi_full_instruction *inst)
 {
    int ch, s1_reg[4], s2_reg[4], d_reg[4], one_reg;
-   bool complement = FALSE;
+   boolean complement = FALSE;
 
    one_reg = get_const_one_reg(gen);
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s1_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      s2_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[1]);
-      d_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      s1_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      s2_reg[ch] = get_src_reg(gen, ch, &inst->Src[1]);
+      d_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
    }
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
@@ -1043,7 +1043,7 @@ emit_inequality(struct codegen *gen, const struct tgsi_full_instruction *inst)
    }
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_reg[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg[ch], ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -1060,10 +1060,10 @@ emit_CMP(struct codegen *gen, const struct tgsi_full_instruction *inst)
    int ch;
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      int s1_reg = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      int s2_reg = get_src_reg(gen, ch, &inst->FullSrcRegisters[1]);
-      int s3_reg = get_src_reg(gen, ch, &inst->FullSrcRegisters[2]);
-      int d_reg = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      int s1_reg = get_src_reg(gen, ch, &inst->Src[0]);
+      int s2_reg = get_src_reg(gen, ch, &inst->Src[1]);
+      int s3_reg = get_src_reg(gen, ch, &inst->Src[2]);
+      int d_reg = get_dst_reg(gen, ch, &inst->Dst[0]);
       int zero_reg = get_itemp(gen);
    
       spe_zero(gen->f, zero_reg);
@@ -1072,7 +1072,7 @@ emit_CMP(struct codegen *gen, const struct tgsi_full_instruction *inst)
       spe_fcgt(gen->f, d_reg, zero_reg, s1_reg);
       spe_selb(gen->f, d_reg, s3_reg, s2_reg, d_reg);
 
-      store_dest_reg(gen, d_reg, ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg, ch, &inst->Dst[0]);
       free_itemps(gen);
    }
 
@@ -1090,8 +1090,8 @@ emit_TRUNC(struct codegen *gen, const struct tgsi_full_instruction *inst)
    int ch, s1_reg[4], d_reg[4];
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s1_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      d_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      s1_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      d_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
    }
 
    /* Convert float to int */
@@ -1105,7 +1105,7 @@ emit_TRUNC(struct codegen *gen, const struct tgsi_full_instruction *inst)
    }
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_reg[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg[ch], ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -1129,8 +1129,8 @@ emit_FLR(struct codegen *gen, const struct tgsi_full_instruction *inst)
    one_reg = get_const_one_reg(gen);
    
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s1_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      d_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      s1_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      d_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
       tmp_reg[ch] = get_itemp(gen);
    }
 
@@ -1156,7 +1156,7 @@ emit_FLR(struct codegen *gen, const struct tgsi_full_instruction *inst)
    }
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_reg[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg[ch], ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -1177,8 +1177,8 @@ emit_FRC(struct codegen *gen, const struct tgsi_full_instruction *inst)
    one_reg = get_const_one_reg(gen);
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s1_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      d_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      s1_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      d_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
       tmp_reg[ch] = get_itemp(gen);
    }
 
@@ -1210,7 +1210,7 @@ emit_FRC(struct codegen *gen, const struct tgsi_full_instruction *inst)
 
    /* store result */
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_reg[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg[ch], ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -1272,7 +1272,7 @@ emit_function_call(struct codegen *gen,
 
    if (scalar) {
       for (a = 0; a < num_args; a++) {
-         s_regs[a] = get_src_reg(gen, CHAN_X, &inst->FullSrcRegisters[a]);
+         s_regs[a] = get_src_reg(gen, CHAN_X, &inst->Src[a]);
       }
       /* we'll call the function, put the return value in this register,
        * then replicate it across all write-enabled components in d_reg.
@@ -1287,11 +1287,11 @@ emit_function_call(struct codegen *gen,
 
       if (!scalar) {
          for (a = 0; a < num_args; a++) {
-            s_regs[a] = get_src_reg(gen, ch, &inst->FullSrcRegisters[a]);
+            s_regs[a] = get_src_reg(gen, ch, &inst->Src[a]);
          }
       }
 
-      d_reg = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      d_reg = get_dst_reg(gen, ch, &inst->Dst[0]);
 
       if (!scalar || !func_called) {
          /* for a scalar function, we'll really only call the function once */
@@ -1336,7 +1336,7 @@ emit_function_call(struct codegen *gen,
          spe_move(gen->f, d_reg, retval_reg);
       }
 
-      store_dest_reg(gen, d_reg, ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg, ch, &inst->Dst[0]);
       free_itemps(gen);
    }
 
@@ -1351,8 +1351,8 @@ emit_function_call(struct codegen *gen,
 static boolean
 emit_TEX(struct codegen *gen, const struct tgsi_full_instruction *inst)
 {
-   const uint target = inst->InstructionExtTexture.Texture;
-   const uint unit = inst->FullSrcRegisters[1].SrcRegister.Index;
+   const uint target = inst->Texture.Texture;
+   const uint unit = inst->Src[1].Register.Index;
    uint addr;
    int ch;
    int coord_regs[4], d_regs[4];
@@ -1373,14 +1373,14 @@ emit_TEX(struct codegen *gen, const struct tgsi_full_instruction *inst)
       return FALSE;
    }
 
-   assert(inst->FullSrcRegisters[1].SrcRegister.File == TGSI_FILE_SAMPLER);
+   assert(inst->Src[1].Register.File == TGSI_FILE_SAMPLER);
 
    spe_comment(gen->f, -4, "CALL tex:");
 
    /* get src/dst reg info */
    for (ch = 0; ch < 4; ch++) {
-      coord_regs[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      d_regs[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      coord_regs[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      d_regs[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
    }
 
    {
@@ -1425,7 +1425,7 @@ emit_TEX(struct codegen *gen, const struct tgsi_full_instruction *inst)
    }
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_regs[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_regs[ch], ch, &inst->Dst[0]);
       free_itemps(gen);
    }
 
@@ -1452,7 +1452,7 @@ emit_KIL(struct codegen *gen, const struct tgsi_full_instruction *inst)
 
    /* get src regs */
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s_regs[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
+      s_regs[ch] = get_src_reg(gen, ch, &inst->Src[0]);
    }
 
    /* test if any src regs are < 0 */
@@ -1500,9 +1500,9 @@ emit_MIN_MAX(struct codegen *gen, const struct tgsi_full_instruction *inst)
    int ch, s0_reg[4], s1_reg[4], d_reg[4], tmp_reg[4];
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      s0_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      s1_reg[ch] = get_src_reg(gen, ch, &inst->FullSrcRegisters[1]);
-      d_reg[ch] = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      s0_reg[ch] = get_src_reg(gen, ch, &inst->Src[0]);
+      s1_reg[ch] = get_src_reg(gen, ch, &inst->Src[1]);
+      d_reg[ch] = get_dst_reg(gen, ch, &inst->Dst[0]);
       tmp_reg[ch] = get_itemp(gen);         
    }
 
@@ -1518,7 +1518,7 @@ emit_MIN_MAX(struct codegen *gen, const struct tgsi_full_instruction *inst)
    }
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      store_dest_reg(gen, d_reg[ch], ch, &inst->FullDstRegisters[0]);
+      store_dest_reg(gen, d_reg[ch], ch, &inst->Dst[0]);
    }
 
    free_itemps(gen);
@@ -1575,7 +1575,7 @@ emit_IF(struct codegen *gen, const struct tgsi_full_instruction *inst)
 
    /* update conditional execution mask with the predicate register */
    int tmp_reg = get_itemp(gen);
-   int s1_reg = get_src_reg(gen, channel, &inst->FullSrcRegisters[0]);
+   int s1_reg = get_src_reg(gen, channel, &inst->Src[0]);
 
    /* tmp = (s1_reg == 0) */
    spe_ceqi(gen->f, tmp_reg, s1_reg, 0);
@@ -1699,8 +1699,8 @@ emit_DDX_DDY(struct codegen *gen, const struct tgsi_full_instruction *inst,
    int ch;
 
    FOR_EACH_ENABLED_CHANNEL(inst, ch) {
-      int s_reg = get_src_reg(gen, ch, &inst->FullSrcRegisters[0]);
-      int d_reg = get_dst_reg(gen, ch, &inst->FullDstRegisters[0]);
+      int s_reg = get_src_reg(gen, ch, &inst->Src[0]);
+      int d_reg = get_dst_reg(gen, ch, &inst->Dst[0]);
 
       int t1_reg = get_itemp(gen);
       int t2_reg = get_itemp(gen);
@@ -1909,8 +1909,8 @@ emit_declaration(struct cell_context *cell,
 
    switch (decl->Declaration.File) {
    case TGSI_FILE_TEMPORARY:
-      for (i = decl->DeclarationRange.First;
-           i <= decl->DeclarationRange.Last;
+      for (i = decl->Range.First;
+           i <= decl->Range.Last;
            i++) {
          assert(i < MAX_TEMPS);
          for (ch = 0; ch < 4; ch++) {
index ccd0fef6e84ef479702a91ad2b2cf43d849e35f8..c18a5d0635e9264713f0cc9268995546f1264cd9 100644 (file)
@@ -383,10 +383,10 @@ cell_init_state_functions(struct cell_context *cell)
    cell->pipe.delete_blend_state = cell_delete_blend_state;
 
    cell->pipe.create_sampler_state = cell_create_sampler_state;
-   cell->pipe.bind_sampler_states = cell_bind_sampler_states;
+   cell->pipe.bind_fragment_sampler_states = cell_bind_sampler_states;
    cell->pipe.delete_sampler_state = cell_delete_sampler_state;
 
-   cell->pipe.set_sampler_textures = cell_set_sampler_textures;
+   cell->pipe.set_fragment_sampler_textures = cell_set_sampler_textures;
 
    cell->pipe.create_depth_stencil_alpha_state = cell_create_depth_stencil_alpha_state;
    cell->pipe.bind_depth_stencil_alpha_state   = cell_bind_depth_stencil_alpha_state;
index efc4f78364b440dcf4e35f91a5a3f8d0b310aa1b..b723e794e71cf25cdbd73966b4030a21f6177f64 100644 (file)
@@ -66,7 +66,7 @@ calculate_vertex_layout( struct cell_context *cell )
    vinfo->num_attribs = 0;
 
    /* we always want to emit vertex pos */
-   src = draw_find_vs_output(cell->draw, TGSI_SEMANTIC_POSITION, 0);
+   src = draw_find_shader_output(cell->draw, TGSI_SEMANTIC_POSITION, 0);
    assert(src >= 0);
    draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_POS, src);
 
@@ -82,14 +82,14 @@ calculate_vertex_layout( struct cell_context *cell )
          break;
 
       case TGSI_SEMANTIC_COLOR:
-         src = draw_find_vs_output(cell->draw, TGSI_SEMANTIC_COLOR, 
+         src = draw_find_shader_output(cell->draw, TGSI_SEMANTIC_COLOR, 
                                    fs->info.input_semantic_index[i]);
          assert(src >= 0);
          draw_emit_vertex_attr(vinfo, EMIT_4F, colorInterp, src);
          break;
 
       case TGSI_SEMANTIC_FOG:
-         src = draw_find_vs_output(cell->draw, TGSI_SEMANTIC_FOG, 0);
+         src = draw_find_shader_output(cell->draw, TGSI_SEMANTIC_FOG, 0);
 #if 1
          if (src < 0) /* XXX temp hack, try demos/fogcoord.c with this */
             src = 0;
@@ -100,7 +100,7 @@ calculate_vertex_layout( struct cell_context *cell )
 
       case TGSI_SEMANTIC_GENERIC:
          /* this includes texcoords and varying vars */
-         src = draw_find_vs_output(cell->draw, TGSI_SEMANTIC_GENERIC,
+         src = draw_find_shader_output(cell->draw, TGSI_SEMANTIC_GENERIC,
                               fs->info.input_semantic_index[i]);
          assert(src >= 0);
          draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_PERSPECTIVE, src);
index 9479c0898fd76c2f6a3b3a205b6d1ecde3d0f764..5b87286d4c5ea7428d23b5c3311dbdc63feddba7 100644 (file)
@@ -27,6 +27,7 @@
 
 #include "pipe/p_inlines.h"
 #include "util/u_memory.h"
+#include "util/u_math.h"
 #include "cell_context.h"
 #include "cell_gen_fragment.h"
 #include "cell_state.h"
@@ -299,9 +300,9 @@ cell_emit_state(struct cell_context *cell)
                for (level = 0; level < CELL_MAX_TEXTURE_LEVELS; level++) {
                   texture->start[level] = (ct->mapped +
                                            ct->level_offset[level]);
-                  texture->width[level] = ct->base.width[level];
-                  texture->height[level] = ct->base.height[level];
-                  texture->depth[level] = ct->base.depth[level];
+                  texture->width[level] = u_minify(ct->base.width0, level);
+                  texture->height[level] = u_minify(ct->base.height0, level);
+                  texture->depth[level] = u_minify(ct->base.depth0, level);
                }
                texture->target = ct->base.target;
             }
@@ -330,7 +331,7 @@ cell_emit_state(struct cell_context *cell)
       const struct draw_context *const draw = cell->draw;
       struct cell_shader_info info;
 
-      info.num_outputs = draw_num_vs_outputs(draw);
+      info.num_outputs = draw_num_shader_outputs(draw);
       info.declarations = (uintptr_t) draw->vs.machine.Declarations;
       info.num_declarations = draw->vs.machine.NumDeclarations;
       info.instructions = (uintptr_t) draw->vs.machine.Instructions;
index ae4c61efb3bee91e072a545578a6de91ae13c442..998944f77a399717a86e06d5b50f1b794f3432ce 100644 (file)
@@ -35,6 +35,8 @@
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
 #include "pipe/internal/p_winsys_screen.h"
+
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -49,9 +51,9 @@ cell_texture_layout(struct cell_texture *ct)
 {
    struct pipe_texture *pt = &ct->base;
    unsigned level;
-   unsigned width = pt->width[0];
-   unsigned height = pt->height[0];
-   unsigned depth = pt->depth[0];
+   unsigned width = pt->width0;
+   unsigned height = pt->height0;
+   unsigned depth = pt->depth0;
 
    ct->buffer_size = 0;
 
@@ -65,17 +67,11 @@ cell_texture_layout(struct cell_texture *ct)
       w_tile = align(width, TILE_SIZE);
       h_tile = align(height, TILE_SIZE);
 
-      pt->width[level] = width;
-      pt->height[level] = height;
-      pt->depth[level] = depth;
-      pt->nblocksx[level] = pf_get_nblocksx(&pt->block, w_tile);  
-      pt->nblocksy[level] = pf_get_nblocksy(&pt->block, h_tile);  
-
-      ct->stride[level] = pt->nblocksx[level] * pt->block.size;
+      ct->stride[level] = util_format_get_stride(pt->format, w_tile);
 
       ct->level_offset[level] = ct->buffer_size;
 
-      size = pt->nblocksx[level] * pt->nblocksy[level] * pt->block.size;
+      size = ct->stride[level] * util_format_get_nblocksy(pt->format, h_tile);
       if (pt->target == PIPE_TEXTURE_CUBE)
          size *= 6;
       else
@@ -83,9 +79,9 @@ cell_texture_layout(struct cell_texture *ct)
 
       ct->buffer_size += size;
 
-      width = minify(width);
-      height = minify(height);
-      depth = minify(depth);
+      width = u_minify(width, 1);
+      height = u_minify(height, 1);
+      depth = u_minify(depth, 1);
    }
 }
 
@@ -276,8 +272,8 @@ cell_get_tex_surface(struct pipe_screen *screen,
       pipe_reference_init(&ps->reference, 1);
       pipe_texture_reference(&ps->texture, pt);
       ps->format = pt->format;
-      ps->width = pt->width[level];
-      ps->height = pt->height[level];
+      ps->width = u_minify(pt->width0, level);
+      ps->height = u_minify(pt->height0, level);
       ps->offset = ct->level_offset[level];
       /* XXX may need to override usage flags (see sp_texture.c) */
       ps->usage = usage;
@@ -286,10 +282,12 @@ cell_get_tex_surface(struct pipe_screen *screen,
       ps->zslice = zslice;
 
       if (pt->target == PIPE_TEXTURE_CUBE) {
-         ps->offset += face * pt->nblocksy[level] * ct->stride[level];
+         unsigned h_tile = align(ps->height, TILE_SIZE);
+         ps->offset += face * util_format_get_nblocksy(ps->format, h_tile) * ct->stride[level];
       }
       else if (pt->target == PIPE_TEXTURE_3D) {
-         ps->offset += zslice * pt->nblocksy[level] * ct->stride[level];
+         unsigned h_tile = align(ps->height, TILE_SIZE);
+         ps->offset += zslice * util_format_get_nblocksy(ps->format, h_tile) * ct->stride[level];
       }
       else {
          assert(face == 0);
@@ -330,14 +328,10 @@ cell_get_tex_transfer(struct pipe_screen *screen,
    if (ctrans) {
       struct pipe_transfer *pt = &ctrans->base;
       pipe_texture_reference(&pt->texture, texture);
-      pt->format = texture->format;
-      pt->block = texture->block;
       pt->x = x;
       pt->y = y;
       pt->width = w;
       pt->height = h;
-      pt->nblocksx = texture->nblocksx[level];
-      pt->nblocksy = texture->nblocksy[level];
       pt->stride = ct->stride[level];
       pt->usage = usage;
       pt->face = face;
@@ -347,10 +341,12 @@ cell_get_tex_transfer(struct pipe_screen *screen,
       ctrans->offset = ct->level_offset[level];
 
       if (texture->target == PIPE_TEXTURE_CUBE) {
-         ctrans->offset += face * pt->nblocksy * pt->stride;
+         unsigned h_tile = align(u_minify(texture->height0, level), TILE_SIZE);
+         ctrans->offset += face * util_format_get_nblocksy(texture->format, h_tile) * pt->stride;
       }
       else if (texture->target == PIPE_TEXTURE_3D) {
-         ctrans->offset += zslice * pt->nblocksy * pt->stride;
+         unsigned h_tile = align(u_minify(texture->height0, level), TILE_SIZE);
+         ctrans->offset += zslice * util_format_get_nblocksy(texture->format, h_tile) * pt->stride;
       }
       else {
          assert(face == 0);
@@ -386,8 +382,8 @@ cell_transfer_map(struct pipe_screen *screen, struct pipe_transfer *transfer)
    struct pipe_texture *pt = transfer->texture;
    struct cell_texture *ct = cell_texture(pt);
    const uint level = ctrans->base.level;
-   const uint texWidth = pt->width[level];
-   const uint texHeight = pt->height[level];
+   const uint texWidth = u_minify(pt->width0, level);
+   const uint texHeight = u_minify(pt->height0, level);
    const uint stride = ct->stride[level];
    unsigned size;
 
@@ -403,7 +399,8 @@ cell_transfer_map(struct pipe_screen *screen, struct pipe_transfer *transfer)
     * Create a buffer of ordinary memory for the linear texture.
     * This is the memory that the user will read/write.
     */
-   size = pt->nblocksx[level] * pt->nblocksy[level] * pt->block.size;
+   size = util_format_get_stride(pt->format, align(texWidth, TILE_SIZE)) *
+          util_format_get_nblocksy(pt->format, align(texHeight, TILE_SIZE));
 
    ctrans->map = align_malloc(size, 16);
    if (!ctrans->map)
@@ -411,7 +408,7 @@ cell_transfer_map(struct pipe_screen *screen, struct pipe_transfer *transfer)
 
    if (transfer->usage & PIPE_TRANSFER_READ) {
       /* need to untwiddle the texture to make a linear version */
-      const uint bpp = pf_get_size(ct->base.format);
+      const uint bpp = util_format_get_blocksize(ct->base.format);
       if (bpp == 4) {
          const uint *src = (uint *) (ct->mapped + ctrans->offset);
          uint *dst = ctrans->map;
@@ -440,8 +437,8 @@ cell_transfer_unmap(struct pipe_screen *screen,
    struct pipe_texture *pt = transfer->texture;
    struct cell_texture *ct = cell_texture(pt);
    const uint level = ctrans->base.level;
-   const uint texWidth = pt->width[level];
-   const uint texHeight = pt->height[level];
+   const uint texWidth = u_minify(pt->width0, level);
+   const uint texHeight = u_minify(pt->height0, level);
    const uint stride = ct->stride[level];
 
    if (!ct->mapped) {
@@ -454,7 +451,7 @@ cell_transfer_unmap(struct pipe_screen *screen,
       /* The user wrote new texture data into the mapped buffer.
        * We need to convert the new linear data into the twiddled/tiled format.
        */
-      const uint bpp = pf_get_size(ct->base.format);
+      const uint bpp = util_format_get_blocksize(ct->base.format);
       if (bpp == 4) {
          const uint *src = ctrans->map;
          uint *dst = (uint *) (ct->mapped + ctrans->offset);
index 4c32b2d06d78bb5ecd22015a9e46f0bfc3f0841b..5ed330aa6ec117abb8879b5fca9bb3b34f07715c 100644 (file)
    for (CHAN = 0; CHAN < 4; CHAN++)
 
 #define IS_CHANNEL_ENABLED(INST, CHAN)\
-   ((INST).FullDstRegisters[0].DstRegister.WriteMask & (1 << (CHAN)))
+   ((INST).Dst[0].Register.WriteMask & (1 << (CHAN)))
 
 #define IS_CHANNEL_ENABLED2(INST, CHAN)\
-   ((INST).FullDstRegisters[1].DstRegister.WriteMask & (1 << (CHAN)))
+   ((INST).Dst[1].Register.WriteMask & (1 << (CHAN)))
 
 #define FOR_EACH_ENABLED_CHANNEL(INST, CHAN)\
    FOR_EACH_CHANNEL( CHAN )\
@@ -431,22 +431,22 @@ fetch_source(
    index.i[0] =
    index.i[1] =
    index.i[2] =
-   index.i[3] = reg->SrcRegister.Index;
+   index.i[3] = reg->Register.Index;
 
-   if (reg->SrcRegister.Indirect) {
+   if (reg->Register.Indirect) {
       union spu_exec_channel index2;
       union spu_exec_channel indir_index;
 
       index2.i[0] =
       index2.i[1] =
       index2.i[2] =
-      index2.i[3] = reg->SrcRegisterInd.Index;
+      index2.i[3] = reg->Indirect.Index;
 
-      swizzle = tgsi_util_get_src_register_swizzle(&reg->SrcRegisterInd,
+      swizzle = tgsi_util_get_src_register_swizzle(&reg->Indirect,
                                                    CHAN_X);
       fetch_src_file_channel(
          mach,
-         reg->SrcRegisterInd.File,
+         reg->Indirect.File,
          swizzle,
          &index2,
          &indir_index );
@@ -454,8 +454,8 @@ fetch_source(
       index.q = si_a(index.q, indir_index.q);
    }
 
-   if( reg->SrcRegister.Dimension ) {
-      switch( reg->SrcRegister.File ) {
+   if( reg->Register.Dimension ) {
+      switch( reg->Register.File ) {
       case TGSI_FILE_INPUT:
          index.q = si_mpyi(index.q, 17);
          break;
@@ -466,24 +466,24 @@ fetch_source(
          ASSERT( 0 );
       }
 
-      index.i[0] += reg->SrcRegisterDim.Index;
-      index.i[1] += reg->SrcRegisterDim.Index;
-      index.i[2] += reg->SrcRegisterDim.Index;
-      index.i[3] += reg->SrcRegisterDim.Index;
+      index.i[0] += reg->Dimension.Index;
+      index.i[1] += reg->Dimension.Index;
+      index.i[2] += reg->Dimension.Index;
+      index.i[3] += reg->Dimension.Index;
 
-      if (reg->SrcRegisterDim.Indirect) {
+      if (reg->Dimension.Indirect) {
          union spu_exec_channel index2;
          union spu_exec_channel indir_index;
 
          index2.i[0] =
          index2.i[1] =
          index2.i[2] =
-         index2.i[3] = reg->SrcRegisterDimInd.Index;
+         index2.i[3] = reg->DimIndirect.Index;
 
-         swizzle = tgsi_util_get_src_register_swizzle( &reg->SrcRegisterDimInd, CHAN_X );
+         swizzle = tgsi_util_get_src_register_swizzle( &reg->DimIndirect, CHAN_X );
          fetch_src_file_channel(
             mach,
-            reg->SrcRegisterDimInd.File,
+            reg->DimIndirect.File,
             swizzle,
             &index2,
             &indir_index );
@@ -495,7 +495,7 @@ fetch_source(
    swizzle = tgsi_util_get_full_src_register_swizzle( reg, chan_index );
    fetch_src_file_channel(
       mach,
-      reg->SrcRegister.File,
+      reg->Register.File,
       swizzle,
       &index,
       chan );
@@ -517,7 +517,7 @@ fetch_source(
       break;
    }
 
-   if (reg->SrcRegisterExtMod.Complement) {
+   if (reg->RegisterExtMod.Complement) {
       chan->q = si_fs(mach->Temps[TEMP_1_I].xyzw[TEMP_1_C].q, chan->q);
    }
 }
@@ -532,21 +532,21 @@ store_dest(
 {
    union spu_exec_channel *dst;
 
-   switch( reg->DstRegister.File ) {
+   switch( reg->Register.File ) {
    case TGSI_FILE_NULL:
       return;
 
    case TGSI_FILE_OUTPUT:
       dst = &mach->Outputs[mach->Temps[TEMP_OUTPUT_I].xyzw[TEMP_OUTPUT_C].u[0]
-                           + reg->DstRegister.Index].xyzw[chan_index];
+                           + reg->Register.Index].xyzw[chan_index];
       break;
 
    case TGSI_FILE_TEMPORARY:
-      dst = &mach->Temps[reg->DstRegister.Index].xyzw[chan_index];
+      dst = &mach->Temps[reg->Register.Index].xyzw[chan_index];
       break;
 
    case TGSI_FILE_ADDRESS:
-      dst = &mach->Addrs[reg->DstRegister.Index].xyzw[chan_index];
+      dst = &mach->Addrs[reg->Register.Index].xyzw[chan_index];
       break;
 
    default:
@@ -583,10 +583,10 @@ store_dest(
 }
 
 #define FETCH(VAL,INDEX,CHAN)\
-    fetch_source (mach, VAL, &inst->FullSrcRegisters[INDEX], CHAN)
+    fetch_source (mach, VAL, &inst->Src[INDEX], CHAN)
 
 #define STORE(VAL,INDEX,CHAN)\
-    store_dest (mach, VAL, &inst->FullDstRegisters[INDEX], inst, CHAN )
+    store_dest (mach, VAL, &inst->Dst[INDEX], inst, CHAN )
 
 
 /**
@@ -612,7 +612,7 @@ exec_kil(struct spu_exec_machine *mach,
 
       /* unswizzle channel */
       swizzle = tgsi_util_get_full_src_register_swizzle (
-                        &inst->FullSrcRegisters[0],
+                        &inst->Src[0],
                         chan_index);
 
       /* check if the component has not been already tested */
@@ -677,7 +677,7 @@ exec_tex(struct spu_exec_machine *mach,
          const struct tgsi_full_instruction *inst,
          boolean biasLod, boolean projected)
 {
-   const uint unit = inst->FullSrcRegisters[1].SrcRegister.Index;
+   const uint unit = inst->Src[1].Register.Index;
    union spu_exec_channel r[8];
    uint chan_index;
    float lodBias;
@@ -833,8 +833,8 @@ exec_declaration(struct spu_exec_machine *mach,
          unsigned first, last, mask;
          interpolation_func interp;
 
-         first = decl->DeclarationRange.First;
-         last = decl->DeclarationRange.Last;
+         first = decl->Range.First;
+         last = decl->Range.Last;
          mask = decl->Declaration.UsageMask;
 
          switch( decl->Declaration.Interpolate ) {
index c2c32b22d5a9dec51f133492e3952f95cc1bb8fb..24057e29e36317dcb42ee1a56beaf10cd082fee4 100644 (file)
@@ -33,7 +33,7 @@ tgsi_util_get_full_src_register_swizzle(
    unsigned component )
 {
    return tgsi_util_get_src_register_swizzle(
-      reg->SrcRegister,
+      reg->Register,
       component );
 }
 
@@ -45,10 +45,10 @@ tgsi_util_get_full_src_register_sign_mode(
 {
    unsigned sign_mode;
 
-   if( reg->SrcRegisterExtMod.Absolute ) {
+   if( reg->RegisterExtMod.Absolute ) {
       /* Consider only the post-abs negation. */
 
-      if( reg->SrcRegisterExtMod.Negate ) {
+      if( reg->RegisterExtMod.Negate ) {
          sign_mode = TGSI_UTIL_SIGN_SET;
       }
       else {
@@ -60,8 +60,8 @@ tgsi_util_get_full_src_register_sign_mode(
 
       unsigned negate;
 
-      negate = reg->SrcRegister.Negate;
-      if( reg->SrcRegisterExtMod.Negate ) {
+      negate = reg->Register.Negate;
+      if( reg->RegisterExtMod.Negate ) {
          negate = !negate;
       }
 
index 9ba86ba86649462b3bd050f4d9194451b07bb13a..149393712a3c90aff853f6fd6ee55a77e82971a2 100644 (file)
@@ -72,6 +72,7 @@ struct failover_context {
     */
    const struct fo_state     *blend;
    const struct fo_state     *sampler[PIPE_MAX_SAMPLERS];
+   const struct fo_state     *vertex_samplers[PIPE_MAX_VERTEX_SAMPLERS];
    const struct fo_state     *depth_stencil;
    const struct fo_state     *rasterizer;
    const struct fo_state     *fragment_shader;
@@ -83,6 +84,7 @@ struct failover_context {
    struct pipe_poly_stipple poly_stipple;
    struct pipe_scissor_state scissor;
    struct pipe_texture *texture[PIPE_MAX_SAMPLERS];
+   struct pipe_texture *vertex_textures[PIPE_MAX_VERTEX_SAMPLERS];
    struct pipe_viewport_state viewport;
    struct pipe_vertex_buffer vertex_buffers[PIPE_MAX_ATTRIBS];
    struct pipe_vertex_element vertex_elements[PIPE_MAX_ATTRIBS];
@@ -92,11 +94,15 @@ struct failover_context {
 
    void *sw_sampler_state[PIPE_MAX_SAMPLERS];
    void *hw_sampler_state[PIPE_MAX_SAMPLERS];
+   void *sw_vertex_sampler_state[PIPE_MAX_VERTEX_SAMPLERS];
+   void *hw_vertex_sampler_state[PIPE_MAX_VERTEX_SAMPLERS];
 
    unsigned dirty;
 
    unsigned num_samplers;
+   unsigned num_vertex_samplers;
    unsigned num_textures;
+   unsigned num_vertex_textures;
 
    unsigned mode;
    struct pipe_context *hw;
index c8eb92629943cbdc509c0355228a048682f25c47..3f5f5560323260612e3d0c99ec052b83716b8fa0 100644 (file)
@@ -322,8 +322,9 @@ failover_create_sampler_state(struct pipe_context *pipe,
 }
 
 static void
-failover_bind_sampler_states(struct pipe_context *pipe,
-                             unsigned num, void **sampler)
+failover_bind_fragment_sampler_states(struct pipe_context *pipe,
+                                      unsigned num,
+                                      void **sampler)
 {
    struct failover_context *failover = failover_context(pipe);
    struct fo_state *state = (struct fo_state*)sampler;
@@ -339,10 +340,40 @@ failover_bind_sampler_states(struct pipe_context *pipe,
    }
    failover->dirty |= FO_NEW_SAMPLER;
    failover->num_samplers = num;
-   failover->sw->bind_sampler_states(failover->sw, num,
-                                     failover->sw_sampler_state);
-   failover->hw->bind_sampler_states(failover->hw, num,
-                                     failover->hw_sampler_state);
+   failover->sw->bind_fragment_sampler_states(failover->sw, num,
+                                              failover->sw_sampler_state);
+   failover->hw->bind_fragment_sampler_states(failover->hw, num,
+                                              failover->hw_sampler_state);
+}
+
+static void
+failover_bind_vertex_sampler_states(struct pipe_context *pipe,
+                                    unsigned num_samplers,
+                                    void **samplers)
+{
+   struct failover_context *failover = failover_context(pipe);
+   struct fo_state *state = (struct fo_state*)samplers;
+   uint i;
+
+   assert(num_samplers <= PIPE_MAX_VERTEX_SAMPLERS);
+
+   /* Check for no-op */
+   if (num_samplers == failover->num_vertex_samplers &&
+       !memcmp(failover->vertex_samplers, samplers, num_samplers * sizeof(void *))) {
+      return;
+   }
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      failover->sw_vertex_sampler_state[i] = i < num_samplers ? state[i].sw_state : NULL;
+      failover->hw_vertex_sampler_state[i] = i < num_samplers ? state[i].hw_state : NULL;
+   }
+   failover->dirty |= FO_NEW_SAMPLER;
+   failover->num_vertex_samplers = num_samplers;
+   failover->sw->bind_vertex_sampler_states(failover->sw,
+                                            num_samplers,
+                                            failover->sw_vertex_sampler_state);
+   failover->hw->bind_vertex_sampler_states(failover->hw,
+                                            num_samplers,
+                                            failover->hw_vertex_sampler_state);
 }
 
 static void
@@ -360,9 +391,9 @@ failover_delete_sampler_state(struct pipe_context *pipe, void *sampler)
 
 
 static void
-failover_set_sampler_textures(struct pipe_context *pipe,
-                              unsigned num,
-                              struct pipe_texture **texture)
+failover_set_fragment_sampler_textures(struct pipe_context *pipe,
+                                       unsigned num,
+                                       struct pipe_texture **texture)
 {
    struct failover_context *failover = failover_context(pipe);
    uint i;
@@ -381,8 +412,38 @@ failover_set_sampler_textures(struct pipe_context *pipe,
                              NULL);
    failover->dirty |= FO_NEW_TEXTURE;
    failover->num_textures = num;
-   failover->sw->set_sampler_textures( failover->sw, num, texture );
-   failover->hw->set_sampler_textures( failover->hw, num, texture );
+   failover->sw->set_fragment_sampler_textures( failover->sw, num, texture );
+   failover->hw->set_fragment_sampler_textures( failover->hw, num, texture );
+}
+
+
+static void
+failover_set_vertex_sampler_textures(struct pipe_context *pipe,
+                                     unsigned num_textures,
+                                     struct pipe_texture **textures)
+{
+   struct failover_context *failover = failover_context(pipe);
+   uint i;
+
+   assert(num_textures <= PIPE_MAX_VERTEX_SAMPLERS);
+
+   /* Check for no-op */
+   if (num_textures == failover->num_vertex_textures &&
+       !memcmp(failover->vertex_textures, textures, num_textures * sizeof(struct pipe_texture *))) {
+      return;
+   }
+   for (i = 0; i < num_textures; i++) {
+      pipe_texture_reference((struct pipe_texture **)&failover->vertex_textures[i],
+                             textures[i]);
+   }
+   for (i = num_textures; i < failover->num_vertex_textures; i++) {
+      pipe_texture_reference((struct pipe_texture **)&failover->vertex_textures[i],
+                             NULL);
+   }
+   failover->dirty |= FO_NEW_TEXTURE;
+   failover->num_vertex_textures = num_textures;
+   failover->sw->set_vertex_sampler_textures(failover->sw, num_textures, textures);
+   failover->hw->set_vertex_sampler_textures(failover->hw, num_textures, textures);
 }
 
 
@@ -453,7 +514,8 @@ failover_init_state_functions( struct failover_context *failover )
    failover->pipe.bind_blend_state   = failover_bind_blend_state;
    failover->pipe.delete_blend_state = failover_delete_blend_state;
    failover->pipe.create_sampler_state = failover_create_sampler_state;
-   failover->pipe.bind_sampler_states  = failover_bind_sampler_states;
+   failover->pipe.bind_fragment_sampler_states  = failover_bind_fragment_sampler_states;
+   failover->pipe.bind_vertex_sampler_states  = failover_bind_vertex_sampler_states;
    failover->pipe.delete_sampler_state = failover_delete_sampler_state;
    failover->pipe.create_depth_stencil_alpha_state = failover_create_depth_stencil_state;
    failover->pipe.bind_depth_stencil_alpha_state   = failover_bind_depth_stencil_state;
@@ -473,7 +535,8 @@ failover_init_state_functions( struct failover_context *failover )
    failover->pipe.set_framebuffer_state = failover_set_framebuffer_state;
    failover->pipe.set_polygon_stipple = failover_set_polygon_stipple;
    failover->pipe.set_scissor_state = failover_set_scissor_state;
-   failover->pipe.set_sampler_textures = failover_set_sampler_textures;
+   failover->pipe.set_fragment_sampler_textures = failover_set_fragment_sampler_textures;
+   failover->pipe.set_vertex_sampler_textures = failover_set_vertex_sampler_textures;
    failover->pipe.set_viewport_state = failover_set_viewport_state;
    failover->pipe.set_vertex_buffers = failover_set_vertex_buffers;
    failover->pipe.set_vertex_elements = failover_set_vertex_elements;
index bd4fce9d2090a52b2667f46985d0205b59ac6af0..a3341e33f807102fda034fc8f1290e9e8d2fd8a4 100644 (file)
@@ -92,13 +92,19 @@ failover_state_emit( struct failover_context *failover )
       failover->sw->set_viewport_state( failover->sw, &failover->viewport );
 
    if (failover->dirty & FO_NEW_SAMPLER) {
-      failover->sw->bind_sampler_states( failover->sw, failover->num_samplers,
-                                         failover->sw_sampler_state );
+      failover->sw->bind_fragment_sampler_states( failover->sw, failover->num_samplers,
+                                                  failover->sw_sampler_state );
+      failover->sw->bind_vertex_sampler_states(failover->sw,
+                                               failover->num_vertex_samplers,
+                                               failover->sw_vertex_sampler_state);
    }
 
    if (failover->dirty & FO_NEW_TEXTURE) {
-      failover->sw->set_sampler_textures( failover->sw, failover->num_textures, 
-                                          failover->texture );
+      failover->sw->set_fragment_sampler_textures( failover->sw, failover->num_textures, 
+                                                   failover->texture );
+      failover->sw->set_vertex_sampler_textures(failover->sw,
+                                                failover->num_vertex_textures, 
+                                                failover->vertex_textures);
    }
 
    if (failover->dirty & FO_NEW_VERTEX_BUFFER) {
index 94c8aee30feeebcc9c2ebdfbfaab7cae45ca8a4e..949f04635015ab508d8d5dce93a345f0ffa1a086 100644 (file)
@@ -84,7 +84,7 @@ i915_draw_range_elements(struct pipe_context *pipe,
    }
 
 
-   draw_set_mapped_constant_buffer(draw,
+   draw_set_mapped_constant_buffer(draw, PIPE_SHADER_VERTEX,
                                    i915->current.constants[PIPE_SHADER_VERTEX],
                                    (i915->current.num_user_constants[PIPE_SHADER_VERTEX] * 
                                       4 * sizeof(float)));
index 379d47e79a35d41d53b5199ef0809010b0a30edc..25c53210be8eb2552440c4e1c74575459b2e2c62 100644 (file)
@@ -143,12 +143,12 @@ static uint
 src_vector(struct i915_fp_compile *p,
            const struct tgsi_full_src_register *source)
 {
-   uint index = source->SrcRegister.Index;
+   uint index = source->Register.Index;
    uint src = 0, sem_name, sem_ind;
 
-   switch (source->SrcRegister.File) {
+   switch (source->Register.File) {
    case TGSI_FILE_TEMPORARY:
-      if (source->SrcRegister.Index >= I915_MAX_TEMPORARY) {
+      if (source->Register.Index >= I915_MAX_TEMPORARY) {
          i915_program_error(p, "Exceeded max temporary reg");
          return 0;
       }
@@ -215,26 +215,25 @@ src_vector(struct i915_fp_compile *p,
    }
 
    src = swizzle(src,
-                source->SrcRegister.SwizzleX,
-                source->SrcRegister.SwizzleY,
-                source->SrcRegister.SwizzleZ,
-                source->SrcRegister.SwizzleW);
+                source->Register.SwizzleX,
+                source->Register.SwizzleY,
+                source->Register.SwizzleZ,
+                source->Register.SwizzleW);
 
 
    /* There's both negate-all-components and per-component negation.
     * Try to handle both here.
     */
    {
-      int n = source->SrcRegister.Negate;
+      int n = source->Register.Negate;
       src = negate(src, n, n, n, n);
    }
 
-   /* no abs() or post-abs negation */
+   /* no abs() */
 #if 0
    /* XXX assertions disabled to allow arbfplight.c to run */
    /* XXX enable these assertions, or fix things */
-   assert(!source->SrcRegisterExtMod.Absolute);
-   assert(!source->SrcRegisterExtMod.Negate);
+   assert(!source->Register.Absolute);
 #endif
    return src;
 }
@@ -247,10 +246,10 @@ static uint
 get_result_vector(struct i915_fp_compile *p,
                   const struct tgsi_full_dst_register *dest)
 {
-   switch (dest->DstRegister.File) {
+   switch (dest->Register.File) {
    case TGSI_FILE_OUTPUT:
       {
-         uint sem_name = p->shader->info.output_semantic_name[dest->DstRegister.Index];
+         uint sem_name = p->shader->info.output_semantic_name[dest->Register.Index];
          switch (sem_name) {
          case TGSI_SEMANTIC_POSITION:
             return UREG(REG_TYPE_OD, 0);
@@ -262,7 +261,7 @@ get_result_vector(struct i915_fp_compile *p,
          }
       }
    case TGSI_FILE_TEMPORARY:
-      return UREG(REG_TYPE_R, dest->DstRegister.Index);
+      return UREG(REG_TYPE_R, dest->Register.Index);
    default:
       i915_program_error(p, "Bad inst->DstReg.File");
       return 0;
@@ -277,7 +276,7 @@ static uint
 get_result_flags(const struct tgsi_full_instruction *inst)
 {
    const uint writeMask
-      = inst->FullDstRegisters[0].DstRegister.WriteMask;
+      = inst->Dst[0].Register.WriteMask;
    uint flags = 0x0;
 
    if (inst->Instruction.Saturate == TGSI_SAT_ZERO_ONE)
@@ -339,14 +338,14 @@ emit_tex(struct i915_fp_compile *p,
          const struct tgsi_full_instruction *inst,
          uint opcode)
 {
-   uint texture = inst->InstructionExtTexture.Texture;
-   uint unit = inst->FullSrcRegisters[1].SrcRegister.Index;
+   uint texture = inst->Texture.Texture;
+   uint unit = inst->Src[1].Register.Index;
    uint tex = translate_tex_src_target( p, texture );
    uint sampler = i915_emit_decl(p, REG_TYPE_S, unit, tex);
-   uint coord = src_vector( p, &inst->FullSrcRegisters[0]);
+   uint coord = src_vector( p, &inst->Src[0]);
 
    i915_emit_texld( p,
-                    get_result_vector( p, &inst->FullDstRegisters[0] ),
+                    get_result_vector( p, &inst->Dst[0] ),
                     get_result_flags( inst ),
                     sampler,
                     coord,
@@ -368,13 +367,13 @@ emit_simple_arith(struct i915_fp_compile *p,
 
    assert(numArgs <= 3);
 
-   arg1 = (numArgs < 1) ? 0 : src_vector( p, &inst->FullSrcRegisters[0] );
-   arg2 = (numArgs < 2) ? 0 : src_vector( p, &inst->FullSrcRegisters[1] );
-   arg3 = (numArgs < 3) ? 0 : src_vector( p, &inst->FullSrcRegisters[2] );
+   arg1 = (numArgs < 1) ? 0 : src_vector( p, &inst->Src[0] );
+   arg2 = (numArgs < 2) ? 0 : src_vector( p, &inst->Src[1] );
+   arg3 = (numArgs < 3) ? 0 : src_vector( p, &inst->Src[2] );
 
    i915_emit_arith( p,
                     opcode,
-                    get_result_vector( p, &inst->FullDstRegisters[0]),
+                    get_result_vector( p, &inst->Dst[0]),
                     get_result_flags( inst ), 0,
                     arg1,
                     arg2,
@@ -394,8 +393,8 @@ emit_simple_arith_swap2(struct i915_fp_compile *p,
 
    /* transpose first two registers */
    inst2 = *inst;
-   inst2.FullSrcRegisters[0] = inst->FullSrcRegisters[1];
-   inst2.FullSrcRegisters[1] = inst->FullSrcRegisters[0];
+   inst2.Src[0] = inst->Src[1];
+   inst2.Src[1] = inst->Src[0];
 
    emit_simple_arith(p, &inst2, opcode, numArgs);
 }
@@ -424,10 +423,10 @@ i915_translate_instruction(struct i915_fp_compile *p,
 
    switch (inst->Instruction.Opcode) {
    case TGSI_OPCODE_ABS:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
+      src0 = src_vector(p, &inst->Src[0]);
       i915_emit_arith(p,
                       A0_MAX,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       src0, negate(src0, 1, 1, 1, 1), 0);
       break;
@@ -437,17 +436,17 @@ i915_translate_instruction(struct i915_fp_compile *p,
       break;
 
    case TGSI_OPCODE_CMP:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
-      src1 = src_vector(p, &inst->FullSrcRegisters[1]);
-      src2 = src_vector(p, &inst->FullSrcRegisters[2]);
+      src0 = src_vector(p, &inst->Src[0]);
+      src1 = src_vector(p, &inst->Src[1]);
+      src2 = src_vector(p, &inst->Src[2]);
       i915_emit_arith(p, A0_CMP, 
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 
                       0, src0, src2, src1);   /* NOTE: order of src2, src1 */
       break;
 
    case TGSI_OPCODE_COS:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
+      src0 = src_vector(p, &inst->Src[0]);
       tmp = i915_get_utemp(p);
 
       i915_emit_arith(p,
@@ -490,7 +489,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
 
       i915_emit_arith(p,
                       A0_DP4,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       swizzle(tmp, ONE, Z, Y, X),
                       i915_emit_const4fv(p, cos_constants), 0);
@@ -505,19 +504,19 @@ i915_translate_instruction(struct i915_fp_compile *p,
       break;
 
    case TGSI_OPCODE_DPH:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
-      src1 = src_vector(p, &inst->FullSrcRegisters[1]);
+      src0 = src_vector(p, &inst->Src[0]);
+      src1 = src_vector(p, &inst->Src[1]);
 
       i915_emit_arith(p,
                       A0_DP4,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       swizzle(src0, X, Y, Z, ONE), src1, 0);
       break;
 
    case TGSI_OPCODE_DST:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
-      src1 = src_vector(p, &inst->FullSrcRegisters[1]);
+      src0 = src_vector(p, &inst->Src[0]);
+      src1 = src_vector(p, &inst->Src[1]);
 
       /* result[0] = 1    * 1;
        * result[1] = a[1] * b[1];
@@ -526,7 +525,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
        */
       i915_emit_arith(p,
                       A0_MUL,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       swizzle(src0, ONE, Y, Z, ONE),
                       swizzle(src1, ONE, Y, ONE, W), 0);
@@ -537,11 +536,11 @@ i915_translate_instruction(struct i915_fp_compile *p,
       break;
 
    case TGSI_OPCODE_EX2:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
+      src0 = src_vector(p, &inst->Src[0]);
 
       i915_emit_arith(p,
                       A0_EXP,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       swizzle(src0, X, X, X, X), 0, 0);
       break;
@@ -556,7 +555,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
 
    case TGSI_OPCODE_KIL:
       /* kill if src[0].x < 0 || src[0].y < 0 ... */
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
+      src0 = src_vector(p, &inst->Src[0]);
       tmp = i915_get_utemp(p);
 
       i915_emit_texld(p,
@@ -572,17 +571,17 @@ i915_translate_instruction(struct i915_fp_compile *p,
       break;
 
    case TGSI_OPCODE_LG2:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
+      src0 = src_vector(p, &inst->Src[0]);
 
       i915_emit_arith(p,
                       A0_LOG,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       swizzle(src0, X, X, X, X), 0, 0);
       break;
 
    case TGSI_OPCODE_LIT:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
+      src0 = src_vector(p, &inst->Src[0]);
       tmp = i915_get_utemp(p);
 
       /* tmp = max( a.xyzw, a.00zw )
@@ -606,7 +605,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
                       swizzle(tmp, Y, Y, Y, Y), 0, 0);
 
       i915_emit_arith(p, A0_CMP,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       negate(swizzle(tmp, ONE, ONE, X, ONE), 0, 0, 1, 0),
                       swizzle(tmp, ONE, X, ZERO, ONE),
@@ -615,9 +614,9 @@ i915_translate_instruction(struct i915_fp_compile *p,
       break;
 
    case TGSI_OPCODE_LRP:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
-      src1 = src_vector(p, &inst->FullSrcRegisters[1]);
-      src2 = src_vector(p, &inst->FullSrcRegisters[2]);
+      src0 = src_vector(p, &inst->Src[0]);
+      src1 = src_vector(p, &inst->Src[1]);
+      src2 = src_vector(p, &inst->Src[2]);
       flags = get_result_flags(inst);
       tmp = i915_get_utemp(p);
 
@@ -632,7 +631,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
                       flags & A0_DEST_CHANNEL_ALL, 0, src1, src0, src2);
 
       i915_emit_arith(p, A0_MAD,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       flags, 0, negate(src2, 1, 1, 1, 1), src0, tmp);
       break;
 
@@ -645,8 +644,8 @@ i915_translate_instruction(struct i915_fp_compile *p,
       break;
 
    case TGSI_OPCODE_MIN:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
-      src1 = src_vector(p, &inst->FullSrcRegisters[1]);
+      src0 = src_vector(p, &inst->Src[0]);
+      src1 = src_vector(p, &inst->Src[1]);
       tmp = i915_get_utemp(p);
       flags = get_result_flags(inst);
 
@@ -658,7 +657,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
 
       i915_emit_arith(p,
                       A0_MOV,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       flags, 0, negate(tmp, 1, 1, 1, 1), 0, 0);
       break;
 
@@ -671,8 +670,8 @@ i915_translate_instruction(struct i915_fp_compile *p,
       break;
 
    case TGSI_OPCODE_POW:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
-      src1 = src_vector(p, &inst->FullSrcRegisters[1]);
+      src0 = src_vector(p, &inst->Src[0]);
+      src1 = src_vector(p, &inst->Src[1]);
       tmp = i915_get_utemp(p);
       flags = get_result_flags(inst);
 
@@ -687,7 +686,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
 
       i915_emit_arith(p,
                       A0_EXP,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       flags, 0, swizzle(tmp, X, X, X, X), 0, 0);
       break;
       
@@ -696,27 +695,27 @@ i915_translate_instruction(struct i915_fp_compile *p,
       break;
       
    case TGSI_OPCODE_RCP:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
+      src0 = src_vector(p, &inst->Src[0]);
 
       i915_emit_arith(p,
                       A0_RCP,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                          get_result_flags(inst), 0,
                       swizzle(src0, X, X, X, X), 0, 0);
       break;
 
    case TGSI_OPCODE_RSQ:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
+      src0 = src_vector(p, &inst->Src[0]);
 
       i915_emit_arith(p,
                       A0_RSQ,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       swizzle(src0, X, X, X, X), 0, 0);
       break;
 
    case TGSI_OPCODE_SCS:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
+      src0 = src_vector(p, &inst->Src[0]);
       tmp = i915_get_utemp(p);
 
       /* 
@@ -739,7 +738,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
                       swizzle(tmp, X, Y, X, Y),
                       swizzle(tmp, X, X, ONE, ONE), 0);
 
-      writemask = inst->FullDstRegisters[0].DstRegister.WriteMask;
+      writemask = inst->Dst[0].Register.WriteMask;
 
       if (writemask & TGSI_WRITEMASK_Y) {
          uint tmp1;
@@ -757,7 +756,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
 
          i915_emit_arith(p,
                          A0_DP4,
-                         get_result_vector(p, &inst->FullDstRegisters[0]),
+                         get_result_vector(p, &inst->Dst[0]),
                          A0_DEST_CHANNEL_Y, 0,
                          swizzle(tmp1, W, Z, Y, X),
                          i915_emit_const4fv(p, sin_constants), 0);
@@ -772,7 +771,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
 
          i915_emit_arith(p,
                          A0_DP4,
-                         get_result_vector(p, &inst->FullDstRegisters[0]),
+                         get_result_vector(p, &inst->Dst[0]),
                          A0_DEST_CHANNEL_X, 0,
                          swizzle(tmp, ONE, Z, Y, X),
                          i915_emit_const4fv(p, cos_constants), 0);
@@ -789,7 +788,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
       break;
 
    case TGSI_OPCODE_SIN:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
+      src0 = src_vector(p, &inst->Src[0]);
       tmp = i915_get_utemp(p);
 
       i915_emit_arith(p,
@@ -832,7 +831,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
 
       i915_emit_arith(p,
                       A0_DP4,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       swizzle(tmp, W, Z, Y, X),
                       i915_emit_const4fv(p, sin_constants), 0);
@@ -848,12 +847,12 @@ i915_translate_instruction(struct i915_fp_compile *p,
       break;
 
    case TGSI_OPCODE_SUB:
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
-      src1 = src_vector(p, &inst->FullSrcRegisters[1]);
+      src0 = src_vector(p, &inst->Src[0]);
+      src1 = src_vector(p, &inst->Src[1]);
 
       i915_emit_arith(p,
                       A0_ADD,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       src0, negate(src1, 1, 1, 1, 1), 0);
       break;
@@ -877,8 +876,8 @@ i915_translate_instruction(struct i915_fp_compile *p,
        *      result.z = src0.x * src1.y - src0.y * src1.x;
        *      result.w = undef;
        */
-      src0 = src_vector(p, &inst->FullSrcRegisters[0]);
-      src1 = src_vector(p, &inst->FullSrcRegisters[1]);
+      src0 = src_vector(p, &inst->Src[0]);
+      src1 = src_vector(p, &inst->Src[1]);
       tmp = i915_get_utemp(p);
 
       i915_emit_arith(p,
@@ -889,7 +888,7 @@ i915_translate_instruction(struct i915_fp_compile *p,
 
       i915_emit_arith(p,
                       A0_MAD,
-                      get_result_vector(p, &inst->FullDstRegisters[0]),
+                      get_result_vector(p, &inst->Dst[0]),
                       get_result_flags(inst), 0,
                       swizzle(src0, Y, Z, X, ONE),
                       swizzle(src1, Z, X, Y, ONE),
@@ -929,8 +928,8 @@ i915_translate_instructions(struct i915_fp_compile *p,
          if (parse.FullToken.FullDeclaration.Declaration.File
                   == TGSI_FILE_CONSTANT) {
             uint i;
-            for (i = parse.FullToken.FullDeclaration.DeclarationRange.First;
-                 i <= parse.FullToken.FullDeclaration.DeclarationRange.Last;
+            for (i = parse.FullToken.FullDeclaration.Range.First;
+                 i <= parse.FullToken.FullDeclaration.Range.Last;
                  i++) {
                assert(ifs->constant_flags[i] == 0x0);
                ifs->constant_flags[i] = I915_CONSTFLAG_USER;
@@ -940,8 +939,8 @@ i915_translate_instructions(struct i915_fp_compile *p,
          else if (parse.FullToken.FullDeclaration.Declaration.File
                   == TGSI_FILE_TEMPORARY) {
             uint i;
-            for (i = parse.FullToken.FullDeclaration.DeclarationRange.First;
-                 i <= parse.FullToken.FullDeclaration.DeclarationRange.Last;
+            for (i = parse.FullToken.FullDeclaration.Range.First;
+                 i <= parse.FullToken.FullDeclaration.Range.Last;
                  i++) {
                assert(i < I915_MAX_TEMPORARY);
                /* XXX just use shader->info->file_mask[TGSI_FILE_TEMPORARY] */
index 71f00bc346a98b674e40974f805b7d414b312c40..e580b6c0f7fb6a0c2ecb62adbd2b7272dfee77f8 100644 (file)
@@ -752,22 +752,15 @@ static void i915_set_vertex_elements(struct pipe_context *pipe,
 }
 
 
-static void i915_set_edgeflags(struct pipe_context *pipe,
-                               const unsigned *bitfield)
-{
-   /* TODO do something here */
-}
-
 void
 i915_init_state_functions( struct i915_context *i915 )
 {
-   i915->base.set_edgeflags = i915_set_edgeflags;
    i915->base.create_blend_state = i915_create_blend_state;
    i915->base.bind_blend_state = i915_bind_blend_state;
    i915->base.delete_blend_state = i915_delete_blend_state;
 
    i915->base.create_sampler_state = i915_create_sampler_state;
-   i915->base.bind_sampler_states = i915_bind_sampler_states;
+   i915->base.bind_fragment_sampler_states = i915_bind_sampler_states;
    i915->base.delete_sampler_state = i915_delete_sampler_state;
 
    i915->base.create_depth_stencil_alpha_state = i915_create_depth_stencil_state;
@@ -791,7 +784,7 @@ i915_init_state_functions( struct i915_context *i915 )
 
    i915->base.set_polygon_stipple = i915_set_polygon_stipple;
    i915->base.set_scissor_state = i915_set_scissor_state;
-   i915->base.set_sampler_textures = i915_set_sampler_textures;
+   i915->base.set_fragment_sampler_textures = i915_set_sampler_textures;
    i915->base.set_viewport_state = i915_set_viewport_state;
    i915->base.set_vertex_buffers = i915_set_vertex_buffers;
    i915->base.set_vertex_elements = i915_set_vertex_elements;
index 178d4e8781d7b261add57e9a39240efc033df5a1..03dd5091a611ba63de9eed9779ffabfff85e6e86 100644 (file)
@@ -84,7 +84,7 @@ static void calculate_vertex_layout( struct i915_context *i915 )
 
    
    /* pos */
-   src = draw_find_vs_output(i915->draw, TGSI_SEMANTIC_POSITION, 0);
+   src = draw_find_shader_output(i915->draw, TGSI_SEMANTIC_POSITION, 0);
    if (needW) {
       draw_emit_vertex_attr(&vinfo, EMIT_4F, INTERP_LINEAR, src);
       vinfo.hwfmt[0] |= S4_VFMT_XYZW;
@@ -101,21 +101,21 @@ static void calculate_vertex_layout( struct i915_context *i915 )
 
    /* primary color */
    if (colors[0]) {
-      src = draw_find_vs_output(i915->draw, TGSI_SEMANTIC_COLOR, 0);
+      src = draw_find_shader_output(i915->draw, TGSI_SEMANTIC_COLOR, 0);
       draw_emit_vertex_attr(&vinfo, EMIT_4UB, colorInterp, src);
       vinfo.hwfmt[0] |= S4_VFMT_COLOR;
    }
 
    /* secondary color */
    if (colors[1]) {
-      src = draw_find_vs_output(i915->draw, TGSI_SEMANTIC_COLOR, 1);
+      src = draw_find_shader_output(i915->draw, TGSI_SEMANTIC_COLOR, 1);
       draw_emit_vertex_attr(&vinfo, EMIT_4UB, colorInterp, src);
       vinfo.hwfmt[0] |= S4_VFMT_SPEC_FOG;
    }
 
    /* fog coord, not fog blend factor */
    if (fog) {
-      src = draw_find_vs_output(i915->draw, TGSI_SEMANTIC_FOG, 0);
+      src = draw_find_shader_output(i915->draw, TGSI_SEMANTIC_FOG, 0);
       draw_emit_vertex_attr(&vinfo, EMIT_1F, INTERP_PERSPECTIVE, src);
       vinfo.hwfmt[0] |= S4_VFMT_FOG_PARAM;
    }
@@ -125,7 +125,7 @@ static void calculate_vertex_layout( struct i915_context *i915 )
       uint hwtc;
       if (texCoords[i]) {
          hwtc = TEXCOORDFMT_4D;
-         src = draw_find_vs_output(i915->draw, TGSI_SEMANTIC_GENERIC, i);
+         src = draw_find_shader_output(i915->draw, TGSI_SEMANTIC_GENERIC, i);
          draw_emit_vertex_attr(&vinfo, EMIT_4F, INTERP_PERSPECTIVE, src);
       }
       else {
index c5e9084d12e7072f4902164ff24001e73cad1649..cbac4175c8f78fb5652718cc3a630771aa9b3eb3 100644 (file)
@@ -231,7 +231,7 @@ i915_update_texture(struct i915_context *i915,
 {
    const struct pipe_texture *pt = &tex->base;
    uint format, pitch;
-   const uint width = pt->width[0], height = pt->height[0], depth = pt->depth[0];
+   const uint width = pt->width0, height = pt->height0, depth = pt->depth0;
    const uint num_levels = pt->last_level;
    unsigned max_lod = num_levels * 4;
    unsigned tiled = MS3_USE_FENCE_REGS;
index ab8331f3e64a276d372b0a9254a501bcc54531f2..c693eb30e87e3a1e9dcde7b1a4b06a1aa815204a 100644 (file)
@@ -32,6 +32,7 @@
 #include "pipe/p_inlines.h"
 #include "pipe/p_inlines.h"
 #include "pipe/internal/p_winsys_screen.h"
+#include "util/u_format.h"
 #include "util/u_tile.h"
 #include "util/u_rect.h"
 
@@ -48,17 +49,19 @@ i915_surface_copy(struct pipe_context *pipe,
 {
    struct i915_texture *dst_tex = (struct i915_texture *)dst->texture;
    struct i915_texture *src_tex = (struct i915_texture *)src->texture;
+   struct pipe_texture *dpt = &dst_tex->base;
+   struct pipe_texture *spt = &src_tex->base;
 
    assert( dst != src );
-   assert( dst_tex->base.block.size == src_tex->base.block.size );
-   assert( dst_tex->base.block.width == src_tex->base.block.height );
-   assert( dst_tex->base.block.height == src_tex->base.block.height );
-   assert( dst_tex->base.block.width == 1 );
-   assert( dst_tex->base.block.height == 1 );
+   assert( util_format_get_blocksize(dpt->format) == util_format_get_blocksize(spt->format) );
+   assert( util_format_get_blockwidth(dpt->format) == util_format_get_blockwidth(spt->format) );
+   assert( util_format_get_blockheight(dpt->format) == util_format_get_blockheight(spt->format) );
+   assert( util_format_get_blockwidth(dpt->format) == 1 );
+   assert( util_format_get_blockheight(dpt->format) == 1 );
 
    i915_copy_blit( i915_context(pipe),
                    FALSE,
-                   dst_tex->base.block.size,
+                   util_format_get_blocksize(dpt->format),
                    (unsigned short) src_tex->stride, src_tex->buffer, src->offset,
                    (unsigned short) dst_tex->stride, dst_tex->buffer, dst->offset,
                    (short) srcx, (short) srcy, (short) dstx, (short) dsty, (short) width, (short) height );
@@ -72,12 +75,13 @@ i915_surface_fill(struct pipe_context *pipe,
                  unsigned width, unsigned height, unsigned value)
 {
    struct i915_texture *tex = (struct i915_texture *)dst->texture;
+   struct pipe_texture *pt = &tex->base;
 
-   assert(tex->base.block.width == 1);
-   assert(tex->base.block.height == 1);
+   assert(util_format_get_blockwidth(pt->format) == 1);
+   assert(util_format_get_blockheight(pt->format) == 1);
 
    i915_fill_blit( i915_context(pipe),
-                   tex->base.block.size,
+                   util_format_get_blocksize(pt->format),
                    (unsigned short) tex->stride,
                    tex->buffer, dst->offset,
                    (short) dstx, (short) dsty,
index 286c9ace8e5736a2ab161eca5dbb2a7be936b913..50a9e19094b8906ca1fe96f2d2f208368f1e1f9b 100644 (file)
@@ -35,6 +35,7 @@
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
 #include "pipe/internal/p_winsys_screen.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -74,6 +75,9 @@ static const int step_offsets[6][2] = {
    {-1, 1}
 };
 
+/* XXX really need twice the size if x is already pot?
+   Otherwise just use util_next_power_of_two?
+*/
 static unsigned
 power_of_two(unsigned x)
 {
@@ -83,13 +87,6 @@ power_of_two(unsigned x)
    return value;
 }
 
-static unsigned
-round_up(unsigned n, unsigned multiple)
-{
-   return (n + multiple - 1) & ~(multiple - 1);
-}
-
-
 /*
  * More advanced helper funcs
  */
@@ -101,17 +98,8 @@ i915_miptree_set_level_info(struct i915_texture *tex,
                              unsigned nr_images,
                              unsigned w, unsigned h, unsigned d)
 {
-   struct pipe_texture *pt = &tex->base;
-
    assert(level < PIPE_MAX_TEXTURE_LEVELS);
 
-   pt->width[level] = w;
-   pt->height[level] = h;
-   pt->depth[level] = d;
-   
-   pt->nblocksx[level] = pf_get_nblocksx(&pt->block, w);
-   pt->nblocksy[level] = pf_get_nblocksy(&pt->block, h);
-
    tex->nr_images[level] = nr_images;
 
    /*
@@ -142,7 +130,7 @@ i915_miptree_set_image_offset(struct i915_texture *tex,
 
    assert(img < tex->nr_images[level]);
 
-   tex->image_offset[level][img] = y * tex->stride + x * tex->base.block.size;
+   tex->image_offset[level][img] = y * tex->stride + x * util_format_get_blocksize(tex->base.format);
 
    /*
    printf("%s level %d img %d pos %d,%d image_offset %x\n",
@@ -164,28 +152,28 @@ i915_scanout_layout(struct i915_texture *tex)
 {
    struct pipe_texture *pt = &tex->base;
 
-   if (pt->last_level > 0 || pt->block.size != 4)
+   if (pt->last_level > 0 || util_format_get_blocksize(pt->format) != 4)
       return FALSE;
 
    i915_miptree_set_level_info(tex, 0, 1,
-                               tex->base.width[0],
-                               tex->base.height[0],
+                               pt->width0,
+                               pt->height0,
                                1);
    i915_miptree_set_image_offset(tex, 0, 0, 0, 0);
 
-   if (tex->base.width[0] >= 240) {
-      tex->stride = power_of_two(tex->base.nblocksx[0] * pt->block.size);
-      tex->total_nblocksy = round_up(tex->base.nblocksy[0], 8);
+   if (pt->width0 >= 240) {
+      tex->stride = power_of_two(util_format_get_stride(pt->format, pt->width0));
+      tex->total_nblocksy = align(util_format_get_nblocksy(pt->format, pt->height0), 8);
       tex->hw_tiled = INTEL_TILE_X;
-   } else if (tex->base.width[0] == 64 && tex->base.height[0] == 64) {
-      tex->stride = power_of_two(tex->base.nblocksx[0] * pt->block.size);
-      tex->total_nblocksy = round_up(tex->base.nblocksy[0], 8);
+   } else if (pt->width0 == 64 && pt->height0 == 64) {
+      tex->stride = power_of_two(util_format_get_stride(pt->format, pt->width0));
+      tex->total_nblocksy = align(util_format_get_nblocksy(pt->format, pt->height0), 8);
    } else {
       return FALSE;
    }
 
    debug_printf("%s size: %d,%d,%d offset %d,%d (0x%x)\n", __FUNCTION__,
-      tex->base.width[0], tex->base.height[0], pt->block.size,
+      pt->width0, pt->height0, util_format_get_blocksize(pt->format),
       tex->stride, tex->total_nblocksy, tex->stride * tex->total_nblocksy);
 
    return TRUE;
@@ -199,25 +187,25 @@ i915_display_target_layout(struct i915_texture *tex)
 {
    struct pipe_texture *pt = &tex->base;
 
-   if (pt->last_level > 0 || pt->block.size != 4)
+   if (pt->last_level > 0 || util_format_get_blocksize(pt->format) != 4)
       return FALSE;
 
    /* fallback to normal textures for small textures */
-   if (tex->base.width[0] < 240)
+   if (pt->width0 < 240)
       return FALSE;
 
    i915_miptree_set_level_info(tex, 0, 1,
-                               tex->base.width[0],
-                               tex->base.height[0],
+                               pt->width0,
+                               pt->height0,
                                1);
    i915_miptree_set_image_offset(tex, 0, 0, 0, 0);
 
-   tex->stride = power_of_two(tex->base.nblocksx[0] * pt->block.size);
-   tex->total_nblocksy = round_up(tex->base.nblocksy[0], 8);
+   tex->stride = power_of_two(util_format_get_stride(pt->format, pt->width0));
+   tex->total_nblocksy = align(util_format_get_nblocksy(pt->format, pt->height0), 8);
    tex->hw_tiled = INTEL_TILE_X;
 
    debug_printf("%s size: %d,%d,%d offset %d,%d (0x%x)\n", __FUNCTION__,
-      tex->base.width[0], tex->base.height[0], pt->block.size,
+      pt->width0, pt->height0, util_format_get_blocksize(pt->format),
       tex->stride, tex->total_nblocksy, tex->stride * tex->total_nblocksy);
 
    return TRUE;
@@ -228,36 +216,34 @@ i915_miptree_layout_2d(struct i915_texture *tex)
 {
    struct pipe_texture *pt = &tex->base;
    unsigned level;
-   unsigned width = pt->width[0];
-   unsigned height = pt->height[0];
-   unsigned nblocksx = pt->nblocksx[0];
-   unsigned nblocksy = pt->nblocksy[0];
+   unsigned width = pt->width0;
+   unsigned height = pt->height0;
+   unsigned nblocksy = util_format_get_nblocksy(pt->format, pt->width0);
 
    /* used for scanouts that need special layouts */
-   if (tex->base.tex_usage & PIPE_TEXTURE_USAGE_PRIMARY)
+   if (pt->tex_usage & PIPE_TEXTURE_USAGE_PRIMARY)
       if (i915_scanout_layout(tex))
          return;
 
    /* for shared buffers we use some very like scanout */
-   if (tex->base.tex_usage & PIPE_TEXTURE_USAGE_DISPLAY_TARGET)
+   if (pt->tex_usage & PIPE_TEXTURE_USAGE_DISPLAY_TARGET)
       if (i915_display_target_layout(tex))
          return;
 
-   tex->stride = round_up(pt->nblocksx[0] * pt->block.size, 4);
+   tex->stride = align(util_format_get_stride(pt->format, pt->width0), 4);
    tex->total_nblocksy = 0;
 
    for (level = 0; level <= pt->last_level; level++) {
       i915_miptree_set_level_info(tex, level, 1, width, height, 1);
       i915_miptree_set_image_offset(tex, level, 0, 0, tex->total_nblocksy);
 
-      nblocksy = round_up(MAX2(2, nblocksy), 2);
+      nblocksy = align(MAX2(2, nblocksy), 2);
 
       tex->total_nblocksy += nblocksy;
 
-      width = minify(width);
-      height = minify(height);
-      nblocksx = pf_get_nblocksx(&pt->block, width);
-      nblocksy = pf_get_nblocksy(&pt->block, height);
+      width = u_minify(width, 1);
+      height = u_minify(height, 1);
+      nblocksy = util_format_get_nblocksy(pt->format, height);
    }
 }
 
@@ -267,16 +253,15 @@ i915_miptree_layout_3d(struct i915_texture *tex)
    struct pipe_texture *pt = &tex->base;
    unsigned level;
 
-   unsigned width = pt->width[0];
-   unsigned height = pt->height[0];
-   unsigned depth = pt->depth[0];
-   unsigned nblocksx = pt->nblocksx[0];
-   unsigned nblocksy = pt->nblocksy[0];
+   unsigned width = pt->width0;
+   unsigned height = pt->height0;
+   unsigned depth = pt->depth0;
+   unsigned nblocksy = util_format_get_nblocksy(pt->format, pt->height0);
    unsigned stack_nblocksy = 0;
 
    /* Calculate the size of a single slice. 
     */
-   tex->stride = round_up(pt->nblocksx[0] * pt->block.size, 4);
+   tex->stride = align(util_format_get_stride(pt->format, pt->width0), 4);
 
    /* XXX: hardware expects/requires 9 levels at minimum.
     */
@@ -285,44 +270,41 @@ i915_miptree_layout_3d(struct i915_texture *tex)
 
       stack_nblocksy += MAX2(2, nblocksy);
 
-      width = minify(width);
-      height = minify(height);
-      depth = minify(depth);
-      nblocksx = pf_get_nblocksx(&pt->block, width);
-      nblocksy = pf_get_nblocksy(&pt->block, height);
+      width = u_minify(width, 1);
+      height = u_minify(height, 1);
+      nblocksy = util_format_get_nblocksy(pt->format, height);
    }
 
    /* Fixup depth image_offsets: 
     */
-   depth = pt->depth[0];
    for (level = 0; level <= pt->last_level; level++) {
       unsigned i;
       for (i = 0; i < depth; i++) 
          i915_miptree_set_image_offset(tex, level, i, 0, i * stack_nblocksy);
 
-      depth = minify(depth);
+      depth = u_minify(depth, 1);
    }
 
    /* Multiply slice size by texture depth for total size.  It's
     * remarkable how wasteful of memory the i915 texture layouts
     * are.  They are largely fixed in the i945.
     */
-   tex->total_nblocksy = stack_nblocksy * pt->depth[0];
+   tex->total_nblocksy = stack_nblocksy * pt->depth0;
 }
 
 static void
 i915_miptree_layout_cube(struct i915_texture *tex)
 {
    struct pipe_texture *pt = &tex->base;
-   unsigned width = pt->width[0], height = pt->height[0];
-   const unsigned nblocks = pt->nblocksx[0];
+   unsigned width = pt->width0, height = pt->height0;
+   const unsigned nblocks = util_format_get_nblocksx(pt->format, pt->width0);
    unsigned level;
    unsigned face;
 
    assert(width == height); /* cubemap images are square */
 
    /* double pitch for cube layouts */
-   tex->stride = round_up(nblocks * pt->block.size * 2, 4);
+   tex->stride = align(nblocks * util_format_get_blocksize(pt->format) * 2, 4);
    tex->total_nblocksy = nblocks * 4;
 
    for (level = 0; level <= pt->last_level; level++) {
@@ -383,10 +365,10 @@ i945_miptree_layout_2d(struct i915_texture *tex)
    unsigned level;
    unsigned x = 0;
    unsigned y = 0;
-   unsigned width = pt->width[0];
-   unsigned height = pt->height[0];
-   unsigned nblocksx = pt->nblocksx[0];
-   unsigned nblocksy = pt->nblocksy[0];
+   unsigned width = pt->width0;
+   unsigned height = pt->height0;
+   unsigned nblocksx = util_format_get_nblocksx(pt->format, pt->width0);
+   unsigned nblocksy = util_format_get_nblocksy(pt->format, pt->height0);
 
    /* used for scanouts that need special layouts */
    if (tex->base.tex_usage & PIPE_TEXTURE_USAGE_PRIMARY)
@@ -398,7 +380,7 @@ i945_miptree_layout_2d(struct i915_texture *tex)
       if (i915_display_target_layout(tex))
          return;
 
-   tex->stride = round_up(pt->nblocksx[0] * pt->block.size, 4);
+   tex->stride = align(util_format_get_stride(pt->format, pt->width0), 4);
 
    /* May need to adjust pitch to accomodate the placement of
     * the 2nd mipmap level.  This occurs when the alignment
@@ -407,11 +389,11 @@ i945_miptree_layout_2d(struct i915_texture *tex)
     */
    if (pt->last_level > 0) {
       unsigned mip1_nblocksx 
-         = align(pf_get_nblocksx(&pt->block, minify(width)), align_x)
-         + pf_get_nblocksx(&pt->block, minify(minify(width)));
+         = align(util_format_get_nblocksx(pt->format, u_minify(width, 1)), align_x)
+         + util_format_get_nblocksx(pt->format, u_minify(width, 2));
 
       if (mip1_nblocksx > nblocksx)
-         tex->stride = mip1_nblocksx * pt->block.size;
+         tex->stride = mip1_nblocksx * util_format_get_blocksize(pt->format);
    }
 
    /* Pitch must be a whole number of dwords
@@ -439,10 +421,10 @@ i945_miptree_layout_2d(struct i915_texture *tex)
          y += nblocksy;
       }
 
-      width  = minify(width);
-      height = minify(height);
-      nblocksx = pf_get_nblocksx(&pt->block, width);
-      nblocksy = pf_get_nblocksy(&pt->block, height);
+      width  = u_minify(width, 1);
+      height = u_minify(height, 1);
+      nblocksx = util_format_get_nblocksx(pt->format, width);
+      nblocksy = util_format_get_nblocksy(pt->format, height);
    }
 }
 
@@ -450,20 +432,19 @@ static void
 i945_miptree_layout_3d(struct i915_texture *tex)
 {
    struct pipe_texture *pt = &tex->base;
-   unsigned width = pt->width[0];
-   unsigned height = pt->height[0];
-   unsigned depth = pt->depth[0];
-   unsigned nblocksx = pt->nblocksx[0];
-   unsigned nblocksy = pt->nblocksy[0];
+   unsigned width = pt->width0;
+   unsigned height = pt->height0;
+   unsigned depth = pt->depth0;
+   unsigned nblocksy = util_format_get_nblocksy(pt->format, pt->width0);
    unsigned pack_x_pitch, pack_x_nr;
    unsigned pack_y_pitch;
    unsigned level;
 
-   tex->stride = round_up(pt->nblocksx[0] * pt->block.size, 4);
+   tex->stride = align(util_format_get_stride(pt->format, pt->width0), 4);
    tex->total_nblocksy = 0;
 
-   pack_y_pitch = MAX2(pt->nblocksy[0], 2);
-   pack_x_pitch = tex->stride / pt->block.size;
+   pack_y_pitch = MAX2(nblocksy, 2);
+   pack_x_pitch = tex->stride / util_format_get_blocksize(pt->format);
    pack_x_nr = 1;
 
    for (level = 0; level <= pt->last_level; level++) {
@@ -488,18 +469,17 @@ i945_miptree_layout_3d(struct i915_texture *tex)
       if (pack_x_pitch > 4) {
          pack_x_pitch >>= 1;
          pack_x_nr <<= 1;
-         assert(pack_x_pitch * pack_x_nr * pt->block.size <= tex->stride);
+         assert(pack_x_pitch * pack_x_nr * util_format_get_blocksize(pt->format) <= tex->stride);
       }
 
       if (pack_y_pitch > 2) {
          pack_y_pitch >>= 1;
       }
 
-      width = minify(width);
-      height = minify(height);
-      depth = minify(depth);
-      nblocksx = pf_get_nblocksx(&pt->block, width);
-      nblocksy = pf_get_nblocksy(&pt->block, height);
+      width = u_minify(width, 1);
+      height = u_minify(height, 1);
+      depth = u_minify(depth, 1);
+      nblocksy = util_format_get_nblocksy(pt->format, height);
    }
 }
 
@@ -509,13 +489,13 @@ i945_miptree_layout_cube(struct i915_texture *tex)
    struct pipe_texture *pt = &tex->base;
    unsigned level;
 
-   const unsigned nblocks = pt->nblocksx[0];
+   const unsigned nblocks = util_format_get_nblocksx(pt->format, pt->width0);
    unsigned face;
-   unsigned width = pt->width[0];
-   unsigned height = pt->height[0];
+   unsigned width = pt->width0;
+   unsigned height = pt->height0;
 
    /*
-   printf("%s %i, %i\n", __FUNCTION__, pt->width[0], pt->height[0]);
+   printf("%s %i, %i\n", __FUNCTION__, pt->width0, pt->height0);
    */
 
    assert(width == height); /* cubemap images are square */
@@ -529,9 +509,9 @@ i945_miptree_layout_cube(struct i915_texture *tex)
     * or the final row of 4x4, 2x2 and 1x1 faces below this.
     */
    if (nblocks > 32)
-      tex->stride = round_up(nblocks * pt->block.size * 2, 4);
+      tex->stride = align(nblocks * util_format_get_blocksize(pt->format) * 2, 4);
    else
-      tex->stride = 14 * 8 * pt->block.size;
+      tex->stride = 14 * 8 * util_format_get_blocksize(pt->format);
 
    tex->total_nblocksy = nblocks * 4;
 
@@ -651,9 +631,6 @@ i915_texture_create(struct pipe_screen *screen,
    pipe_reference_init(&tex->base.reference, 1);
    tex->base.screen = screen;
 
-   tex->base.nblocksx[0] = pf_get_nblocksx(&tex->base.block, tex->base.width[0]);
-   tex->base.nblocksy[0] = pf_get_nblocksy(&tex->base.block, tex->base.height[0]);
-   
    if (is->is_i945) {
       if (!i945_miptree_layout(tex))
          goto fail;
@@ -667,7 +644,7 @@ i915_texture_create(struct pipe_screen *screen,
 
 
    /* for scanouts and cursors, cursors arn't scanouts */
-   if (templat->tex_usage & PIPE_TEXTURE_USAGE_PRIMARY && templat->width[0] != 64)
+   if (templat->tex_usage & PIPE_TEXTURE_USAGE_PRIMARY && templat->width0 != 64)
       buf_usage = INTEL_NEW_SCANOUT;
    else
       buf_usage = INTEL_NEW_TEXTURE;
@@ -710,7 +687,7 @@ i915_texture_blanket(struct pipe_screen * screen,
    /* Only supports one type */
    if (base->target != PIPE_TEXTURE_2D ||
        base->last_level != 0 ||
-       base->depth[0] != 1) {
+       base->depth0 != 1) {
       return NULL;
    }
 
@@ -724,7 +701,7 @@ i915_texture_blanket(struct pipe_screen * screen,
 
    tex->stride = stride[0];
 
-   i915_miptree_set_level_info(tex, 0, 1, base->width[0], base->height[0], 1);
+   i915_miptree_set_level_info(tex, 0, 1, base->width0, base->height0, 1);
    i915_miptree_set_image_offset(tex, 0, 0, 0, 0);
 
    pipe_buffer_reference(&tex->buffer, buffer);
@@ -788,8 +765,8 @@ i915_get_tex_surface(struct pipe_screen *screen,
       pipe_reference_init(&ps->reference, 1);
       pipe_texture_reference(&ps->texture, pt);
       ps->format = pt->format;
-      ps->width = pt->width[level];
-      ps->height = pt->height[level];
+      ps->width = u_minify(pt->width0, level);
+      ps->height = u_minify(pt->height0, level);
       ps->offset = offset;
       ps->usage = flags;
    }
@@ -835,14 +812,10 @@ i915_get_tex_transfer(struct pipe_screen *screen,
    trans = CALLOC_STRUCT(i915_transfer);
    if (trans) {
       pipe_texture_reference(&trans->base.texture, texture);
-      trans->base.format = trans->base.format;
       trans->base.x = x;
       trans->base.y = y;
       trans->base.width = w;
       trans->base.height = h;
-      trans->base.block = texture->block;
-      trans->base.nblocksx = texture->nblocksx[level];
-      trans->base.nblocksy = texture->nblocksy[level];
       trans->base.stride = tex->stride;
       trans->offset = offset;
       trans->base.usage = usage;
@@ -858,6 +831,7 @@ i915_transfer_map(struct pipe_screen *screen,
    struct intel_winsys *iws = i915_screen(tex->base.screen)->iws;
    char *map;
    boolean write = FALSE;
+   enum pipe_format format = tex->base.format;
 
    if (transfer->usage & PIPE_TRANSFER_WRITE)
       write = TRUE;
@@ -867,8 +841,8 @@ i915_transfer_map(struct pipe_screen *screen,
       return NULL;
 
    return map + i915_transfer(transfer)->offset +
-      transfer->y / transfer->block.height * transfer->stride +
-      transfer->x / transfer->block.width * transfer->block.size;
+      transfer->y / util_format_get_blockheight(format) * transfer->stride +
+      transfer->x / util_format_get_blockwidth(format) * util_format_get_blocksize(format);
 }
 
 static void
@@ -919,7 +893,7 @@ i915_texture_blanket_intel(struct pipe_screen *screen,
    /* Only supports one type */
    if (base->target != PIPE_TEXTURE_2D ||
        base->last_level != 0 ||
-       base->depth[0] != 1) {
+       base->depth0 != 1) {
       return NULL;
    }
 
@@ -933,7 +907,7 @@ i915_texture_blanket_intel(struct pipe_screen *screen,
 
    tex->stride = stride;
 
-   i915_miptree_set_level_info(tex, 0, 1, base->width[0], base->height[0], 1);
+   i915_miptree_set_level_info(tex, 0, 1, base->width0, base->height0, 1);
    i915_miptree_set_image_offset(tex, 0, 0, 0, 0);
 
    tex->buffer = buffer;
diff --git a/src/gallium/drivers/i965/Makefile b/src/gallium/drivers/i965/Makefile
new file mode 100644 (file)
index 0000000..95fd3cd
--- /dev/null
@@ -0,0 +1,74 @@
+TOP = ../../../..
+include $(TOP)/configs/current
+
+LIBNAME = i965
+
+C_SOURCES = \
+       brw_cc.c \
+       brw_clip.c \
+       brw_clip_line.c \
+       brw_clip_point.c \
+       brw_clip_state.c \
+       brw_clip_tri.c \
+       brw_clip_unfilled.c \
+       brw_clip_util.c \
+       brw_context.c \
+       brw_curbe.c \
+       brw_disasm.c \
+       brw_draw.c \
+       brw_draw_upload.c \
+       brw_eu.c \
+       brw_eu_debug.c \
+       brw_eu_emit.c \
+       brw_eu_util.c \
+       brw_gs.c \
+       brw_gs_emit.c \
+       brw_gs_state.c \
+       brw_misc_state.c \
+       brw_pipe_blend.c \
+       brw_pipe_depth.c \
+       brw_pipe_fb.c \
+       brw_pipe_query.c \
+       brw_pipe_shader.c \
+       brw_pipe_flush.c \
+       brw_pipe_misc.c \
+       brw_pipe_sampler.c \
+       brw_pipe_vertex.c \
+       brw_pipe_clear.c \
+       brw_pipe_rast.c \
+       brw_sf.c \
+       brw_sf_emit.c \
+       brw_sf_state.c \
+       brw_state_batch.c \
+       brw_state_debug.c \
+       brw_state_cache.c \
+       brw_state_upload.c \
+       brw_structs_dump.c \
+       brw_swtnl.c \
+       brw_urb.c \
+       brw_util.c \
+       brw_vs.c \
+       brw_vs_emit.c \
+       brw_vs_state.c \
+       brw_vs_surface_state.c \
+       brw_wm.c \
+       brw_wm_debug.c \
+       brw_wm_emit.c \
+       brw_wm_fp.c \
+       brw_wm_iz.c \
+       brw_wm_pass0.c \
+       brw_wm_pass1.c \
+       brw_wm_pass2.c \
+       brw_wm_sampler_state.c \
+       brw_wm_state.c \
+       brw_wm_surface_state.c \
+       brw_screen.c \
+       brw_screen_buffers.c \
+       brw_screen_tex_layout.c \
+       brw_screen_texture.c \
+       brw_screen_surface.c \
+       brw_batchbuffer.c \
+       brw_winsys_debug.c \
+       intel_decode.c
+
+include ../../Makefile.template
diff --git a/src/gallium/drivers/i965/SConscript b/src/gallium/drivers/i965/SConscript
new file mode 100644 (file)
index 0000000..9c2faaf
--- /dev/null
@@ -0,0 +1,77 @@
+Import('*')
+
+env = env.Clone()
+
+i965 = env.ConvenienceLibrary(
+       target = 'i965',
+       source = [
+               'brw_batchbuffer.c',
+               'brw_cc.c',
+               'brw_clip.c',
+               'brw_clip_line.c',
+               'brw_clip_point.c',
+               'brw_clip_state.c',
+               'brw_clip_tri.c',
+               'brw_clip_unfilled.c',
+               'brw_clip_util.c',
+               'brw_context.c',
+               'brw_curbe.c',
+               'brw_disasm.c',
+               'brw_draw.c',
+               'brw_draw_upload.c',
+               'brw_eu.c',
+               'brw_eu_debug.c',
+               'brw_eu_emit.c',
+               'brw_eu_util.c',
+               'brw_gs.c',
+               'brw_gs_emit.c',
+               'brw_gs_state.c',
+               'brw_misc_state.c',
+               'brw_pipe_blend.c',
+               'brw_pipe_clear.c',
+               'brw_pipe_depth.c',
+               'brw_pipe_fb.c',
+               'brw_pipe_flush.c',
+               'brw_pipe_misc.c',
+               'brw_pipe_query.c',
+               'brw_pipe_rast.c',
+               'brw_pipe_sampler.c',
+               'brw_pipe_shader.c',
+               'brw_pipe_vertex.c',
+               'brw_screen_buffers.c',
+               'brw_screen.c',
+               'brw_screen_surface.c',
+               'brw_screen_tex_layout.c',
+               'brw_screen_texture.c',
+               'brw_structs_dump.c',
+               'brw_sf.c',
+               'brw_sf_emit.c',
+               'brw_sf_state.c',
+               'brw_state_batch.c',
+               'brw_state_cache.c',
+#              'brw_state_debug.c',
+               'brw_state_upload.c',
+               'brw_swtnl.c',
+               'brw_urb.c',
+               'brw_util.c',
+               'brw_vs.c',
+               'brw_vs_emit.c',
+               'brw_vs_state.c',
+               'brw_vs_surface_state.c',
+               'brw_wm.c',
+#              'brw_wm_constant_buffer.c',
+               'brw_wm_debug.c',
+               'brw_wm_emit.c',
+               'brw_wm_fp.c',
+#              'brw_wm_glsl.c',
+               'brw_wm_iz.c',
+               'brw_wm_pass0.c',
+               'brw_wm_pass1.c',
+               'brw_wm_pass2.c',
+               'brw_wm_sampler_state.c',
+               'brw_wm_state.c',
+               'brw_wm_surface_state.c',
+               'intel_decode.c',
+       ])
+
+Export('i965')
diff --git a/src/gallium/drivers/i965/brw_batchbuffer.c b/src/gallium/drivers/i965/brw_batchbuffer.c
new file mode 100644 (file)
index 0000000..22607dc
--- /dev/null
@@ -0,0 +1,202 @@
+/**************************************************************************
+ * 
+ * Copyright 2006 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include "util/u_memory.h"
+
+#include "brw_batchbuffer.h"
+#include "brw_reg.h"
+#include "brw_winsys.h"
+#include "brw_debug.h"
+#include "brw_structs.h"
+
+#define ALWAYS_EMIT_MI_FLUSH 1
+
+enum pipe_error
+brw_batchbuffer_reset(struct brw_batchbuffer *batch)
+{
+   enum pipe_error ret;
+
+   ret = batch->sws->bo_alloc( batch->sws,
+                               BRW_BUFFER_TYPE_BATCH,
+                               BRW_BATCH_SIZE, 4096,
+                               &batch->buf );
+   if (ret)
+      return ret;
+
+   batch->size = BRW_BATCH_SIZE;
+
+   /* With map_range semantics, the winsys can decide whether to
+    * inject a malloc'ed bounce buffer instead of mapping directly.
+    */
+   batch->map = batch->sws->bo_map(batch->buf,
+                                   BRW_DATA_BATCH_BUFFER,
+                                   0, batch->size,
+                                   GL_TRUE,
+                                   GL_TRUE,
+                                   GL_TRUE);
+
+   batch->ptr = batch->map;
+   return PIPE_OK;
+}
+
+struct brw_batchbuffer *
+brw_batchbuffer_alloc(struct brw_winsys_screen *sws,
+                      struct brw_chipset chipset)
+{
+   struct brw_batchbuffer *batch = CALLOC_STRUCT(brw_batchbuffer);
+
+   batch->sws = sws;
+   batch->chipset = chipset;
+   brw_batchbuffer_reset(batch);
+
+   return batch;
+}
+
+void
+brw_batchbuffer_free(struct brw_batchbuffer *batch)
+{
+   if (batch->map) {
+      batch->sws->bo_unmap(batch->buf);
+      batch->map = NULL;
+   }
+
+   bo_reference(&batch->buf, NULL);
+   FREE(batch);
+}
+
+
+void
+_brw_batchbuffer_flush(struct brw_batchbuffer *batch, 
+                      const char *file,
+                      int line)
+{
+   GLuint used = batch->ptr - batch->map;
+
+   if (used == 0)
+      return;
+
+   /* Post-swap throttling done by the state tracker.
+    */
+
+   if (BRW_DEBUG & DEBUG_BATCH)
+      debug_printf("%s:%d: Batchbuffer flush with %db used\n", 
+                  file, line, used);
+
+   if (ALWAYS_EMIT_MI_FLUSH) {
+      *(GLuint *) (batch->ptr) = MI_FLUSH | BRW_FLUSH_STATE_CACHE;
+      batch->ptr += 4;
+      used = batch->ptr - batch->map;
+   }
+
+   /* Round batchbuffer usage to 2 DWORDs. 
+    */
+   if ((used & 4) == 0) {
+      *(GLuint *) (batch->ptr) = 0; /* noop */
+      batch->ptr += 4;
+      used = batch->ptr - batch->map;
+   }
+
+   /* Mark the end of the buffer. 
+    */
+   *(GLuint *) (batch->ptr) = MI_BATCH_BUFFER_END;
+   batch->ptr += 4;
+   used = batch->ptr - batch->map;
+
+   batch->sws->bo_flush_range(batch->buf, 0, used);
+   batch->sws->bo_unmap(batch->buf);
+   batch->map = NULL;
+   batch->ptr = NULL;
+      
+   batch->sws->bo_exec(batch->buf, used );
+
+   if (BRW_DEBUG & DEBUG_SYNC) {
+      /* Abuse map/unmap to achieve wait-for-fence.
+       *
+       * XXX: hide this inside the winsys and export a fence
+       * interface.
+       */
+      debug_printf("waiting for idle\n");
+      batch->sws->bo_wait_idle(batch->buf);
+   }
+
+   /* Reset the buffer:
+    */
+   brw_batchbuffer_reset(batch);
+}
+
+
+/* The OUT_RELOC() macro ends up here, generating a relocation within
+ * the batch buffer.
+ */
+enum pipe_error
+brw_batchbuffer_emit_reloc(struct brw_batchbuffer *batch,
+                          struct brw_winsys_buffer *buffer,
+                          uint32_t usage,
+                          uint32_t delta)
+{
+   int ret;
+
+   if (batch->ptr - batch->map > batch->buf->size) {
+      debug_printf("bad relocation ptr %p map %p offset %d size %d\n",
+                  batch->ptr, batch->map, batch->ptr - batch->map, batch->buf->size);
+
+      return PIPE_ERROR_OUT_OF_MEMORY;
+   }
+
+   ret = batch->sws->bo_emit_reloc(batch->buf,
+                                  usage,
+                                  delta, 
+                                  batch->ptr - batch->map,
+                                  buffer);
+   if (ret != 0)
+      return ret;
+
+   /* bo_emit_reloc was resposible for writing a zero into the
+    * batchbuffer if necessary.  Just need to update our pointer.
+    */
+   batch->ptr += 4;
+
+   return 0;
+}
+
+enum pipe_error
+brw_batchbuffer_data(struct brw_batchbuffer *batch,
+                       const void *data, GLuint bytes,
+                      enum cliprect_mode cliprect_mode)
+{
+   enum pipe_error ret;
+
+   assert((bytes & 3) == 0);
+
+   ret = brw_batchbuffer_require_space(batch, bytes);
+   if (ret)
+      return ret;
+
+   memcpy(batch->ptr, data, bytes);
+   batch->ptr += bytes;
+   return 0;
+}
diff --git a/src/gallium/drivers/i965/brw_batchbuffer.h b/src/gallium/drivers/i965/brw_batchbuffer.h
new file mode 100644 (file)
index 0000000..7473f5b
--- /dev/null
@@ -0,0 +1,148 @@
+#ifndef BRW_BATCHBUFFER_H
+#define BRW_BATCHBUFFER_H
+
+#include "util/u_debug.h"
+
+#include "brw_types.h"
+#include "brw_winsys.h"
+#include "brw_reg.h"
+
+#define BATCH_SZ 16384
+#define BATCH_RESERVED 16
+
+/* All ignored:
+ */
+enum cliprect_mode {
+   IGNORE_CLIPRECTS,
+   LOOP_CLIPRECTS,
+   NO_LOOP_CLIPRECTS,
+   REFERENCES_CLIPRECTS
+};
+
+
+
+
+struct brw_batchbuffer {
+
+   struct brw_winsys_screen *sws;
+   struct brw_winsys_buffer *buf;
+   struct brw_chipset chipset;
+
+   /**
+    * Values exported to speed up the writing the batchbuffer,
+    * instead of having to go trough a accesor function for
+    * each dword written.
+    */
+   /*{@*/
+   uint8_t *map;
+   uint8_t *ptr;
+   size_t size;
+   struct {
+      uint8_t *end_ptr;
+   } emit;
+
+
+   size_t relocs;
+   size_t max_relocs;
+   /*@}*/
+};
+
+struct brw_batchbuffer *brw_batchbuffer_alloc( struct brw_winsys_screen *sws,
+                                               struct brw_chipset chipset );
+
+void brw_batchbuffer_free(struct brw_batchbuffer *batch);
+
+void _brw_batchbuffer_flush(struct brw_batchbuffer *batch,
+                             const char *file, int line);
+
+
+enum pipe_error
+brw_batchbuffer_reset(struct brw_batchbuffer *batch);
+
+
+/* Unlike bmBufferData, this currently requires the buffer be mapped.
+ * Consider it a convenience function wrapping multple
+ * intel_buffer_dword() calls.
+ */
+int brw_batchbuffer_data(struct brw_batchbuffer *batch,
+                            const void *data, GLuint bytes,
+                           enum cliprect_mode cliprect_mode);
+
+
+int brw_batchbuffer_emit_reloc(struct brw_batchbuffer *batch,
+                              struct brw_winsys_buffer *buffer,
+                              enum brw_buffer_usage usage,
+                              uint32_t offset);
+
+/* Inline functions - might actually be better off with these
+ * non-inlined.  Certainly better off switching all command packets to
+ * be passed as structs rather than dwords, but that's a little bit of
+ * work...
+ */
+static INLINE GLint
+brw_batchbuffer_space(struct brw_batchbuffer *batch)
+{
+   return (batch->size - BATCH_RESERVED) - (batch->ptr - batch->map);
+}
+
+
+static INLINE void
+brw_batchbuffer_emit_dword(struct brw_batchbuffer *batch, GLuint dword)
+{
+   assert(batch->map);
+   assert(brw_batchbuffer_space(batch) >= 4);
+   *(GLuint *) (batch->ptr) = dword;
+   batch->ptr += 4;
+}
+
+static INLINE enum pipe_error
+brw_batchbuffer_require_space(struct brw_batchbuffer *batch,
+                                GLuint sz)
+{
+   assert(sz < batch->size - 8);
+   if (brw_batchbuffer_space(batch) < sz) {
+      assert(0);
+      return PIPE_ERROR_OUT_OF_MEMORY;
+   }
+#ifdef DEBUG
+   batch->emit.end_ptr = batch->ptr + sz;
+#endif
+   return 0;
+}
+
+/* Here are the crusty old macros, to be removed:
+ */
+#define BEGIN_BATCH(n, cliprect_mode) do {                             \
+      brw_batchbuffer_require_space(brw->batch, (n)*4);                        \
+   } while (0)
+
+#define OUT_BATCH(d) brw_batchbuffer_emit_dword(brw->batch, d)
+
+#define OUT_RELOC(buf, usage, delta) do {                              \
+      assert((unsigned) (delta) < buf->size);                          \
+      brw_batchbuffer_emit_reloc(brw->batch, buf,                      \
+                                usage, delta);                         \
+   } while (0)
+
+#ifdef DEBUG
+#define ADVANCE_BATCH() do {                                           \
+      unsigned int _n = brw->batch->ptr - brw->batch->emit.end_ptr;    \
+      if (_n != 0) {                                                   \
+        debug_printf("%s: %d too many bytes emitted to batch\n",       \
+                     __FUNCTION__, _n);                                \
+        abort();                                                       \
+      }                                                                        \
+      brw->batch->emit.end_ptr = NULL;                                 \
+   } while(0)
+#else
+#define ADVANCE_BATCH()
+#endif
+
+static INLINE void
+brw_batchbuffer_emit_mi_flush(struct brw_batchbuffer *batch)
+{
+   brw_batchbuffer_require_space(batch, 4);
+   brw_batchbuffer_emit_dword(batch, MI_FLUSH);
+}
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_cc.c b/src/gallium/drivers/i965/brw_cc.c
new file mode 100644 (file)
index 0000000..3e070f5
--- /dev/null
@@ -0,0 +1,111 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+
+
+static enum pipe_error prepare_cc_vp( struct brw_context *brw )
+{
+   return brw_cache_data( &brw->cache, 
+                         BRW_CC_VP,
+                         &brw->curr.ccv,
+                         NULL, 0,
+                         &brw->cc.reloc[CC_RELOC_VP].bo );
+}
+
+const struct brw_tracked_state brw_cc_vp = {
+   .dirty = {
+      .mesa = PIPE_NEW_VIEWPORT,
+      .brw = BRW_NEW_CONTEXT,
+      .cache = 0
+   },
+   .prepare = prepare_cc_vp
+};
+
+
+/* A long-winded way to OR two unsigned integers together:
+ */
+static INLINE struct brw_cc3
+combine_cc3( struct brw_cc3 a, struct brw_cc3 b )
+{
+   union { struct brw_cc3 cc3; unsigned i; } ca, cb;
+   ca.cc3 = a;
+   cb.cc3 = b;
+   ca.i |= cb.i;
+   return ca.cc3;
+}
+
+
+static int prepare_cc_unit( struct brw_context *brw )
+{
+   brw->cc.cc.cc0 = brw->curr.zstencil->cc0;
+   brw->cc.cc.cc1 = brw->curr.zstencil->cc1;
+   brw->cc.cc.cc2 = brw->curr.zstencil->cc2;
+   brw->cc.cc.cc3 = combine_cc3( brw->curr.zstencil->cc3, brw->curr.blend->cc3 );
+   
+   brw->cc.cc.cc5 = brw->curr.blend->cc5;
+   brw->cc.cc.cc6 = brw->curr.blend->cc6;
+   brw->cc.cc.cc7 = brw->curr.zstencil->cc7;
+
+   return brw_cache_data_sz(&brw->cache, BRW_CC_UNIT,
+                           &brw->cc.cc, sizeof(brw->cc.cc),
+                           brw->cc.reloc, 1,
+                           &brw->cc.state_bo);
+}
+
+const struct brw_tracked_state brw_cc_unit = {
+   .dirty = {
+      .mesa = PIPE_NEW_DEPTH_STENCIL_ALPHA | PIPE_NEW_BLEND,
+      .brw = 0,
+      .cache = CACHE_NEW_CC_VP
+   },
+   .prepare = prepare_cc_unit,
+};
+
+
+void brw_hw_cc_init( struct brw_context *brw )
+{
+   make_reloc(&brw->cc.reloc[0],
+              BRW_USAGE_STATE,
+              0,
+              offsetof(struct brw_cc_unit_state, cc4),
+              NULL);
+}
+
+
+void brw_hw_cc_cleanup( struct brw_context *brw )
+{
+   bo_reference(&brw->cc.state_bo, NULL);
+   bo_reference(&brw->cc.reloc[0].bo, NULL);
+}
diff --git a/src/gallium/drivers/i965/brw_clip.c b/src/gallium/drivers/i965/brw_clip.c
new file mode 100644 (file)
index 0000000..d67a1a6
--- /dev/null
@@ -0,0 +1,224 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "pipe/p_state.h"
+
+#include "util/u_math.h"
+
+#include "brw_screen.h"
+#include "brw_batchbuffer.h"
+#include "brw_defines.h"
+#include "brw_context.h"
+#include "brw_eu.h"
+#include "brw_util.h"
+#include "brw_state.h"
+#include "brw_pipe_rast.h"
+#include "brw_clip.h"
+
+
+#define FRONT_UNFILLED_BIT  0x1
+#define BACK_UNFILLED_BIT   0x2
+
+
+static enum pipe_error
+compile_clip_prog( struct brw_context *brw,
+                   struct brw_clip_prog_key *key,
+                   struct brw_winsys_buffer **bo_out )
+{
+   enum pipe_error ret;
+   struct brw_clip_compile c;
+   const GLuint *program;
+   GLuint program_size;
+   GLuint delta;
+
+   memset(&c, 0, sizeof(c));
+   
+   /* Begin the compilation:
+    */
+   brw_init_compile(brw, &c.func);
+
+   c.func.single_program_flow = 1;
+
+   c.chipset = brw->chipset;
+   c.key = *key;
+   c.need_ff_sync = c.chipset.is_igdng;
+
+   /* Need to locate the two positions present in vertex + header.
+    * These are currently hardcoded:
+    */
+   c.header_position_offset = ATTR_SIZE;
+
+   if (c.chipset.is_igdng)
+       delta = 3 * REG_SIZE;
+   else
+       delta = REG_SIZE;
+
+   c.offset_hpos = delta + c.key.output_hpos * ATTR_SIZE;
+
+   if (c.key.output_color0 != BRW_OUTPUT_NOT_PRESENT)
+      c.offset_color0 = delta + c.key.output_color0 * ATTR_SIZE;
+
+   if (c.key.output_color1 != BRW_OUTPUT_NOT_PRESENT)
+      c.offset_color1 = delta + c.key.output_color1 * ATTR_SIZE;
+
+   if (c.key.output_bfc0 != BRW_OUTPUT_NOT_PRESENT)
+      c.offset_bfc0 = delta + c.key.output_bfc0 * ATTR_SIZE;
+
+   if (c.key.output_bfc1 != BRW_OUTPUT_NOT_PRESENT)
+      c.offset_bfc1 = delta + c.key.output_bfc1 * ATTR_SIZE;
+
+   if (c.key.output_edgeflag != BRW_OUTPUT_NOT_PRESENT)
+      c.offset_edgeflag = delta + c.key.output_edgeflag * ATTR_SIZE;
+   
+   if (BRW_IS_IGDNG(brw))
+       c.nr_regs = (c.key.nr_attrs + 1) / 2 + 3;  /* are vertices packed, or reg-aligned? */
+   else
+       c.nr_regs = (c.key.nr_attrs + 1) / 2 + 1;  /* are vertices packed, or reg-aligned? */
+
+   c.nr_bytes = c.nr_regs * REG_SIZE;
+
+   c.prog_data.clip_mode = c.key.clip_mode; /* XXX */
+
+   /* For some reason the thread is spawned with only 4 channels
+    * unmasked.  
+    */
+   brw_set_mask_control(&c.func, BRW_MASK_DISABLE);
+
+
+   /* Would ideally have the option of producing a program which could
+    * do all three:
+    */
+   switch (key->primitive) {
+   case PIPE_PRIM_TRIANGLES: 
+      if (key->do_unfilled)
+        brw_emit_unfilled_clip( &c );
+      else
+        brw_emit_tri_clip( &c );
+      break;
+   case PIPE_PRIM_LINES:
+      brw_emit_line_clip( &c );
+      break;
+   case PIPE_PRIM_POINTS:
+      brw_emit_point_clip( &c );
+      break;
+   default:
+      assert(0);
+      return PIPE_ERROR_BAD_INPUT;
+   }
+
+        
+
+   /* get the program
+    */
+   ret = brw_get_program(&c.func, &program, &program_size);
+   if (ret)
+      return ret;
+
+   /* Upload
+    */
+   ret = brw_upload_cache( &brw->cache,
+                           BRW_CLIP_PROG,
+                           &c.key, sizeof(c.key),
+                           NULL, 0,
+                           program, program_size,
+                           &c.prog_data,
+                           &brw->clip.prog_data,
+                           bo_out );
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+/* Calculate interpolants for triangle and line rasterization.
+ */
+static enum pipe_error
+upload_clip_prog(struct brw_context *brw)
+{
+   const struct brw_vertex_shader *vs = brw->curr.vertex_shader;
+   struct brw_clip_prog_key key;
+   enum pipe_error ret;
+
+   /* Populate the key, starting from the almost-complete version from
+    * the rast state. 
+    */
+
+   /* PIPE_NEW_RAST */
+   key = brw->curr.rast->clip_key;
+   
+   /* BRW_NEW_REDUCED_PRIMITIVE */
+   key.primitive = brw->reduced_primitive;
+
+   /* XXX: if edgeflag is moved to a proper TGSI vs output, can remove
+    * dependency on CACHE_NEW_VS_PROG
+    */
+   /* CACHE_NEW_VS_PROG */
+   key.nr_attrs        = brw->vs.prog_data->nr_outputs;
+
+   /* PIPE_NEW_VS */
+   key.output_hpos     = vs->output_hpos;
+   key.output_color0   = vs->output_color0;
+   key.output_color1   = vs->output_color1;
+   key.output_bfc0     = vs->output_bfc0;
+   key.output_bfc1     = vs->output_bfc1;
+   key.output_edgeflag = vs->output_edgeflag;
+
+   /* PIPE_NEW_CLIP */
+   key.nr_userclip = brw->curr.ucp.nr;
+
+   /* Already cached?
+    */
+   if (brw_search_cache(&brw->cache, BRW_CLIP_PROG,
+                        &key, sizeof(key),
+                        NULL, 0,
+                        &brw->clip.prog_data,
+                        &brw->clip.prog_bo))
+      return PIPE_OK;
+
+   /* Compile new program:
+    */
+   ret = compile_clip_prog( brw, &key, &brw->clip.prog_bo );
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+const struct brw_tracked_state brw_clip_prog = {
+   .dirty = {
+      .mesa  = (PIPE_NEW_RAST | 
+               PIPE_NEW_CLIP),
+      .brw   = (BRW_NEW_REDUCED_PRIMITIVE),
+      .cache = CACHE_NEW_VS_PROG
+   },
+   .prepare = upload_clip_prog
+};
diff --git a/src/gallium/drivers/i965/brw_clip.h b/src/gallium/drivers/i965/brw_clip.h
new file mode 100644 (file)
index 0000000..80e3a11
--- /dev/null
@@ -0,0 +1,199 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#ifndef BRW_CLIP_H
+#define BRW_CLIP_H
+
+#include "pipe/p_state.h"
+#include "brw_reg.h"
+#include "brw_eu.h"
+
+#define MAX_VERTS (3+6+6)      
+
+/* Note that if unfilled primitives are being emitted, we have to fix
+ * up polygon offset and flatshading at this point:
+ */
+struct brw_clip_prog_key {
+   GLuint nr_attrs:6;
+   GLuint primitive:4;
+   GLuint nr_userclip:3;
+   GLuint do_flat_shading:1;
+   GLuint do_unfilled:1;
+   GLuint fill_cw:2;           /* includes cull information */
+   GLuint fill_ccw:2;          /* includes cull information */
+   GLuint offset_cw:1;
+   GLuint offset_ccw:1;
+   GLuint copy_bfc_cw:1;
+   GLuint copy_bfc_ccw:1;
+   GLuint clip_mode:3;
+   GLuint output_hpos:6;        /* not always zero? */
+
+   GLuint output_color0:6;
+   GLuint output_color1:6;
+   GLuint output_bfc0:6;
+   GLuint output_bfc1:6;
+   GLuint output_edgeflag:6;
+   GLuint pad1:2;
+   
+   GLfloat offset_factor;
+   GLfloat offset_units;
+};
+
+struct brw_clip_prog_data {
+   GLuint curb_read_length;    /* user planes? */
+   GLuint clip_mode;
+   GLuint urb_read_length;
+   GLuint total_grf;
+};
+
+#define CLIP_LINE   0
+#define CLIP_POINT  1
+#define CLIP_FILL   2
+#define CLIP_CULL   3
+
+
+#define PRIM_MASK  (0x1f)
+
+struct brw_clip_compile {
+   struct brw_compile func;
+   struct brw_clip_prog_key key;
+   struct brw_clip_prog_data prog_data;
+   
+   struct {
+      struct brw_reg R0;
+      struct brw_reg vertex[MAX_VERTS];
+
+      struct brw_reg t;
+      struct brw_reg t0, t1;
+      struct brw_reg dp0, dp1;
+
+      struct brw_reg dpPrev;
+      struct brw_reg dp;
+      struct brw_reg loopcount;
+      struct brw_reg nr_verts;
+      struct brw_reg planemask;
+
+      struct brw_reg inlist;
+      struct brw_reg outlist;
+      struct brw_reg freelist;
+
+      struct brw_reg dir;
+      struct brw_reg tmp0, tmp1;
+      struct brw_reg offset;
+      
+      struct brw_reg fixed_planes;
+      struct brw_reg plane_equation;
+       
+      struct brw_reg ff_sync;
+   } reg;
+
+   /* 3 different ways of expressing vertex size, including
+    * key.nr_attrs.
+    */
+   GLuint nr_regs;
+   GLuint nr_bytes;
+
+   GLuint first_tmp;
+   GLuint last_tmp;
+
+   GLboolean need_direction;
+   struct brw_chipset chipset;
+
+   GLuint last_mrf;
+
+   GLuint header_position_offset;
+   GLboolean need_ff_sync;
+
+   GLuint nr_color_attrs;
+   GLuint offset_color0;
+   GLuint offset_color1;
+   GLuint offset_bfc0;
+   GLuint offset_bfc1;
+
+   GLuint offset_hpos;
+   GLuint offset_edgeflag;
+};
+
+#define ATTR_SIZE  (4*4)
+
+/* Points are only culled, so no need for a clip routine, however it
+ * works out easier to have a dummy one.
+ */
+void brw_emit_unfilled_clip( struct brw_clip_compile *c );
+void brw_emit_tri_clip( struct brw_clip_compile *c );
+void brw_emit_line_clip( struct brw_clip_compile *c );
+void brw_emit_point_clip( struct brw_clip_compile *c );
+
+/* brw_clip_tri.c, for use by the unfilled clip routine:
+ */
+void brw_clip_tri_init_vertices( struct brw_clip_compile *c );
+void brw_clip_tri_flat_shade( struct brw_clip_compile *c );
+void brw_clip_tri( struct brw_clip_compile *c );
+void brw_clip_tri_emit_polygon( struct brw_clip_compile *c );
+void brw_clip_tri_alloc_regs( struct brw_clip_compile *c, 
+                             GLuint nr_verts );
+
+
+/* Utils:
+ */
+
+void brw_clip_interp_vertex( struct brw_clip_compile *c,
+                            struct brw_indirect dest_ptr,
+                            struct brw_indirect v0_ptr, /* from */
+                            struct brw_indirect v1_ptr, /* to */
+                            struct brw_reg t0,
+                            GLboolean force_edgeflag );
+
+void brw_clip_init_planes( struct brw_clip_compile *c );
+
+void brw_clip_emit_vue(struct brw_clip_compile *c, 
+                      struct brw_indirect vert,
+                      GLboolean allocate,
+                      GLboolean eot,
+                      GLuint header);
+
+void brw_clip_kill_thread(struct brw_clip_compile *c);
+
+struct brw_reg brw_clip_plane_stride( struct brw_clip_compile *c );
+struct brw_reg brw_clip_plane0_address( struct brw_clip_compile *c );
+
+void brw_clip_copy_colors( struct brw_clip_compile *c,
+                          GLuint to, GLuint from );
+
+void brw_clip_init_clipmask( struct brw_clip_compile *c );
+
+struct brw_reg get_tmp( struct brw_clip_compile *c );
+
+void brw_clip_project_position(struct brw_clip_compile *c,
+             struct brw_reg pos );
+void brw_clip_ff_sync(struct brw_clip_compile *c);
+void brw_clip_init_ff_sync(struct brw_clip_compile *c);
+#endif
diff --git a/src/gallium/drivers/i965/brw_clip_line.c b/src/gallium/drivers/i965/brw_clip_line.c
new file mode 100644 (file)
index 0000000..54282d9
--- /dev/null
@@ -0,0 +1,271 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "util/u_debug.h"
+
+#include "brw_defines.h"
+#include "brw_eu.h"
+#include "brw_util.h"
+#include "brw_clip.h"
+
+
+
+
+static void brw_clip_line_alloc_regs( struct brw_clip_compile *c )
+{
+   GLuint i = 0,j;
+
+   /* Register usage is static, precompute here:
+    */
+   c->reg.R0 = retype(brw_vec8_grf(i, 0), BRW_REGISTER_TYPE_UD); i++;
+
+   if (c->key.nr_userclip) {
+      c->reg.fixed_planes = brw_vec4_grf(i, 0);
+      i += (6 + c->key.nr_userclip + 1) / 2;
+
+      c->prog_data.curb_read_length = (6 + c->key.nr_userclip + 1) / 2;
+   }
+   else
+      c->prog_data.curb_read_length = 0;
+
+
+   /* Payload vertices plus space for more generated vertices:
+    */
+   for (j = 0; j < 4; j++) {
+      c->reg.vertex[j] = brw_vec4_grf(i, 0);
+      i += c->nr_regs;
+   }
+
+   c->reg.t           = brw_vec1_grf(i, 0);
+   c->reg.t0          = brw_vec1_grf(i, 1);
+   c->reg.t1          = brw_vec1_grf(i, 2);
+   c->reg.planemask   = retype(brw_vec1_grf(i, 3), BRW_REGISTER_TYPE_UD);
+   c->reg.plane_equation = brw_vec4_grf(i, 4);
+   i++;
+
+   c->reg.dp0         = brw_vec1_grf(i, 0); /* fixme - dp4 will clobber r.1,2,3 */
+   c->reg.dp1         = brw_vec1_grf(i, 4);
+   i++;
+
+   if (!c->key.nr_userclip) {
+      c->reg.fixed_planes = brw_vec8_grf(i, 0); 
+      i++;
+   }
+
+   if (c->need_ff_sync) {
+      c->reg.ff_sync = retype(brw_vec1_grf(i, 0), BRW_REGISTER_TYPE_UD);
+      i++;
+   }
+
+   c->first_tmp = i;
+   c->last_tmp = i;
+
+   c->prog_data.urb_read_length = c->nr_regs; /* ? */
+   c->prog_data.total_grf = i;
+}
+
+
+
+/* Line clipping, more or less following the following algorithm:
+ *
+ *  for (p=0;p<MAX_PLANES;p++) {
+ *     if (clipmask & (1 << p)) {
+ *        GLfloat dp0 = DOTPROD( vtx0, plane[p] );
+ *        GLfloat dp1 = DOTPROD( vtx1, plane[p] );
+ *
+ *        if (IS_NEGATIVE(dp1)) {
+ *           GLfloat t = dp1 / (dp1 - dp0);
+ *           if (t > t1) t1 = t;
+ *        } else {
+ *           GLfloat t = dp0 / (dp0 - dp1);
+ *           if (t > t0) t0 = t;
+ *        }
+ *  
+ *        if (t0 + t1 >= 1.0)
+ *           return;
+ *     }
+ *  }
+ *
+ *  interp( ctx, newvtx0, vtx0, vtx1, t0 );
+ *  interp( ctx, newvtx1, vtx1, vtx0, t1 );
+ *
+ */
+static void clip_and_emit_line( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_indirect vtx0     = brw_indirect(0, 0);
+   struct brw_indirect vtx1      = brw_indirect(1, 0);
+   struct brw_indirect newvtx0   = brw_indirect(2, 0);
+   struct brw_indirect newvtx1   = brw_indirect(3, 0);
+   struct brw_indirect plane_ptr = brw_indirect(4, 0);
+   struct brw_instruction *plane_loop;
+   struct brw_instruction *plane_active;
+   struct brw_instruction *is_negative;
+   struct brw_instruction *is_neg2 = NULL;
+   struct brw_instruction *not_culled;
+   struct brw_reg v1_null_ud = retype(vec1(brw_null_reg()), BRW_REGISTER_TYPE_UD);
+
+   brw_MOV(p, get_addr_reg(vtx0),      brw_address(c->reg.vertex[0]));
+   brw_MOV(p, get_addr_reg(vtx1),      brw_address(c->reg.vertex[1]));
+   brw_MOV(p, get_addr_reg(newvtx0),   brw_address(c->reg.vertex[2]));
+   brw_MOV(p, get_addr_reg(newvtx1),   brw_address(c->reg.vertex[3]));
+   brw_MOV(p, get_addr_reg(plane_ptr), brw_clip_plane0_address(c));
+
+   /* Note: init t0, t1 together: 
+    */
+   brw_MOV(p, vec2(c->reg.t0), brw_imm_f(0));
+
+   brw_clip_init_planes(c);
+   brw_clip_init_clipmask(c);
+
+   /* -ve rhw workaround */
+   if (c->chipset.is_965) {
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+      brw_AND(p, brw_null_reg(), get_element_ud(c->reg.R0, 2),
+              brw_imm_ud(1<<20));
+      brw_OR(p, c->reg.planemask, c->reg.planemask, brw_imm_ud(0x3f));
+   }
+
+   brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+   plane_loop = brw_DO(p, BRW_EXECUTE_1);
+   {
+      /* if (planemask & 1)
+       */
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+      brw_AND(p, v1_null_ud, c->reg.planemask, brw_imm_ud(1));
+      
+      plane_active = brw_IF(p, BRW_EXECUTE_1);
+      {
+        if (c->key.nr_userclip)
+           brw_MOV(p, c->reg.plane_equation, deref_4f(plane_ptr, 0));
+        else
+           brw_MOV(p, c->reg.plane_equation, deref_4b(plane_ptr, 0));
+
+        /* dp = DP4(vtx->position, plane) 
+         */
+        brw_DP4(p, vec4(c->reg.dp0), deref_4f(vtx0, c->offset_hpos), c->reg.plane_equation);
+
+        /* if (IS_NEGATIVE(dp1)) 
+         */
+        brw_set_conditionalmod(p, BRW_CONDITIONAL_L);
+        brw_DP4(p, vec4(c->reg.dp1), deref_4f(vtx1, c->offset_hpos), c->reg.plane_equation);
+        is_negative = brw_IF(p, BRW_EXECUTE_1);
+        {
+             /*
+              * Both can be negative on GM965/G965 due to RHW workaround
+              * if so, this object should be rejected.
+              */
+             if (c->chipset.is_965) {
+                 brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_LE, c->reg.dp0, brw_imm_f(0.0));
+                 is_neg2 = brw_IF(p, BRW_EXECUTE_1);
+                 {
+                     brw_clip_kill_thread(c);
+                 }
+                 brw_ENDIF(p, is_neg2);
+             }
+
+             brw_ADD(p, c->reg.t, c->reg.dp1, negate(c->reg.dp0));
+             brw_math_invert(p, c->reg.t, c->reg.t);
+             brw_MUL(p, c->reg.t, c->reg.t, c->reg.dp1);
+
+             brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_G, c->reg.t, c->reg.t1 );
+             brw_MOV(p, c->reg.t1, c->reg.t);
+             brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+        } 
+        is_negative = brw_ELSE(p, is_negative);
+        {
+             /* Coming back in.  We know that both cannot be negative
+              * because the line would have been culled in that case.
+              */
+
+             /* If both are positive, do nothing */
+             /* Only on GM965/G965 */
+             if (c->chipset.is_965) {
+                 brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_L, c->reg.dp0, brw_imm_f(0.0));
+                 is_neg2 = brw_IF(p, BRW_EXECUTE_1);
+             }
+
+             {
+                 brw_ADD(p, c->reg.t, c->reg.dp0, negate(c->reg.dp1));
+                 brw_math_invert(p, c->reg.t, c->reg.t);
+                 brw_MUL(p, c->reg.t, c->reg.t, c->reg.dp0);
+
+                 brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_G, c->reg.t, c->reg.t0 );
+                 brw_MOV(p, c->reg.t0, c->reg.t);
+                 brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+             }
+
+             if (c->chipset.is_965) {
+                 brw_ENDIF(p, is_neg2);
+             }
+         }
+        brw_ENDIF(p, is_negative);      
+      }
+      brw_ENDIF(p, plane_active);
+      
+      /* plane_ptr++;
+       */
+      brw_ADD(p, get_addr_reg(plane_ptr), get_addr_reg(plane_ptr), brw_clip_plane_stride(c));
+
+      /* while (planemask>>=1) != 0
+       */
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+      brw_SHR(p, c->reg.planemask, c->reg.planemask, brw_imm_ud(1));
+   }
+   brw_WHILE(p, plane_loop);
+
+   brw_ADD(p, c->reg.t, c->reg.t0, c->reg.t1);
+   brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_L, c->reg.t, brw_imm_f(1.0));
+   not_culled = brw_IF(p, BRW_EXECUTE_1);
+   {
+      brw_clip_interp_vertex(c, newvtx0, vtx0, vtx1, c->reg.t0, FALSE);
+      brw_clip_interp_vertex(c, newvtx1, vtx1, vtx0, c->reg.t1, FALSE);
+
+      brw_clip_emit_vue(c, newvtx0, 1, 0, (_3DPRIM_LINESTRIP << 2) | R02_PRIM_START);
+      brw_clip_emit_vue(c, newvtx1, 0, 1, (_3DPRIM_LINESTRIP << 2) | R02_PRIM_END); 
+   }
+   brw_ENDIF(p, not_culled);
+   brw_clip_kill_thread(c);
+}
+
+
+
+void brw_emit_line_clip( struct brw_clip_compile *c )
+{
+   brw_clip_line_alloc_regs(c);
+   brw_clip_init_ff_sync(c);
+
+   if (c->key.do_flat_shading)
+      brw_clip_copy_colors(c, 0, 1);
+                
+   clip_and_emit_line(c);
+}
diff --git a/src/gallium/drivers/i965/brw_clip_point.c b/src/gallium/drivers/i965/brw_clip_point.c
new file mode 100644 (file)
index 0000000..e0a5330
--- /dev/null
@@ -0,0 +1,48 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "brw_defines.h"
+#include "brw_eu.h"
+#include "brw_util.h"
+#include "brw_clip.h"
+
+
+/* Point clipping, nothing to do?
+ */
+void brw_emit_point_clip( struct brw_clip_compile *c )
+{
+   /* Send an empty message to kill the thread:
+    */
+   brw_clip_tri_alloc_regs(c, 0);
+   brw_clip_init_ff_sync(c);
+
+   brw_clip_kill_thread(c);
+}
diff --git a/src/gallium/drivers/i965/brw_clip_state.c b/src/gallium/drivers/i965/brw_clip_state.c
new file mode 100644 (file)
index 0000000..5c3ccfd
--- /dev/null
@@ -0,0 +1,209 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "util/u_math.h"
+
+#include "brw_context.h"
+#include "brw_clip.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+#include "brw_debug.h"
+
+struct brw_clip_unit_key {
+   unsigned int total_grf;
+   unsigned int urb_entry_read_length;
+   unsigned int curb_entry_read_length;
+   unsigned int clip_mode;
+
+   unsigned int curbe_offset;
+
+   unsigned int nr_urb_entries, urb_size;
+
+   GLboolean depth_clamp;
+};
+
+static void
+clip_unit_populate_key(struct brw_context *brw, struct brw_clip_unit_key *key)
+{
+   memset(key, 0, sizeof(*key));
+
+   /* CACHE_NEW_CLIP_PROG */
+   key->total_grf = brw->clip.prog_data->total_grf;
+   key->urb_entry_read_length = brw->clip.prog_data->urb_read_length;
+   key->curb_entry_read_length = brw->clip.prog_data->curb_read_length;
+   key->clip_mode = brw->clip.prog_data->clip_mode;
+
+   /* BRW_NEW_CURBE_OFFSETS */
+   key->curbe_offset = brw->curbe.clip_start;
+
+   /* BRW_NEW_URB_FENCE */
+   key->nr_urb_entries = brw->urb.nr_clip_entries;
+   key->urb_size = brw->urb.vsize;
+
+   /*  */
+   key->depth_clamp = 0; /* XXX: add this to gallium: ctx->Transform.DepthClamp; */
+}
+
+static enum pipe_error
+clip_unit_create_from_key(struct brw_context *brw,
+                          struct brw_clip_unit_key *key,
+                          struct brw_winsys_reloc *reloc,
+                          struct brw_winsys_buffer **bo_out)
+{
+   struct brw_clip_unit_state clip;
+   enum pipe_error ret;
+
+   memset(&clip, 0, sizeof(clip));
+
+   clip.thread0.grf_reg_count = align(key->total_grf, 16) / 16 - 1;
+   /* reloc */
+   clip.thread0.kernel_start_pointer = 0;
+
+   clip.thread1.floating_point_mode = BRW_FLOATING_POINT_NON_IEEE_754;
+   clip.thread1.single_program_flow = 1;
+
+   clip.thread3.urb_entry_read_length = key->urb_entry_read_length;
+   clip.thread3.const_urb_entry_read_length = key->curb_entry_read_length;
+   clip.thread3.const_urb_entry_read_offset = key->curbe_offset * 2;
+   clip.thread3.dispatch_grf_start_reg = 1;
+   clip.thread3.urb_entry_read_offset = 0;
+
+   clip.thread4.nr_urb_entries = key->nr_urb_entries;
+   clip.thread4.urb_entry_allocation_size = key->urb_size - 1;
+   /* If we have enough clip URB entries to run two threads, do so.
+    */
+   if (key->nr_urb_entries >= 10) {
+      /* Half of the URB entries go to each thread, and it has to be an
+       * even number.
+       */
+      assert(key->nr_urb_entries % 2 == 0);
+      
+      /* Although up to 16 concurrent Clip threads are allowed on IGDNG, 
+       * only 2 threads can output VUEs at a time.
+       */
+      if (BRW_IS_IGDNG(brw))
+         clip.thread4.max_threads = 16 - 1;        
+      else
+         clip.thread4.max_threads = 2 - 1;
+   } else {
+      assert(key->nr_urb_entries >= 5);
+      clip.thread4.max_threads = 1 - 1;
+   }
+
+   if (BRW_DEBUG & DEBUG_SINGLE_THREAD)
+      clip.thread4.max_threads = 0;
+
+   if (BRW_DEBUG & DEBUG_STATS)
+      clip.thread4.stats_enable = 1;
+
+   clip.clip5.userclip_enable_flags = 0x7f;
+   clip.clip5.userclip_must_clip = 1;
+   clip.clip5.guard_band_enable = 0;
+   if (!key->depth_clamp)
+      clip.clip5.viewport_z_clip_enable = 1;
+   clip.clip5.viewport_xy_clip_enable = 1;
+   clip.clip5.vertex_position_space = BRW_CLIP_NDCSPACE;
+   clip.clip5.api_mode = BRW_CLIP_API_OGL;
+   clip.clip5.clip_mode = key->clip_mode;
+
+   if (BRW_IS_G4X(brw))
+      clip.clip5.negative_w_clip_test = 1;
+
+   clip.clip6.clipper_viewport_state_ptr = 0;
+   clip.viewport_xmin = -1;
+   clip.viewport_xmax = 1;
+   clip.viewport_ymin = -1;
+   clip.viewport_ymax = 1;
+
+   ret = brw_upload_cache(&brw->cache, BRW_CLIP_UNIT,
+                          key, sizeof(*key),
+                          reloc, 1,
+                          &clip, sizeof(clip),
+                          NULL, NULL,
+                          bo_out);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+static int upload_clip_unit( struct brw_context *brw )
+{
+   struct brw_clip_unit_key key;
+   struct brw_winsys_reloc reloc[1];
+   unsigned grf_reg_count;
+   enum pipe_error ret;
+
+   clip_unit_populate_key(brw, &key);
+
+   grf_reg_count = align(key.total_grf, 16) / 16 - 1;
+
+   /* clip program relocation
+    *
+    * XXX: these reloc structs are long lived and only need to be
+    * updated when the bound BO changes.  Hopefully the stuff mixed in
+    * in the delta's is non-orthogonal.
+    */
+   assert(brw->clip.prog_bo);
+   make_reloc(&reloc[0],
+              BRW_USAGE_STATE,
+              grf_reg_count << 1,
+              offsetof(struct brw_clip_unit_state, thread0),
+              brw->clip.prog_bo);
+
+
+   if (brw_search_cache(&brw->cache, BRW_CLIP_UNIT,
+                        &key, sizeof(key),
+                        reloc, 1,
+                        NULL,
+                        &brw->clip.state_bo))
+      return PIPE_OK;
+      
+   /* Create new:
+    */
+   ret = clip_unit_create_from_key(brw, &key, 
+                                   reloc,
+                                   &brw->clip.state_bo);
+   if (ret)
+      return ret;
+   
+   return PIPE_OK;
+}
+
+const struct brw_tracked_state brw_clip_unit = {
+   .dirty = {
+      .mesa  = 0,
+      .brw   = (BRW_NEW_CURBE_OFFSETS |
+               BRW_NEW_URB_FENCE),
+      .cache = CACHE_NEW_CLIP_PROG
+   },
+   .prepare = upload_clip_unit,
+};
diff --git a/src/gallium/drivers/i965/brw_clip_tri.c b/src/gallium/drivers/i965/brw_clip_tri.c
new file mode 100644 (file)
index 0000000..4cde729
--- /dev/null
@@ -0,0 +1,595 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "brw_defines.h"
+#include "brw_eu.h"
+#include "brw_util.h"
+#include "brw_clip.h"
+
+static void release_tmps( struct brw_clip_compile *c )
+{
+   c->last_tmp = c->first_tmp;
+}
+
+
+void brw_clip_tri_alloc_regs( struct brw_clip_compile *c, 
+                             GLuint nr_verts )
+{
+   GLuint i = 0,j;
+
+   /* Register usage is static, precompute here:
+    */
+   c->reg.R0 = retype(brw_vec8_grf(i, 0), BRW_REGISTER_TYPE_UD); i++;
+
+   if (c->key.nr_userclip) {
+      c->reg.fixed_planes = brw_vec4_grf(i, 0);
+      i += (6 + c->key.nr_userclip + 1) / 2;
+
+      c->prog_data.curb_read_length = (6 + c->key.nr_userclip + 1) / 2;
+   }
+   else
+      c->prog_data.curb_read_length = 0;
+
+
+   /* Payload vertices plus space for more generated vertices:
+    */
+   for (j = 0; j < nr_verts; j++) {
+      c->reg.vertex[j] = brw_vec4_grf(i, 0);
+      i += c->nr_regs;
+   }
+
+   if (c->key.nr_attrs & 1) {
+      for (j = 0; j < 3; j++) {
+        GLuint delta = c->key.nr_attrs*16 + 32;
+
+         if (c->chipset.is_igdng)
+             delta = c->key.nr_attrs * 16 + 32 * 3;
+
+        brw_MOV(&c->func, byte_offset(c->reg.vertex[j], delta), brw_imm_f(0));
+      }
+   }
+
+   c->reg.t          = brw_vec1_grf(i, 0);
+   c->reg.loopcount  = retype(brw_vec1_grf(i, 1), BRW_REGISTER_TYPE_D);
+   c->reg.nr_verts   = retype(brw_vec1_grf(i, 2), BRW_REGISTER_TYPE_UD);
+   c->reg.planemask  = retype(brw_vec1_grf(i, 3), BRW_REGISTER_TYPE_UD);
+   c->reg.plane_equation = brw_vec4_grf(i, 4);
+   i++;
+
+   c->reg.dpPrev     = brw_vec1_grf(i, 0); /* fixme - dp4 will clobber r.1,2,3 */
+   c->reg.dp         = brw_vec1_grf(i, 4);
+   i++;
+
+   c->reg.inlist     = brw_uw16_reg(BRW_GENERAL_REGISTER_FILE, i, 0);
+   i++;
+
+   c->reg.outlist    = brw_uw16_reg(BRW_GENERAL_REGISTER_FILE, i, 0);
+   i++;
+
+   c->reg.freelist   = brw_uw16_reg(BRW_GENERAL_REGISTER_FILE, i, 0);
+   i++;
+
+   if (!c->key.nr_userclip) {
+      c->reg.fixed_planes = brw_vec8_grf(i, 0); 
+      i++;
+   }
+
+   if (c->key.do_unfilled) {
+      c->reg.dir     = brw_vec4_grf(i, 0);
+      c->reg.offset  = brw_vec4_grf(i, 4);
+      i++;
+      c->reg.tmp0    = brw_vec4_grf(i, 0);
+      c->reg.tmp1    = brw_vec4_grf(i, 4);
+      i++;
+   }
+
+   if (c->need_ff_sync) {
+      c->reg.ff_sync = retype(brw_vec1_grf(i, 0), BRW_REGISTER_TYPE_UD);
+      i++;
+   }
+
+   c->first_tmp = i;
+   c->last_tmp = i;
+
+   c->prog_data.urb_read_length = c->nr_regs; /* ? */
+   c->prog_data.total_grf = i;
+}
+
+
+
+void brw_clip_tri_init_vertices( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg tmp0 = c->reg.loopcount; /* handy temporary */
+   struct brw_instruction *is_rev;
+
+   /* Initial list of indices for incoming vertexes:
+    */
+   brw_AND(p, tmp0, get_element_ud(c->reg.R0, 2), brw_imm_ud(PRIM_MASK)); 
+   brw_CMP(p, 
+          vec1(brw_null_reg()), 
+          BRW_CONDITIONAL_EQ, 
+          tmp0,
+          brw_imm_ud(_3DPRIM_TRISTRIP_REVERSE));
+
+   /* XXX: Is there an easier way to do this?  Need to reverse every
+    * second tristrip element:  Can ignore sometimes?
+    */
+   is_rev = brw_IF(p, BRW_EXECUTE_1);
+   {   
+      brw_MOV(p, get_element(c->reg.inlist, 0),  brw_address(c->reg.vertex[1]) );
+      brw_MOV(p, get_element(c->reg.inlist, 1),  brw_address(c->reg.vertex[0]) );
+      if (c->need_direction)
+        brw_MOV(p, c->reg.dir, brw_imm_f(-1));
+   }
+   is_rev = brw_ELSE(p, is_rev);
+   {
+      brw_MOV(p, get_element(c->reg.inlist, 0),  brw_address(c->reg.vertex[0]) );
+      brw_MOV(p, get_element(c->reg.inlist, 1),  brw_address(c->reg.vertex[1]) );
+      if (c->need_direction)
+        brw_MOV(p, c->reg.dir, brw_imm_f(1));
+   }
+   brw_ENDIF(p, is_rev);
+
+   brw_MOV(p, get_element(c->reg.inlist, 2),  brw_address(c->reg.vertex[2]) );
+   brw_MOV(p, brw_vec8_grf(c->reg.outlist.nr, 0), brw_imm_f(0));
+   brw_MOV(p, c->reg.nr_verts, brw_imm_ud(3));
+}
+
+
+
+void brw_clip_tri_flat_shade( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *is_poly;
+   struct brw_reg tmp0 = c->reg.loopcount; /* handy temporary */
+
+   brw_AND(p, tmp0, get_element_ud(c->reg.R0, 2), brw_imm_ud(PRIM_MASK)); 
+   brw_CMP(p, 
+          vec1(brw_null_reg()), 
+          BRW_CONDITIONAL_EQ, 
+          tmp0,
+          brw_imm_ud(_3DPRIM_POLYGON));
+
+   is_poly = brw_IF(p, BRW_EXECUTE_1);
+   {   
+      brw_clip_copy_colors(c, 1, 0);
+      brw_clip_copy_colors(c, 2, 0);
+   }
+   is_poly = brw_ELSE(p, is_poly);
+   {
+      brw_clip_copy_colors(c, 0, 2);
+      brw_clip_copy_colors(c, 1, 2);
+   }
+   brw_ENDIF(p, is_poly);
+}
+
+
+
+/* Use mesa's clipping algorithms, translated to GEN4 assembly.
+ */
+void brw_clip_tri( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_indirect vtx = brw_indirect(0, 0);
+   struct brw_indirect vtxPrev = brw_indirect(1, 0);
+   struct brw_indirect vtxOut = brw_indirect(2, 0);
+   struct brw_indirect plane_ptr = brw_indirect(3, 0);
+   struct brw_indirect inlist_ptr = brw_indirect(4, 0);
+   struct brw_indirect outlist_ptr = brw_indirect(5, 0);
+   struct brw_indirect freelist_ptr = brw_indirect(6, 0);
+   struct brw_instruction *plane_loop;
+   struct brw_instruction *plane_active;
+   struct brw_instruction *vertex_loop;
+   struct brw_instruction *next_test;
+   struct brw_instruction *prev_test;
+   
+   brw_MOV(p, get_addr_reg(vtxPrev),     brw_address(c->reg.vertex[2]) );
+   brw_MOV(p, get_addr_reg(plane_ptr),   brw_clip_plane0_address(c));
+   brw_MOV(p, get_addr_reg(inlist_ptr),  brw_address(c->reg.inlist));
+   brw_MOV(p, get_addr_reg(outlist_ptr), brw_address(c->reg.outlist));
+
+   brw_MOV(p, get_addr_reg(freelist_ptr), brw_address(c->reg.vertex[3]) );
+
+   plane_loop = brw_DO(p, BRW_EXECUTE_1);
+   {
+      /* if (planemask & 1)
+       */
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+      brw_AND(p, vec1(brw_null_reg()), c->reg.planemask, brw_imm_ud(1));
+      
+      plane_active = brw_IF(p, BRW_EXECUTE_1);
+      {
+        /* vtxOut = freelist_ptr++ 
+         */
+        brw_MOV(p, get_addr_reg(vtxOut),       get_addr_reg(freelist_ptr) );
+        brw_ADD(p, get_addr_reg(freelist_ptr), get_addr_reg(freelist_ptr), brw_imm_uw(c->nr_regs * REG_SIZE));
+
+        if (c->key.nr_userclip)
+           brw_MOV(p, c->reg.plane_equation, deref_4f(plane_ptr, 0));
+        else
+           brw_MOV(p, c->reg.plane_equation, deref_4b(plane_ptr, 0));
+           
+        brw_MOV(p, c->reg.loopcount, c->reg.nr_verts);
+        brw_MOV(p, c->reg.nr_verts, brw_imm_ud(0));
+
+        vertex_loop = brw_DO(p, BRW_EXECUTE_1);
+        {
+           /* vtx = *input_ptr;
+            */
+           brw_MOV(p, get_addr_reg(vtx), deref_1uw(inlist_ptr, 0));
+
+           /* IS_NEGATIVE(prev) */
+           brw_set_conditionalmod(p, BRW_CONDITIONAL_L);
+           brw_DP4(p, vec4(c->reg.dpPrev), deref_4f(vtxPrev, c->offset_hpos), c->reg.plane_equation);
+           prev_test = brw_IF(p, BRW_EXECUTE_1);
+           {
+              /* IS_POSITIVE(next)
+               */
+              brw_set_conditionalmod(p, BRW_CONDITIONAL_GE);
+              brw_DP4(p, vec4(c->reg.dp), deref_4f(vtx, c->offset_hpos), c->reg.plane_equation);
+              next_test = brw_IF(p, BRW_EXECUTE_1);
+              {
+
+                 /* Coming back in.
+                  */
+                 brw_ADD(p, c->reg.t, c->reg.dpPrev, negate(c->reg.dp));
+                 brw_math_invert(p, c->reg.t, c->reg.t);
+                 brw_MUL(p, c->reg.t, c->reg.t, c->reg.dpPrev);
+
+                 /* If (vtxOut == 0) vtxOut = vtxPrev
+                  */
+                 brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_EQ, get_addr_reg(vtxOut), brw_imm_uw(0) );
+                 brw_MOV(p, get_addr_reg(vtxOut), get_addr_reg(vtxPrev) );
+                 brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+                 brw_clip_interp_vertex(c, vtxOut, vtxPrev, vtx, c->reg.t, GL_FALSE);
+
+                 /* *outlist_ptr++ = vtxOut;
+                  * nr_verts++; 
+                  * vtxOut = 0;
+                  */
+                 brw_MOV(p, deref_1uw(outlist_ptr, 0), get_addr_reg(vtxOut));
+                 brw_ADD(p, get_addr_reg(outlist_ptr), get_addr_reg(outlist_ptr), brw_imm_uw(sizeof(short)));
+                 brw_ADD(p, c->reg.nr_verts, c->reg.nr_verts, brw_imm_ud(1));
+                 brw_MOV(p, get_addr_reg(vtxOut), brw_imm_uw(0) );
+              }
+              brw_ENDIF(p, next_test);
+              
+           }
+           prev_test = brw_ELSE(p, prev_test);
+           {
+              /* *outlist_ptr++ = vtxPrev;
+               * nr_verts++;
+               */
+              brw_MOV(p, deref_1uw(outlist_ptr, 0), get_addr_reg(vtxPrev));
+              brw_ADD(p, get_addr_reg(outlist_ptr), get_addr_reg(outlist_ptr), brw_imm_uw(sizeof(short)));
+              brw_ADD(p, c->reg.nr_verts, c->reg.nr_verts, brw_imm_ud(1));
+
+              /* IS_NEGATIVE(next)
+               */
+              brw_set_conditionalmod(p, BRW_CONDITIONAL_L);
+              brw_DP4(p, vec4(c->reg.dp), deref_4f(vtx, c->offset_hpos), c->reg.plane_equation);
+              next_test = brw_IF(p, BRW_EXECUTE_1);
+              {
+                 /* Going out of bounds.  Avoid division by zero as we
+                  * know dp != dpPrev from DIFFERENT_SIGNS, above.
+                  */
+                 brw_ADD(p, c->reg.t, c->reg.dp, negate(c->reg.dpPrev));
+                 brw_math_invert(p, c->reg.t, c->reg.t);
+                 brw_MUL(p, c->reg.t, c->reg.t, c->reg.dp);
+
+                 /* If (vtxOut == 0) vtxOut = vtx
+                  */
+                 brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_EQ, get_addr_reg(vtxOut), brw_imm_uw(0) );
+                 brw_MOV(p, get_addr_reg(vtxOut), get_addr_reg(vtx) );
+                 brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+                 brw_clip_interp_vertex(c, vtxOut, vtx, vtxPrev, c->reg.t, GL_TRUE);             
+
+                 /* *outlist_ptr++ = vtxOut;
+                  * nr_verts++; 
+                  * vtxOut = 0;
+                  */
+                 brw_MOV(p, deref_1uw(outlist_ptr, 0), get_addr_reg(vtxOut));
+                 brw_ADD(p, get_addr_reg(outlist_ptr), get_addr_reg(outlist_ptr), brw_imm_uw(sizeof(short)));
+                 brw_ADD(p, c->reg.nr_verts, c->reg.nr_verts, brw_imm_ud(1));
+                 brw_MOV(p, get_addr_reg(vtxOut), brw_imm_uw(0) );
+              }               
+              brw_ENDIF(p, next_test);
+           }
+           brw_ENDIF(p, prev_test);
+           
+           /* vtxPrev = vtx;
+            * inlist_ptr++;
+            */
+           brw_MOV(p, get_addr_reg(vtxPrev), get_addr_reg(vtx));
+           brw_ADD(p, get_addr_reg(inlist_ptr), get_addr_reg(inlist_ptr), brw_imm_uw(sizeof(short)));
+
+           /* while (--loopcount != 0)
+            */
+           brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+           brw_ADD(p, c->reg.loopcount, c->reg.loopcount, brw_imm_d(-1));
+        } 
+        brw_WHILE(p, vertex_loop);
+
+        /* vtxPrev = *(outlist_ptr-1)  OR: outlist[nr_verts-1]
+         * inlist = outlist
+         * inlist_ptr = &inlist[0]
+         * outlist_ptr = &outlist[0]
+         */
+        brw_ADD(p, get_addr_reg(outlist_ptr), get_addr_reg(outlist_ptr), brw_imm_w(-2));
+        brw_MOV(p, get_addr_reg(vtxPrev), deref_1uw(outlist_ptr, 0));
+        brw_MOV(p, brw_vec8_grf(c->reg.inlist.nr, 0), brw_vec8_grf(c->reg.outlist.nr, 0));
+        brw_MOV(p, get_addr_reg(inlist_ptr), brw_address(c->reg.inlist));
+        brw_MOV(p, get_addr_reg(outlist_ptr), brw_address(c->reg.outlist));
+      }
+      brw_ENDIF(p, plane_active);
+      
+      /* plane_ptr++;
+       */
+      brw_ADD(p, get_addr_reg(plane_ptr), get_addr_reg(plane_ptr), brw_clip_plane_stride(c));
+
+      /* nr_verts >= 3 
+       */
+      brw_CMP(p,
+             vec1(brw_null_reg()),
+             BRW_CONDITIONAL_GE,
+             c->reg.nr_verts,
+             brw_imm_ud(3));
+   
+      /* && (planemask>>=1) != 0
+       */
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+      brw_SHR(p, c->reg.planemask, c->reg.planemask, brw_imm_ud(1));
+   }
+   brw_WHILE(p, plane_loop);
+}
+
+
+
+void brw_clip_tri_emit_polygon(struct brw_clip_compile *c)
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *loop, *if_insn;
+
+   /* for (loopcount = nr_verts-2; loopcount > 0; loopcount--)
+    */
+   brw_set_conditionalmod(p, BRW_CONDITIONAL_G);
+   brw_ADD(p,
+          c->reg.loopcount,
+          c->reg.nr_verts,
+          brw_imm_d(-2));
+
+   if_insn = brw_IF(p, BRW_EXECUTE_1);
+   {
+      struct brw_indirect v0 = brw_indirect(0, 0);
+      struct brw_indirect vptr = brw_indirect(1, 0);
+
+      brw_MOV(p, get_addr_reg(vptr), brw_address(c->reg.inlist));
+      brw_MOV(p, get_addr_reg(v0), deref_1uw(vptr, 0));
+
+      brw_clip_emit_vue(c, v0, 1, 0, ((_3DPRIM_TRIFAN << 2) | R02_PRIM_START));
+      
+      brw_ADD(p, get_addr_reg(vptr), get_addr_reg(vptr), brw_imm_uw(2));
+      brw_MOV(p, get_addr_reg(v0), deref_1uw(vptr, 0));
+
+      loop = brw_DO(p, BRW_EXECUTE_1);
+      {
+        brw_clip_emit_vue(c, v0, 1, 0, (_3DPRIM_TRIFAN << 2));
+  
+        brw_ADD(p, get_addr_reg(vptr), get_addr_reg(vptr), brw_imm_uw(2));
+        brw_MOV(p, get_addr_reg(v0), deref_1uw(vptr, 0));
+
+        brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+        brw_ADD(p, c->reg.loopcount, c->reg.loopcount, brw_imm_d(-1));
+      }
+      brw_WHILE(p, loop);
+
+      brw_clip_emit_vue(c, v0, 0, 1, ((_3DPRIM_TRIFAN << 2) | R02_PRIM_END));
+   }
+   brw_ENDIF(p, if_insn);
+}
+
+static void do_clip_tri( struct brw_clip_compile *c )
+{
+   brw_clip_init_planes(c);
+
+   brw_clip_tri(c);
+}
+
+
+static void maybe_do_clip_tri( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *do_clip;
+
+   brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_NZ, c->reg.planemask, brw_imm_ud(0));
+   do_clip = brw_IF(p, BRW_EXECUTE_1);
+   {
+      do_clip_tri(c);
+   }
+   brw_ENDIF(p, do_clip);
+}
+
+static void brw_clip_test( struct brw_clip_compile *c )
+{
+    struct brw_reg t = retype(get_tmp(c), BRW_REGISTER_TYPE_UD);
+    struct brw_reg t1 = retype(get_tmp(c), BRW_REGISTER_TYPE_UD);
+    struct brw_reg t2 = retype(get_tmp(c), BRW_REGISTER_TYPE_UD);
+    struct brw_reg t3 = retype(get_tmp(c), BRW_REGISTER_TYPE_UD);
+
+    struct brw_reg v0 = get_tmp(c);
+    struct brw_reg v1 = get_tmp(c);
+    struct brw_reg v2 = get_tmp(c);
+
+    struct brw_indirect vt0 = brw_indirect(0, 0);
+    struct brw_indirect vt1 = brw_indirect(1, 0);
+    struct brw_indirect vt2 = brw_indirect(2, 0);
+
+    struct brw_compile *p = &c->func;
+    struct brw_instruction *is_outside;
+    struct brw_reg tmp0 = c->reg.loopcount; /* handy temporary */
+
+    brw_MOV(p, get_addr_reg(vt0), brw_address(c->reg.vertex[0]));
+    brw_MOV(p, get_addr_reg(vt1), brw_address(c->reg.vertex[1]));
+    brw_MOV(p, get_addr_reg(vt2), brw_address(c->reg.vertex[2]));
+    brw_MOV(p, v0, deref_4f(vt0, c->offset_hpos));
+    brw_MOV(p, v1, deref_4f(vt1, c->offset_hpos));
+    brw_MOV(p, v2, deref_4f(vt2, c->offset_hpos));
+    brw_AND(p, c->reg.planemask, c->reg.planemask, brw_imm_ud(~0x3f));
+
+    /* test nearz, xmin, ymin plane */
+    /* clip.xyz < -clip.w */
+    brw_CMP(p, t1, BRW_CONDITIONAL_L, v0, negate(get_element(v0, 3))); 
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+    brw_CMP(p, t2, BRW_CONDITIONAL_L, v1, negate(get_element(v1, 3))); 
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+    brw_CMP(p, t3, BRW_CONDITIONAL_L, v2, negate(get_element(v2, 3))); 
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+    /* All vertices are outside of a plane, rejected */
+    brw_AND(p, t, t1, t2);
+    brw_AND(p, t, t, t3);
+    brw_OR(p, tmp0, get_element(t, 0), get_element(t, 1));
+    brw_OR(p, tmp0, tmp0, get_element(t, 2));
+    brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+    brw_AND(p, brw_null_reg(), tmp0, brw_imm_ud(0x1));
+    is_outside = brw_IF(p, BRW_EXECUTE_1);
+    {
+        brw_clip_kill_thread(c);
+    }
+    brw_ENDIF(p, is_outside);
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+    /* some vertices are inside a plane, some are outside,need to clip */
+    brw_XOR(p, t, t1, t2);
+    brw_XOR(p, t1, t2, t3);
+    brw_OR(p, t, t, t1);
+    brw_AND(p, t, t, brw_imm_ud(0x1));
+    brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_NZ,
+            get_element(t, 0), brw_imm_ud(0));
+    brw_OR(p, c->reg.planemask, c->reg.planemask, brw_imm_ud((1<<5)));
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+    brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_NZ,
+            get_element(t, 1), brw_imm_ud(0));
+    brw_OR(p, c->reg.planemask, c->reg.planemask, brw_imm_ud((1<<3)));
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+    brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_NZ,
+            get_element(t, 2), brw_imm_ud(0));
+    brw_OR(p, c->reg.planemask, c->reg.planemask, brw_imm_ud((1<<1)));
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+    /* test farz, xmax, ymax plane */
+    /* clip.xyz > clip.w */
+    brw_CMP(p, t1, BRW_CONDITIONAL_G, v0, get_element(v0, 3)); 
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+    brw_CMP(p, t2, BRW_CONDITIONAL_G, v1, get_element(v1, 3)); 
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+    brw_CMP(p, t3, BRW_CONDITIONAL_G, v2, get_element(v2, 3)); 
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+    /* All vertices are outside of a plane, rejected */
+    brw_AND(p, t, t1, t2);
+    brw_AND(p, t, t, t3);
+    brw_OR(p, tmp0, get_element(t, 0), get_element(t, 1));
+    brw_OR(p, tmp0, tmp0, get_element(t, 2));
+    brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+    brw_AND(p, brw_null_reg(), tmp0, brw_imm_ud(0x1));
+    is_outside = brw_IF(p, BRW_EXECUTE_1);
+    {
+        brw_clip_kill_thread(c);
+    }
+    brw_ENDIF(p, is_outside);
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+    /* some vertices are inside a plane, some are outside,need to clip */
+    brw_XOR(p, t, t1, t2);
+    brw_XOR(p, t1, t2, t3);
+    brw_OR(p, t, t, t1);
+    brw_AND(p, t, t, brw_imm_ud(0x1));
+    brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_NZ,
+            get_element(t, 0), brw_imm_ud(0));
+    brw_OR(p, c->reg.planemask, c->reg.planemask, brw_imm_ud((1<<4)));
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+    brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_NZ,
+            get_element(t, 1), brw_imm_ud(0));
+    brw_OR(p, c->reg.planemask, c->reg.planemask, brw_imm_ud((1<<2)));
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+    brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_NZ,
+            get_element(t, 2), brw_imm_ud(0));
+    brw_OR(p, c->reg.planemask, c->reg.planemask, brw_imm_ud((1<<0)));
+    brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+    release_tmps(c);
+}
+
+
+void brw_emit_tri_clip( struct brw_clip_compile *c )
+{
+   struct brw_instruction *neg_rhw;
+   struct brw_compile *p = &c->func;
+   brw_clip_tri_alloc_regs(c, 3 + c->key.nr_userclip + 6);
+   brw_clip_tri_init_vertices(c);
+   brw_clip_init_clipmask(c);
+   brw_clip_init_ff_sync(c);
+
+   /* if -ve rhw workaround bit is set, 
+      do cliptest */
+   if (c->chipset.is_965) {
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+      brw_AND(p, brw_null_reg(), get_element_ud(c->reg.R0, 2), 
+              brw_imm_ud(1<<20));
+      neg_rhw = brw_IF(p, BRW_EXECUTE_1); 
+      {
+         brw_clip_test(c);
+      }
+      brw_ENDIF(p, neg_rhw);
+   }
+   /* Can't push into do_clip_tri because with polygon (or quad)
+    * flatshading, need to apply the flatshade here because we don't
+    * respect the PV when converting to trifan for emit:
+    */
+   if (c->key.do_flat_shading) 
+      brw_clip_tri_flat_shade(c); 
+      
+   if ((c->key.clip_mode == BRW_CLIPMODE_NORMAL) ||
+       (c->key.clip_mode == BRW_CLIPMODE_KERNEL_CLIP))
+      do_clip_tri(c);
+   else 
+      maybe_do_clip_tri(c);
+
+   brw_clip_tri_emit_polygon(c);
+
+   /* Send an empty message to kill the thread:
+    */
+   brw_clip_kill_thread(c);
+}
diff --git a/src/gallium/drivers/i965/brw_clip_unfilled.c b/src/gallium/drivers/i965/brw_clip_unfilled.c
new file mode 100644 (file)
index 0000000..aec835b
--- /dev/null
@@ -0,0 +1,497 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "brw_defines.h"
+#include "brw_eu.h"
+#include "brw_util.h"
+#include "brw_clip.h"
+
+
+
+/* This is performed against the original triangles, so no indirection
+ * required:
+BZZZT!
+ */
+static void compute_tri_direction( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg e = c->reg.tmp0;
+   struct brw_reg f = c->reg.tmp1;
+   struct brw_reg v0 = byte_offset(c->reg.vertex[0], c->offset_hpos); 
+   struct brw_reg v1 = byte_offset(c->reg.vertex[1], c->offset_hpos); 
+   struct brw_reg v2 = byte_offset(c->reg.vertex[2], c->offset_hpos); 
+
+
+   struct brw_reg v0n = get_tmp(c);
+   struct brw_reg v1n = get_tmp(c);
+   struct brw_reg v2n = get_tmp(c);
+
+   /* Convert to NDC.
+    * NOTE: We can't modify the original vertex coordinates,
+    * as it may impact further operations.
+    * So, we have to keep normalized coordinates in temp registers.
+    *
+    * TBD-KC
+    * Try to optimize unnecessary MOV's.
+    */
+   brw_MOV(p, v0n, v0);
+   brw_MOV(p, v1n, v1);
+   brw_MOV(p, v2n, v2);
+
+   brw_clip_project_position(c, v0n);
+   brw_clip_project_position(c, v1n);
+   brw_clip_project_position(c, v2n);
+
+   /* Calculate the vectors of two edges of the triangle:
+    */
+   brw_ADD(p, e, v0n, negate(v2n)); 
+   brw_ADD(p, f, v1n, negate(v2n)); 
+
+   /* Take their crossproduct:
+    */
+   brw_set_access_mode(p, BRW_ALIGN_16);
+   brw_MUL(p, vec4(brw_null_reg()), brw_swizzle(e, 1,2,0,3),  brw_swizzle(f,2,0,1,3));
+   brw_MAC(p, vec4(e),  negate(brw_swizzle(e, 2,0,1,3)), brw_swizzle(f,1,2,0,3));
+   brw_set_access_mode(p, BRW_ALIGN_1);
+
+   brw_MUL(p, c->reg.dir, c->reg.dir, vec4(e));
+}
+
+
+static void cull_direction( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *ccw;
+   GLuint conditional;
+
+   assert (!(c->key.fill_ccw == CLIP_CULL &&
+            c->key.fill_cw == CLIP_CULL));
+
+   if (c->key.fill_ccw == CLIP_CULL)
+      conditional = BRW_CONDITIONAL_GE;
+   else
+      conditional = BRW_CONDITIONAL_L;
+
+   brw_CMP(p,
+          vec1(brw_null_reg()),
+          conditional,
+          get_element(c->reg.dir, 2),
+          brw_imm_f(0));
+   
+   ccw = brw_IF(p, BRW_EXECUTE_1);
+   {
+      brw_clip_kill_thread(c);
+   }
+   brw_ENDIF(p, ccw);
+}
+
+
+
+static void copy_bfc( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *ccw;
+   GLuint conditional;
+
+   /* Do we have any colors to copy? 
+    */
+   if ((c->offset_color0 == 0 || c->offset_bfc0 == 0) &&
+       (c->offset_color1 == 0 || c->offset_bfc1 == 0))
+      return;
+
+   /* In some wierd degnerate cases we can end up testing the
+    * direction twice, once for culling and once for bfc copying.  Oh
+    * well, that's what you get for setting wierd GL state.
+    */
+   if (c->key.copy_bfc_ccw)
+      conditional = BRW_CONDITIONAL_GE;
+   else
+      conditional = BRW_CONDITIONAL_L;
+
+   brw_CMP(p,
+          vec1(brw_null_reg()),
+          conditional,
+          get_element(c->reg.dir, 2),
+          brw_imm_f(0));
+   
+   ccw = brw_IF(p, BRW_EXECUTE_1);
+   {
+      GLuint i;
+
+      for (i = 0; i < 3; i++) {
+        if (c->offset_color0 && c->offset_bfc0)
+           brw_MOV(p, 
+                   byte_offset(c->reg.vertex[i], c->offset_color0),
+                   byte_offset(c->reg.vertex[i], c->offset_bfc0));
+
+        if (c->offset_color1 && c->offset_bfc1)
+           brw_MOV(p, 
+                   byte_offset(c->reg.vertex[i], c->offset_color0),
+                   byte_offset(c->reg.vertex[i], c->offset_bfc0));
+      }
+   }
+   brw_ENDIF(p, ccw);
+}
+
+
+
+
+/*
+  GLfloat iz   = 1.0 / dir.z;
+  GLfloat ac   = dir.x * iz;
+  GLfloat bc   = dir.y * iz;
+  offset = ctx->Polygon.OffsetUnits * DEPTH_SCALE;
+  offset += MAX2( abs(ac), abs(bc) ) * ctx->Polygon.OffsetFactor;
+  offset *= MRD;
+*/
+static void compute_offset( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg off = c->reg.offset;
+   struct brw_reg dir = c->reg.dir;
+   
+   brw_math_invert(p, get_element(off, 2), get_element(dir, 2));
+   brw_MUL(p, vec2(off), dir, get_element(off, 2));
+
+   brw_CMP(p, 
+          vec1(brw_null_reg()), 
+          BRW_CONDITIONAL_GE,
+          brw_abs(get_element(off, 0)), 
+          brw_abs(get_element(off, 1)));
+
+   brw_SEL(p, vec1(off), brw_abs(get_element(off, 0)), brw_abs(get_element(off, 1)));
+   brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+   brw_MUL(p, vec1(off), off, brw_imm_f(c->key.offset_factor));
+   brw_ADD(p, vec1(off), off, brw_imm_f(c->key.offset_units));
+}
+
+
+static void merge_edgeflags( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *is_poly;
+   struct brw_reg tmp0 = get_element_ud(c->reg.tmp0, 0);
+
+   brw_AND(p, tmp0, get_element_ud(c->reg.R0, 2), brw_imm_ud(PRIM_MASK)); 
+   brw_CMP(p, 
+          vec1(brw_null_reg()), 
+          BRW_CONDITIONAL_EQ, 
+          tmp0,
+          brw_imm_ud(_3DPRIM_POLYGON));
+
+   /* Get away with using reg.vertex because we know that this is not
+    * a _3DPRIM_TRISTRIP_REVERSE:
+    */
+   is_poly = brw_IF(p, BRW_EXECUTE_1);
+   {   
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_EQ);
+      brw_AND(p, vec1(brw_null_reg()), get_element_ud(c->reg.R0, 2), brw_imm_ud(1<<8));
+      brw_MOV(p, byte_offset(c->reg.vertex[0], c->offset_edgeflag), brw_imm_f(0));
+      brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_EQ);
+      brw_AND(p, vec1(brw_null_reg()), get_element_ud(c->reg.R0, 2), brw_imm_ud(1<<9));
+      brw_MOV(p, byte_offset(c->reg.vertex[2], c->offset_edgeflag), brw_imm_f(0));
+      brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+   }
+   brw_ENDIF(p, is_poly);
+}
+
+
+
+static void apply_one_offset( struct brw_clip_compile *c,
+                         struct brw_indirect vert )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg z = deref_1f(vert, c->header_position_offset +
+                              2 * type_sz(BRW_REGISTER_TYPE_F));
+
+   brw_ADD(p, z, z, vec1(c->reg.offset));
+}
+
+
+
+/***********************************************************************
+ * Output clipped polygon as an unfilled primitive:
+ */
+static void emit_lines(struct brw_clip_compile *c,
+                      GLboolean do_offset)
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *loop;
+   struct brw_instruction *draw_edge;
+   struct brw_indirect v0 = brw_indirect(0, 0);
+   struct brw_indirect v1 = brw_indirect(1, 0);
+   struct brw_indirect v0ptr = brw_indirect(2, 0);
+   struct brw_indirect v1ptr = brw_indirect(3, 0);
+
+   /* Need a seperate loop for offset:
+    */
+   if (do_offset) {
+      brw_MOV(p, c->reg.loopcount, c->reg.nr_verts);
+      brw_MOV(p, get_addr_reg(v0ptr), brw_address(c->reg.inlist));
+
+      loop = brw_DO(p, BRW_EXECUTE_1);
+      {
+        brw_MOV(p, get_addr_reg(v0), deref_1uw(v0ptr, 0));
+        brw_ADD(p, get_addr_reg(v0ptr), get_addr_reg(v0ptr), brw_imm_uw(2));
+           
+        apply_one_offset(c, v0);
+           
+        brw_set_conditionalmod(p, BRW_CONDITIONAL_G);
+        brw_ADD(p, c->reg.loopcount, c->reg.loopcount, brw_imm_d(-1));
+      }
+      brw_WHILE(p, loop);
+   }
+
+   /* v1ptr = &inlist[nr_verts]
+    * *v1ptr = v0
+    */
+   brw_MOV(p, c->reg.loopcount, c->reg.nr_verts);
+   brw_MOV(p, get_addr_reg(v0ptr), brw_address(c->reg.inlist));
+   brw_ADD(p, get_addr_reg(v1ptr), get_addr_reg(v0ptr), retype(c->reg.nr_verts, BRW_REGISTER_TYPE_UW));
+   brw_ADD(p, get_addr_reg(v1ptr), get_addr_reg(v1ptr), retype(c->reg.nr_verts, BRW_REGISTER_TYPE_UW));
+   brw_MOV(p, deref_1uw(v1ptr, 0), deref_1uw(v0ptr, 0));
+
+   loop = brw_DO(p, BRW_EXECUTE_1);
+   {
+      brw_MOV(p, get_addr_reg(v0), deref_1uw(v0ptr, 0));
+      brw_MOV(p, get_addr_reg(v1), deref_1uw(v0ptr, 2));
+      brw_ADD(p, get_addr_reg(v0ptr), get_addr_reg(v0ptr), brw_imm_uw(2));
+
+      /* draw edge if edgeflag != 0 */
+      brw_CMP(p, 
+             vec1(brw_null_reg()), BRW_CONDITIONAL_NZ, 
+             deref_1f(v0, c->offset_edgeflag),
+             brw_imm_f(0));
+      draw_edge = brw_IF(p, BRW_EXECUTE_1);
+      {
+        brw_clip_emit_vue(c, v0, 1, 0, (_3DPRIM_LINESTRIP << 2) | R02_PRIM_START);
+        brw_clip_emit_vue(c, v1, 1, 0, (_3DPRIM_LINESTRIP << 2) | R02_PRIM_END);
+      }
+      brw_ENDIF(p, draw_edge);
+
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+      brw_ADD(p, c->reg.loopcount, c->reg.loopcount, brw_imm_d(-1));
+   }
+   brw_WHILE(p, loop);
+}
+
+
+
+static void emit_points(struct brw_clip_compile *c,
+                       GLboolean do_offset )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *loop;
+   struct brw_instruction *draw_point;
+
+   struct brw_indirect v0 = brw_indirect(0, 0);
+   struct brw_indirect v0ptr = brw_indirect(2, 0);
+
+   brw_MOV(p, c->reg.loopcount, c->reg.nr_verts);
+   brw_MOV(p, get_addr_reg(v0ptr), brw_address(c->reg.inlist));
+
+   loop = brw_DO(p, BRW_EXECUTE_1);
+   {
+      brw_MOV(p, get_addr_reg(v0), deref_1uw(v0ptr, 0));
+      brw_ADD(p, get_addr_reg(v0ptr), get_addr_reg(v0ptr), brw_imm_uw(2));
+
+      /* draw if edgeflag != 0 
+       */
+      brw_CMP(p, 
+             vec1(brw_null_reg()), BRW_CONDITIONAL_NZ, 
+             deref_1f(v0, c->offset_edgeflag),
+             brw_imm_f(0));
+      draw_point = brw_IF(p, BRW_EXECUTE_1);
+      {
+        if (do_offset)
+           apply_one_offset(c, v0);
+
+        brw_clip_emit_vue(c, v0, 1, 0, (_3DPRIM_POINTLIST << 2) | R02_PRIM_START | R02_PRIM_END);
+      }
+      brw_ENDIF(p, draw_point);
+
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+      brw_ADD(p, c->reg.loopcount, c->reg.loopcount, brw_imm_d(-1));
+   }
+   brw_WHILE(p, loop);
+}
+
+
+
+
+
+
+
+static void emit_primitives( struct brw_clip_compile *c,
+                            GLuint mode, 
+                            GLboolean do_offset )
+{
+   switch (mode) {
+   case CLIP_FILL:
+      brw_clip_tri_emit_polygon(c);
+      break;
+
+   case CLIP_LINE:
+      emit_lines(c, do_offset);
+      break;
+
+   case CLIP_POINT:
+      emit_points(c, do_offset);
+      break;
+
+   case CLIP_CULL:
+      assert(0);
+      break;
+   }
+} 
+
+
+
+static void emit_unfilled_primitives( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *ccw;
+
+   /* Direction culling has already been done.
+    */
+   if (c->key.fill_ccw != c->key.fill_cw &&
+       c->key.fill_ccw != CLIP_CULL &&
+       c->key.fill_cw != CLIP_CULL)
+   {
+      brw_CMP(p,
+             vec1(brw_null_reg()),
+             BRW_CONDITIONAL_GE,
+             get_element(c->reg.dir, 2),
+             brw_imm_f(0));
+   
+      ccw = brw_IF(p, BRW_EXECUTE_1);
+      {
+        emit_primitives(c, c->key.fill_ccw, c->key.offset_ccw);
+      }
+      ccw = brw_ELSE(p, ccw);
+      {
+        emit_primitives(c, c->key.fill_cw, c->key.offset_cw);
+      }
+      brw_ENDIF(p, ccw);
+   }
+   else if (c->key.fill_cw != CLIP_CULL) {
+      emit_primitives(c, c->key.fill_cw, c->key.offset_cw);
+   }
+   else if (c->key.fill_ccw != CLIP_CULL) { 
+      emit_primitives(c, c->key.fill_ccw, c->key.offset_ccw);
+   }
+}
+
+
+
+
+static void check_nr_verts( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *if_insn;
+
+   brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_L, c->reg.nr_verts, brw_imm_d(3));      
+   if_insn = brw_IF(p, BRW_EXECUTE_1);
+   {
+      brw_clip_kill_thread(c);
+   }
+   brw_ENDIF(p, if_insn);
+}
+
+
+void brw_emit_unfilled_clip( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *do_clip;
+   
+
+   c->need_direction = ((c->key.offset_ccw || c->key.offset_cw) ||
+                       (c->key.fill_ccw != c->key.fill_cw) ||
+                       c->key.fill_ccw == CLIP_CULL ||
+                       c->key.fill_cw == CLIP_CULL ||
+                       c->key.copy_bfc_cw ||
+                       c->key.copy_bfc_ccw);
+
+   brw_clip_tri_alloc_regs(c, 3 + c->key.nr_userclip + 6);
+   brw_clip_tri_init_vertices(c);
+   brw_clip_init_ff_sync(c);
+
+   assert(c->offset_edgeflag);
+
+   if (c->key.fill_ccw == CLIP_CULL &&
+       c->key.fill_cw == CLIP_CULL) {
+      brw_clip_kill_thread(c);
+      return;
+   }
+
+   merge_edgeflags(c);
+
+   /* Need to use the inlist indirection here: 
+    */
+   if (c->need_direction) 
+      compute_tri_direction(c);
+   
+   if (c->key.fill_ccw == CLIP_CULL ||
+       c->key.fill_cw == CLIP_CULL)
+      cull_direction(c);
+
+   if (c->key.offset_ccw ||
+       c->key.offset_cw)
+      compute_offset(c);
+
+   if (c->key.copy_bfc_ccw ||
+       c->key.copy_bfc_cw)
+      copy_bfc(c);
+
+   /* Need to do this whether we clip or not:
+    */
+   if (c->key.do_flat_shading)
+      brw_clip_tri_flat_shade(c);
+   
+   brw_clip_init_clipmask(c);
+   brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_NZ, c->reg.planemask, brw_imm_ud(0));
+   do_clip = brw_IF(p, BRW_EXECUTE_1);
+   {
+      brw_clip_init_planes(c);
+      brw_clip_tri(c);
+      check_nr_verts(c);
+   }
+   brw_ENDIF(p, do_clip);
+   
+   emit_unfilled_primitives(c);
+   brw_clip_kill_thread(c);
+}
+
+
+
diff --git a/src/gallium/drivers/i965/brw_clip_util.c b/src/gallium/drivers/i965/brw_clip_util.c
new file mode 100644 (file)
index 0000000..97a5710
--- /dev/null
@@ -0,0 +1,388 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+
+#include "brw_defines.h"
+#include "brw_eu.h"
+#include "brw_util.h"
+#include "brw_clip.h"
+
+
+
+
+struct brw_reg get_tmp( struct brw_clip_compile *c )
+{
+   struct brw_reg tmp = brw_vec4_grf(c->last_tmp, 0);
+
+   if (++c->last_tmp > c->prog_data.total_grf)
+      c->prog_data.total_grf = c->last_tmp;
+
+   return tmp;
+}
+
+static void release_tmp( struct brw_clip_compile *c, struct brw_reg tmp )
+{
+   if (tmp.nr == c->last_tmp-1)
+      c->last_tmp--;
+}
+
+
+static struct brw_reg make_plane_ud(GLuint x, GLuint y, GLuint z, GLuint w)
+{
+   return brw_imm_ud((w<<24) | (z<<16) | (y<<8) | x);
+}
+
+
+void brw_clip_init_planes( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+
+   if (!c->key.nr_userclip) {
+      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 0), make_plane_ud( 0,    0, 0xff, 1));
+      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 1), make_plane_ud( 0,    0,    1, 1));
+      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 2), make_plane_ud( 0, 0xff,    0, 1));
+      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 3), make_plane_ud( 0,    1,    0, 1));
+      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 4), make_plane_ud(0xff,  0,    0, 1));
+      brw_MOV(p, get_element_ud(c->reg.fixed_planes, 5), make_plane_ud( 1,    0,    0, 1));
+   }
+}
+
+
+
+#define W 3
+
+/* Project 'pos' to screen space (or back again), overwrite with results:
+ */
+void brw_clip_project_position(struct brw_clip_compile *c, struct brw_reg pos )
+{
+   struct brw_compile *p = &c->func;
+
+   /* calc rhw 
+    */
+   brw_math_invert(p, get_element(pos, W), get_element(pos, W));
+
+   /* value.xyz *= value.rhw
+    */
+   brw_set_access_mode(p, BRW_ALIGN_16);
+   brw_MUL(p, brw_writemask(pos, BRW_WRITEMASK_XYZ), pos, brw_swizzle1(pos, W));
+   brw_set_access_mode(p, BRW_ALIGN_1);
+}
+
+
+static void brw_clip_project_vertex( struct brw_clip_compile *c, 
+                                    struct brw_indirect vert_addr )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg tmp = get_tmp(c);
+
+   /* Fixup position.  Extract from the original vertex and re-project
+    * to screen space:
+    */
+   brw_MOV(p, tmp, deref_4f(vert_addr, c->offset_hpos));
+   brw_clip_project_position(c, tmp);
+   brw_MOV(p, deref_4f(vert_addr, c->header_position_offset), tmp);
+        
+   release_tmp(c, tmp);
+}
+
+
+
+
+/* Interpolate between two vertices and put the result into a0.0.  
+ * Increment a0.0 accordingly.
+ */
+void brw_clip_interp_vertex( struct brw_clip_compile *c,
+                            struct brw_indirect dest_ptr,
+                            struct brw_indirect v0_ptr, /* from */
+                            struct brw_indirect v1_ptr, /* to */
+                            struct brw_reg t0,
+                            GLboolean force_edgeflag)
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg tmp = get_tmp(c);
+   GLuint i;
+
+   /* Just copy the vertex header:
+    */
+   /*
+    * After CLIP stage, only first 256 bits of the VUE are read
+    * back on IGDNG, so needn't change it
+    */
+   brw_copy_indirect_to_indirect(p, dest_ptr, v0_ptr, 1);
+      
+   /* Iterate over each attribute (could be done in pairs?)
+    */
+   for (i = 0; i < c->key.nr_attrs; i++) {
+      GLuint delta = i*16 + 32;
+
+      if (c->chipset.is_igdng)
+          delta = i * 16 + 32 * 3;
+
+      if (delta == c->offset_edgeflag) {
+        if (force_edgeflag) 
+           brw_MOV(p, deref_4f(dest_ptr, delta), brw_imm_f(1));
+        else
+           brw_MOV(p, deref_4f(dest_ptr, delta), deref_4f(v0_ptr, delta));
+      }
+      else {
+        /* Interpolate: 
+         *
+         *        New = attr0 + t*attr1 - t*attr0
+         */
+        brw_MUL(p, 
+                vec4(brw_null_reg()),
+                deref_4f(v1_ptr, delta),
+                t0);
+
+        brw_MAC(p, 
+                tmp,         
+                negate(deref_4f(v0_ptr, delta)),
+                t0); 
+             
+        brw_ADD(p,
+                deref_4f(dest_ptr, delta), 
+                deref_4f(v0_ptr, delta),
+                tmp);
+      }
+   }
+
+   if (i & 1) {
+      GLuint delta = i*16 + 32;
+
+      if (c->chipset.is_igdng)
+          delta = i * 16 + 32 * 3;
+
+      brw_MOV(p, deref_4f(dest_ptr, delta), brw_imm_f(0));
+   }
+
+   release_tmp(c, tmp);
+
+   /* Recreate the projected (NDC) coordinate in the new vertex
+    * header:
+    */
+   brw_clip_project_vertex(c, dest_ptr );
+}
+
+
+
+
+#define MAX_MRF 16
+
+void brw_clip_emit_vue(struct brw_clip_compile *c, 
+                      struct brw_indirect vert,
+                      GLboolean allocate,
+                      GLboolean eot,
+                      GLuint header)
+{
+   struct brw_compile *p = &c->func;
+   GLuint start = c->last_mrf;
+
+   brw_clip_ff_sync(c);
+
+   assert(!(allocate && eot));
+   
+   /* Cycle through mrf regs - probably futile as we have to wait for
+    * the allocation response anyway.  Also, the order this function
+    * is invoked doesn't correspond to the order the instructions will
+    * be executed, so it won't have any effect in many cases.
+    */
+#if 0
+   if (start + c->nr_regs + 1 >= MAX_MRF)
+      start = 0;
+
+   c->last_mrf = start + c->nr_regs + 1;
+#endif
+       
+   /* Copy the vertex from vertn into m1..mN+1:
+    */
+   brw_copy_from_indirect(p, brw_message_reg(start+1), vert, c->nr_regs);
+
+   /* Overwrite PrimType and PrimStart in the message header, for
+    * each vertex in turn:
+    */
+   brw_MOV(p, get_element_ud(c->reg.R0, 2), brw_imm_ud(header));
+
+
+   /* Send each vertex as a seperate write to the urb.  This
+    * is different to the concept in brw_sf_emit.c, where
+    * subsequent writes are used to build up a single urb
+    * entry.  Each of these writes instantiates a seperate
+    * urb entry - (I think... what about 'allocate'?)
+    */
+   brw_urb_WRITE(p, 
+                allocate ? c->reg.R0 : retype(brw_null_reg(), BRW_REGISTER_TYPE_UD),
+                start,
+                c->reg.R0,
+                allocate,
+                1,             /* used */
+                c->nr_regs + 1, /* msg length */
+                allocate ? 1 : 0, /* response_length */ 
+                eot,           /* eot */
+                1,             /* writes_complete */
+                0,             /* urb offset */
+                BRW_URB_SWIZZLE_NONE);
+}
+
+
+
+void brw_clip_kill_thread(struct brw_clip_compile *c)
+{
+   struct brw_compile *p = &c->func;
+
+   brw_clip_ff_sync(c);
+   /* Send an empty message to kill the thread and release any
+    * allocated urb entry:
+    */
+   brw_urb_WRITE(p, 
+                retype(brw_null_reg(), BRW_REGISTER_TYPE_UD),
+                0,
+                c->reg.R0,
+                0,             /* allocate */
+                0,             /* used */
+                1,             /* msg len */
+                0,             /* response len */
+                1,             /* eot */
+                1,             /* writes complete */
+                0,
+                BRW_URB_SWIZZLE_NONE);
+}
+
+
+
+
+struct brw_reg brw_clip_plane0_address( struct brw_clip_compile *c )
+{
+   return brw_address(c->reg.fixed_planes);
+}
+
+
+struct brw_reg brw_clip_plane_stride( struct brw_clip_compile *c )
+{
+   if (c->key.nr_userclip) {
+      return brw_imm_uw(16);
+   }
+   else {
+      return brw_imm_uw(4);
+   }
+}
+
+
+/* If flatshading, distribute color from provoking vertex prior to
+ * clipping.
+ */
+void brw_clip_copy_colors( struct brw_clip_compile *c,
+                          GLuint to, GLuint from )
+{
+   struct brw_compile *p = &c->func;
+
+   if (c->offset_color0)
+      brw_MOV(p, 
+             byte_offset(c->reg.vertex[to], c->offset_color0),
+             byte_offset(c->reg.vertex[from], c->offset_color0));
+
+   if (c->offset_color1)
+      brw_MOV(p, 
+             byte_offset(c->reg.vertex[to], c->offset_color1),
+             byte_offset(c->reg.vertex[from], c->offset_color1));
+
+   if (c->offset_bfc0)
+      brw_MOV(p, 
+             byte_offset(c->reg.vertex[to], c->offset_bfc0),
+             byte_offset(c->reg.vertex[from], c->offset_bfc0));
+
+   if (c->offset_bfc1)
+      brw_MOV(p, 
+             byte_offset(c->reg.vertex[to], c->offset_bfc1),
+             byte_offset(c->reg.vertex[from], c->offset_bfc1));
+}
+
+
+
+void brw_clip_init_clipmask( struct brw_clip_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg incoming = get_element_ud(c->reg.R0, 2);
+   
+   /* Shift so that lowest outcode bit is rightmost: 
+    */
+   brw_SHR(p, c->reg.planemask, incoming, brw_imm_ud(26));
+
+   if (c->key.nr_userclip) {
+      struct brw_reg tmp = retype(vec1(get_tmp(c)), BRW_REGISTER_TYPE_UD);
+
+      /* Rearrange userclip outcodes so that they come directly after
+       * the fixed plane bits.
+       */
+      brw_AND(p, tmp, incoming, brw_imm_ud(0x3f<<14));
+      brw_SHR(p, tmp, tmp, brw_imm_ud(8));
+      brw_OR(p, c->reg.planemask, c->reg.planemask, tmp);
+      
+      release_tmp(c, tmp);
+   }
+}
+
+void brw_clip_ff_sync(struct brw_clip_compile *c)
+{
+    if (c->need_ff_sync) {
+        struct brw_compile *p = &c->func;
+        struct brw_instruction *need_ff_sync;
+
+        brw_set_conditionalmod(p, BRW_CONDITIONAL_Z);
+        brw_AND(p, brw_null_reg(), c->reg.ff_sync, brw_imm_ud(0x1));
+        need_ff_sync = brw_IF(p, BRW_EXECUTE_1);
+        {
+            brw_OR(p, c->reg.ff_sync, c->reg.ff_sync, brw_imm_ud(0x1));
+            brw_ff_sync(p, 
+                    c->reg.R0,
+                    0,
+                    c->reg.R0,
+                    1, 
+                    1,         /* used */
+                    1,         /* msg length */
+                    1,         /* response length */
+                    0,         /* eot */
+                    1,         /* write compelete */
+                    0,         /* urb offset */
+                    BRW_URB_SWIZZLE_NONE);
+        }
+        brw_ENDIF(p, need_ff_sync);
+        brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+    }
+}
+
+void brw_clip_init_ff_sync(struct brw_clip_compile *c)
+{
+    if (c->need_ff_sync) {
+       struct brw_compile *p = &c->func;
+        
+        brw_MOV(p, c->reg.ff_sync, brw_imm_ud(0));
+    }
+}
diff --git a/src/gallium/drivers/i965/brw_context.c b/src/gallium/drivers/i965/brw_context.c
new file mode 100644 (file)
index 0000000..e675518
--- /dev/null
@@ -0,0 +1,154 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+
+#include "pipe/p_context.h"
+#include "util/u_simple_list.h"
+
+#include "brw_context.h"
+#include "brw_defines.h"
+#include "brw_draw.h"
+#include "brw_state.h"
+#include "brw_batchbuffer.h"
+#include "brw_winsys.h"
+#include "brw_screen.h"
+
+
+static void brw_destroy_context( struct pipe_context *pipe )
+{
+   struct brw_context *brw = brw_context(pipe);
+   int i;
+
+   brw_context_flush( brw );
+   brw_batchbuffer_free( brw->batch );
+   brw_destroy_state(brw);
+
+   brw_draw_cleanup( brw );
+
+   brw_pipe_blend_cleanup( brw );
+   brw_pipe_depth_stencil_cleanup( brw );
+   brw_pipe_framebuffer_cleanup( brw );
+   brw_pipe_flush_cleanup( brw );
+   brw_pipe_misc_cleanup( brw );
+   brw_pipe_query_cleanup( brw );
+   brw_pipe_rast_cleanup( brw );
+   brw_pipe_sampler_cleanup( brw );
+   brw_pipe_shader_cleanup( brw );
+   brw_pipe_vertex_cleanup( brw );
+   brw_pipe_clear_cleanup( brw );
+
+   brw_hw_cc_cleanup( brw );
+
+
+   FREE(brw->wm.compile_data);
+
+   for (i = 0; i < brw->curr.fb.nr_cbufs; i++)
+      pipe_surface_reference(&brw->curr.fb.cbufs[i], NULL);
+   brw->curr.fb.nr_cbufs = 0;
+   pipe_surface_reference(&brw->curr.fb.zsbuf, NULL);
+
+   bo_reference(&brw->curbe.curbe_bo, NULL);
+   bo_reference(&brw->vs.prog_bo, NULL);
+   bo_reference(&brw->vs.state_bo, NULL);
+   bo_reference(&brw->vs.bind_bo, NULL);
+   bo_reference(&brw->gs.prog_bo, NULL);
+   bo_reference(&brw->gs.state_bo, NULL);
+   bo_reference(&brw->clip.prog_bo, NULL);
+   bo_reference(&brw->clip.state_bo, NULL);
+   bo_reference(&brw->clip.vp_bo, NULL);
+   bo_reference(&brw->sf.prog_bo, NULL);
+   bo_reference(&brw->sf.state_bo, NULL);
+   bo_reference(&brw->sf.vp_bo, NULL);
+
+   for (i = 0; i < Elements(brw->wm.sdc_bo); i++)
+      bo_reference(&brw->wm.sdc_bo[i], NULL);
+
+   bo_reference(&brw->wm.bind_bo, NULL);
+
+   for (i = 0; i < Elements(brw->wm.surf_bo); i++)
+      bo_reference(&brw->wm.surf_bo[i], NULL);
+
+   bo_reference(&brw->wm.sampler_bo, NULL);
+   bo_reference(&brw->wm.prog_bo, NULL);
+   bo_reference(&brw->wm.state_bo, NULL);
+}
+
+
+struct pipe_context *brw_create_context(struct pipe_screen *screen)
+{
+   struct brw_context *brw = (struct brw_context *) CALLOC_STRUCT(brw_context);
+
+   if (!brw) {
+      debug_printf("%s: failed to alloc context\n", __FUNCTION__);
+      return NULL;
+   }
+
+   brw->base.screen = screen;
+   brw->base.destroy = brw_destroy_context;
+   brw->sws = brw_screen(screen)->sws;
+   brw->chipset = brw_screen(screen)->chipset;
+
+   brw_pipe_blend_init( brw );
+   brw_pipe_depth_stencil_init( brw );
+   brw_pipe_framebuffer_init( brw );
+   brw_pipe_flush_init( brw );
+   brw_pipe_misc_init( brw );
+   brw_pipe_query_init( brw );
+   brw_pipe_rast_init( brw );
+   brw_pipe_sampler_init( brw );
+   brw_pipe_shader_init( brw );
+   brw_pipe_vertex_init( brw );
+   brw_pipe_clear_init( brw );
+
+   brw_hw_cc_init( brw );
+
+   brw_init_state( brw );
+   brw_draw_init( brw );
+
+   brw->state.dirty.mesa = ~0;
+   brw->state.dirty.brw = ~0;
+
+   brw->flags.always_emit_state = 0;
+
+   make_empty_list(&brw->query.active_head);
+
+   brw->batch = brw_batchbuffer_alloc( brw->sws, brw->chipset );
+   if (brw->batch == NULL)
+      goto fail;
+
+   return &brw->base;
+
+fail:
+   if (brw->batch)
+      brw_batchbuffer_free( brw->batch );
+   return NULL;
+}
+
diff --git a/src/gallium/drivers/i965/brw_context.h b/src/gallium/drivers/i965/brw_context.h
new file mode 100644 (file)
index 0000000..8c006bb
--- /dev/null
@@ -0,0 +1,858 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+
+#ifndef BRWCONTEXT_INC
+#define BRWCONTEXT_INC
+
+#include "brw_structs.h"
+#include "brw_winsys.h"
+#include "brw_reg.h"
+#include "pipe/p_state.h"
+#include "pipe/p_context.h"
+#include "tgsi/tgsi_scan.h"
+
+
+/* Glossary:
+ *
+ * URB - uniform resource buffer.  A mid-sized buffer which is
+ * partitioned between the fixed function units and used for passing
+ * values (vertices, primitives, constants) between them.
+ *
+ * CURBE - constant URB entry.  An urb region (entry) used to hold
+ * constant values which the fixed function units can be instructed to
+ * preload into the GRF when spawning a thread.
+ *
+ * VUE - vertex URB entry.  An urb entry holding a vertex and usually
+ * a vertex header.  The header contains control information and
+ * things like primitive type, Begin/end flags and clip codes.  
+ *
+ * PUE - primitive URB entry.  An urb entry produced by the setup (SF)
+ * unit holding rasterization and interpolation parameters.
+ *
+ * GRF - general register file.  One of several register files
+ * addressable by programmed threads.  The inputs (r0, payload, curbe,
+ * urb) of the thread are preloaded to this area before the thread is
+ * spawned.  The registers are individually 8 dwords wide and suitable
+ * for general usage.  Registers holding thread input values are not
+ * special and may be overwritten.
+ *
+ * MRF - message register file.  Threads communicate (and terminate)
+ * by sending messages.  Message parameters are placed in contiguous
+ * MRF registers.  All program output is via these messages.  URB
+ * entries are populated by sending a message to the shared URB
+ * function containing the new data, together with a control word,
+ * often an unmodified copy of R0.
+ *
+ * R0 - GRF register 0.  Typically holds control information used when
+ * sending messages to other threads.
+ *
+ * EU or GEN4 EU: The name of the programmable subsystem of the
+ * i965 hardware.  Threads are executed by the EU, the registers
+ * described above are part of the EU architecture.
+ *
+ * Fixed function units:
+ *
+ * CS - Command streamer.  Notional first unit, little software
+ * interaction.  Holds the URB entries used for constant data, ie the
+ * CURBEs.
+ *
+ * VF/VS - Vertex Fetch / Vertex Shader.  The fixed function part of
+ * this unit is responsible for pulling vertices out of vertex buffers
+ * in vram and injecting them into the processing pipe as VUEs.  If
+ * enabled, it first passes them to a VS thread which is a good place
+ * for the driver to implement any active vertex shader.
+ *
+ * GS - Geometry Shader.  This corresponds to a new DX10 concept.  If
+ * enabled, incoming strips etc are passed to GS threads in individual
+ * line/triangle/point units.  The GS thread may perform arbitary
+ * computation and emit whatever primtives with whatever vertices it
+ * chooses.  This makes GS an excellent place to implement GL's
+ * unfilled polygon modes, though of course it is capable of much
+ * more.  Additionally, GS is used to translate away primitives not
+ * handled by latter units, including Quads and Lineloops.
+ *
+ * CS - Clipper.  Mesa's clipping algorithms are imported to run on
+ * this unit.  The fixed function part performs cliptesting against
+ * the 6 fixed clipplanes and makes decisions on whether or not the
+ * incoming primitive needs to be passed to a thread for clipping.
+ * User clip planes are handled via cooperation with the VS thread.
+ *
+ * SF - Strips Fans or Setup: Triangles are prepared for
+ * rasterization.  Interpolation coefficients are calculated.
+ * Flatshading and two-side lighting usually performed here.
+ *
+ * WM - Windower.  Interpolation of vertex attributes performed here.
+ * Fragment shader implemented here.  SIMD aspects of EU taken full
+ * advantage of, as pixels are processed in blocks of 16.
+ *
+ * CC - Color Calculator.  No EU threads associated with this unit.
+ * Handles blending and (presumably) depth and stencil testing.
+ */
+
+#define BRW_MAX_CURBE                    (32*16)
+
+
+/* Need a value to say a particular vertex shader output isn't
+ * present.  Limits us to 63 outputs currently.
+ */
+#define BRW_OUTPUT_NOT_PRESENT           ((1<<6)-1)
+
+
+struct brw_context;
+
+struct brw_depth_stencil_state {
+   /* Precalculated hardware state:
+    */
+   struct brw_cc0 cc0;
+   struct brw_cc1 cc1;
+   struct brw_cc2 cc2;
+   struct brw_cc3 cc3;
+   struct brw_cc7 cc7;
+
+   unsigned iz_lookup;
+};
+
+
+struct brw_blend_state {
+   /* Precalculated hardware state:
+    */
+   struct brw_cc2 cc2;
+   struct brw_cc3 cc3;
+   struct brw_cc5 cc5;
+   struct brw_cc6 cc6;
+
+   struct brw_surf_ss0 ss0;
+};
+
+
+struct brw_rasterizer_state;
+
+struct brw_immediate_data {
+   unsigned nr;
+   float (*data)[4];
+};
+
+struct brw_vertex_shader {
+   const struct tgsi_token *tokens;
+   struct brw_winsys_buffer *const_buffer;    /** Program constant buffer/surface */
+
+   struct tgsi_shader_info info;
+   struct brw_immediate_data immediates;
+
+   GLuint has_flow_control:1;
+   GLuint use_const_buffer:1;
+
+   /* Offsets of special vertex shader outputs required for clipping.
+    */
+   GLuint output_hpos:6;        /* not always zero? */
+   GLuint output_color0:6;
+   GLuint output_color1:6;
+   GLuint output_bfc0:6;
+   GLuint output_bfc1:6;
+   GLuint output_edgeflag:6;
+
+   unsigned id;
+};
+
+struct brw_fs_signature {
+   GLuint nr_inputs;
+   struct {
+      GLuint interp:3;          /* TGSI_INTERPOLATE_x */
+      GLuint semantic:5;        /* TGSI_SEMANTIC_x */
+      GLuint semantic_index:24;
+   } input[PIPE_MAX_SHADER_INPUTS];
+};
+
+#define brw_fs_signature_size(s) (offsetof(struct brw_fs_signature, input) + \
+                                  ((s)->nr_inputs * sizeof (s)->input[0])) 
+
+
+struct brw_fragment_shader {
+   const struct tgsi_token *tokens;
+   struct tgsi_shader_info info;
+
+   struct brw_fs_signature signature;
+   struct brw_immediate_data immediates;
+
+   unsigned iz_lookup;
+   /*unsigned wm_lookup;*/
+   
+   unsigned  uses_depth:1;
+   unsigned  has_flow_control:1;
+
+   unsigned id;
+   struct brw_winsys_buffer *const_buffer;    /** Program constant buffer/surface */
+   GLboolean use_const_buffer;
+};
+
+
+struct brw_sampler {
+   struct brw_ss0 ss0;
+   struct brw_ss1 ss1;
+   float border_color[4];
+   struct brw_ss3 ss3;
+};
+
+
+
+#define PIPE_NEW_DEPTH_STENCIL_ALPHA    0x1
+#define PIPE_NEW_RAST                   0x2
+#define PIPE_NEW_BLEND                  0x4
+#define PIPE_NEW_VIEWPORT               0x8
+#define PIPE_NEW_SAMPLERS               0x10
+#define PIPE_NEW_VERTEX_BUFFER          0x20
+#define PIPE_NEW_VERTEX_ELEMENT         0x40
+#define PIPE_NEW_FRAGMENT_SHADER        0x80
+#define PIPE_NEW_VERTEX_SHADER          0x100
+#define PIPE_NEW_FRAGMENT_CONSTANTS     0x200
+#define PIPE_NEW_VERTEX_CONSTANTS       0x400
+#define PIPE_NEW_CLIP                   0x800
+#define PIPE_NEW_INDEX_BUFFER           0x1000
+#define PIPE_NEW_INDEX_RANGE            0x2000
+#define PIPE_NEW_BLEND_COLOR            0x4000
+#define PIPE_NEW_POLYGON_STIPPLE        0x8000
+#define PIPE_NEW_FRAMEBUFFER_DIMENSIONS 0x10000
+#define PIPE_NEW_DEPTH_BUFFER           0x20000
+#define PIPE_NEW_COLOR_BUFFERS          0x40000
+#define PIPE_NEW_QUERY                  0x80000
+#define PIPE_NEW_SCISSOR                0x100000
+#define PIPE_NEW_BOUND_TEXTURES         0x200000
+#define PIPE_NEW_NR_CBUFS               0x400000
+#define PIPE_NEW_FRAGMENT_SIGNATURE     0x800000
+
+
+
+#define BRW_NEW_URB_FENCE               0x1
+#define BRW_NEW_FRAGMENT_PROGRAM        0x2
+#define BRW_NEW_VERTEX_PROGRAM          0x4
+#define BRW_NEW_INPUT_DIMENSIONS        0x8
+#define BRW_NEW_CURBE_OFFSETS           0x10
+#define BRW_NEW_REDUCED_PRIMITIVE       0x20
+#define BRW_NEW_PRIMITIVE               0x40
+#define BRW_NEW_CONTEXT                 0x80
+#define BRW_NEW_WM_INPUT_DIMENSIONS     0x100
+#define BRW_NEW_PSP                     0x800
+#define BRW_NEW_WM_SURFACES            0x1000
+#define BRW_NEW_xxx                     0x2000 /* was FENCE */
+#define BRW_NEW_INDICES                        0x4000
+
+/**
+ * Used for any batch entry with a relocated pointer that will be used
+ * by any 3D rendering.  Need to re-emit these fresh in each
+ * batchbuffer as the referenced buffers may be relocated in the
+ * meantime.
+ */
+#define BRW_NEW_BATCH                  0x10000
+#define BRW_NEW_NR_WM_SURFACES         0x40000
+#define BRW_NEW_NR_VS_SURFACES         0x80000
+#define BRW_NEW_INDEX_BUFFER           0x100000
+
+struct brw_state_flags {
+   /** State update flags signalled by mesa internals */
+   GLuint mesa;
+   /**
+    * State update flags signalled as the result of brw_tracked_state updates
+    */
+   GLuint brw;
+   /** State update flags signalled by brw_state_cache.c searches */
+   GLuint cache;
+};
+
+
+
+/* Data about a particular attempt to compile a program.  Note that
+ * there can be many of these, each in a different GL state
+ * corresponding to a different brw_wm_prog_key struct, with different
+ * compiled programs:
+ */
+struct brw_wm_prog_data {
+   GLuint curb_read_length;
+   GLuint urb_read_length;
+
+   GLuint first_curbe_grf;
+   GLuint total_grf;
+   GLuint total_scratch;
+
+   GLuint nr_params;       /**< number of float params/constants */
+   GLboolean error;
+
+   /* Pointer to tracked values (only valid once
+    * _mesa_load_state_parameters has been called at runtime).
+    */
+   const GLfloat *param[BRW_MAX_CURBE];
+};
+
+struct brw_sf_prog_data {
+   GLuint urb_read_length;
+   GLuint total_grf;
+
+   /* Each vertex may have upto 12 attributes, 4 components each,
+    * except WPOS which requires only 2.  (11*4 + 2) == 44 ==> 11
+    * rows.
+    *
+    * Actually we use 4 for each, so call it 12 rows.
+    */
+   GLuint urb_entry_size;
+};
+
+
+struct brw_clip_prog_data;
+
+struct brw_gs_prog_data {
+   GLuint urb_read_length;
+   GLuint total_grf;
+};
+
+struct brw_vs_prog_data {
+   GLuint curb_read_length;
+   GLuint urb_read_length;
+   GLuint total_grf;
+
+   GLuint nr_outputs;
+   GLuint nr_inputs;
+
+   GLuint nr_params;       /**< number of TGSI_FILE_CONSTANT's */
+
+   GLboolean writes_psiz;
+
+   /* Used for calculating urb partitions:
+    */
+   GLuint urb_entry_size;
+};
+
+
+/* Size == 0 if output either not written, or always [0,0,0,1]
+ */
+struct brw_vs_ouput_sizes {
+   GLubyte output_size[PIPE_MAX_SHADER_OUTPUTS];
+};
+
+
+/** Number of texture sampler units */
+#define BRW_MAX_TEX_UNIT 16
+
+/** Max number of render targets in a shader */
+#define BRW_MAX_DRAW_BUFFERS 4
+
+/**
+ * Size of our surface binding table for the WM.
+ * This contains pointers to the drawing surfaces and current texture
+ * objects and shader constant buffers (+2).
+ */
+#define BRW_WM_MAX_SURF (BRW_MAX_DRAW_BUFFERS + BRW_MAX_TEX_UNIT + 1)
+
+/**
+ * Helpers to convert drawing buffers, textures and constant buffers
+ * to surface binding table indexes, for WM.
+ */
+#define BTI_COLOR_BUF(d)          (d)
+#define BTI_FRAGMENT_CONSTANTS    (BRW_MAX_DRAW_BUFFERS) 
+#define BTI_TEXTURE(t)            (BRW_MAX_DRAW_BUFFERS + 1 + (t))
+
+/**
+ * Size of surface binding table for the VS.
+ * Only one constant buffer for now.
+ */
+#define BRW_VS_MAX_SURF 1
+
+/**
+ * Only a VS constant buffer
+ */
+#define SURF_INDEX_VERT_CONST_BUFFER 0
+
+
+/* Bit of a hack to align these with the winsys buffer_data_type enum.
+ */
+enum brw_cache_id {
+   BRW_CC_VP         = BRW_DATA_GS_CC_VP,
+   BRW_CC_UNIT       = BRW_DATA_GS_CC_UNIT,
+   BRW_WM_PROG       = BRW_DATA_GS_WM_PROG,
+   BRW_SAMPLER_DEFAULT_COLOR    = BRW_DATA_GS_SAMPLER_DEFAULT_COLOR,
+   BRW_SAMPLER       = BRW_DATA_GS_SAMPLER,
+   BRW_WM_UNIT       = BRW_DATA_GS_WM_UNIT,
+   BRW_SF_PROG       = BRW_DATA_GS_SF_PROG,
+   BRW_SF_VP         = BRW_DATA_GS_SF_VP,
+   BRW_SF_UNIT       = BRW_DATA_GS_SF_UNIT,
+   BRW_VS_UNIT       = BRW_DATA_GS_VS_UNIT,
+   BRW_VS_PROG       = BRW_DATA_GS_VS_PROG,
+   BRW_GS_UNIT       = BRW_DATA_GS_GS_UNIT,
+   BRW_GS_PROG       = BRW_DATA_GS_GS_PROG,
+   BRW_CLIP_VP       = BRW_DATA_GS_CLIP_VP,
+   BRW_CLIP_UNIT     = BRW_DATA_GS_CLIP_UNIT,
+   BRW_CLIP_PROG     = BRW_DATA_GS_CLIP_PROG,
+   BRW_SS_SURFACE    = BRW_DATA_SS_SURFACE,
+   BRW_SS_SURF_BIND  = BRW_DATA_SS_SURF_BIND,
+
+   BRW_MAX_CACHE
+};
+
+struct brw_cache_item {
+   /**
+    * Effectively part of the key, cache_id identifies what kind of state
+    * buffer is involved, and also which brw->state.dirty.cache flag should
+    * be set when this cache item is chosen.
+    */
+   enum brw_cache_id cache_id;
+   /** 32-bit hash of the key data */
+   GLuint hash;
+   GLuint key_size;            /* for variable-sized keys */
+   const void *key;
+   struct brw_winsys_reloc *relocs;
+   GLuint nr_relocs;
+
+   struct brw_winsys_buffer *bo;
+   GLuint data_size;
+
+   struct brw_cache_item *next;
+};   
+
+
+
+struct brw_cache {
+   struct brw_context *brw;
+   struct brw_winsys_screen *sws;
+
+   struct brw_cache_item **items;
+   GLuint size, n_items;
+
+   enum brw_buffer_type buffer_type;
+
+   GLuint key_size[BRW_MAX_CACHE];             /* for fixed-size keys */
+   GLuint aux_size[BRW_MAX_CACHE];
+   char *name[BRW_MAX_CACHE];
+   
+
+   /* Record of the last BOs chosen for each cache_id.  Used to set
+    * brw->state.dirty.cache when a new cache item is chosen.
+    */
+   struct brw_winsys_buffer *last_bo[BRW_MAX_CACHE];
+};
+
+
+struct brw_tracked_state {
+   struct brw_state_flags dirty;
+   int (*prepare)( struct brw_context *brw );
+   int (*emit)( struct brw_context *brw );
+};
+
+/* Flags for brw->state.cache.
+ */
+#define CACHE_NEW_CC_VP                  (1<<BRW_CC_VP)
+#define CACHE_NEW_CC_UNIT                (1<<BRW_CC_UNIT)
+#define CACHE_NEW_WM_PROG                (1<<BRW_WM_PROG)
+#define CACHE_NEW_SAMPLER_DEFAULT_COLOR  (1<<BRW_SAMPLER_DEFAULT_COLOR)
+#define CACHE_NEW_SAMPLER                (1<<BRW_SAMPLER)
+#define CACHE_NEW_WM_UNIT                (1<<BRW_WM_UNIT)
+#define CACHE_NEW_SF_PROG                (1<<BRW_SF_PROG)
+#define CACHE_NEW_SF_VP                  (1<<BRW_SF_VP)
+#define CACHE_NEW_SF_UNIT                (1<<BRW_SF_UNIT)
+#define CACHE_NEW_VS_UNIT                (1<<BRW_VS_UNIT)
+#define CACHE_NEW_VS_PROG                (1<<BRW_VS_PROG)
+#define CACHE_NEW_GS_UNIT                (1<<BRW_GS_UNIT)
+#define CACHE_NEW_GS_PROG                (1<<BRW_GS_PROG)
+#define CACHE_NEW_CLIP_VP                (1<<BRW_CLIP_VP)
+#define CACHE_NEW_CLIP_UNIT              (1<<BRW_CLIP_UNIT)
+#define CACHE_NEW_CLIP_PROG              (1<<BRW_CLIP_PROG)
+#define CACHE_NEW_SURFACE                (1<<BRW_SS_SURFACE)
+#define CACHE_NEW_SURF_BIND              (1<<BRW_SS_SURF_BIND)
+
+struct brw_cached_batch_item {
+   struct header *header;
+   GLuint sz;
+   struct brw_cached_batch_item *next;
+};
+   
+
+
+/* Protect against a future where VERT_ATTRIB_MAX > 32.  Wouldn't life
+ * be easier if C allowed arrays of packed elements?
+ */
+#define VS_INPUT_BITMASK_DWORDS  ((PIPE_MAX_SHADER_INPUTS+31)/32)
+
+
+
+
+struct brw_vertex_info {
+   GLuint sizes[VS_INPUT_BITMASK_DWORDS * 2]; /* sizes:2[VERT_ATTRIB_MAX] */
+};
+
+
+struct brw_query_object {
+   /** Doubly linked list of active query objects in the context. */
+   struct brw_query_object *prev, *next;
+
+   /** Last query BO associated with this query. */
+   struct brw_winsys_buffer *bo;
+   /** First index in bo with query data for this object. */
+   int first_index;
+   /** Last index in bo with query data for this object. */
+   int last_index;
+
+   /* Total count of pixels from previous BOs */
+   uint64_t result;
+};
+
+#define CC_RELOC_VP 0
+
+
+/**
+ * brw_context is derived from pipe_context
+ */
+struct brw_context 
+{
+   struct pipe_context base;
+   struct brw_chipset chipset;
+
+   struct brw_winsys_screen *sws;
+
+   struct brw_batchbuffer *batch;
+
+   GLuint primitive;
+   GLuint reduced_primitive;
+
+   /* Active state from the state tracker: 
+    */
+   struct {
+      struct brw_vertex_shader *vertex_shader;
+      struct brw_fragment_shader *fragment_shader;
+      const struct brw_blend_state *blend;
+      const struct brw_rasterizer_state *rast;
+      const struct brw_depth_stencil_state *zstencil;
+
+      const struct brw_sampler *sampler[PIPE_MAX_SAMPLERS];
+      unsigned num_samplers;
+
+      struct pipe_texture *texture[PIPE_MAX_SAMPLERS];
+      struct pipe_vertex_buffer vertex_buffer[PIPE_MAX_ATTRIBS];
+      struct pipe_vertex_element vertex_element[PIPE_MAX_ATTRIBS];
+      unsigned num_vertex_elements;
+      unsigned num_textures;
+      unsigned num_vertex_buffers;
+
+      struct pipe_scissor_state scissor;
+      struct pipe_viewport_state viewport;
+      struct pipe_framebuffer_state fb;
+      struct pipe_clip_state ucp;
+      struct pipe_buffer *vertex_constants;
+      struct pipe_buffer *fragment_constants;
+
+      struct brw_blend_constant_color bcc;
+      struct brw_polygon_stipple bps;
+      struct brw_cc_viewport ccv;
+
+      /**
+       * Index buffer for this draw_prims call.
+       *
+       * Updates are signaled by PIPE_NEW_INDEX_BUFFER.
+       */
+      struct pipe_buffer *index_buffer;
+      unsigned index_size;
+
+      /* Updates are signalled by PIPE_NEW_INDEX_RANGE:
+       */
+      unsigned min_index;
+      unsigned max_index;
+
+   } curr;
+
+   struct {
+      struct brw_state_flags dirty;
+
+      /**
+       * List of buffers accumulated in brw_validate_state to receive
+       * dri_bo_check_aperture treatment before exec, so we can know if we
+       * should flush the batch and try again before emitting primitives.
+       *
+       * This can be a fixed number as we only have a limited number of
+       * objects referenced from the batchbuffer in a primitive emit,
+       * consisting of the vertex buffers, pipelined state pointers,
+       * the CURBE, the depth buffer, and a query BO.
+       */
+      struct brw_winsys_buffer *validated_bos[PIPE_MAX_SHADER_INPUTS + 16];
+      int validated_bo_count;
+   } state;
+
+   struct brw_cache cache;  /** non-surface items */
+   struct brw_cache surface_cache;  /* surface items */
+   struct brw_cached_batch_item *cached_batch_items;
+
+   struct {
+      struct u_upload_mgr *upload_vertex;
+      struct u_upload_mgr *upload_index;
+      
+      /* Information on uploaded vertex buffers:
+       */
+      struct {
+        unsigned stride;       /* in bytes between successive vertices */
+        unsigned offset;       /* in bytes, of first vertex in bo */
+        unsigned vertex_count; /* count of valid vertices which may be accessed */
+        struct brw_winsys_buffer *bo;
+      } vb[PIPE_MAX_ATTRIBS];
+
+      unsigned nr_vb;          /* currently the same as curr.num_vertex_buffers */
+   } vb;
+
+   struct {
+      /* Updates to these fields are signaled by BRW_NEW_INDEX_BUFFER. */
+      struct brw_winsys_buffer *bo;
+      unsigned int offset;
+      unsigned int size;
+      /* Offset to index buffer index to use in CMD_3D_PRIM so that we can
+       * avoid re-uploading the IB packet over and over if we're actually
+       * referencing the same index buffer.
+       */
+      unsigned int start_vertex_offset;
+   } ib;
+
+
+   /* BRW_NEW_URB_ALLOCATIONS:
+    */
+   struct {
+      GLuint vsize;            /* vertex size plus header in urb registers */
+      GLuint csize;            /* constant buffer size in urb registers */
+      GLuint sfsize;           /* setup data size in urb registers */
+
+      GLboolean constrained;
+
+      GLuint nr_vs_entries;
+      GLuint nr_gs_entries;
+      GLuint nr_clip_entries;
+      GLuint nr_sf_entries;
+      GLuint nr_cs_entries;
+
+      GLuint vs_start;
+      GLuint gs_start;
+      GLuint clip_start;
+      GLuint sf_start;
+      GLuint cs_start;
+   } urb;
+
+   
+   /* BRW_NEW_CURBE_OFFSETS: 
+    */
+   struct {
+      GLuint wm_start;  /**< pos of first wm const in CURBE buffer */
+      GLuint wm_size;   /**< number of float[4] consts, multiple of 16 */
+      GLuint clip_start;
+      GLuint clip_size;
+      GLuint vs_start;
+      GLuint vs_size;
+      GLuint total_size;
+
+      struct brw_winsys_buffer *curbe_bo;
+      /** Offset within curbe_bo of space for current curbe entry */
+      GLuint curbe_offset;
+      /** Offset within curbe_bo of space for next curbe entry */
+      GLuint curbe_next_offset;
+
+      GLfloat *last_buf;
+      GLuint last_bufsz;
+      /**
+       *  Whether we should create a new bo instead of reusing the old one
+       * (if we just dispatch the batch pointing at the old one.
+       */
+      GLboolean need_new_bo;
+   } curbe;
+
+   struct {
+      struct brw_vs_prog_data *prog_data;
+
+      struct brw_winsys_buffer *prog_bo;
+      struct brw_winsys_buffer *state_bo;
+
+      /** Binding table of pointers to surf_bo entries */
+      struct brw_winsys_buffer *bind_bo;
+      struct brw_winsys_buffer *surf_bo[BRW_VS_MAX_SURF];
+      GLuint nr_surfaces;      
+   } vs;
+
+   struct {
+      struct brw_gs_prog_data *prog_data;
+
+      GLboolean prog_active;
+      struct brw_winsys_buffer *prog_bo;
+      struct brw_winsys_buffer *state_bo;
+   } gs;
+
+   struct {
+      struct brw_clip_prog_data *prog_data;
+
+      struct brw_winsys_buffer *prog_bo;
+      struct brw_winsys_buffer *state_bo;
+      struct brw_winsys_buffer *vp_bo;
+   } clip;
+
+
+   struct {
+      struct brw_sf_prog_data *prog_data;
+
+      struct brw_winsys_buffer *prog_bo;
+      struct brw_winsys_buffer *state_bo;
+      struct brw_winsys_buffer *vp_bo;
+   } sf;
+
+   struct {
+      struct brw_wm_prog_data *prog_data;
+      struct brw_wm_compile *compile_data;
+
+      /** Input sizes, calculated from active vertex program.
+       * One bit per fragment program input attribute.
+       */
+      /*GLbitfield input_size_masks[4];*/
+
+      /** Array of surface default colors (texture border color) */
+      struct brw_winsys_buffer *sdc_bo[BRW_MAX_TEX_UNIT];
+
+      GLuint render_surf;
+      GLuint nr_surfaces;      
+
+      GLuint max_threads;
+      struct brw_winsys_buffer *scratch_bo;
+
+      GLuint sampler_count;
+      struct brw_winsys_buffer *sampler_bo;
+
+      /** Binding table of pointers to surf_bo entries */
+      struct brw_winsys_buffer *bind_bo;
+      struct brw_winsys_buffer *surf_bo[BRW_WM_MAX_SURF];
+
+      struct brw_winsys_buffer *prog_bo;
+      struct brw_winsys_buffer *state_bo;
+   } wm;
+
+
+   struct {
+      struct brw_winsys_buffer *state_bo;
+
+      struct brw_cc_unit_state cc;
+      struct brw_winsys_reloc reloc[1];
+   } cc;
+
+   struct {
+      struct brw_query_object active_head;
+      struct brw_winsys_buffer *bo;
+      int index;
+      GLboolean active;
+      int stats_wm;
+   } query;
+
+   struct {
+      unsigned always_emit_state:1;
+      unsigned always_flush_batch:1;
+      unsigned force_swtnl:1;
+      unsigned no_swtnl:1;
+   } flags;
+
+   /* Used to give every program string a unique id
+    */
+   GLuint program_id;
+};
+
+
+
+/*======================================================================
+ * brw_queryobj.c
+ */
+void brw_init_query(struct brw_context *brw);
+enum pipe_error brw_prepare_query_begin(struct brw_context *brw);
+void brw_emit_query_begin(struct brw_context *brw);
+void brw_emit_query_end(struct brw_context *brw);
+
+/*======================================================================
+ * brw_state_dump.c
+ */
+void brw_debug_batch(struct brw_context *intel);
+
+
+/*======================================================================
+ * brw_pipe_*.c
+ */
+void brw_pipe_blend_init( struct brw_context *brw );
+void brw_pipe_depth_stencil_init( struct brw_context *brw );
+void brw_pipe_framebuffer_init( struct brw_context *brw );
+void brw_pipe_flush_init( struct brw_context *brw );
+void brw_pipe_misc_init( struct brw_context *brw );
+void brw_pipe_query_init( struct brw_context *brw );
+void brw_pipe_rast_init( struct brw_context *brw );
+void brw_pipe_sampler_init( struct brw_context *brw );
+void brw_pipe_shader_init( struct brw_context *brw );
+void brw_pipe_vertex_init( struct brw_context *brw );
+void brw_pipe_clear_init( struct brw_context *brw );
+
+
+void brw_pipe_blend_cleanup( struct brw_context *brw );
+void brw_pipe_depth_stencil_cleanup( struct brw_context *brw );
+void brw_pipe_framebuffer_cleanup( struct brw_context *brw );
+void brw_pipe_flush_cleanup( struct brw_context *brw );
+void brw_pipe_misc_cleanup( struct brw_context *brw );
+void brw_pipe_query_cleanup( struct brw_context *brw );
+void brw_pipe_rast_cleanup( struct brw_context *brw );
+void brw_pipe_sampler_cleanup( struct brw_context *brw );
+void brw_pipe_shader_cleanup( struct brw_context *brw );
+void brw_pipe_vertex_cleanup( struct brw_context *brw );
+void brw_pipe_clear_cleanup( struct brw_context *brw );
+
+void brw_hw_cc_init( struct brw_context *brw );
+void brw_hw_cc_cleanup( struct brw_context *brw );
+
+
+
+void brw_context_flush( struct brw_context *brw );
+
+
+/* brw_urb.c
+ */
+int brw_upload_urb_fence(struct brw_context *brw);
+
+/* brw_curbe.c
+ */
+int brw_upload_cs_urb_state(struct brw_context *brw);
+
+
+/*======================================================================
+ * Inline conversion functions.  These are better-typed than the
+ * macros used previously:
+ */
+static INLINE struct brw_context *
+brw_context( struct pipe_context *ctx )
+{
+   return (struct brw_context *)ctx;
+}
+
+
+#define BRW_IS_965(brw)    ((brw)->chipset.is_965)
+#define BRW_IS_IGDNG(brw)  ((brw)->chipset.is_igdng)
+#define BRW_IS_G4X(brw)    ((brw)->chipset.is_g4x)
+
+
+#endif
+
diff --git a/src/gallium/drivers/i965/brw_curbe.c b/src/gallium/drivers/i965/brw_curbe.c
new file mode 100644 (file)
index 0000000..3f03157
--- /dev/null
@@ -0,0 +1,390 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "util/u_memory.h"
+#include "util/u_math.h"
+
+#include "brw_batchbuffer.h"
+#include "brw_context.h"
+#include "brw_defines.h"
+#include "brw_state.h"
+#include "brw_util.h"
+#include "brw_debug.h"
+#include "brw_screen.h"
+
+
+/**
+ * Partition the CURBE between the various users of constant values:
+ * Note that vertex and fragment shaders can now fetch constants out
+ * of constant buffers.  We no longer allocatea block of the GRF for
+ * constants.  That greatly reduces the demand for space in the CURBE.
+ * Some of the comments within are dated...
+ */
+static int calculate_curbe_offsets( struct brw_context *brw )
+{
+   /* CACHE_NEW_WM_PROG */
+   const GLuint nr_fp_regs = brw->wm.prog_data->curb_read_length;
+   
+   /* BRW_NEW_VERTEX_PROGRAM */
+   const GLuint nr_vp_regs = brw->vs.prog_data->curb_read_length;
+   GLuint nr_clip_regs = 0;
+   GLuint total_regs;
+
+   /* PIPE_NEW_CLIP */
+   if (brw->curr.ucp.nr) {
+      GLuint nr_planes = 6 + brw->curr.ucp.nr;
+      nr_clip_regs = (nr_planes * 4 + 15) / 16;
+   }
+
+
+   total_regs = nr_fp_regs + nr_vp_regs + nr_clip_regs;
+
+   /* When this is > 32, want to use a true constant buffer to hold
+    * the extra constants.
+    */
+   assert(total_regs <= 32);
+
+   /* Lazy resize:
+    */
+   if (nr_fp_regs > brw->curbe.wm_size ||
+       nr_vp_regs > brw->curbe.vs_size ||
+       nr_clip_regs != brw->curbe.clip_size ||
+       (total_regs < brw->curbe.total_size / 4 &&
+       brw->curbe.total_size > 16)) {
+
+      GLuint reg = 0;
+
+      /* Calculate a new layout: 
+       */
+      reg = 0;
+      brw->curbe.wm_start = reg;
+      brw->curbe.wm_size = nr_fp_regs; reg += nr_fp_regs;
+      brw->curbe.clip_start = reg;
+      brw->curbe.clip_size = nr_clip_regs; reg += nr_clip_regs;
+      brw->curbe.vs_start = reg;
+      brw->curbe.vs_size = nr_vp_regs; reg += nr_vp_regs;
+      brw->curbe.total_size = reg;
+
+      if (BRW_DEBUG & DEBUG_CURBE)
+        debug_printf("curbe wm %d+%d clip %d+%d vs %d+%d\n",
+                     brw->curbe.wm_start,
+                     brw->curbe.wm_size,
+                     brw->curbe.clip_start,
+                     brw->curbe.clip_size,
+                     brw->curbe.vs_start,
+                     brw->curbe.vs_size );
+
+      brw->state.dirty.brw |= BRW_NEW_CURBE_OFFSETS;
+   }
+
+   return 0;
+}
+
+
+const struct brw_tracked_state brw_curbe_offsets = {
+   .dirty = {
+      .mesa = PIPE_NEW_CLIP,
+      .brw  = BRW_NEW_VERTEX_PROGRAM,
+      .cache = CACHE_NEW_WM_PROG
+   },
+   .prepare = calculate_curbe_offsets
+};
+
+
+
+
+/* Define the number of curbes within CS's urb allocation.  Multiple
+ * urb entries -> multiple curbes.  These will be used by
+ * fixed-function hardware in a double-buffering scheme to avoid a
+ * pipeline stall each time the contents of the curbe is changed.
+ */
+int brw_upload_cs_urb_state(struct brw_context *brw)
+{
+   struct brw_cs_urb_state cs_urb;
+   memset(&cs_urb, 0, sizeof(cs_urb));
+
+   /* It appears that this is the state packet for the CS unit, ie. the
+    * urb entries detailed here are housed in the CS range from the
+    * URB_FENCE command.
+    */
+   cs_urb.header.opcode = CMD_CS_URB_STATE;
+   cs_urb.header.length = sizeof(cs_urb)/4 - 2;
+
+   /* BRW_NEW_URB_FENCE */
+   cs_urb.bits0.nr_urb_entries = brw->urb.nr_cs_entries;
+   cs_urb.bits0.urb_entry_size = brw->urb.csize - 1;
+
+   assert(brw->urb.nr_cs_entries);
+   BRW_CACHED_BATCH_STRUCT(brw, &cs_urb);
+   return 0;
+}
+
+static GLfloat fixed_plane[6][4] = {
+   { 0,    0,   -1, 1 },
+   { 0,    0,    1, 1 },
+   { 0,   -1,    0, 1 },
+   { 0,    1,    0, 1 },
+   {-1,    0,    0, 1 },
+   { 1,    0,    0, 1 }
+};
+
+/* Upload a new set of constants.  Too much variability to go into the
+ * cache mechanism, but maybe would benefit from a comparison against
+ * the current uploaded set of constants.
+ */
+static enum pipe_error prepare_curbe_buffer(struct brw_context *brw)
+{
+   struct pipe_screen *screen = brw->base.screen;
+   const GLuint sz = brw->curbe.total_size;
+   const GLuint bufsz = sz * 16 * sizeof(GLfloat);
+   enum pipe_error ret;
+   GLfloat *buf;
+   GLuint i;
+
+   if (sz == 0) {
+      if (brw->curbe.last_buf) {
+        free(brw->curbe.last_buf);
+        brw->curbe.last_buf = NULL;
+        brw->curbe.last_bufsz  = 0;
+      }
+      return 0;
+   }
+
+   buf = (GLfloat *) CALLOC(bufsz, 1);
+
+   /* fragment shader constants */
+   if (brw->curbe.wm_size) {
+      const struct brw_fragment_shader *fs = brw->curr.fragment_shader;
+      GLuint offset = brw->curbe.wm_start * 16;
+      GLuint nr_immediate, nr_const;
+
+      nr_immediate = fs->immediates.nr;
+      if (nr_immediate) {
+         memcpy(&buf[offset], 
+                fs->immediates.data,
+                nr_immediate * 4 * sizeof(float));
+
+         offset += nr_immediate * 4;
+      }
+
+      nr_const = fs->info.file_max[TGSI_FILE_CONSTANT] + 1;
+/*      nr_const = brw->wm.prog_data->nr_params; */
+      if (nr_const) {
+         const GLfloat *value = screen->buffer_map( screen,
+                                                    brw->curr.fragment_constants,
+                                                    PIPE_BUFFER_USAGE_CPU_READ);
+
+         memcpy(&buf[offset], value,
+                nr_const * 4 * sizeof(float));
+         
+         screen->buffer_unmap( screen, 
+                               brw->curr.fragment_constants );
+      }
+   }
+
+
+   /* The clipplanes are actually delivered to both CLIP and VS units.
+    * VS uses them to calculate the outcode bitmasks.
+    */
+   if (brw->curbe.clip_size) {
+      GLuint offset = brw->curbe.clip_start * 16;
+      GLuint j;
+
+      /* If any planes are going this way, send them all this way:
+       */
+      for (i = 0; i < 6; i++) {
+        buf[offset + i * 4 + 0] = fixed_plane[i][0];
+        buf[offset + i * 4 + 1] = fixed_plane[i][1];
+        buf[offset + i * 4 + 2] = fixed_plane[i][2];
+        buf[offset + i * 4 + 3] = fixed_plane[i][3];
+      }
+
+      /* Clip planes:
+       */
+      assert(brw->curr.ucp.nr <= 6);
+      for (j = 0; j < brw->curr.ucp.nr; j++) {
+        buf[offset + i * 4 + 0] = brw->curr.ucp.ucp[j][0];
+        buf[offset + i * 4 + 1] = brw->curr.ucp.ucp[j][1];
+        buf[offset + i * 4 + 2] = brw->curr.ucp.ucp[j][2];
+        buf[offset + i * 4 + 3] = brw->curr.ucp.ucp[j][3];
+        i++;
+      }
+   }
+
+   /* vertex shader constants */
+   if (brw->curbe.vs_size) {
+      GLuint offset = brw->curbe.vs_start * 16;
+      const struct brw_vertex_shader *vs = brw->curr.vertex_shader;
+      GLuint nr_immediate, nr_const;
+
+      nr_immediate = vs->immediates.nr;
+      if (nr_immediate) {
+         memcpy(&buf[offset], 
+                vs->immediates.data,
+                nr_immediate * 4 * sizeof(float));
+
+         offset += nr_immediate * 4;
+      }
+
+      nr_const = vs->info.file_max[TGSI_FILE_CONSTANT] + 1;
+      if (nr_const) {
+         /* XXX: note that constant buffers are currently *already* in
+          * buffer objects.  If we want to keep on putting them into the
+          * curbe, makes sense to treat constbuf's specially with malloc.
+          */
+         const GLfloat *value = screen->buffer_map( screen,
+                                                    brw->curr.vertex_constants,
+                                                    PIPE_BUFFER_USAGE_CPU_READ);
+         
+         /* XXX: what if user's constant buffer is too small?
+          */
+         memcpy(&buf[offset], value, nr_const * 4 * sizeof(float));
+         
+         screen->buffer_unmap( screen, brw->curr.vertex_constants );
+      }
+   }
+
+   if (BRW_DEBUG & DEBUG_CURBE) {
+      for (i = 0; i < sz*16; i+=4) 
+        debug_printf("curbe %d.%d: %f %f %f %f\n", i/8, i&4,
+                     buf[i+0], buf[i+1], buf[i+2], buf[i+3]);
+
+      debug_printf("last_buf %p buf %p sz %d/%d cmp %d\n",
+                  (void *)brw->curbe.last_buf, (void *)buf,
+                  bufsz, brw->curbe.last_bufsz,
+                  brw->curbe.last_buf ? memcmp(buf, brw->curbe.last_buf, bufsz) : -1);
+   }
+
+   if (brw->curbe.curbe_bo != NULL &&
+       brw->curbe.last_buf &&
+       bufsz == brw->curbe.last_bufsz &&
+       memcmp(buf, brw->curbe.last_buf, bufsz) == 0) {
+      /* constants have not changed */
+      FREE(buf);
+   } 
+   else {
+      /* constants have changed */
+      FREE(brw->curbe.last_buf);
+
+      brw->curbe.last_buf = buf;
+      brw->curbe.last_bufsz = bufsz;
+
+      if (brw->curbe.curbe_bo != NULL &&
+         (brw->curbe.need_new_bo ||
+          brw->curbe.curbe_next_offset + bufsz > brw->curbe.curbe_bo->size))
+      {
+        bo_reference(&brw->curbe.curbe_bo, NULL);
+      }
+
+      if (brw->curbe.curbe_bo == NULL) {
+        /* Allocate a single page for CURBE entries for this
+         * batchbuffer.  They're generally around 64b.  We will
+         * discard the curbe buffer after the batch is flushed to
+         * avoid synchronous updates.
+         */
+        ret = brw->sws->bo_alloc(brw->sws, 
+                                  BRW_BUFFER_TYPE_CURBE,
+                                  4096, 1 << 6,
+                                  &brw->curbe.curbe_bo);
+         if (ret)
+            return ret;
+
+        brw->curbe.curbe_next_offset = 0;
+      }
+
+      brw->curbe.curbe_offset = brw->curbe.curbe_next_offset;
+      brw->curbe.curbe_next_offset += bufsz;
+      brw->curbe.curbe_next_offset = align(brw->curbe.curbe_next_offset, 64);
+
+      /* Copy data to the buffer:
+       */
+      brw->sws->bo_subdata(brw->curbe.curbe_bo,
+                           BRW_DATA_CONSTANT_BUFFER,
+                          brw->curbe.curbe_offset,
+                          bufsz,
+                          buf,
+                           NULL, 0);
+   }
+
+   brw_add_validated_bo(brw, brw->curbe.curbe_bo);
+
+   /* Because this provokes an action (ie copy the constants into the
+    * URB), it shouldn't be shortcircuited if identical to the
+    * previous time - because eg. the urb destination may have
+    * changed, or the urb contents different to last time.
+    *
+    * Note that the data referred to is actually copied internally,
+    * not just used in place according to passed pointer.
+    *
+    * It appears that the CS unit takes care of using each available
+    * URB entry (Const URB Entry == CURBE) in turn, and issuing
+    * flushes as necessary when doublebuffering of CURBEs isn't
+    * possible.
+    */
+
+   return 0;
+}
+
+static enum pipe_error emit_curbe_buffer(struct brw_context *brw)
+{
+   GLuint sz = brw->curbe.total_size;
+
+   BEGIN_BATCH(2, IGNORE_CLIPRECTS);
+   if (sz == 0) {
+      OUT_BATCH((CMD_CONST_BUFFER << 16) | (2 - 2));
+      OUT_BATCH(0);
+   } else {
+      OUT_BATCH((CMD_CONST_BUFFER << 16) | (1 << 8) | (2 - 2));
+      OUT_RELOC(brw->curbe.curbe_bo,
+               BRW_USAGE_STATE,
+               (sz - 1) + brw->curbe.curbe_offset);
+   }
+   ADVANCE_BATCH();
+   return 0;
+}
+
+const struct brw_tracked_state brw_curbe_buffer = {
+   .dirty = {
+      .mesa = (PIPE_NEW_FRAGMENT_CONSTANTS |
+              PIPE_NEW_VERTEX_CONSTANTS |
+              PIPE_NEW_CLIP),
+      .brw  = (BRW_NEW_FRAGMENT_PROGRAM |
+              BRW_NEW_VERTEX_PROGRAM |
+              BRW_NEW_URB_FENCE | /* Implicit - hardware requires this, not used above */
+              BRW_NEW_PSP | /* Implicit - hardware requires this, not used above */
+              BRW_NEW_CURBE_OFFSETS |
+              BRW_NEW_BATCH),
+      .cache = (CACHE_NEW_WM_PROG) 
+   },
+   .prepare = prepare_curbe_buffer,
+   .emit = emit_curbe_buffer,
+};
+
diff --git a/src/gallium/drivers/i965/brw_debug.h b/src/gallium/drivers/i965/brw_debug.h
new file mode 100644 (file)
index 0000000..ae8e925
--- /dev/null
@@ -0,0 +1,43 @@
+#ifndef BRW_DEBUG_H
+#define BRW_DEBUG_H
+
+/* ================================================================
+ * Debugging:
+ */
+
+#define DEBUG_TEXTURE          0x1
+#define DEBUG_STATE            0x2
+#define DEBUG_IOCTL            0x4
+#define DEBUG_BLIT             0x8
+#define DEBUG_CURBE             0x10
+#define DEBUG_FALLBACKS                0x20
+#define DEBUG_VERBOSE          0x40
+#define DEBUG_BATCH             0x80
+#define DEBUG_PIXEL             0x100
+#define DEBUG_WINSYS            0x200
+#define DEBUG_MIN_URB           0x400
+#define DEBUG_DISASSEM           0x800
+#define DEBUG_unused3           0x1000
+#define DEBUG_SYNC             0x2000
+#define DEBUG_PRIMS            0x4000
+#define DEBUG_VERTS            0x8000
+#define DEBUG_unused4           0x10000
+#define DEBUG_DMA               0x20000
+#define DEBUG_SANITY            0x40000
+#define DEBUG_SLEEP             0x80000
+#define DEBUG_STATS             0x100000
+#define DEBUG_unused5           0x200000
+#define DEBUG_SINGLE_THREAD     0x400000
+#define DEBUG_WM                0x800000
+#define DEBUG_URB               0x1000000
+#define DEBUG_VS                0x2000000
+
+#ifdef DEBUG
+extern int BRW_DEBUG;
+#else
+#define BRW_DEBUG 0
+#endif
+
+
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_defines.h b/src/gallium/drivers/i965/brw_defines.h
new file mode 100644 (file)
index 0000000..e201ce4
--- /dev/null
@@ -0,0 +1,847 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#ifndef BRW_DEFINES_H
+#define BRW_DEFINES_H
+
+/* 3D state:
+ */
+#define _3DOP_3DSTATE_PIPELINED       0x0
+#define _3DOP_3DSTATE_NONPIPELINED    0x1
+#define _3DOP_3DCONTROL               0x2
+#define _3DOP_3DPRIMITIVE             0x3
+
+#define _3DSTATE_PIPELINED_POINTERS       0x00
+#define _3DSTATE_BINDING_TABLE_POINTERS   0x01
+#define _3DSTATE_VERTEX_BUFFERS           0x08
+#define _3DSTATE_VERTEX_ELEMENTS          0x09
+#define _3DSTATE_INDEX_BUFFER             0x0A
+#define _3DSTATE_VF_STATISTICS            0x0B
+#define _3DSTATE_DRAWING_RECTANGLE            0x00
+#define _3DSTATE_CONSTANT_COLOR               0x01
+#define _3DSTATE_SAMPLER_PALETTE_LOAD         0x02
+#define _3DSTATE_CHROMA_KEY                   0x04
+#define _3DSTATE_DEPTH_BUFFER                 0x05
+#define _3DSTATE_POLY_STIPPLE_OFFSET          0x06
+#define _3DSTATE_POLY_STIPPLE_PATTERN         0x07
+#define _3DSTATE_LINE_STIPPLE                 0x08
+#define _3DSTATE_GLOBAL_DEPTH_OFFSET_CLAMP    0x09
+#define _3DCONTROL    0x00
+
+#define PIPE_CONTROL_NOWRITE          0x00
+#define PIPE_CONTROL_WRITEIMMEDIATE   0x01
+#define PIPE_CONTROL_WRITEDEPTH       0x02
+#define PIPE_CONTROL_WRITETIMESTAMP   0x03
+
+#define PIPE_CONTROL_GTTWRITE_PROCESS_LOCAL 0x00
+#define PIPE_CONTROL_GTTWRITE_GLOBAL        0x01
+
+#define _3DPRIM_POINTLIST         0x01
+#define _3DPRIM_LINELIST          0x02
+#define _3DPRIM_LINESTRIP         0x03
+#define _3DPRIM_TRILIST           0x04
+#define _3DPRIM_TRISTRIP          0x05
+#define _3DPRIM_TRIFAN            0x06
+#define _3DPRIM_QUADLIST          0x07
+#define _3DPRIM_QUADSTRIP         0x08
+#define _3DPRIM_LINELIST_ADJ      0x09
+#define _3DPRIM_LINESTRIP_ADJ     0x0A
+#define _3DPRIM_TRILIST_ADJ       0x0B
+#define _3DPRIM_TRISTRIP_ADJ      0x0C
+#define _3DPRIM_TRISTRIP_REVERSE  0x0D
+#define _3DPRIM_POLYGON           0x0E
+#define _3DPRIM_RECTLIST          0x0F
+#define _3DPRIM_LINELOOP          0x10
+#define _3DPRIM_POINTLIST_BF      0x11
+#define _3DPRIM_LINESTRIP_CONT    0x12
+#define _3DPRIM_LINESTRIP_BF      0x13
+#define _3DPRIM_LINESTRIP_CONT_BF 0x14
+#define _3DPRIM_TRIFAN_NOSTIPPLE  0x15
+
+#define _3DPRIM_VERTEXBUFFER_ACCESS_SEQUENTIAL 0
+#define _3DPRIM_VERTEXBUFFER_ACCESS_RANDOM     1
+
+#define BRW_ANISORATIO_2     0 
+#define BRW_ANISORATIO_4     1 
+#define BRW_ANISORATIO_6     2 
+#define BRW_ANISORATIO_8     3 
+#define BRW_ANISORATIO_10    4 
+#define BRW_ANISORATIO_12    5 
+#define BRW_ANISORATIO_14    6 
+#define BRW_ANISORATIO_16    7
+
+#define BRW_BLENDFACTOR_ONE                 0x1
+#define BRW_BLENDFACTOR_SRC_COLOR           0x2
+#define BRW_BLENDFACTOR_SRC_ALPHA           0x3
+#define BRW_BLENDFACTOR_DST_ALPHA           0x4
+#define BRW_BLENDFACTOR_DST_COLOR           0x5
+#define BRW_BLENDFACTOR_SRC_ALPHA_SATURATE  0x6
+#define BRW_BLENDFACTOR_CONST_COLOR         0x7
+#define BRW_BLENDFACTOR_CONST_ALPHA         0x8
+#define BRW_BLENDFACTOR_SRC1_COLOR          0x9
+#define BRW_BLENDFACTOR_SRC1_ALPHA          0x0A
+#define BRW_BLENDFACTOR_ZERO                0x11
+#define BRW_BLENDFACTOR_INV_SRC_COLOR       0x12
+#define BRW_BLENDFACTOR_INV_SRC_ALPHA       0x13
+#define BRW_BLENDFACTOR_INV_DST_ALPHA       0x14
+#define BRW_BLENDFACTOR_INV_DST_COLOR       0x15
+#define BRW_BLENDFACTOR_INV_CONST_COLOR     0x17
+#define BRW_BLENDFACTOR_INV_CONST_ALPHA     0x18
+#define BRW_BLENDFACTOR_INV_SRC1_COLOR      0x19
+#define BRW_BLENDFACTOR_INV_SRC1_ALPHA      0x1A
+
+#define BRW_BLENDFUNCTION_ADD               0
+#define BRW_BLENDFUNCTION_SUBTRACT          1
+#define BRW_BLENDFUNCTION_REVERSE_SUBTRACT  2
+#define BRW_BLENDFUNCTION_MIN               3
+#define BRW_BLENDFUNCTION_MAX               4
+
+#define BRW_ALPHATEST_FORMAT_UNORM8         0
+#define BRW_ALPHATEST_FORMAT_FLOAT32        1
+
+#define BRW_CHROMAKEY_KILL_ON_ANY_MATCH  0
+#define BRW_CHROMAKEY_REPLACE_BLACK      1
+
+#define BRW_CLIP_API_OGL     0
+#define BRW_CLIP_API_DX      1
+
+#define BRW_CLIPMODE_NORMAL              0
+#define BRW_CLIPMODE_CLIP_ALL            1
+#define BRW_CLIPMODE_CLIP_NON_REJECTED   2
+#define BRW_CLIPMODE_REJECT_ALL          3
+#define BRW_CLIPMODE_ACCEPT_ALL          4
+#define BRW_CLIPMODE_KERNEL_CLIP         5
+
+#define BRW_CLIP_NDCSPACE     0
+#define BRW_CLIP_SCREENSPACE  1
+
+#define BRW_COMPAREFUNCTION_ALWAYS       0
+#define BRW_COMPAREFUNCTION_NEVER        1
+#define BRW_COMPAREFUNCTION_LESS         2
+#define BRW_COMPAREFUNCTION_EQUAL        3
+#define BRW_COMPAREFUNCTION_LEQUAL       4
+#define BRW_COMPAREFUNCTION_GREATER      5
+#define BRW_COMPAREFUNCTION_NOTEQUAL     6
+#define BRW_COMPAREFUNCTION_GEQUAL       7
+
+#define BRW_COVERAGE_PIXELS_HALF     0
+#define BRW_COVERAGE_PIXELS_1        1
+#define BRW_COVERAGE_PIXELS_2        2
+#define BRW_COVERAGE_PIXELS_4        3
+
+#define BRW_CULLMODE_BOTH        0
+#define BRW_CULLMODE_NONE        1
+#define BRW_CULLMODE_FRONT       2
+#define BRW_CULLMODE_BACK        3
+
+#define BRW_DEFAULTCOLOR_R8G8B8A8_UNORM      0
+#define BRW_DEFAULTCOLOR_R32G32B32A32_FLOAT  1
+
+#define BRW_DEPTHFORMAT_D32_FLOAT_S8X24_UINT     0
+#define BRW_DEPTHFORMAT_D32_FLOAT                1
+#define BRW_DEPTHFORMAT_D24_UNORM_S8_UINT        2
+#define BRW_DEPTHFORMAT_D16_UNORM                5
+
+#define BRW_FLOATING_POINT_IEEE_754        0
+#define BRW_FLOATING_POINT_NON_IEEE_754    1
+
+#define BRW_FRONTWINDING_CW      0
+#define BRW_FRONTWINDING_CCW     1
+
+#define BRW_SPRITE_POINT_ENABLE  16
+
+#define BRW_INDEX_BYTE     0
+#define BRW_INDEX_WORD     1
+#define BRW_INDEX_DWORD    2
+
+#define BRW_LOGICOPFUNCTION_CLEAR            0
+#define BRW_LOGICOPFUNCTION_NOR              1
+#define BRW_LOGICOPFUNCTION_AND_INVERTED     2
+#define BRW_LOGICOPFUNCTION_COPY_INVERTED    3
+#define BRW_LOGICOPFUNCTION_AND_REVERSE      4
+#define BRW_LOGICOPFUNCTION_INVERT           5
+#define BRW_LOGICOPFUNCTION_XOR              6
+#define BRW_LOGICOPFUNCTION_NAND             7
+#define BRW_LOGICOPFUNCTION_AND              8
+#define BRW_LOGICOPFUNCTION_EQUIV            9
+#define BRW_LOGICOPFUNCTION_NOOP             10
+#define BRW_LOGICOPFUNCTION_OR_INVERTED      11
+#define BRW_LOGICOPFUNCTION_COPY             12
+#define BRW_LOGICOPFUNCTION_OR_REVERSE       13
+#define BRW_LOGICOPFUNCTION_OR               14
+#define BRW_LOGICOPFUNCTION_SET              15  
+
+#define BRW_MAPFILTER_NEAREST        0x0 
+#define BRW_MAPFILTER_LINEAR         0x1 
+#define BRW_MAPFILTER_ANISOTROPIC    0x2
+
+#define BRW_MIPFILTER_NONE        0   
+#define BRW_MIPFILTER_NEAREST     1   
+#define BRW_MIPFILTER_LINEAR      3
+
+#define BRW_POLYGON_FRONT_FACING     0
+#define BRW_POLYGON_BACK_FACING      1
+
+#define BRW_PREFILTER_ALWAYS     0x0 
+#define BRW_PREFILTER_NEVER      0x1
+#define BRW_PREFILTER_LESS       0x2
+#define BRW_PREFILTER_EQUAL      0x3
+#define BRW_PREFILTER_LEQUAL     0x4
+#define BRW_PREFILTER_GREATER    0x5
+#define BRW_PREFILTER_NOTEQUAL   0x6
+#define BRW_PREFILTER_GEQUAL     0x7
+
+#define BRW_PROVOKING_VERTEX_0    0
+#define BRW_PROVOKING_VERTEX_1    1 
+#define BRW_PROVOKING_VERTEX_2    2
+
+#define BRW_RASTRULE_UPPER_LEFT  0    
+#define BRW_RASTRULE_UPPER_RIGHT 1
+/* These are listed as "Reserved, but not seen as useful"
+ * in Intel documentation (page 212, "Point Rasterization Rule",
+ * section 7.4 "SF Pipeline State Summary", of document
+ * "Intel® 965 Express Chipset Family and Intel® G35 Express
+ * Chipset Graphics Controller Programmer's Reference Manual,
+ * Volume 2: 3D/Media", Revision 1.0b as of January 2008,
+ * available at 
+ *     http://intellinuxgraphics.org/documentation.html
+ * at the time of this writing).
+ *
+ * These appear to be supported on at least some
+ * i965-family devices, and the BRW_RASTRULE_LOWER_RIGHT
+ * is useful when using OpenGL to render to a FBO
+ * (which has the pixel coordinate Y orientation inverted
+ * with respect to the normal OpenGL pixel coordinate system).
+ */
+#define BRW_RASTRULE_LOWER_LEFT  2
+#define BRW_RASTRULE_LOWER_RIGHT 3
+
+#define BRW_RENDERTARGET_CLAMPRANGE_UNORM    0
+#define BRW_RENDERTARGET_CLAMPRANGE_SNORM    1
+#define BRW_RENDERTARGET_CLAMPRANGE_FORMAT   2
+
+#define BRW_STENCILOP_KEEP               0
+#define BRW_STENCILOP_ZERO               1
+#define BRW_STENCILOP_REPLACE            2
+#define BRW_STENCILOP_INCRSAT            3
+#define BRW_STENCILOP_DECRSAT            4
+#define BRW_STENCILOP_INCR               5
+#define BRW_STENCILOP_DECR               6
+#define BRW_STENCILOP_INVERT             7
+
+#define BRW_SURFACE_MIPMAPLAYOUT_BELOW   0
+#define BRW_SURFACE_MIPMAPLAYOUT_RIGHT   1
+
+#define BRW_SURFACEFORMAT_R32G32B32A32_FLOAT             0x000 
+#define BRW_SURFACEFORMAT_R32G32B32A32_SINT              0x001 
+#define BRW_SURFACEFORMAT_R32G32B32A32_UINT              0x002 
+#define BRW_SURFACEFORMAT_R32G32B32A32_UNORM             0x003 
+#define BRW_SURFACEFORMAT_R32G32B32A32_SNORM             0x004 
+#define BRW_SURFACEFORMAT_R64G64_FLOAT                   0x005 
+#define BRW_SURFACEFORMAT_R32G32B32X32_FLOAT             0x006 
+#define BRW_SURFACEFORMAT_R32G32B32A32_SSCALED           0x007
+#define BRW_SURFACEFORMAT_R32G32B32A32_USCALED           0x008
+#define BRW_SURFACEFORMAT_R32G32B32_FLOAT                0x040 
+#define BRW_SURFACEFORMAT_R32G32B32_SINT                 0x041 
+#define BRW_SURFACEFORMAT_R32G32B32_UINT                 0x042 
+#define BRW_SURFACEFORMAT_R32G32B32_UNORM                0x043 
+#define BRW_SURFACEFORMAT_R32G32B32_SNORM                0x044 
+#define BRW_SURFACEFORMAT_R32G32B32_SSCALED              0x045 
+#define BRW_SURFACEFORMAT_R32G32B32_USCALED              0x046 
+#define BRW_SURFACEFORMAT_R16G16B16A16_UNORM             0x080 
+#define BRW_SURFACEFORMAT_R16G16B16A16_SNORM             0x081 
+#define BRW_SURFACEFORMAT_R16G16B16A16_SINT              0x082 
+#define BRW_SURFACEFORMAT_R16G16B16A16_UINT              0x083 
+#define BRW_SURFACEFORMAT_R16G16B16A16_FLOAT             0x084 
+#define BRW_SURFACEFORMAT_R32G32_FLOAT                   0x085 
+#define BRW_SURFACEFORMAT_R32G32_SINT                    0x086 
+#define BRW_SURFACEFORMAT_R32G32_UINT                    0x087 
+#define BRW_SURFACEFORMAT_R32_FLOAT_X8X24_TYPELESS       0x088 
+#define BRW_SURFACEFORMAT_X32_TYPELESS_G8X24_UINT        0x089 
+#define BRW_SURFACEFORMAT_L32A32_FLOAT                   0x08A 
+#define BRW_SURFACEFORMAT_R32G32_UNORM                   0x08B 
+#define BRW_SURFACEFORMAT_R32G32_SNORM                   0x08C 
+#define BRW_SURFACEFORMAT_R64_FLOAT                      0x08D 
+#define BRW_SURFACEFORMAT_R16G16B16X16_UNORM             0x08E 
+#define BRW_SURFACEFORMAT_R16G16B16X16_FLOAT             0x08F 
+#define BRW_SURFACEFORMAT_A32X32_FLOAT                   0x090 
+#define BRW_SURFACEFORMAT_L32X32_FLOAT                   0x091 
+#define BRW_SURFACEFORMAT_I32X32_FLOAT                   0x092 
+#define BRW_SURFACEFORMAT_R16G16B16A16_SSCALED           0x093
+#define BRW_SURFACEFORMAT_R16G16B16A16_USCALED           0x094
+#define BRW_SURFACEFORMAT_R32G32_SSCALED                 0x095
+#define BRW_SURFACEFORMAT_R32G32_USCALED                 0x096
+#define BRW_SURFACEFORMAT_B8G8R8A8_UNORM                 0x0C0 
+#define BRW_SURFACEFORMAT_B8G8R8A8_UNORM_SRGB            0x0C1 
+#define BRW_SURFACEFORMAT_R10G10B10A2_UNORM              0x0C2 
+#define BRW_SURFACEFORMAT_R10G10B10A2_UNORM_SRGB         0x0C3 
+#define BRW_SURFACEFORMAT_R10G10B10A2_UINT               0x0C4 
+#define BRW_SURFACEFORMAT_R10G10B10_SNORM_A2_UNORM       0x0C5 
+#define BRW_SURFACEFORMAT_R8G8B8A8_UNORM                 0x0C7 
+#define BRW_SURFACEFORMAT_R8G8B8A8_UNORM_SRGB            0x0C8 
+#define BRW_SURFACEFORMAT_R8G8B8A8_SNORM                 0x0C9 
+#define BRW_SURFACEFORMAT_R8G8B8A8_SINT                  0x0CA 
+#define BRW_SURFACEFORMAT_R8G8B8A8_UINT                  0x0CB 
+#define BRW_SURFACEFORMAT_R16G16_UNORM                   0x0CC 
+#define BRW_SURFACEFORMAT_R16G16_SNORM                   0x0CD 
+#define BRW_SURFACEFORMAT_R16G16_SINT                    0x0CE 
+#define BRW_SURFACEFORMAT_R16G16_UINT                    0x0CF 
+#define BRW_SURFACEFORMAT_R16G16_FLOAT                   0x0D0 
+#define BRW_SURFACEFORMAT_B10G10R10A2_UNORM              0x0D1 
+#define BRW_SURFACEFORMAT_B10G10R10A2_UNORM_SRGB         0x0D2 
+#define BRW_SURFACEFORMAT_R11G11B10_FLOAT                0x0D3 
+#define BRW_SURFACEFORMAT_R32_SINT                       0x0D6 
+#define BRW_SURFACEFORMAT_R32_UINT                       0x0D7 
+#define BRW_SURFACEFORMAT_R32_FLOAT                      0x0D8 
+#define BRW_SURFACEFORMAT_R24_UNORM_X8_TYPELESS          0x0D9 
+#define BRW_SURFACEFORMAT_X24_TYPELESS_G8_UINT           0x0DA 
+#define BRW_SURFACEFORMAT_L16A16_UNORM                   0x0DF 
+#define BRW_SURFACEFORMAT_I24X8_UNORM                    0x0E0 
+#define BRW_SURFACEFORMAT_L24X8_UNORM                    0x0E1 
+#define BRW_SURFACEFORMAT_A24X8_UNORM                    0x0E2 
+#define BRW_SURFACEFORMAT_I32_FLOAT                      0x0E3 
+#define BRW_SURFACEFORMAT_L32_FLOAT                      0x0E4 
+#define BRW_SURFACEFORMAT_A32_FLOAT                      0x0E5 
+#define BRW_SURFACEFORMAT_B8G8R8X8_UNORM                 0x0E9 
+#define BRW_SURFACEFORMAT_B8G8R8X8_UNORM_SRGB            0x0EA 
+#define BRW_SURFACEFORMAT_R8G8B8X8_UNORM                 0x0EB 
+#define BRW_SURFACEFORMAT_R8G8B8X8_UNORM_SRGB            0x0EC 
+#define BRW_SURFACEFORMAT_R9G9B9E5_SHAREDEXP             0x0ED 
+#define BRW_SURFACEFORMAT_B10G10R10X2_UNORM              0x0EE 
+#define BRW_SURFACEFORMAT_L16A16_FLOAT                   0x0F0 
+#define BRW_SURFACEFORMAT_R32_UNORM                      0x0F1 
+#define BRW_SURFACEFORMAT_R32_SNORM                      0x0F2 
+#define BRW_SURFACEFORMAT_R10G10B10X2_USCALED            0x0F3
+#define BRW_SURFACEFORMAT_R8G8B8A8_SSCALED               0x0F4
+#define BRW_SURFACEFORMAT_R8G8B8A8_USCALED               0x0F5
+#define BRW_SURFACEFORMAT_R16G16_SSCALED                 0x0F6
+#define BRW_SURFACEFORMAT_R16G16_USCALED                 0x0F7
+#define BRW_SURFACEFORMAT_R32_SSCALED                    0x0F8
+#define BRW_SURFACEFORMAT_R32_USCALED                    0x0F9
+#define BRW_SURFACEFORMAT_B5G6R5_UNORM                   0x100 
+#define BRW_SURFACEFORMAT_B5G6R5_UNORM_SRGB              0x101 
+#define BRW_SURFACEFORMAT_B5G5R5A1_UNORM                 0x102 
+#define BRW_SURFACEFORMAT_B5G5R5A1_UNORM_SRGB            0x103 
+#define BRW_SURFACEFORMAT_B4G4R4A4_UNORM                 0x104 
+#define BRW_SURFACEFORMAT_B4G4R4A4_UNORM_SRGB            0x105 
+#define BRW_SURFACEFORMAT_R8G8_UNORM                     0x106 
+#define BRW_SURFACEFORMAT_R8G8_SNORM                     0x107 
+#define BRW_SURFACEFORMAT_R8G8_SINT                      0x108 
+#define BRW_SURFACEFORMAT_R8G8_UINT                      0x109 
+#define BRW_SURFACEFORMAT_R16_UNORM                      0x10A 
+#define BRW_SURFACEFORMAT_R16_SNORM                      0x10B 
+#define BRW_SURFACEFORMAT_R16_SINT                       0x10C 
+#define BRW_SURFACEFORMAT_R16_UINT                       0x10D 
+#define BRW_SURFACEFORMAT_R16_FLOAT                      0x10E 
+#define BRW_SURFACEFORMAT_I16_UNORM                      0x111 
+#define BRW_SURFACEFORMAT_L16_UNORM                      0x112 
+#define BRW_SURFACEFORMAT_A16_UNORM                      0x113 
+#define BRW_SURFACEFORMAT_L8A8_UNORM                     0x114 
+#define BRW_SURFACEFORMAT_I16_FLOAT                      0x115
+#define BRW_SURFACEFORMAT_L16_FLOAT                      0x116
+#define BRW_SURFACEFORMAT_A16_FLOAT                      0x117
+#define BRW_SURFACEFORMAT_L8A8_UNORM_SRGB                0x118
+#define BRW_SURFACEFORMAT_R5G5_SNORM_B6_UNORM            0x119
+#define BRW_SURFACEFORMAT_B5G5R5X1_UNORM                 0x11A
+#define BRW_SURFACEFORMAT_B5G5R5X1_UNORM_SRGB            0x11B
+#define BRW_SURFACEFORMAT_R8G8_SSCALED                   0x11C
+#define BRW_SURFACEFORMAT_R8G8_USCALED                   0x11D
+#define BRW_SURFACEFORMAT_R16_SSCALED                    0x11E
+#define BRW_SURFACEFORMAT_R16_USCALED                    0x11F
+#define BRW_SURFACEFORMAT_R8_UNORM                       0x140 
+#define BRW_SURFACEFORMAT_R8_SNORM                       0x141 
+#define BRW_SURFACEFORMAT_R8_SINT                        0x142 
+#define BRW_SURFACEFORMAT_R8_UINT                        0x143 
+#define BRW_SURFACEFORMAT_A8_UNORM                       0x144 
+#define BRW_SURFACEFORMAT_I8_UNORM                       0x145 
+#define BRW_SURFACEFORMAT_L8_UNORM                       0x146 
+#define BRW_SURFACEFORMAT_P4A4_UNORM                     0x147 
+#define BRW_SURFACEFORMAT_A4P4_UNORM                     0x148
+#define BRW_SURFACEFORMAT_R8_SSCALED                     0x149
+#define BRW_SURFACEFORMAT_R8_USCALED                     0x14A
+#define BRW_SURFACEFORMAT_L8_UNORM_SRGB                  0x14C
+#define BRW_SURFACEFORMAT_R1_UINT                        0x181 
+#define BRW_SURFACEFORMAT_YCRCB_NORMAL                   0x182 
+#define BRW_SURFACEFORMAT_YCRCB_SWAPUVY                  0x183 
+#define BRW_SURFACEFORMAT_BC1_UNORM                      0x186 
+#define BRW_SURFACEFORMAT_BC2_UNORM                      0x187 
+#define BRW_SURFACEFORMAT_BC3_UNORM                      0x188 
+#define BRW_SURFACEFORMAT_BC4_UNORM                      0x189 
+#define BRW_SURFACEFORMAT_BC5_UNORM                      0x18A 
+#define BRW_SURFACEFORMAT_BC1_UNORM_SRGB                 0x18B 
+#define BRW_SURFACEFORMAT_BC2_UNORM_SRGB                 0x18C 
+#define BRW_SURFACEFORMAT_BC3_UNORM_SRGB                 0x18D 
+#define BRW_SURFACEFORMAT_MONO8                          0x18E 
+#define BRW_SURFACEFORMAT_YCRCB_SWAPUV                   0x18F 
+#define BRW_SURFACEFORMAT_YCRCB_SWAPY                    0x190 
+#define BRW_SURFACEFORMAT_DXT1_RGB                       0x191 
+#define BRW_SURFACEFORMAT_FXT1                           0x192 
+#define BRW_SURFACEFORMAT_R8G8B8_UNORM                   0x193 
+#define BRW_SURFACEFORMAT_R8G8B8_SNORM                   0x194 
+#define BRW_SURFACEFORMAT_R8G8B8_SSCALED                 0x195 
+#define BRW_SURFACEFORMAT_R8G8B8_USCALED                 0x196 
+#define BRW_SURFACEFORMAT_R64G64B64A64_FLOAT             0x197 
+#define BRW_SURFACEFORMAT_R64G64B64_FLOAT                0x198 
+#define BRW_SURFACEFORMAT_BC4_SNORM                      0x199 
+#define BRW_SURFACEFORMAT_BC5_SNORM                      0x19A 
+#define BRW_SURFACEFORMAT_R16G16B16_UNORM                0x19C 
+#define BRW_SURFACEFORMAT_R16G16B16_SNORM                0x19D 
+#define BRW_SURFACEFORMAT_R16G16B16_SSCALED              0x19E 
+#define BRW_SURFACEFORMAT_R16G16B16_USCALED              0x19F
+#define BRW_SURFACEFORMAT_INVALID                        0xFFF
+
+#define BRW_SURFACERETURNFORMAT_FLOAT32  0
+#define BRW_SURFACERETURNFORMAT_S1       1
+
+#define BRW_SURFACE_1D      0
+#define BRW_SURFACE_2D      1
+#define BRW_SURFACE_3D      2
+#define BRW_SURFACE_CUBE    3
+#define BRW_SURFACE_BUFFER  4
+#define BRW_SURFACE_NULL    7
+
+#define BRW_TEXCOORDMODE_WRAP            0
+#define BRW_TEXCOORDMODE_MIRROR          1
+#define BRW_TEXCOORDMODE_CLAMP           2
+#define BRW_TEXCOORDMODE_CUBE            3
+#define BRW_TEXCOORDMODE_CLAMP_BORDER    4
+#define BRW_TEXCOORDMODE_MIRROR_ONCE     5
+
+#define BRW_THREAD_PRIORITY_NORMAL   0
+#define BRW_THREAD_PRIORITY_HIGH     1
+
+#define BRW_TILEWALK_XMAJOR                 0
+#define BRW_TILEWALK_YMAJOR                 1
+
+#define BRW_VERTEX_SUBPIXEL_PRECISION_8BITS  0
+#define BRW_VERTEX_SUBPIXEL_PRECISION_4BITS  1
+
+/* Execution Unit (EU) defines
+ */
+
+#define BRW_ALIGN_1   0
+#define BRW_ALIGN_16  1
+
+#define BRW_ADDRESS_DIRECT                        0
+#define BRW_ADDRESS_REGISTER_INDIRECT_REGISTER    1
+
+#define BRW_CHANNEL_X     0
+#define BRW_CHANNEL_Y     1
+#define BRW_CHANNEL_Z     2
+#define BRW_CHANNEL_W     3
+
+#define BRW_COMPRESSION_NONE          0
+#define BRW_COMPRESSION_2NDHALF       1
+#define BRW_COMPRESSION_COMPRESSED    2
+
+#define BRW_CONDITIONAL_NONE  0
+#define BRW_CONDITIONAL_Z     1
+#define BRW_CONDITIONAL_NZ    2
+#define BRW_CONDITIONAL_EQ    1        /* Z */
+#define BRW_CONDITIONAL_NEQ   2        /* NZ */
+#define BRW_CONDITIONAL_G     3
+#define BRW_CONDITIONAL_GE    4
+#define BRW_CONDITIONAL_L     5
+#define BRW_CONDITIONAL_LE    6
+#define BRW_CONDITIONAL_R     7
+#define BRW_CONDITIONAL_O     8
+#define BRW_CONDITIONAL_U     9
+
+#define BRW_DEBUG_NONE        0
+#define BRW_DEBUG_BREAKPOINT  1
+
+#define BRW_DEPENDENCY_NORMAL         0
+#define BRW_DEPENDENCY_NOTCLEARED     1
+#define BRW_DEPENDENCY_NOTCHECKED     2
+#define BRW_DEPENDENCY_DISABLE        3
+
+#define BRW_EXECUTE_1     0
+#define BRW_EXECUTE_2     1
+#define BRW_EXECUTE_4     2
+#define BRW_EXECUTE_8     3
+#define BRW_EXECUTE_16    4
+#define BRW_EXECUTE_32    5
+
+#define BRW_HORIZONTAL_STRIDE_0   0
+#define BRW_HORIZONTAL_STRIDE_1   1
+#define BRW_HORIZONTAL_STRIDE_2   2
+#define BRW_HORIZONTAL_STRIDE_4   3
+
+#define BRW_INSTRUCTION_NORMAL    0
+#define BRW_INSTRUCTION_SATURATE  1
+
+#define BRW_MASK_ENABLE   0
+#define BRW_MASK_DISABLE  1
+
+#define BRW_OPCODE_MOV        1
+#define BRW_OPCODE_SEL        2
+#define BRW_OPCODE_NOT        4
+#define BRW_OPCODE_AND        5
+#define BRW_OPCODE_OR         6
+#define BRW_OPCODE_XOR        7
+#define BRW_OPCODE_SHR        8
+#define BRW_OPCODE_SHL        9
+#define BRW_OPCODE_RSR        10
+#define BRW_OPCODE_RSL        11
+#define BRW_OPCODE_ASR        12
+#define BRW_OPCODE_CMP        16
+#define BRW_OPCODE_CMPN       17
+#define BRW_OPCODE_JMPI       32
+#define BRW_OPCODE_IF         34
+#define BRW_OPCODE_IFF        35
+#define BRW_OPCODE_ELSE       36
+#define BRW_OPCODE_ENDIF      37
+#define BRW_OPCODE_DO         38
+#define BRW_OPCODE_WHILE      39
+#define BRW_OPCODE_BREAK      40
+#define BRW_OPCODE_CONTINUE   41
+#define BRW_OPCODE_HALT       42
+#define BRW_OPCODE_MSAVE      44
+#define BRW_OPCODE_MRESTORE   45
+#define BRW_OPCODE_PUSH       46
+#define BRW_OPCODE_POP        47
+#define BRW_OPCODE_WAIT       48
+#define BRW_OPCODE_SEND       49
+#define BRW_OPCODE_ADD        64
+#define BRW_OPCODE_MUL        65
+#define BRW_OPCODE_AVG        66
+#define BRW_OPCODE_FRC        67
+#define BRW_OPCODE_RNDU       68
+#define BRW_OPCODE_RNDD       69
+#define BRW_OPCODE_RNDE       70
+#define BRW_OPCODE_RNDZ       71
+#define BRW_OPCODE_MAC        72
+#define BRW_OPCODE_MACH       73
+#define BRW_OPCODE_LZD        74
+#define BRW_OPCODE_SAD2       80
+#define BRW_OPCODE_SADA2      81
+#define BRW_OPCODE_DP4        84
+#define BRW_OPCODE_DPH        85
+#define BRW_OPCODE_DP3        86
+#define BRW_OPCODE_DP2        87
+#define BRW_OPCODE_DPA2       88
+#define BRW_OPCODE_LINE       89
+#define BRW_OPCODE_NOP        126
+
+#define BRW_PREDICATE_NONE             0
+#define BRW_PREDICATE_NORMAL           1
+#define BRW_PREDICATE_ALIGN1_ANYV             2
+#define BRW_PREDICATE_ALIGN1_ALLV             3
+#define BRW_PREDICATE_ALIGN1_ANY2H            4
+#define BRW_PREDICATE_ALIGN1_ALL2H            5
+#define BRW_PREDICATE_ALIGN1_ANY4H            6
+#define BRW_PREDICATE_ALIGN1_ALL4H            7
+#define BRW_PREDICATE_ALIGN1_ANY8H            8
+#define BRW_PREDICATE_ALIGN1_ALL8H            9
+#define BRW_PREDICATE_ALIGN1_ANY16H           10
+#define BRW_PREDICATE_ALIGN1_ALL16H           11
+#define BRW_PREDICATE_ALIGN16_REPLICATE_X     2
+#define BRW_PREDICATE_ALIGN16_REPLICATE_Y     3
+#define BRW_PREDICATE_ALIGN16_REPLICATE_Z     4
+#define BRW_PREDICATE_ALIGN16_REPLICATE_W     5
+#define BRW_PREDICATE_ALIGN16_ANY4H           6
+#define BRW_PREDICATE_ALIGN16_ALL4H           7
+
+#define BRW_ARCHITECTURE_REGISTER_FILE    0
+#define BRW_GENERAL_REGISTER_FILE         1
+#define BRW_MESSAGE_REGISTER_FILE         2
+#define BRW_IMMEDIATE_VALUE               3
+
+#define BRW_REGISTER_TYPE_UD  0
+#define BRW_REGISTER_TYPE_D   1
+#define BRW_REGISTER_TYPE_UW  2
+#define BRW_REGISTER_TYPE_W   3
+#define BRW_REGISTER_TYPE_UB  4
+#define BRW_REGISTER_TYPE_B   5
+#define BRW_REGISTER_TYPE_VF  5        /* packed float vector, immediates only? */
+#define BRW_REGISTER_TYPE_HF  6
+#define BRW_REGISTER_TYPE_V   6        /* packed int vector, immediates only, uword dest only */
+#define BRW_REGISTER_TYPE_F   7
+
+#define BRW_ARF_NULL                  0x00
+#define BRW_ARF_ADDRESS               0x10
+#define BRW_ARF_ACCUMULATOR           0x20   
+#define BRW_ARF_FLAG                  0x30
+#define BRW_ARF_MASK                  0x40
+#define BRW_ARF_MASK_STACK            0x50
+#define BRW_ARF_MASK_STACK_DEPTH      0x60
+#define BRW_ARF_STATE                 0x70
+#define BRW_ARF_CONTROL               0x80
+#define BRW_ARF_NOTIFICATION_COUNT    0x90
+#define BRW_ARF_IP                    0xA0
+
+#define BRW_AMASK   0
+#define BRW_IMASK   1
+#define BRW_LMASK   2
+#define BRW_CMASK   3
+
+
+
+#define BRW_THREAD_NORMAL     0
+#define BRW_THREAD_ATOMIC     1
+#define BRW_THREAD_SWITCH     2
+
+#define BRW_VERTICAL_STRIDE_0                 0
+#define BRW_VERTICAL_STRIDE_1                 1
+#define BRW_VERTICAL_STRIDE_2                 2
+#define BRW_VERTICAL_STRIDE_4                 3
+#define BRW_VERTICAL_STRIDE_8                 4
+#define BRW_VERTICAL_STRIDE_16                5
+#define BRW_VERTICAL_STRIDE_32                6
+#define BRW_VERTICAL_STRIDE_64                7
+#define BRW_VERTICAL_STRIDE_128               8
+#define BRW_VERTICAL_STRIDE_256               9
+#define BRW_VERTICAL_STRIDE_ONE_DIMENSIONAL   0xF
+
+#define BRW_WIDTH_1       0
+#define BRW_WIDTH_2       1
+#define BRW_WIDTH_4       2
+#define BRW_WIDTH_8       3
+#define BRW_WIDTH_16      4
+
+#define BRW_STATELESS_BUFFER_BOUNDARY_1K      0
+#define BRW_STATELESS_BUFFER_BOUNDARY_2K      1
+#define BRW_STATELESS_BUFFER_BOUNDARY_4K      2
+#define BRW_STATELESS_BUFFER_BOUNDARY_8K      3
+#define BRW_STATELESS_BUFFER_BOUNDARY_16K     4
+#define BRW_STATELESS_BUFFER_BOUNDARY_32K     5
+#define BRW_STATELESS_BUFFER_BOUNDARY_64K     6
+#define BRW_STATELESS_BUFFER_BOUNDARY_128K    7
+#define BRW_STATELESS_BUFFER_BOUNDARY_256K    8
+#define BRW_STATELESS_BUFFER_BOUNDARY_512K    9
+#define BRW_STATELESS_BUFFER_BOUNDARY_1M      10
+#define BRW_STATELESS_BUFFER_BOUNDARY_2M      11
+
+#define BRW_POLYGON_FACING_FRONT      0
+#define BRW_POLYGON_FACING_BACK       1
+
+#define BRW_MESSAGE_TARGET_NULL               0
+#define BRW_MESSAGE_TARGET_MATH               1
+#define BRW_MESSAGE_TARGET_SAMPLER            2
+#define BRW_MESSAGE_TARGET_GATEWAY            3
+#define BRW_MESSAGE_TARGET_DATAPORT_READ      4
+#define BRW_MESSAGE_TARGET_DATAPORT_WRITE     5
+#define BRW_MESSAGE_TARGET_URB                6
+#define BRW_MESSAGE_TARGET_THREAD_SPAWNER     7
+
+#define BRW_SAMPLER_RETURN_FORMAT_FLOAT32     0
+#define BRW_SAMPLER_RETURN_FORMAT_UINT32      2
+#define BRW_SAMPLER_RETURN_FORMAT_SINT32      3
+
+#define BRW_SAMPLER_MESSAGE_SIMD8_SAMPLE              0
+#define BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE             0
+#define BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_BIAS        0
+#define BRW_SAMPLER_MESSAGE_SIMD8_KILLPIX             1
+#define BRW_SAMPLER_MESSAGE_SIMD4X2_SAMPLE_LOD        1
+#define BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_LOD         1
+#define BRW_SAMPLER_MESSAGE_SIMD4X2_SAMPLE_GRADIENTS  2
+#define BRW_SAMPLER_MESSAGE_SIMD8_SAMPLE_GRADIENTS    2
+#define BRW_SAMPLER_MESSAGE_SIMD4X2_SAMPLE_COMPARE    0
+#define BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_COMPARE     2
+#define BRW_SAMPLER_MESSAGE_SIMD4X2_RESINFO           2
+#define BRW_SAMPLER_MESSAGE_SIMD8_RESINFO             2
+#define BRW_SAMPLER_MESSAGE_SIMD16_RESINFO            2
+#define BRW_SAMPLER_MESSAGE_SIMD4X2_LD                3
+#define BRW_SAMPLER_MESSAGE_SIMD8_LD                  3
+#define BRW_SAMPLER_MESSAGE_SIMD16_LD                 3
+
+#define BRW_SAMPLER_MESSAGE_SIMD8_SAMPLE_IGDNG            0
+#define BRW_SAMPLER_MESSAGE_SIMD4X2_SAMPLE_IGDNG          0
+#define BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_IGDNG           0
+#define BRW_SAMPLER_MESSAGE_SIMD8_SAMPLE_BIAS_IGDNG       1
+#define BRW_SAMPLER_MESSAGE_SIMD4X2_SAMPLE_BIAS_IGDNG     1
+#define BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_BIAS_IGDNG      1
+#define BRW_SAMPLER_MESSAGE_SIMD8_SAMPLE_LOD_IGDNG        2
+#define BRW_SAMPLER_MESSAGE_SIMD4X2_SAMPLE_LOD_IGDNG      2
+#define BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_LOD_IGDNG       2
+#define BRW_SAMPLER_MESSAGE_SIMD8_SAMPLE_COMPARE_IGDNG    3
+#define BRW_SAMPLER_MESSAGE_SIMD4X2_SAMPLE_COMPARE_IGDNG  3
+#define BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_COMPARE_IGDNG   3
+
+/* for IGDNG only */
+#define BRW_SAMPLER_SIMD_MODE_SIMD4X2                   0
+#define BRW_SAMPLER_SIMD_MODE_SIMD8                     1
+#define BRW_SAMPLER_SIMD_MODE_SIMD16                    2
+#define BRW_SAMPLER_SIMD_MODE_SIMD32_64                 3
+
+#define BRW_DATAPORT_OWORD_BLOCK_1_OWORDLOW   0
+#define BRW_DATAPORT_OWORD_BLOCK_1_OWORDHIGH  1
+#define BRW_DATAPORT_OWORD_BLOCK_2_OWORDS     2
+#define BRW_DATAPORT_OWORD_BLOCK_4_OWORDS     3
+#define BRW_DATAPORT_OWORD_BLOCK_8_OWORDS     4
+
+#define BRW_DATAPORT_OWORD_DUAL_BLOCK_1OWORD     0
+#define BRW_DATAPORT_OWORD_DUAL_BLOCK_4OWORDS    2
+
+#define BRW_DATAPORT_DWORD_SCATTERED_BLOCK_8DWORDS   2
+#define BRW_DATAPORT_DWORD_SCATTERED_BLOCK_16DWORDS  3
+
+#define BRW_DATAPORT_READ_MESSAGE_OWORD_BLOCK_READ          0
+#define BRW_DATAPORT_READ_MESSAGE_OWORD_DUAL_BLOCK_READ     1
+#define BRW_DATAPORT_READ_MESSAGE_DWORD_BLOCK_READ          2
+#define BRW_DATAPORT_READ_MESSAGE_DWORD_SCATTERED_READ      3
+
+#define BRW_DATAPORT_READ_TARGET_DATA_CACHE      0
+#define BRW_DATAPORT_READ_TARGET_RENDER_CACHE    1
+#define BRW_DATAPORT_READ_TARGET_SAMPLER_CACHE   2
+
+#define BRW_DATAPORT_RENDER_TARGET_WRITE_SIMD16_SINGLE_SOURCE                0
+#define BRW_DATAPORT_RENDER_TARGET_WRITE_SIMD16_SINGLE_SOURCE_REPLICATED     1
+#define BRW_DATAPORT_RENDER_TARGET_WRITE_SIMD8_DUAL_SOURCE_SUBSPAN01         2
+#define BRW_DATAPORT_RENDER_TARGET_WRITE_SIMD8_DUAL_SOURCE_SUBSPAN23         3
+#define BRW_DATAPORT_RENDER_TARGET_WRITE_SIMD8_SINGLE_SOURCE_SUBSPAN01       4
+
+#define BRW_DATAPORT_WRITE_MESSAGE_OWORD_BLOCK_WRITE                0
+#define BRW_DATAPORT_WRITE_MESSAGE_OWORD_DUAL_BLOCK_WRITE           1
+#define BRW_DATAPORT_WRITE_MESSAGE_DWORD_BLOCK_WRITE                2
+#define BRW_DATAPORT_WRITE_MESSAGE_DWORD_SCATTERED_WRITE            3
+#define BRW_DATAPORT_WRITE_MESSAGE_RENDER_TARGET_WRITE              4
+#define BRW_DATAPORT_WRITE_MESSAGE_STREAMED_VERTEX_BUFFER_WRITE     5
+#define BRW_DATAPORT_WRITE_MESSAGE_FLUSH_RENDER_CACHE               7
+
+#define BRW_MATH_FUNCTION_INV                              1
+#define BRW_MATH_FUNCTION_LOG                              2
+#define BRW_MATH_FUNCTION_EXP                              3
+#define BRW_MATH_FUNCTION_SQRT                             4
+#define BRW_MATH_FUNCTION_RSQ                              5
+#define BRW_MATH_FUNCTION_SIN                              6 /* was 7 */
+#define BRW_MATH_FUNCTION_COS                              7 /* was 8 */
+#define BRW_MATH_FUNCTION_SINCOS                           8 /* was 6 */
+#define BRW_MATH_FUNCTION_TAN                              9
+#define BRW_MATH_FUNCTION_POW                              10
+#define BRW_MATH_FUNCTION_INT_DIV_QUOTIENT_AND_REMAINDER   11
+#define BRW_MATH_FUNCTION_INT_DIV_QUOTIENT                 12
+#define BRW_MATH_FUNCTION_INT_DIV_REMAINDER                13
+
+#define BRW_MATH_INTEGER_UNSIGNED     0
+#define BRW_MATH_INTEGER_SIGNED       1
+
+#define BRW_MATH_PRECISION_FULL        0
+#define BRW_MATH_PRECISION_PARTIAL     1
+
+#define BRW_MATH_SATURATE_NONE         0
+#define BRW_MATH_SATURATE_SATURATE     1
+
+#define BRW_MATH_DATA_VECTOR  0
+#define BRW_MATH_DATA_SCALAR  1
+
+#define BRW_URB_OPCODE_WRITE  0
+
+#define BRW_URB_SWIZZLE_NONE          0
+#define BRW_URB_SWIZZLE_INTERLEAVE    1
+#define BRW_URB_SWIZZLE_TRANSPOSE     2
+
+#define BRW_SCRATCH_SPACE_SIZE_1K     0
+#define BRW_SCRATCH_SPACE_SIZE_2K     1
+#define BRW_SCRATCH_SPACE_SIZE_4K     2
+#define BRW_SCRATCH_SPACE_SIZE_8K     3
+#define BRW_SCRATCH_SPACE_SIZE_16K    4
+#define BRW_SCRATCH_SPACE_SIZE_32K    5
+#define BRW_SCRATCH_SPACE_SIZE_64K    6
+#define BRW_SCRATCH_SPACE_SIZE_128K   7
+#define BRW_SCRATCH_SPACE_SIZE_256K   8
+#define BRW_SCRATCH_SPACE_SIZE_512K   9
+#define BRW_SCRATCH_SPACE_SIZE_1M     10
+#define BRW_SCRATCH_SPACE_SIZE_2M     11
+
+
+
+
+#define CMD_URB_FENCE                 0x6000
+#define CMD_CS_URB_STATE              0x6001
+#define CMD_CONST_BUFFER              0x6002
+
+#define CMD_STATE_BASE_ADDRESS        0x6101
+#define CMD_STATE_INSN_POINTER        0x6102
+#define CMD_PIPELINE_SELECT_965       0x6104
+#define CMD_PIPELINE_SELECT_GM45      0x6904
+
+#define CMD_PIPELINED_STATE_POINTERS  0x7800
+#define CMD_BINDING_TABLE_PTRS        0x7801
+
+#define CMD_VERTEX_BUFFER             0x7808
+# define BRW_VB0_INDEX_SHIFT           27
+# define BRW_VB0_ACCESS_VERTEXDATA     (0 << 26)
+# define BRW_VB0_ACCESS_INSTANCEDATA   (1 << 26)
+# define BRW_VB0_PITCH_SHIFT           0
+
+#define CMD_VERTEX_ELEMENT            0x7809
+# define BRW_VE0_INDEX_SHIFT           27
+# define BRW_VE0_FORMAT_SHIFT          16
+# define BRW_VE0_VALID                 (1 << 26)
+# define BRW_VE0_SRC_OFFSET_SHIFT      0
+# define BRW_VE1_COMPONENT_NOSTORE     0
+# define BRW_VE1_COMPONENT_STORE_SRC   1
+# define BRW_VE1_COMPONENT_STORE_0     2
+# define BRW_VE1_COMPONENT_STORE_1_FLT 3
+# define BRW_VE1_COMPONENT_STORE_1_INT 4
+# define BRW_VE1_COMPONENT_STORE_VID   5
+# define BRW_VE1_COMPONENT_STORE_IID   6
+# define BRW_VE1_COMPONENT_STORE_PID   7
+# define BRW_VE1_COMPONENT_0_SHIFT     28
+# define BRW_VE1_COMPONENT_1_SHIFT     24
+# define BRW_VE1_COMPONENT_2_SHIFT     20
+# define BRW_VE1_COMPONENT_3_SHIFT     16
+# define BRW_VE1_DST_OFFSET_SHIFT      0
+
+#define CMD_INDEX_BUFFER              0x780a
+#define CMD_VF_STATISTICS_965         0x780b
+#define CMD_VF_STATISTICS_GM45        0x680b
+
+#define CMD_DRAW_RECT                 0x7900
+#define CMD_BLEND_CONSTANT_COLOR      0x7901
+#define CMD_CHROMA_KEY                0x7904
+#define CMD_DEPTH_BUFFER              0x7905
+#define CMD_POLY_STIPPLE_OFFSET       0x7906
+#define CMD_POLY_STIPPLE_PATTERN      0x7907
+#define CMD_LINE_STIPPLE_PATTERN      0x7908
+#define CMD_GLOBAL_DEPTH_OFFSET_CLAMP 0x7909
+#define CMD_AA_LINE_PARAMETERS        0x790a
+
+#define CMD_PIPE_CONTROL              0x7a00
+
+#define CMD_3D_PRIM                   0x7b00
+
+#define CMD_MI_FLUSH                  0x0200
+
+
+/* Various values from the R0 vertex header:
+ */
+#define R02_PRIM_END    0x1
+#define R02_PRIM_START  0x2
+
+#define URB_SIZES(brw)                  (BRW_IS_IGDNG(brw) ? 1024 : \
+                                         (BRW_IS_G4X(brw) ? 384 : 256))  /* 512 bit units */
+
+
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_disasm.c b/src/gallium/drivers/i965/brw_disasm.c
new file mode 100644 (file)
index 0000000..65db272
--- /dev/null
@@ -0,0 +1,922 @@
+/*
+ * Copyright Â© 2008 Keith Packard
+ *
+ * Permission to use, copy, modify, distribute, and sell this software and its
+ * documentation for any purpose is hereby granted without fee, provided that
+ * the above copyright notice appear in all copies and that both that copyright
+ * notice and this permission notice appear in supporting documentation, and
+ * that the name of the copyright holders not be used in advertising or
+ * publicity pertaining to distribution of the software without specific,
+ * written prior permission.  The copyright holders make no representations
+ * about the suitability of this software for any purpose.  It is provided "as
+ * is" without express or implied warranty.
+ *
+ * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
+ * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
+ * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
+ * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
+ * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
+ * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
+ * OF THIS SOFTWARE.
+ */
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <getopt.h>
+#include <unistd.h>
+#include <stdarg.h>
+
+#include "brw_disasm.h"
+#include "brw_structs.h"
+#include "brw_reg.h"
+#include "brw_defines.h"
+
+struct {
+    char    *name;
+    int            nsrc;
+    int            ndst;
+} opcode[128] = {
+    [BRW_OPCODE_MOV] = { .name = "mov", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_FRC] = { .name = "frc", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_RNDU] = { .name = "rndu", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_RNDD] = { .name = "rndd", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_RNDE] = { .name = "rnde", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_RNDZ] = { .name = "rndz", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_NOT] = { .name = "not", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_LZD] = { .name = "lzd", .nsrc = 1, .ndst = 1 },
+
+    [BRW_OPCODE_MUL] = { .name = "mul", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_MAC] = { .name = "mac", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_MACH] = { .name = "mach", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_LINE] = { .name = "line", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_SAD2] = { .name = "sad2", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_SADA2] = { .name = "sada2", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_DP4] = { .name = "dp4", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_DPH] = { .name = "dph", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_DP3] = { .name = "dp3", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_DP2] = { .name = "dp2", .nsrc = 2, .ndst = 1 },
+
+    [BRW_OPCODE_AVG] = { .name = "avg", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_ADD] = { .name = "add", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_SEL] = { .name = "sel", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_AND] = { .name = "and", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_OR] = { .name = "or", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_XOR] = { .name = "xor", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_SHR] = { .name = "shr", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_SHL] = { .name = "shl", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_ASR] = { .name = "asr", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_CMP] = { .name = "cmp", .nsrc = 2, .ndst = 1 },
+    [BRW_OPCODE_CMPN] = { .name = "cmpn", .nsrc = 2, .ndst = 1 },
+
+    [BRW_OPCODE_SEND] = { .name = "send", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_NOP] = { .name = "nop", .nsrc = 0, .ndst = 0 },
+    [BRW_OPCODE_JMPI] = { .name = "jmpi", .nsrc = 1, .ndst = 0 },
+    [BRW_OPCODE_IF] = { .name = "if", .nsrc = 2, .ndst = 0 },
+    [BRW_OPCODE_IFF] = { .name = "iff", .nsrc = 1, .ndst = 01 },
+    [BRW_OPCODE_WHILE] = { .name = "while", .nsrc = 1, .ndst = 0 },
+    [BRW_OPCODE_ELSE] = { .name = "else", .nsrc = 2, .ndst = 0 },
+    [BRW_OPCODE_BREAK] = { .name = "break", .nsrc = 1, .ndst = 0 },
+    [BRW_OPCODE_CONTINUE] = { .name = "cont", .nsrc = 1, .ndst = 0 },
+    [BRW_OPCODE_HALT] = { .name = "halt", .nsrc = 1, .ndst = 0 },
+    [BRW_OPCODE_MSAVE] = { .name = "msave", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_PUSH] = { .name = "push", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_MRESTORE] = { .name = "mrest", .nsrc = 1, .ndst = 1 },
+    [BRW_OPCODE_POP] = { .name = "pop", .nsrc = 2, .ndst = 0 },
+    [BRW_OPCODE_WAIT] = { .name = "wait", .nsrc = 1, .ndst = 0 },
+    [BRW_OPCODE_DO] = { .name = "do", .nsrc = 0, .ndst = 0 },
+    [BRW_OPCODE_ENDIF] = { .name = "endif", .nsrc = 2, .ndst = 0 },
+};
+
+char *conditional_modifier[16] = {
+    [BRW_CONDITIONAL_NONE] = "",
+    [BRW_CONDITIONAL_Z] = ".e",
+    [BRW_CONDITIONAL_NZ] = ".ne",
+    [BRW_CONDITIONAL_G] = ".g",
+    [BRW_CONDITIONAL_GE] = ".ge",
+    [BRW_CONDITIONAL_L] = ".l",
+    [BRW_CONDITIONAL_LE] = ".le",
+    [BRW_CONDITIONAL_R] = ".r",
+    [BRW_CONDITIONAL_O] = ".o",
+    [BRW_CONDITIONAL_U] = ".u",
+};
+
+char *negate[2] = {
+    [0] = "",
+    [1] = "-",
+};
+
+char *_abs[2] = {
+    [0] = "",
+    [1] = "(abs)",
+};
+
+char *vert_stride[16] = {
+    [0] = "0",
+    [1] = "1",
+    [2] = "2",
+    [3] = "4",
+    [4] = "8",
+    [5] = "16",
+    [6] = "32",
+    [15] = "VxH",
+};
+
+char *width[8] = {
+    [0] = "1",
+    [1] = "2",
+    [2] = "4",
+    [3] = "8",
+    [4] = "16",
+};
+
+char *horiz_stride[4] = {
+    [0] = "0",
+    [1] = "1",
+    [2] = "2",
+    [3] = "4"
+};
+
+char *chan_sel[4] = {
+    [0] = "x",
+    [1] = "y",
+    [2] = "z",
+    [3] = "w",
+};
+
+char *dest_condmod[16] = {
+   [0] = NULL
+};
+
+char *debug_ctrl[2] = {
+    [0] = "",
+    [1] = ".breakpoint"
+};
+
+char *saturate[2] = {
+    [0] = "",
+    [1] = ".sat"
+};
+
+char *exec_size[8] = {
+    [0] = "1",
+    [1] = "2",
+    [2] = "4",
+    [3] = "8",
+    [4] = "16",
+    [5] = "32"
+};
+
+char *pred_inv[2] = {
+    [0] = "+",
+    [1] = "-"
+};
+
+char *pred_ctrl_align16[16] = {
+    [1] = "",
+    [2] = ".x",
+    [3] = ".y",
+    [4] = ".z",
+    [5] = ".w",
+    [6] = ".any4h",
+    [7] = ".all4h",
+};
+
+char *pred_ctrl_align1[16] = {
+    [1] = "",
+    [2] = ".anyv",
+    [3] = ".allv",
+    [4] = ".any2h",
+    [5] = ".all2h",
+    [6] = ".any4h",
+    [7] = ".all4h",
+    [8] = ".any8h",
+    [9] = ".all8h",
+    [10] = ".any16h",
+    [11] = ".all16h",
+};
+
+char *thread_ctrl[4] = {
+    [0] = "",
+    [2] = "switch"
+};
+
+char *compr_ctrl[4] = {
+    [0] = "",
+    [1] = "sechalf",
+    [2] = "compr",
+};
+
+char *dep_ctrl[4] = {
+    [0] = "",
+    [1] = "NoDDClr",
+    [2] = "NoDDChk",
+    [3] = "NoDDClr,NoDDChk",
+};
+
+char *mask_ctrl[4] = {
+    [0] = "",
+    [1] = "nomask",
+};
+
+char *access_mode[2] = {
+    [0] = "align1",
+    [1] = "align16",
+};
+
+char *reg_encoding[8] = {
+    [0] = "UD",
+    [1] = "D",
+    [2] = "UW",
+    [3] = "W",
+    [4] = "UB",
+    [5] = "B",
+    [7] = "F"
+};
+
+char *imm_encoding[8] = {
+    [0] = "UD",
+    [1] = "D",
+    [2] = "UW",
+    [3] = "W",
+    [5] = "VF",
+    [5] = "V",
+    [7] = "F"
+};
+
+char *reg_file[4] = {
+    [0] = "A",
+    [1] = "g",
+    [2] = "m",
+    [3] = "imm",
+};
+
+char *writemask[16] = {
+    [0x0] = ".",
+    [0x1] = ".x",
+    [0x2] = ".y",
+    [0x3] = ".xy",
+    [0x4] = ".z",
+    [0x5] = ".xz",
+    [0x6] = ".yz",
+    [0x7] = ".xyz",
+    [0x8] = ".w",
+    [0x9] = ".xw",
+    [0xa] = ".yw",
+    [0xb] = ".xyw",
+    [0xc] = ".zw",
+    [0xd] = ".xzw",
+    [0xe] = ".yzw",
+    [0xf] = "",
+};
+
+char *end_of_thread[2] = {
+    [0] = "",
+    [1] = "EOT"
+};
+
+char *target_function[16] = {
+    [BRW_MESSAGE_TARGET_NULL] = "null",
+    [BRW_MESSAGE_TARGET_MATH] = "math",
+    [BRW_MESSAGE_TARGET_SAMPLER] = "sampler",
+    [BRW_MESSAGE_TARGET_GATEWAY] = "gateway",
+    [BRW_MESSAGE_TARGET_DATAPORT_READ] = "read",
+    [BRW_MESSAGE_TARGET_DATAPORT_WRITE] = "write",
+    [BRW_MESSAGE_TARGET_URB] = "urb",
+    [BRW_MESSAGE_TARGET_THREAD_SPAWNER] = "thread_spawner"
+};
+
+char *math_function[16] = {
+    [BRW_MATH_FUNCTION_INV] = "inv",
+    [BRW_MATH_FUNCTION_LOG] = "log",
+    [BRW_MATH_FUNCTION_EXP] = "exp",
+    [BRW_MATH_FUNCTION_SQRT] = "sqrt",
+    [BRW_MATH_FUNCTION_RSQ] = "rsq",
+    [BRW_MATH_FUNCTION_SIN] = "sin",
+    [BRW_MATH_FUNCTION_COS] = "cos",
+    [BRW_MATH_FUNCTION_SINCOS] = "sincos",
+    [BRW_MATH_FUNCTION_TAN] = "tan",
+    [BRW_MATH_FUNCTION_POW] = "pow",
+    [BRW_MATH_FUNCTION_INT_DIV_QUOTIENT_AND_REMAINDER] = "intdivmod",
+    [BRW_MATH_FUNCTION_INT_DIV_QUOTIENT] = "intmod",
+    [BRW_MATH_FUNCTION_INT_DIV_REMAINDER] = "intdiv",
+};
+
+char *math_saturate[2] = {
+    [0] = "",
+    [1] = "sat"
+};
+
+char *math_signed[2] = {
+    [0] = "",
+    [1] = "signed"
+};
+
+char *math_scalar[2] = {
+    [0] = "",
+    [1] = "scalar"
+};
+
+char *math_precision[2] = {
+    [0] = "",
+    [1] = "partial_precision"
+};
+
+char *urb_swizzle[4] = {
+    [BRW_URB_SWIZZLE_NONE] = "",
+    [BRW_URB_SWIZZLE_INTERLEAVE] = "interleave",
+    [BRW_URB_SWIZZLE_TRANSPOSE] = "transpose",
+};
+
+char *urb_allocate[2] = {
+    [0] = "",
+    [1] = "allocate"
+};
+
+char *urb_used[2] = {
+    [0] = "",
+    [1] = "used"
+};
+
+char *urb_complete[2] = {
+    [0] = "",
+    [1] = "complete"
+};
+
+char *sampler_target_format[4] = {
+    [0] = "F",
+    [2] = "UD",
+    [3] = "D"
+};
+
+
+static int column;
+
+static int string (FILE *file, char *string)
+{
+    fputs (string, file);
+    column += strlen (string);
+    return 0;
+}
+
+static int format (FILE *f, char *format, ...)
+{
+    char    buf[1024];
+    va_list    args;
+    va_start (args, format);
+
+    vsnprintf (buf, sizeof (buf) - 1, format, args);
+    string (f, buf);
+    return 0;
+}
+
+static int newline (FILE *f)
+{
+    putc ('\n', f);
+    column = 0;
+    return 0;
+}
+
+static int pad (FILE *f, int c)
+{
+    do
+       string (f, " ");
+    while (column < c);
+    return 0;
+}
+
+static int control (FILE *file, char *name, char *ctrl[], GLuint id, int *space)
+{
+    if (!ctrl[id]) {
+       fprintf (file, "*** invalid %s value %d ",
+                name, id);
+       return 1;
+    }
+    if (ctrl[id][0])
+    {
+       if (space && *space)
+           string (file, " ");
+       string (file, ctrl[id]);
+       if (space)
+           *space = 1;
+    }
+    return 0;
+}
+
+static int print_opcode (FILE *file, int id)
+{
+    if (!opcode[id].name) {
+       format (file, "*** invalid opcode value %d ", id);
+       return 1;
+    }
+    string (file, opcode[id].name);
+    return 0;
+}
+
+static int reg (FILE *file, GLuint _reg_file, GLuint _reg_nr)
+{
+    int        err = 0;
+    if (_reg_file == BRW_ARCHITECTURE_REGISTER_FILE) {
+       switch (_reg_nr & 0xf0) {
+       case BRW_ARF_NULL:
+           string (file, "null");
+           return -1;
+       case BRW_ARF_ADDRESS:
+           format (file, "a%d", _reg_nr & 0x0f);
+           break;
+       case BRW_ARF_ACCUMULATOR:
+           format (file, "acc%d", _reg_nr & 0x0f);
+           break;
+       case BRW_ARF_MASK:
+           format (file, "mask%d", _reg_nr & 0x0f);
+           break;
+       case BRW_ARF_MASK_STACK:
+           format (file, "msd%d", _reg_nr & 0x0f);
+           break;
+       case BRW_ARF_STATE:
+           format (file, "sr%d", _reg_nr & 0x0f);
+           break;
+       case BRW_ARF_CONTROL:
+           format (file, "cr%d", _reg_nr & 0x0f);
+           break;
+       case BRW_ARF_NOTIFICATION_COUNT:
+           format (file, "n%d", _reg_nr & 0x0f);
+           break;
+       case BRW_ARF_IP:
+           string (file, "ip");
+           return -1;
+           break;
+       default:
+           format (file, "ARF%d", _reg_nr);
+           break;
+       }
+    } else {
+       err  |= control (file, "src reg file", reg_file, _reg_file, NULL);
+       format (file, "%d", _reg_nr);
+    }
+    return err;
+}
+
+static int dest (FILE *file, const struct brw_instruction *inst)
+{
+    int        err = 0;
+
+    if (inst->header.access_mode == BRW_ALIGN_1)
+    {
+       if (inst->bits1.da1.dest_address_mode == BRW_ADDRESS_DIRECT)
+       {
+           err |= reg (file, inst->bits1.da1.dest_reg_file, inst->bits1.da1.dest_reg_nr);
+           if (err == -1)
+               return 0;
+           if (inst->bits1.da1.dest_subreg_nr)
+               format (file, ".%d", inst->bits1.da1.dest_subreg_nr);
+           format (file, "<%d>", inst->bits1.da1.dest_horiz_stride);
+           err |= control (file, "dest reg encoding", reg_encoding, inst->bits1.da1.dest_reg_type, NULL);
+       }
+       else
+       {
+           string (file, "g[a0");
+           if (inst->bits1.ia1.dest_subreg_nr)
+               format (file, ".%d", inst->bits1.ia1.dest_subreg_nr);
+           if (inst->bits1.ia1.dest_indirect_offset)
+               format (file, " %d", inst->bits1.ia1.dest_indirect_offset);
+           string (file, "]");
+           format (file, "<%d>", inst->bits1.ia1.dest_horiz_stride);
+           err |= control (file, "dest reg encoding", reg_encoding, inst->bits1.ia1.dest_reg_type, NULL);
+       }
+    }
+    else
+    {
+       if (inst->bits1.da16.dest_address_mode == BRW_ADDRESS_DIRECT)
+       {
+           err |= reg (file, inst->bits1.da16.dest_reg_file, inst->bits1.da16.dest_reg_nr);
+           if (err == -1)
+               return 0;
+           if (inst->bits1.da16.dest_subreg_nr)
+               format (file, ".%d", inst->bits1.da16.dest_subreg_nr);
+           string (file, "<1>");
+           err |= control (file, "writemask", writemask, inst->bits1.da16.dest_writemask, NULL);
+           err |= control (file, "dest reg encoding", reg_encoding, inst->bits1.da16.dest_reg_type, NULL);
+       }
+       else
+       {
+           err = 1;
+           string (file, "Indirect align16 address mode not supported");
+       }
+    }
+
+    return 0;
+}
+
+static int src_align1_region (FILE *file,
+                             GLuint _vert_stride, GLuint _width, GLuint _horiz_stride)
+{
+    int err = 0;
+    string (file, "<");
+    err |= control (file, "vert stride", vert_stride, _vert_stride, NULL);
+    string (file, ",");
+    err |= control (file, "width", width, _width, NULL);
+    string (file, ",");
+    err |= control (file, "horiz_stride", horiz_stride, _horiz_stride, NULL);
+    string (file, ">");
+    return err;
+}
+
+static int src_da1 (FILE *file, GLuint type, GLuint _reg_file,
+                   GLuint _vert_stride, GLuint _width, GLuint _horiz_stride,
+                   GLuint reg_num, GLuint sub_reg_num, GLuint __abs, GLuint _negate)
+{
+    int err = 0;
+    err |= control (file, "negate", negate, _negate, NULL);
+    err |= control (file, "abs", _abs, __abs, NULL);
+
+    err |= reg (file, _reg_file, reg_num);
+    if (err == -1)
+       return 0;
+    if (sub_reg_num)
+       format (file, ".%d", sub_reg_num);
+    src_align1_region (file, _vert_stride, _width, _horiz_stride);
+    err |= control (file, "src reg encoding", reg_encoding, type, NULL);
+    return err;
+}
+
+static int src_ia1 (FILE *file,
+                   GLuint type,
+                   GLuint _reg_file,
+                   GLint _addr_imm,
+                   GLuint _addr_subreg_nr,
+                   GLuint _negate,
+                   GLuint __abs,
+                   GLuint _addr_mode,
+                   GLuint _horiz_stride,
+                   GLuint _width,
+                   GLuint _vert_stride)
+{
+    int err = 0;
+    err |= control (file, "negate", negate, _negate, NULL);
+    err |= control (file, "abs", _abs, __abs, NULL);
+
+    string (file, "g[a0");
+    if (_addr_subreg_nr)
+       format (file, ".%d", _addr_subreg_nr);
+    if (_addr_imm)
+       format (file, " %d", _addr_imm);
+    string (file, "]");
+    src_align1_region (file, _vert_stride, _width, _horiz_stride);
+    err |= control (file, "src reg encoding", reg_encoding, type, NULL);
+    return err;
+}
+
+static int src_da16 (FILE *file,
+                    GLuint _reg_type,
+                    GLuint _reg_file,
+                    GLuint _vert_stride,
+                    GLuint _reg_nr,
+                    GLuint _subreg_nr,
+                    GLuint __abs,
+                    GLuint _negate,
+                    GLuint swz_x,
+                    GLuint swz_y,
+                    GLuint swz_z,
+                    GLuint swz_w)
+{
+    int err = 0;
+    err |= control (file, "negate", negate, _negate, NULL);
+    err |= control (file, "abs", _abs, __abs, NULL);
+
+    err |= reg (file, _reg_file, _reg_nr);
+    if (err == -1)
+       return 0;
+    if (_subreg_nr)
+       format (file, ".%d", _subreg_nr);
+    string (file, "<");
+    err |= control (file, "vert stride", vert_stride, _vert_stride, NULL);
+    string (file, ",1,1>");
+    err |= control (file, "src da16 reg type", reg_encoding, _reg_type, NULL);
+    /*
+     * Three kinds of swizzle display:
+     *  identity - nothing printed
+     *  1->all  - print the single channel
+     *  1->1     - print the mapping
+     */
+    if (swz_x == BRW_CHANNEL_X &&
+       swz_y == BRW_CHANNEL_Y &&
+       swz_z == BRW_CHANNEL_Z &&
+       swz_w == BRW_CHANNEL_W)
+    {
+       ;
+    }
+    else if (swz_x == swz_y && swz_x == swz_z && swz_x == swz_w)
+    {
+       string (file, ".");
+       err |= control (file, "channel select", chan_sel, swz_x, NULL);
+    }
+    else
+    {
+       string (file, ".");
+       err |= control (file, "channel select", chan_sel, swz_x, NULL);
+       err |= control (file, "channel select", chan_sel, swz_y, NULL);
+       err |= control (file, "channel select", chan_sel, swz_z, NULL);
+       err |= control (file, "channel select", chan_sel, swz_w, NULL);
+    }
+    return err;
+}
+
+
+static int imm (FILE *file, GLuint type, const struct brw_instruction *inst) {
+    switch (type) {
+    case BRW_REGISTER_TYPE_UD:
+       format (file, "0x%08xUD", inst->bits3.ud);
+       break;
+    case BRW_REGISTER_TYPE_D:
+       format (file, "%dD", inst->bits3.d);
+       break;
+    case BRW_REGISTER_TYPE_UW:
+       format (file, "0x%04xUW", (uint16_t) inst->bits3.ud);
+       break;
+    case BRW_REGISTER_TYPE_W:
+       format (file, "%dW", (int16_t) inst->bits3.d);
+       break;
+    case BRW_REGISTER_TYPE_UB:
+       format (file, "0x%02xUB", (int8_t) inst->bits3.ud);
+       break;
+    case BRW_REGISTER_TYPE_VF:
+       format (file, "Vector Float");
+       break;
+    case BRW_REGISTER_TYPE_V:
+       format (file, "0x%08xV", inst->bits3.ud);
+       break;
+    case BRW_REGISTER_TYPE_F:
+       format (file, "%-gF", inst->bits3.f);
+    }
+    return 0;
+}
+
+static int src0 (FILE *file, const struct brw_instruction *inst)
+{
+    if (inst->bits1.da1.src0_reg_file == BRW_IMMEDIATE_VALUE)
+       return imm (file, inst->bits1.da1.src0_reg_type,
+                   inst);
+    else if (inst->header.access_mode == BRW_ALIGN_1)
+    {
+       if (inst->bits2.da1.src0_address_mode == BRW_ADDRESS_DIRECT)
+       {
+           return src_da1 (file,
+                           inst->bits1.da1.src0_reg_type,
+                           inst->bits1.da1.src0_reg_file,
+                           inst->bits2.da1.src0_vert_stride,
+                           inst->bits2.da1.src0_width,
+                           inst->bits2.da1.src0_horiz_stride,
+                           inst->bits2.da1.src0_reg_nr,
+                           inst->bits2.da1.src0_subreg_nr,
+                           inst->bits2.da1.src0_abs,
+                           inst->bits2.da1.src0_negate);
+       }
+       else
+       {
+           return src_ia1 (file,
+                           inst->bits1.ia1.src0_reg_type,
+                           inst->bits1.ia1.src0_reg_file,
+                           inst->bits2.ia1.src0_indirect_offset,
+                           inst->bits2.ia1.src0_subreg_nr,
+                           inst->bits2.ia1.src0_negate,
+                           inst->bits2.ia1.src0_abs,
+                           inst->bits2.ia1.src0_address_mode,
+                           inst->bits2.ia1.src0_horiz_stride,
+                           inst->bits2.ia1.src0_width,
+                           inst->bits2.ia1.src0_vert_stride);
+       }
+    }
+    else
+    {
+       if (inst->bits2.da16.src0_address_mode == BRW_ADDRESS_DIRECT)
+       {
+           return src_da16 (file,
+                            inst->bits1.da16.src0_reg_type,
+                            inst->bits1.da16.src0_reg_file,
+                            inst->bits2.da16.src0_vert_stride,
+                            inst->bits2.da16.src0_reg_nr,
+                            inst->bits2.da16.src0_subreg_nr,
+                            inst->bits2.da16.src0_abs,
+                            inst->bits2.da16.src0_negate,
+                            inst->bits2.da16.src0_swz_x,
+                            inst->bits2.da16.src0_swz_y,
+                            inst->bits2.da16.src0_swz_z,
+                            inst->bits2.da16.src0_swz_w);
+       }
+       else
+       {
+           string (file, "Indirect align16 address mode not supported");
+           return 1;
+       }
+    }
+}
+
+static int src1 (FILE *file, const struct brw_instruction *inst)
+{
+    if (inst->bits1.da1.src1_reg_file == BRW_IMMEDIATE_VALUE)
+       return imm (file, inst->bits1.da1.src1_reg_type,
+                   inst);
+    else if (inst->header.access_mode == BRW_ALIGN_1)
+    {
+       if (inst->bits3.da1.src1_address_mode == BRW_ADDRESS_DIRECT)
+       {
+           return src_da1 (file,
+                           inst->bits1.da1.src1_reg_type,
+                           inst->bits1.da1.src1_reg_file,
+                           inst->bits3.da1.src1_vert_stride,
+                           inst->bits3.da1.src1_width,
+                           inst->bits3.da1.src1_horiz_stride,
+                           inst->bits3.da1.src1_reg_nr,
+                           inst->bits3.da1.src1_subreg_nr,
+                           inst->bits3.da1.src1_abs,
+                           inst->bits3.da1.src1_negate);
+       }
+       else
+       {
+           return src_ia1 (file,
+                           inst->bits1.ia1.src1_reg_type,
+                           inst->bits1.ia1.src1_reg_file,
+                           inst->bits3.ia1.src1_indirect_offset,
+                           inst->bits3.ia1.src1_subreg_nr,
+                           inst->bits3.ia1.src1_negate,
+                           inst->bits3.ia1.src1_abs,
+                           inst->bits3.ia1.src1_address_mode,
+                           inst->bits3.ia1.src1_horiz_stride,
+                           inst->bits3.ia1.src1_width,
+                           inst->bits3.ia1.src1_vert_stride);
+       }
+    }
+    else
+    {
+       if (inst->bits3.da16.src1_address_mode == BRW_ADDRESS_DIRECT)
+       {
+           return src_da16 (file,
+                            inst->bits1.da16.src1_reg_type,
+                            inst->bits1.da16.src1_reg_file,
+                            inst->bits3.da16.src1_vert_stride,
+                            inst->bits3.da16.src1_reg_nr,
+                            inst->bits3.da16.src1_subreg_nr,
+                            inst->bits3.da16.src1_abs,
+                            inst->bits3.da16.src1_negate,
+                            inst->bits3.da16.src1_swz_x,
+                            inst->bits3.da16.src1_swz_y,
+                            inst->bits3.da16.src1_swz_z,
+                            inst->bits3.da16.src1_swz_w);
+       }
+       else
+       {
+           string (file, "Indirect align16 address mode not supported");
+           return 1;
+       }
+    }
+}
+
+int brw_disasm_insn (FILE *file, const struct brw_instruction *inst)
+{
+    int        err = 0;
+    int space = 0;
+
+    if (inst->header.predicate_control) {
+       string (file, "(");
+       err |= control (file, "predicate inverse", pred_inv, inst->header.predicate_inverse, NULL);
+       string (file, "f0");
+       if (inst->bits2.da1.flag_reg_nr)
+           format (file, ".%d", inst->bits2.da1.flag_reg_nr);
+       if (inst->header.access_mode == BRW_ALIGN_1)
+           err |= control (file, "predicate control align1", pred_ctrl_align1,
+                           inst->header.predicate_control, NULL);
+       else
+           err |= control (file, "predicate control align16", pred_ctrl_align16,
+                           inst->header.predicate_control, NULL);
+       string (file, ") ");
+    }
+
+    err |= print_opcode (file, inst->header.opcode);
+    err |= control (file, "saturate", saturate, inst->header.saturate, NULL);
+    err |= control (file, "debug control", debug_ctrl, inst->header.debug_control, NULL);
+
+    if (inst->header.opcode != BRW_OPCODE_SEND)
+       err |= control (file, "conditional modifier", conditional_modifier,
+                       inst->header.destreg__conditionalmod, NULL);
+
+    if (inst->header.opcode != BRW_OPCODE_NOP) {
+       string (file, "(");
+       err |= control (file, "execution size", exec_size, inst->header.execution_size, NULL);
+       string (file, ")");
+    }
+
+    if (inst->header.opcode == BRW_OPCODE_SEND)
+       format (file, " %d", inst->header.destreg__conditionalmod);
+
+    if (opcode[inst->header.opcode].ndst > 0) {
+       pad (file, 16);
+       err |= dest (file, inst);
+    }
+    if (opcode[inst->header.opcode].nsrc > 0) {
+       pad (file, 32);
+       err |= src0 (file, inst);
+    }
+    if (opcode[inst->header.opcode].nsrc > 1) {
+       pad (file, 48);
+       err |= src1 (file, inst);
+    }
+
+    if (inst->header.opcode == BRW_OPCODE_SEND) {
+       newline (file);
+       pad (file, 16);
+       space = 0;
+       err |= control (file, "target function", target_function,
+                       inst->bits3.generic.msg_target, &space);
+       switch (inst->bits3.generic.msg_target) {
+       case BRW_MESSAGE_TARGET_MATH:
+           err |= control (file, "math function", math_function,
+                           inst->bits3.math.function, &space);
+           err |= control (file, "math saturate", math_saturate,
+                           inst->bits3.math.saturate, &space);
+           err |= control (file, "math signed", math_signed,
+                           inst->bits3.math.int_type, &space);
+           err |= control (file, "math scalar", math_scalar,
+                           inst->bits3.math.data_type, &space);
+           err |= control (file, "math precision", math_precision,
+                           inst->bits3.math.precision, &space);
+           break;
+       case BRW_MESSAGE_TARGET_SAMPLER:
+           format (file, " (%d, %d, ",
+                   inst->bits3.sampler.binding_table_index,
+                   inst->bits3.sampler.sampler);
+           err |= control (file, "sampler target format", sampler_target_format,
+                           inst->bits3.sampler.return_format, NULL);
+           string (file, ")");
+           break;
+       case BRW_MESSAGE_TARGET_DATAPORT_WRITE:
+           format (file, " (%d, %d, %d, %d)",
+                   inst->bits3.dp_write.binding_table_index,
+                   (inst->bits3.dp_write.pixel_scoreboard_clear << 3) |
+                   inst->bits3.dp_write.msg_control,
+                   inst->bits3.dp_write.msg_type,
+                   inst->bits3.dp_write.send_commit_msg);
+           break;
+       case BRW_MESSAGE_TARGET_URB:
+           format (file, " %d", inst->bits3.urb.offset);
+           space = 1;
+           err |= control (file, "urb swizzle", urb_swizzle,
+                           inst->bits3.urb.swizzle_control, &space);
+           err |= control (file, "urb allocate", urb_allocate,
+                           inst->bits3.urb.allocate, &space);
+           err |= control (file, "urb used", urb_used,
+                           inst->bits3.urb.used, &space);
+           err |= control (file, "urb complete", urb_complete,
+                           inst->bits3.urb.complete, &space);
+           break;
+       case BRW_MESSAGE_TARGET_THREAD_SPAWNER:
+           break;
+       default:
+           format (file, "unsupported target %d", inst->bits3.generic.msg_target);
+           break;
+       }
+       if (space)
+           string (file, " ");
+       format (file, "mlen %d",
+               inst->bits3.generic.msg_length);
+       format (file, " rlen %d",
+               inst->bits3.generic.response_length);
+    }
+    pad (file, 64);
+    if (inst->header.opcode != BRW_OPCODE_NOP) {
+       string (file, "{");
+       space = 1;
+       err |= control(file, "access mode", access_mode, inst->header.access_mode, &space);
+       err |= control (file, "mask control", mask_ctrl, inst->header.mask_control, &space);
+       err |= control (file, "dependency control", dep_ctrl, inst->header.dependency_control, &space);
+       err |= control (file, "compression control", compr_ctrl, inst->header.compression_control, &space);
+       err |= control (file, "thread control", thread_ctrl, inst->header.thread_control, &space);
+       if (inst->header.opcode == BRW_OPCODE_SEND)
+           err |= control (file, "end of thread", end_of_thread,
+                           inst->bits3.generic.end_of_thread, &space);
+       if (space)
+           string (file, " ");
+       string (file, "}");
+    }
+    string (file, ";");
+    newline (file);
+    return err;
+}
+
+
+int brw_disasm (FILE *file, 
+                const struct brw_instruction *inst,
+                unsigned count)
+{
+   int i, err;
+
+   for (i = 0; i < count; i++) {
+      err = brw_disasm_insn(stderr, &inst[i]);
+      if (err)
+         return err;
+   }
+
+   fprintf(file, "\n");
+   return 0;
+}
+
diff --git a/src/gallium/drivers/i965/brw_disasm.h b/src/gallium/drivers/i965/brw_disasm.h
new file mode 100644 (file)
index 0000000..77d402d
--- /dev/null
@@ -0,0 +1,34 @@
+/*
+ * Copyright Â© 2008 Keith Packard
+ *
+ * Permission to use, copy, modify, distribute, and sell this software and its
+ * documentation for any purpose is hereby granted without fee, provided that
+ * the above copyright notice appear in all copies and that both that copyright
+ * notice and this permission notice appear in supporting documentation, and
+ * that the name of the copyright holders not be used in advertising or
+ * publicity pertaining to distribution of the software without specific,
+ * written prior permission.  The copyright holders make no representations
+ * about the suitability of this software for any purpose.  It is provided "as
+ * is" without express or implied warranty.
+ *
+ * THE COPYRIGHT HOLDERS DISCLAIM ALL WARRANTIES WITH REGARD TO THIS SOFTWARE,
+ * INCLUDING ALL IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS, IN NO
+ * EVENT SHALL THE COPYRIGHT HOLDERS BE LIABLE FOR ANY SPECIAL, INDIRECT OR
+ * CONSEQUENTIAL DAMAGES OR ANY DAMAGES WHATSOEVER RESULTING FROM LOSS OF USE,
+ * DATA OR PROFITS, WHETHER IN AN ACTION OF CONTRACT, NEGLIGENCE OR OTHER
+ * TORTIOUS ACTION, ARISING OUT OF OR IN CONNECTION WITH THE USE OR PERFORMANCE
+ * OF THIS SOFTWARE.
+ */
+
+#ifndef BRW_DISASM_H
+#define BRW_DISASM_H
+
+struct brw_instruction;
+
+int brw_disasm_insn (FILE *file, const struct brw_instruction *inst);
+int brw_disasm (FILE *file, 
+                const struct brw_instruction *inst,
+                unsigned count);
+
+#endif
+
diff --git a/src/gallium/drivers/i965/brw_draw.c b/src/gallium/drivers/i965/brw_draw.c
new file mode 100644 (file)
index 0000000..852fd22
--- /dev/null
@@ -0,0 +1,291 @@
+/**************************************************************************
+ * 
+ * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+
+#include "util/u_prim.h"
+#include "util/u_upload_mgr.h"
+
+#include "brw_draw.h"
+#include "brw_defines.h"
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_debug.h"
+#include "brw_screen.h"
+
+#include "brw_batchbuffer.h"
+
+
+static uint32_t prim_to_hw_prim[PIPE_PRIM_POLYGON+1] = {
+   _3DPRIM_POINTLIST,
+   _3DPRIM_LINELIST,
+   _3DPRIM_LINELOOP,
+   _3DPRIM_LINESTRIP,
+   _3DPRIM_TRILIST,
+   _3DPRIM_TRISTRIP,
+   _3DPRIM_TRIFAN,
+   _3DPRIM_QUADLIST,
+   _3DPRIM_QUADSTRIP,
+   _3DPRIM_POLYGON
+};
+
+
+
+/* When the primitive changes, set a state bit and re-validate.  Not
+ * the nicest and would rather deal with this by having all the
+ * programs be immune to the active primitive (ie. cope with all
+ * possibilities).  That may not be realistic however.
+ */
+static int brw_set_prim(struct brw_context *brw, unsigned prim )
+{
+
+   if (BRW_DEBUG & DEBUG_PRIMS)
+      debug_printf("PRIM: %s\n", u_prim_name(prim));
+   
+   if (prim != brw->primitive) {
+      unsigned reduced_prim;
+
+      brw->primitive = prim;
+      brw->state.dirty.brw |= BRW_NEW_PRIMITIVE;
+
+      reduced_prim = u_reduced_prim(prim);
+      if (reduced_prim != brw->reduced_primitive) {
+        brw->reduced_primitive = reduced_prim;
+        brw->state.dirty.brw |= BRW_NEW_REDUCED_PRIMITIVE;
+      }
+   }
+
+   return prim_to_hw_prim[prim];
+}
+
+
+
+static int brw_emit_prim(struct brw_context *brw,
+                        unsigned start,
+                        unsigned count,
+                        boolean indexed,
+                        uint32_t hw_prim)
+{
+   struct brw_3d_primitive prim_packet;
+   int ret;
+
+   if (BRW_DEBUG & DEBUG_PRIMS)
+      debug_printf("%s start %d count %d indexed %d hw_prim %d\n",
+                   __FUNCTION__, start, count, indexed, hw_prim); 
+
+   prim_packet.header.opcode = CMD_3D_PRIM;
+   prim_packet.header.length = sizeof(prim_packet)/4 - 2;
+   prim_packet.header.pad = 0;
+   prim_packet.header.topology = hw_prim;
+   prim_packet.header.indexed = indexed;
+
+   prim_packet.verts_per_instance = count;
+   prim_packet.start_vert_location = start;
+   if (indexed)
+      prim_packet.start_vert_location += brw->ib.start_vertex_offset;
+   prim_packet.instance_count = 1;
+   prim_packet.start_instance_location = 0;
+   prim_packet.base_vert_location = 0; /* prim->basevertex; XXX: add this to gallium */
+
+
+   /* If we're set to always flush, do it before and after the primitive emit.
+    * We want to catch both missed flushes that hurt instruction/state cache
+    * and missed flushes of the render cache as it heads to other parts of
+    * the besides the draw code.
+    */
+   if (0) {
+      BEGIN_BATCH(1, IGNORE_CLIPRECTS);
+      OUT_BATCH((CMD_MI_FLUSH << 16) | BRW_FLUSH_STATE_CACHE);
+      ADVANCE_BATCH();
+   }
+   if (prim_packet.verts_per_instance) {
+      ret = brw_batchbuffer_data( brw->batch, &prim_packet,
+                                 sizeof(prim_packet), LOOP_CLIPRECTS);
+      if (ret)
+        return ret;
+   }
+   if (0) {
+      BEGIN_BATCH(1, IGNORE_CLIPRECTS);
+      OUT_BATCH((CMD_MI_FLUSH << 16) | BRW_FLUSH_STATE_CACHE);
+      ADVANCE_BATCH();
+   }
+
+   return 0;
+}
+
+
+/* May fail if out of video memory for texture or vbo upload, or on
+ * fallback conditions.
+ */
+static int
+try_draw_range_elements(struct brw_context *brw,
+                       struct pipe_buffer *index_buffer,
+                       unsigned hw_prim, 
+                       unsigned start, unsigned count)
+{
+   int ret;
+
+   ret = brw_validate_state(brw);
+   if (ret)
+      return ret;
+
+   /* Check that we can fit our state in with our existing batchbuffer, or
+    * flush otherwise.
+    */
+   ret = brw->sws->check_aperture_space(brw->sws,
+                                       brw->state.validated_bos,
+                                       brw->state.validated_bo_count);
+   if (ret)
+      return ret;
+
+   ret = brw_upload_state(brw);
+   if (ret)
+      return ret;
+   
+   ret = brw_emit_prim(brw, start, count, index_buffer != NULL, hw_prim);
+   if (ret)
+      return ret;
+
+   if (brw->flags.always_flush_batch)
+      brw_context_flush( brw );
+
+   return 0;
+}
+
+
+static boolean
+brw_draw_range_elements(struct pipe_context *pipe,
+                       struct pipe_buffer *index_buffer,
+                       unsigned index_size,
+                       unsigned min_index,
+                       unsigned max_index,
+                       unsigned mode, unsigned start, unsigned count)
+{
+   struct brw_context *brw = brw_context(pipe);
+   int ret;
+   uint32_t hw_prim;
+
+   hw_prim = brw_set_prim(brw, mode);
+
+   if (BRW_DEBUG & DEBUG_PRIMS)
+      debug_printf("PRIM: %s start %d count %d index_buffer %p\n",
+                   u_prim_name(mode), start, count, (void *)index_buffer);
+
+   /* Potentially trigger upload of new index buffer.
+    *
+    * XXX: do we need to go through state validation to achieve this?
+    * Could just call upload code directly.
+    */
+   if (brw->curr.index_buffer != index_buffer ||
+       brw->curr.index_size != index_size) {
+      pipe_buffer_reference( &brw->curr.index_buffer, index_buffer );
+      brw->curr.index_size = index_size;
+      brw->state.dirty.mesa |= PIPE_NEW_INDEX_BUFFER;
+   }
+
+   /* XXX: do we really care?
+    */
+   if (brw->curr.min_index != min_index ||
+       brw->curr.max_index != max_index) 
+   { 
+      brw->curr.min_index = min_index;
+      brw->curr.max_index = max_index;
+      brw->state.dirty.mesa |= PIPE_NEW_INDEX_RANGE;
+   }
+
+
+   /* Make a first attempt at drawing:
+    */
+   ret = try_draw_range_elements(brw, index_buffer, hw_prim, start, count );
+
+   /* Otherwise, flush and retry:
+    */
+   if (ret != 0) {
+      brw_context_flush( brw );
+      ret = try_draw_range_elements(brw, index_buffer, hw_prim, start, count );
+      assert(ret == 0);
+   }
+
+   return TRUE;
+}
+
+static boolean
+brw_draw_elements(struct pipe_context *pipe,
+                 struct pipe_buffer *index_buffer,
+                 unsigned index_size,
+                 unsigned mode, 
+                 unsigned start, unsigned count)
+{
+   return brw_draw_range_elements( pipe, index_buffer,
+                                  index_size,
+                                  0, 0xffffffff,
+                                  mode, 
+                                  start, count );
+}
+
+static boolean
+brw_draw_arrays(struct pipe_context *pipe, unsigned mode,
+                     unsigned start, unsigned count)
+{
+   return brw_draw_elements(pipe, NULL, 0, mode, start, count);
+}
+
+
+
+boolean brw_draw_init( struct brw_context *brw )
+{
+   /* Register our drawing function: 
+    */
+   brw->base.draw_arrays = brw_draw_arrays;
+   brw->base.draw_elements = brw_draw_elements;
+   brw->base.draw_range_elements = brw_draw_range_elements;
+
+   /* Create helpers for uploading data in user buffers:
+    */
+   brw->vb.upload_vertex = u_upload_create( brw->base.screen,
+                                           128 * 1024,
+                                           64,
+                                           PIPE_BUFFER_USAGE_VERTEX );
+   if (brw->vb.upload_vertex == NULL)
+      return FALSE;
+
+   brw->vb.upload_index = u_upload_create( brw->base.screen,
+                                          32 * 1024,
+                                          64,
+                                          PIPE_BUFFER_USAGE_INDEX );
+   if (brw->vb.upload_index == NULL)
+      return FALSE;
+
+   return TRUE;
+}
+
+void brw_draw_cleanup( struct brw_context *brw )
+{
+   u_upload_destroy( brw->vb.upload_vertex );
+   u_upload_destroy( brw->vb.upload_index );
+
+   bo_reference(&brw->ib.bo, NULL);
+}
diff --git a/src/gallium/drivers/i965/brw_draw.h b/src/gallium/drivers/i965/brw_draw.h
new file mode 100644 (file)
index 0000000..8dc5dbc
--- /dev/null
@@ -0,0 +1,39 @@
+ /**************************************************************************
+ * 
+ * Copyright 2005 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef BRW_DRAW_H
+#define BRW_DRAW_H
+
+#include "brw_types.h"
+
+struct brw_context;
+
+boolean brw_draw_init( struct brw_context *brw );
+void brw_draw_cleanup( struct brw_context *brw );
+
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_draw_upload.c b/src/gallium/drivers/i965/brw_draw_upload.c
new file mode 100644 (file)
index 0000000..a27da5f
--- /dev/null
@@ -0,0 +1,542 @@
+/**************************************************************************
+ * 
+ * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include "pipe/p_context.h"
+
+#include "util/u_upload_mgr.h"
+#include "util/u_math.h"
+
+#include "brw_draw.h"
+#include "brw_defines.h"
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_screen.h"
+#include "brw_batchbuffer.h"
+#include "brw_debug.h"
+
+
+
+
+static unsigned brw_translate_surface_format( unsigned id )
+{
+   switch (id) {
+   case PIPE_FORMAT_R64_FLOAT:
+      return BRW_SURFACEFORMAT_R64_FLOAT;
+   case PIPE_FORMAT_R64G64_FLOAT:
+      return BRW_SURFACEFORMAT_R64G64_FLOAT;
+   case PIPE_FORMAT_R64G64B64_FLOAT:
+      return BRW_SURFACEFORMAT_R64G64B64_FLOAT;
+   case PIPE_FORMAT_R64G64B64A64_FLOAT:
+      return BRW_SURFACEFORMAT_R64G64B64A64_FLOAT;
+
+   case PIPE_FORMAT_R32_FLOAT:
+      return BRW_SURFACEFORMAT_R32_FLOAT;
+   case PIPE_FORMAT_R32G32_FLOAT:
+      return BRW_SURFACEFORMAT_R32G32_FLOAT;
+   case PIPE_FORMAT_R32G32B32_FLOAT:
+      return BRW_SURFACEFORMAT_R32G32B32_FLOAT;
+   case PIPE_FORMAT_R32G32B32A32_FLOAT:
+      return BRW_SURFACEFORMAT_R32G32B32A32_FLOAT;
+
+   case PIPE_FORMAT_R32_UNORM:
+      return BRW_SURFACEFORMAT_R32_UNORM;
+   case PIPE_FORMAT_R32G32_UNORM:
+      return BRW_SURFACEFORMAT_R32G32_UNORM;
+   case PIPE_FORMAT_R32G32B32_UNORM:
+      return BRW_SURFACEFORMAT_R32G32B32_UNORM;
+   case PIPE_FORMAT_R32G32B32A32_UNORM:
+      return BRW_SURFACEFORMAT_R32G32B32A32_UNORM;
+
+   case PIPE_FORMAT_R32_USCALED:
+      return BRW_SURFACEFORMAT_R32_USCALED;
+   case PIPE_FORMAT_R32G32_USCALED:
+      return BRW_SURFACEFORMAT_R32G32_USCALED;
+   case PIPE_FORMAT_R32G32B32_USCALED:
+      return BRW_SURFACEFORMAT_R32G32B32_USCALED;
+   case PIPE_FORMAT_R32G32B32A32_USCALED:
+      return BRW_SURFACEFORMAT_R32G32B32A32_USCALED;
+
+   case PIPE_FORMAT_R32_SNORM:
+      return BRW_SURFACEFORMAT_R32_SNORM;
+   case PIPE_FORMAT_R32G32_SNORM:
+      return BRW_SURFACEFORMAT_R32G32_SNORM;
+   case PIPE_FORMAT_R32G32B32_SNORM:
+      return BRW_SURFACEFORMAT_R32G32B32_SNORM;
+   case PIPE_FORMAT_R32G32B32A32_SNORM:
+      return BRW_SURFACEFORMAT_R32G32B32A32_SNORM;
+
+   case PIPE_FORMAT_R32_SSCALED:
+      return BRW_SURFACEFORMAT_R32_SSCALED;
+   case PIPE_FORMAT_R32G32_SSCALED:
+      return BRW_SURFACEFORMAT_R32G32_SSCALED;
+   case PIPE_FORMAT_R32G32B32_SSCALED:
+      return BRW_SURFACEFORMAT_R32G32B32_SSCALED;
+   case PIPE_FORMAT_R32G32B32A32_SSCALED:
+      return BRW_SURFACEFORMAT_R32G32B32A32_SSCALED;
+
+   case PIPE_FORMAT_R16_UNORM:
+      return BRW_SURFACEFORMAT_R16_UNORM;
+   case PIPE_FORMAT_R16G16_UNORM:
+      return BRW_SURFACEFORMAT_R16G16_UNORM;
+   case PIPE_FORMAT_R16G16B16_UNORM:
+      return BRW_SURFACEFORMAT_R16G16B16_UNORM;
+   case PIPE_FORMAT_R16G16B16A16_UNORM:
+      return BRW_SURFACEFORMAT_R16G16B16A16_UNORM;
+
+   case PIPE_FORMAT_R16_USCALED:
+      return BRW_SURFACEFORMAT_R16_USCALED;
+   case PIPE_FORMAT_R16G16_USCALED:
+      return BRW_SURFACEFORMAT_R16G16_USCALED;
+   case PIPE_FORMAT_R16G16B16_USCALED:
+      return BRW_SURFACEFORMAT_R16G16B16_USCALED;
+   case PIPE_FORMAT_R16G16B16A16_USCALED:
+      return BRW_SURFACEFORMAT_R16G16B16A16_USCALED;
+
+   case PIPE_FORMAT_R16_SNORM:
+      return BRW_SURFACEFORMAT_R16_SNORM;
+   case PIPE_FORMAT_R16G16_SNORM:
+      return BRW_SURFACEFORMAT_R16G16_SNORM;
+   case PIPE_FORMAT_R16G16B16_SNORM:
+      return BRW_SURFACEFORMAT_R16G16B16_SNORM;
+   case PIPE_FORMAT_R16G16B16A16_SNORM:
+      return BRW_SURFACEFORMAT_R16G16B16A16_SNORM;
+
+   case PIPE_FORMAT_R16_SSCALED:
+      return BRW_SURFACEFORMAT_R16_SSCALED;
+   case PIPE_FORMAT_R16G16_SSCALED:
+      return BRW_SURFACEFORMAT_R16G16_SSCALED;
+   case PIPE_FORMAT_R16G16B16_SSCALED:
+      return BRW_SURFACEFORMAT_R16G16B16_SSCALED;
+   case PIPE_FORMAT_R16G16B16A16_SSCALED:
+      return BRW_SURFACEFORMAT_R16G16B16A16_SSCALED;
+
+   case PIPE_FORMAT_R8_UNORM:
+      return BRW_SURFACEFORMAT_R8_UNORM;
+   case PIPE_FORMAT_R8G8_UNORM:
+      return BRW_SURFACEFORMAT_R8G8_UNORM;
+   case PIPE_FORMAT_R8G8B8_UNORM:
+      return BRW_SURFACEFORMAT_R8G8B8_UNORM;
+   case PIPE_FORMAT_R8G8B8A8_UNORM:
+      return BRW_SURFACEFORMAT_R8G8B8A8_UNORM;
+
+   case PIPE_FORMAT_R8_USCALED:
+      return BRW_SURFACEFORMAT_R8_USCALED;
+   case PIPE_FORMAT_R8G8_USCALED:
+      return BRW_SURFACEFORMAT_R8G8_USCALED;
+   case PIPE_FORMAT_R8G8B8_USCALED:
+      return BRW_SURFACEFORMAT_R8G8B8_USCALED;
+   case PIPE_FORMAT_R8G8B8A8_USCALED:
+      return BRW_SURFACEFORMAT_R8G8B8A8_USCALED;
+
+   case PIPE_FORMAT_R8_SNORM:
+      return BRW_SURFACEFORMAT_R8_SNORM;
+   case PIPE_FORMAT_R8G8_SNORM:
+      return BRW_SURFACEFORMAT_R8G8_SNORM;
+   case PIPE_FORMAT_R8G8B8_SNORM:
+      return BRW_SURFACEFORMAT_R8G8B8_SNORM;
+   case PIPE_FORMAT_R8G8B8A8_SNORM:
+      return BRW_SURFACEFORMAT_R8G8B8A8_SNORM;
+
+   case PIPE_FORMAT_R8_SSCALED:
+      return BRW_SURFACEFORMAT_R8_SSCALED;
+   case PIPE_FORMAT_R8G8_SSCALED:
+      return BRW_SURFACEFORMAT_R8G8_SSCALED;
+   case PIPE_FORMAT_R8G8B8_SSCALED:
+      return BRW_SURFACEFORMAT_R8G8B8_SSCALED;
+   case PIPE_FORMAT_R8G8B8A8_SSCALED:
+      return BRW_SURFACEFORMAT_R8G8B8A8_SSCALED;
+
+   default:
+      assert(0);
+      return 0;
+   }
+}
+
+static unsigned get_index_type(int type)
+{
+   switch (type) {
+   case 1: return BRW_INDEX_BYTE;
+   case 2: return BRW_INDEX_WORD;
+   case 4: return BRW_INDEX_DWORD;
+   default: assert(0); return 0;
+   }
+}
+
+
+static int brw_prepare_vertices(struct brw_context *brw)
+{
+   unsigned int min_index = brw->curr.min_index;
+   unsigned int max_index = brw->curr.max_index;
+   GLuint i;
+   int ret;
+
+   if (BRW_DEBUG & DEBUG_VERTS)
+      debug_printf("%s %d..%d\n", __FUNCTION__, min_index, max_index);
+
+
+   for (i = 0; i < brw->curr.num_vertex_buffers; i++) {
+      struct pipe_vertex_buffer *vb = &brw->curr.vertex_buffer[i];
+      struct brw_winsys_buffer *bo;
+      struct pipe_buffer *upload_buf = NULL;
+      unsigned offset;
+      
+      if (BRW_DEBUG & DEBUG_VERTS)
+        debug_printf("%s vb[%d] user:%d offset:0x%x sz:0x%x stride:0x%x\n",
+                     __FUNCTION__, i,
+                     brw_buffer_is_user_buffer(vb->buffer),
+                     vb->buffer_offset,
+                     vb->buffer->size,
+                     vb->stride);
+
+      if (brw_buffer_is_user_buffer(vb->buffer)) {
+
+        /* XXX: simplify this.  Stop the state trackers from generating
+         * zero-stride buffers & have them use additional constants (or
+         * add support for >1 constant buffer) instead.
+         */
+        unsigned size = (vb->stride == 0 ? 
+                         vb->buffer->size - vb->buffer_offset :
+                         MAX2(vb->buffer->size - vb->buffer_offset,
+                              vb->stride * (max_index + 1 - min_index)));
+
+        ret = u_upload_buffer( brw->vb.upload_vertex, 
+                               vb->buffer_offset + min_index * vb->stride,
+                               size,
+                               vb->buffer,
+                               &offset,
+                               &upload_buf );
+        if (ret)
+           return ret;
+
+        bo = brw_buffer(upload_buf)->bo;
+        
+        assert(offset + size <= bo->size);
+      }
+      else
+      {
+        offset = vb->buffer_offset;
+        bo = brw_buffer(vb->buffer)->bo;
+      }
+
+      assert(offset < bo->size);
+      
+      /* Set up post-upload info about this vertex buffer:
+       */
+      brw->vb.vb[i].offset = offset;
+      brw->vb.vb[i].stride = vb->stride;
+      brw->vb.vb[i].vertex_count = (vb->stride == 0 ?
+                                   1 :
+                                   (bo->size - offset) / vb->stride);
+
+      bo_reference( &brw->vb.vb[i].bo,  bo );
+
+      /* Don't need to retain this reference.  We have a reference on
+       * the underlying winsys buffer:
+       */
+      pipe_buffer_reference( &upload_buf, NULL );
+   }
+
+   brw->vb.nr_vb = i;
+   brw_prepare_query_begin(brw);
+
+   for (i = 0; i < brw->vb.nr_vb; i++) {
+      brw_add_validated_bo(brw, brw->vb.vb[i].bo);
+   }
+
+   return 0;
+}
+
+static int brw_emit_vertex_buffers( struct brw_context *brw )
+{
+   int i;
+
+   /* If the VS doesn't read any inputs (calculating vertex position from
+    * a state variable for some reason, for example), just bail.
+    *
+    * The stale VB state stays in place, but they don't do anything unless
+    * a VE loads from them.
+    */
+   if (brw->vb.nr_vb == 0) {
+      if (BRW_DEBUG & DEBUG_VERTS)
+        debug_printf("%s: no active vertex buffers\n", __FUNCTION__);
+
+      return 0;
+   }
+
+   /* Emit VB state packets.
+    */
+   BEGIN_BATCH(1 + brw->vb.nr_vb * 4, IGNORE_CLIPRECTS);
+   OUT_BATCH((CMD_VERTEX_BUFFER << 16) |
+            ((1 + brw->vb.nr_vb * 4) - 2));
+
+   for (i = 0; i < brw->vb.nr_vb; i++) {
+      OUT_BATCH((i << BRW_VB0_INDEX_SHIFT) |
+               BRW_VB0_ACCESS_VERTEXDATA |
+               (brw->vb.vb[i].stride << BRW_VB0_PITCH_SHIFT));
+      OUT_RELOC(brw->vb.vb[i].bo,
+               BRW_USAGE_VERTEX,
+               brw->vb.vb[i].offset);
+      if (BRW_IS_IGDNG(brw)) {
+        OUT_RELOC(brw->vb.vb[i].bo,
+                  BRW_USAGE_VERTEX,
+                  brw->vb.vb[i].bo->size - 1);
+      } else
+        OUT_BATCH(brw->vb.vb[i].stride ? brw->vb.vb[i].vertex_count : 0);
+      OUT_BATCH(0); /* Instance data step rate */
+   }
+   ADVANCE_BATCH();
+   return 0;
+}
+
+
+
+
+static int brw_emit_vertex_elements(struct brw_context *brw)
+{
+   GLuint nr = brw->curr.num_vertex_elements;
+   GLuint i;
+
+   brw_emit_query_begin(brw);
+
+   /* If the VS doesn't read any inputs (calculating vertex position from
+    * a state variable for some reason, for example), emit a single pad
+    * VERTEX_ELEMENT struct and bail.
+    *
+    * The stale VB state stays in place, but they don't do anything unless
+    * a VE loads from them.
+    */
+   if (nr == 0) {
+      BEGIN_BATCH(3, IGNORE_CLIPRECTS);
+      OUT_BATCH((CMD_VERTEX_ELEMENT << 16) | 1);
+      OUT_BATCH((0 << BRW_VE0_INDEX_SHIFT) |
+               BRW_VE0_VALID |
+               (BRW_SURFACEFORMAT_R32G32B32A32_FLOAT << BRW_VE0_FORMAT_SHIFT) |
+               (0 << BRW_VE0_SRC_OFFSET_SHIFT));
+      OUT_BATCH((BRW_VE1_COMPONENT_STORE_0 << BRW_VE1_COMPONENT_0_SHIFT) |
+               (BRW_VE1_COMPONENT_STORE_0 << BRW_VE1_COMPONENT_1_SHIFT) |
+               (BRW_VE1_COMPONENT_STORE_0 << BRW_VE1_COMPONENT_2_SHIFT) |
+               (BRW_VE1_COMPONENT_STORE_1_FLT << BRW_VE1_COMPONENT_3_SHIFT));
+      ADVANCE_BATCH();
+      return 0;
+   }
+
+   /* Now emit vertex element (VEP) state packets.
+    *
+    */
+   BEGIN_BATCH(1 + nr * 2, IGNORE_CLIPRECTS);
+   OUT_BATCH((CMD_VERTEX_ELEMENT << 16) | ((1 + nr * 2) - 2));
+   for (i = 0; i < nr; i++) {
+      const struct pipe_vertex_element *input = &brw->curr.vertex_element[i];
+      uint32_t format = brw_translate_surface_format( input->src_format );
+      uint32_t comp0 = BRW_VE1_COMPONENT_STORE_SRC;
+      uint32_t comp1 = BRW_VE1_COMPONENT_STORE_SRC;
+      uint32_t comp2 = BRW_VE1_COMPONENT_STORE_SRC;
+      uint32_t comp3 = BRW_VE1_COMPONENT_STORE_SRC;
+
+      switch (input->nr_components) {
+      case 0: comp0 = BRW_VE1_COMPONENT_STORE_0;
+      case 1: comp1 = BRW_VE1_COMPONENT_STORE_0;
+      case 2: comp2 = BRW_VE1_COMPONENT_STORE_0;
+      case 3: comp3 = BRW_VE1_COMPONENT_STORE_1_FLT;
+        break;
+      }
+
+      OUT_BATCH((input->vertex_buffer_index << BRW_VE0_INDEX_SHIFT) |
+               BRW_VE0_VALID |
+               (format << BRW_VE0_FORMAT_SHIFT) |
+               (input->src_offset << BRW_VE0_SRC_OFFSET_SHIFT));
+
+      if (BRW_IS_IGDNG(brw))
+          OUT_BATCH((comp0 << BRW_VE1_COMPONENT_0_SHIFT) |
+                    (comp1 << BRW_VE1_COMPONENT_1_SHIFT) |
+                    (comp2 << BRW_VE1_COMPONENT_2_SHIFT) |
+                    (comp3 << BRW_VE1_COMPONENT_3_SHIFT));
+      else
+          OUT_BATCH((comp0 << BRW_VE1_COMPONENT_0_SHIFT) |
+                    (comp1 << BRW_VE1_COMPONENT_1_SHIFT) |
+                    (comp2 << BRW_VE1_COMPONENT_2_SHIFT) |
+                    (comp3 << BRW_VE1_COMPONENT_3_SHIFT) |
+                    ((i * 4) << BRW_VE1_DST_OFFSET_SHIFT));
+   }
+   ADVANCE_BATCH();
+   return 0;
+}
+
+
+static int brw_emit_vertices( struct brw_context *brw )
+{
+   int ret;
+
+   ret = brw_emit_vertex_buffers( brw );
+   if (ret)
+      return ret;
+
+   ret = brw_emit_vertex_elements( brw );
+   if (ret)
+      return ret;
+   
+   return 0;
+}
+
+
+const struct brw_tracked_state brw_vertices = {
+   .dirty = {
+      .mesa = (PIPE_NEW_INDEX_RANGE |
+               PIPE_NEW_VERTEX_BUFFER),
+      .brw = BRW_NEW_BATCH,
+      .cache = 0,
+   },
+   .prepare = brw_prepare_vertices,
+   .emit = brw_emit_vertices,
+};
+
+
+static int brw_prepare_indices(struct brw_context *brw)
+{
+   struct pipe_buffer *index_buffer = brw->curr.index_buffer;
+   struct pipe_buffer *upload_buf = NULL;
+   struct brw_winsys_buffer *bo = NULL;
+   GLuint offset;
+   GLuint index_size;
+   GLuint ib_size;
+   int ret;
+
+   if (index_buffer == NULL)
+      return 0;
+
+   if (BRW_DEBUG & DEBUG_VERTS)
+      debug_printf("%s: index_size:%d index_buffer->size:%d\n",
+                  __FUNCTION__,
+                  brw->curr.index_size,
+                  brw->curr.index_buffer->size);
+
+   ib_size = index_buffer->size;
+   index_size = brw->curr.index_size;
+
+   /* Turn userbuffer into a proper hardware buffer?
+    */
+   if (brw_buffer_is_user_buffer(index_buffer)) {
+
+      ret = u_upload_buffer( brw->vb.upload_index,
+                            0,
+                            ib_size,
+                            index_buffer,
+                            &offset,
+                            &upload_buf );
+      if (ret)
+        return ret;
+
+      bo = brw_buffer(upload_buf)->bo;
+
+      /* XXX: annotate the userbuffer with the upload information so
+       * that successive calls don't get re-uploaded.
+       */
+   }
+   else {
+      bo = brw_buffer(index_buffer)->bo;
+      ib_size = bo->size;
+      offset = 0;
+   }
+
+   /* Use CMD_3D_PRIM's start_vertex_offset to avoid re-uploading the
+    * index buffer state when we're just moving the start index of our
+    * drawing.
+    *
+    * In gallium this will happen in the case where successive draw
+    * calls are made with (distinct?) userbuffers, but the upload_mgr
+    * places the data into a single winsys buffer.
+    * 
+    * This statechange doesn't raise any state flags and is always
+    * just merged into the final draw packet:
+    */
+   if (1) {
+      assert((offset & (index_size - 1)) == 0);
+      brw->ib.start_vertex_offset = offset / index_size;
+   }
+
+   /* These statechanges trigger a new CMD_INDEX_BUFFER packet:
+    */
+   if (brw->ib.bo != bo ||
+       brw->ib.size != ib_size)
+   {
+      bo_reference(&brw->ib.bo, bo);
+      brw->ib.size = ib_size;
+      brw->state.dirty.brw |= BRW_NEW_INDEX_BUFFER;
+   }
+
+   pipe_buffer_reference( &upload_buf, NULL );
+   brw_add_validated_bo(brw, brw->ib.bo);
+   return 0;
+}
+
+const struct brw_tracked_state brw_indices = {
+   .dirty = {
+      .mesa = PIPE_NEW_INDEX_BUFFER,
+      .brw = 0,
+      .cache = 0,
+   },
+   .prepare = brw_prepare_indices,
+};
+
+static int brw_emit_index_buffer(struct brw_context *brw)
+{
+   /* Emit the indexbuffer packet:
+    */
+   if (brw->ib.bo)
+   {
+      struct brw_indexbuffer ib;
+
+      memset(&ib, 0, sizeof(ib));
+
+      ib.header.bits.opcode = CMD_INDEX_BUFFER;
+      ib.header.bits.length = sizeof(ib)/4 - 2;
+      ib.header.bits.index_format = get_index_type(brw->ib.size);
+      ib.header.bits.cut_index_enable = 0;
+
+      BEGIN_BATCH(4, IGNORE_CLIPRECTS);
+      OUT_BATCH( ib.header.dword );
+      OUT_RELOC(brw->ib.bo,
+               BRW_USAGE_VERTEX,
+               brw->ib.offset);
+      OUT_RELOC(brw->ib.bo,
+               BRW_USAGE_VERTEX,
+               brw->ib.offset + brw->ib.size - 1);
+      OUT_BATCH( 0 );
+      ADVANCE_BATCH();
+   }
+
+   return 0;
+}
+
+const struct brw_tracked_state brw_index_buffer = {
+   .dirty = {
+      .mesa = 0,
+      .brw = BRW_NEW_BATCH | BRW_NEW_INDEX_BUFFER,
+      .cache = 0,
+   },
+   .emit = brw_emit_index_buffer,
+};
diff --git a/src/gallium/drivers/i965/brw_eu.c b/src/gallium/drivers/i965/brw_eu.c
new file mode 100644 (file)
index 0000000..a8fcb5f
--- /dev/null
@@ -0,0 +1,262 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+  
+#include "util/u_memory.h"
+
+#include "brw_context.h"
+#include "brw_defines.h"
+#include "brw_eu.h"
+
+
+
+/* How does predicate control work when execution_size != 8?  Do I
+ * need to test/set for 0xffff when execution_size is 16?
+ */
+void brw_set_predicate_control_flag_value( struct brw_compile *p, GLuint value )
+{
+   p->current->header.predicate_control = BRW_PREDICATE_NONE;
+
+   if (value != 0xff) {
+      if (value != p->flag_value) {
+        brw_push_insn_state(p);
+        brw_MOV(p, brw_flag_reg(), brw_imm_uw(value));
+        p->flag_value = value;
+        brw_pop_insn_state(p);
+      }
+
+      p->current->header.predicate_control = BRW_PREDICATE_NORMAL;
+   }   
+}
+
+void brw_set_predicate_control( struct brw_compile *p, GLuint pc )
+{
+   p->current->header.predicate_control = pc;
+}
+
+void brw_set_conditionalmod( struct brw_compile *p, GLuint conditional )
+{
+   p->current->header.destreg__conditionalmod = conditional;
+}
+
+void brw_set_access_mode( struct brw_compile *p, GLuint access_mode )
+{
+   p->current->header.access_mode = access_mode;
+}
+
+void brw_set_compression_control( struct brw_compile *p, GLboolean compression_control )
+{
+   p->current->header.compression_control = compression_control;
+}
+
+void brw_set_mask_control( struct brw_compile *p, GLuint value )
+{
+   p->current->header.mask_control = value;
+}
+
+void brw_set_saturate( struct brw_compile *p, GLuint value )
+{
+   p->current->header.saturate = value;
+}
+
+void brw_push_insn_state( struct brw_compile *p )
+{
+   assert(p->current != &p->stack[BRW_EU_MAX_INSN_STACK-1]);
+   memcpy(p->current+1, p->current, sizeof(struct brw_instruction));
+   p->current++;   
+}
+
+void brw_pop_insn_state( struct brw_compile *p )
+{
+   assert(p->current != p->stack);
+   p->current--;
+}
+
+
+/***********************************************************************
+ */
+void brw_init_compile( struct brw_context *brw, struct brw_compile *p )
+{
+   p->brw = brw;
+   p->nr_insn = 0;
+   p->current = p->stack;
+   memset(p->current, 0, sizeof(p->current[0]));
+
+   /* Some defaults?
+    */
+   brw_set_mask_control(p, BRW_MASK_ENABLE); /* what does this do? */
+   brw_set_saturate(p, 0);
+   brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+   brw_set_predicate_control_flag_value(p, 0xff); 
+}
+
+
+enum pipe_error brw_get_program( struct brw_compile *p,
+                                 const GLuint **data,
+                                 GLuint *sz )
+{
+   GLuint i;
+
+   for (i = 0; i < 8; i++)
+      brw_NOP(p);
+
+   /* Is the generated program malformed for some reason?
+    */
+   if (p->error)
+      return PIPE_ERROR_BAD_INPUT;
+
+   *sz = p->nr_insn * sizeof(struct brw_instruction);
+   *data = (const GLuint *)p->store;
+   return PIPE_OK;
+}
+
+
+
+/**
+ * Subroutine calls require special attention.
+ * Mesa instructions may be expanded into multiple hardware instructions
+ * so the prog_instruction::BranchTarget field can't be used as an index
+ * into the hardware instructions.
+ *
+ * The BranchTarget field isn't needed, however.  Mesa's GLSL compiler
+ * emits CAL and BGNSUB instructions with labels that can be used to map
+ * subroutine calls to actual subroutine code blocks.
+ *
+ * The structures and function here implement patching of CAL instructions
+ * so they jump to the right subroutine code...
+ */
+
+
+/**
+ * For each OPCODE_BGNSUB we create one of these.
+ */
+struct brw_eu_label
+{
+   GLuint label;     /**< the label number */
+   GLuint position;  /**< the position of the brw instruction for this label */
+   struct brw_eu_label *next;  /**< next in linked list */
+};
+
+
+/**
+ * For each OPCODE_CAL we create one of these.
+ */
+struct brw_eu_call
+{
+   GLuint call_inst_pos;  /**< location of the CAL instruction */
+   GLuint label;
+   struct brw_eu_call *next;  /**< next in linked list */
+};
+
+
+/**
+ * Called for each OPCODE_BGNSUB.
+ */
+void
+brw_save_label(struct brw_compile *c, unsigned l, GLuint position)
+{
+   struct brw_eu_label *label = CALLOC_STRUCT(brw_eu_label);
+   label->label = l;
+   label->position = position;
+   label->next = c->first_label;
+   c->first_label = label;
+}
+
+
+/**
+ * Called for each OPCODE_CAL.
+ */
+void
+brw_save_call(struct brw_compile *c, GLuint label, GLuint call_pos)
+{
+   struct brw_eu_call *call = CALLOC_STRUCT(brw_eu_call);
+   call->call_inst_pos = call_pos;
+   call->label = label;
+   call->next = c->first_call;
+   c->first_call = call;
+}
+
+
+/**
+ * Lookup a label, return label's position/offset.
+ */
+static GLuint
+brw_lookup_label(struct brw_compile *c, unsigned l)
+{
+   const struct brw_eu_label *label;
+   for (label = c->first_label; label; label = label->next) {
+      if (l == label->label) {
+         return label->position;
+      }
+   }
+   abort();  /* should never happen */
+   return ~0;
+}
+
+
+/**
+ * When we're done generating code, this function is called to resolve
+ * subroutine calls.
+ */
+void
+brw_resolve_cals(struct brw_compile *c)
+{
+    const struct brw_eu_call *call;
+
+    for (call = c->first_call; call; call = call->next) {
+        const GLuint sub_loc = brw_lookup_label(c, call->label);
+       struct brw_instruction *brw_call_inst = &c->store[call->call_inst_pos];
+       struct brw_instruction *brw_sub_inst = &c->store[sub_loc];
+       GLint offset = brw_sub_inst - brw_call_inst;
+
+       /* patch brw_inst1 to point to brw_inst2 */
+       brw_set_src1(brw_call_inst, brw_imm_d(offset * 16));
+    }
+
+    /* free linked list of calls */
+    {
+        struct brw_eu_call *call, *next;
+        for (call = c->first_call; call; call = next) {
+           next = call->next;
+           FREE(call);
+       }
+       c->first_call = NULL;
+    }
+
+    /* free linked list of labels */
+    {
+        struct brw_eu_label *label, *next;
+       for (label = c->first_label; label; label = next) {
+           next = label->next;
+           FREE(label);
+       }
+       c->first_label = NULL;
+    }
+}
diff --git a/src/gallium/drivers/i965/brw_eu.h b/src/gallium/drivers/i965/brw_eu.h
new file mode 100644 (file)
index 0000000..af509b2
--- /dev/null
@@ -0,0 +1,992 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+   
+
+#ifndef BRW_EU_H
+#define BRW_EU_H
+
+#include "util/u_debug.h"
+#include "pipe/p_defines.h"
+
+#include "brw_structs.h"
+#include "brw_defines.h"
+
+#define BRW_SWIZZLE4(a,b,c,d) (((a)<<0) | ((b)<<2) | ((c)<<4) | ((d)<<6))
+#define BRW_GET_SWZ(swz, idx) (((swz) >> ((idx)*2)) & 0x3)
+
+#define BRW_SWIZZLE_NOOP      BRW_SWIZZLE4(0,1,2,3)
+#define BRW_SWIZZLE_XYZW      BRW_SWIZZLE4(0,1,2,3)
+#define BRW_SWIZZLE_XXXX      BRW_SWIZZLE4(0,0,0,0)
+#define BRW_SWIZZLE_XYXY      BRW_SWIZZLE4(0,1,0,1)
+
+#define BRW_WRITEMASK_NONE     0x00
+#define BRW_WRITEMASK_X        0x01
+#define BRW_WRITEMASK_Y        0x02
+#define BRW_WRITEMASK_XY       0x03
+#define BRW_WRITEMASK_Z        0x04
+#define BRW_WRITEMASK_XZ       0x05
+#define BRW_WRITEMASK_YZ       0x06
+#define BRW_WRITEMASK_XYZ      0x07
+#define BRW_WRITEMASK_W        0x08
+#define BRW_WRITEMASK_XW       0x09
+#define BRW_WRITEMASK_YW       0x0A
+#define BRW_WRITEMASK_XYW      0x0B
+#define BRW_WRITEMASK_ZW       0x0C
+#define BRW_WRITEMASK_XZW      0x0D
+#define BRW_WRITEMASK_YZW      0x0E
+#define BRW_WRITEMASK_XYZW     0x0F
+
+
+#define REG_SIZE (8*4)
+
+
+/* These aren't hardware structs, just something useful for us to pass around:
+ *
+ * Align1 operation has a lot of control over input ranges.  Used in
+ * WM programs to implement shaders decomposed into "channel serial"
+ * or "structure of array" form:
+ */
+struct brw_reg
+{
+   GLuint type:4;
+   GLuint file:2;
+   GLuint nr:8;
+   GLuint subnr:5;             /* :1 in align16 */
+   GLuint negate:1;            /* source only */
+   GLuint abs:1;               /* source only */
+   GLuint vstride:4;           /* source only */
+   GLuint width:3;             /* src only, align1 only */
+   GLuint hstride:2;                   /* align1 only */
+   GLuint address_mode:1;      /* relative addressing, hopefully! */
+   GLuint pad0:1;
+
+   union {      
+      struct {
+        GLuint swizzle:8;              /* src only, align16 only */
+        GLuint writemask:4;            /* dest only, align16 only */
+        GLint  indirect_offset:10;     /* relative addressing offset */
+        GLuint pad1:10;                /* two dwords total */
+      } bits;
+
+      GLfloat f;
+      GLint   d;
+      GLuint ud;
+   } dw1;      
+};
+
+
+struct brw_indirect {
+   GLuint addr_subnr:4;
+   GLint addr_offset:10;
+   GLuint pad:18;
+};
+
+
+struct brw_eu_label;
+struct brw_eu_call;
+
+
+
+#define BRW_EU_MAX_INSN_STACK 5
+#define BRW_EU_MAX_INSN 10000
+
+struct brw_compile {
+   struct brw_instruction store[BRW_EU_MAX_INSN];
+   GLuint nr_insn;
+
+   /* Allow clients to push/pop instruction state:
+    */
+   struct brw_instruction stack[BRW_EU_MAX_INSN_STACK];
+   struct brw_instruction *current;
+
+   GLuint flag_value;
+   GLboolean single_program_flow;
+   struct brw_context *brw;
+
+   struct brw_eu_label *first_label;  /**< linked list of labels */
+   struct brw_eu_call *first_call;    /**< linked list of CALs */
+
+   boolean error;
+};
+
+
+void
+brw_save_label(struct brw_compile *c, unsigned label, GLuint position);
+
+void
+brw_save_call(struct brw_compile *c, unsigned label, GLuint call_pos);
+
+void
+brw_resolve_cals(struct brw_compile *c);
+
+
+
+static INLINE int type_sz( GLuint type )
+{
+   switch( type ) {
+   case BRW_REGISTER_TYPE_UD:
+   case BRW_REGISTER_TYPE_D:
+   case BRW_REGISTER_TYPE_F:
+      return 4;
+   case BRW_REGISTER_TYPE_HF:
+   case BRW_REGISTER_TYPE_UW:
+   case BRW_REGISTER_TYPE_W:
+      return 2;
+   case BRW_REGISTER_TYPE_UB:
+   case BRW_REGISTER_TYPE_B:
+      return 1;
+   default:
+      return 0;
+   }
+}
+
+/**
+ * Construct a brw_reg.
+ * \param file  one of the BRW_x_REGISTER_FILE values
+ * \param nr  register number/index
+ * \param subnr  register sub number
+ * \param type  one of BRW_REGISTER_TYPE_x
+ * \param vstride  one of BRW_VERTICAL_STRIDE_x
+ * \param width  one of BRW_WIDTH_x
+ * \param hstride  one of BRW_HORIZONTAL_STRIDE_x
+ * \param swizzle  one of BRW_SWIZZLE_x
+ * \param writemask  BRW_WRITEMASK_X/Y/Z/W bitfield
+ */
+static INLINE struct brw_reg brw_reg( GLuint file,
+                                      GLuint nr,
+                                      GLuint subnr,
+                                      GLuint type,
+                                      GLuint vstride,
+                                      GLuint width,
+                                      GLuint hstride,
+                                      GLuint swizzle,
+                                      GLuint writemask )
+{
+   struct brw_reg reg;
+   if (type == BRW_GENERAL_REGISTER_FILE)
+      assert(nr < BRW_MAX_GRF);
+   else if (type == BRW_MESSAGE_REGISTER_FILE)
+      assert(nr < BRW_MAX_MRF);
+   else if (type == BRW_ARCHITECTURE_REGISTER_FILE)
+      assert(nr <= BRW_ARF_IP);
+
+   reg.type = type;
+   reg.file = file;
+   reg.nr = nr;
+   reg.subnr = subnr * type_sz(type);
+   reg.negate = 0;
+   reg.abs = 0;
+   reg.vstride = vstride;
+   reg.width = width;
+   reg.hstride = hstride;
+   reg.address_mode = BRW_ADDRESS_DIRECT;
+   reg.pad0 = 0;
+
+   /* Could do better: If the reg is r5.3<0;1,0>, we probably want to
+    * set swizzle and writemask to W, as the lower bits of subnr will
+    * be lost when converted to align16.  This is probably too much to
+    * keep track of as you'd want it adjusted by suboffset(), etc.
+    * Perhaps fix up when converting to align16?
+    */
+   reg.dw1.bits.swizzle = swizzle;
+   reg.dw1.bits.writemask = writemask;
+   reg.dw1.bits.indirect_offset = 0;
+   reg.dw1.bits.pad1 = 0;
+   return reg;
+}
+
+/** Construct float[16] register */
+static INLINE struct brw_reg brw_vec16_reg( GLuint file,
+                                             GLuint nr,
+                                             GLuint subnr )
+{
+   return brw_reg(file,
+                 nr,
+                 subnr,
+                 BRW_REGISTER_TYPE_F,
+                 BRW_VERTICAL_STRIDE_16,
+                 BRW_WIDTH_16,
+                 BRW_HORIZONTAL_STRIDE_1,
+                 BRW_SWIZZLE_XYZW,
+                 BRW_WRITEMASK_XYZW);
+}
+
+/** Construct float[8] register */
+static INLINE struct brw_reg brw_vec8_reg( GLuint file,
+                                            GLuint nr,
+                                            GLuint subnr )
+{
+   return brw_reg(file,
+                 nr,
+                 subnr,
+                 BRW_REGISTER_TYPE_F,
+                 BRW_VERTICAL_STRIDE_8,
+                 BRW_WIDTH_8,
+                 BRW_HORIZONTAL_STRIDE_1,
+                 BRW_SWIZZLE_XYZW,
+                 BRW_WRITEMASK_XYZW);
+}
+
+/** Construct float[4] register */
+static INLINE struct brw_reg brw_vec4_reg( GLuint file,
+                                             GLuint nr,
+                                             GLuint subnr )
+{
+   return brw_reg(file,
+                 nr,
+                 subnr,
+                 BRW_REGISTER_TYPE_F,
+                 BRW_VERTICAL_STRIDE_4,
+                 BRW_WIDTH_4,
+                 BRW_HORIZONTAL_STRIDE_1,
+                 BRW_SWIZZLE_XYZW,
+                 BRW_WRITEMASK_XYZW);
+}
+
+/** Construct float[2] register */
+static INLINE struct brw_reg brw_vec2_reg( GLuint file,
+                                             GLuint nr,
+                                             GLuint subnr )
+{
+   return brw_reg(file,
+                 nr,
+                 subnr,
+                 BRW_REGISTER_TYPE_F,
+                 BRW_VERTICAL_STRIDE_2,
+                 BRW_WIDTH_2,
+                 BRW_HORIZONTAL_STRIDE_1,
+                 BRW_SWIZZLE_XYXY,
+                 BRW_WRITEMASK_XY);
+}
+
+/** Construct float[1] register */
+static INLINE struct brw_reg brw_vec1_reg( GLuint file,
+                                            GLuint nr,
+                                            GLuint subnr )
+{
+   return brw_reg(file,
+                 nr,
+                 subnr,
+                 BRW_REGISTER_TYPE_F,
+                 BRW_VERTICAL_STRIDE_0,
+                 BRW_WIDTH_1,
+                 BRW_HORIZONTAL_STRIDE_0,
+                 BRW_SWIZZLE_XXXX,
+                 BRW_WRITEMASK_X);
+}
+
+
+static INLINE struct brw_reg retype( struct brw_reg reg,
+                                      GLuint type )
+{
+   reg.type = type;
+   return reg;
+}
+
+static INLINE struct brw_reg suboffset( struct brw_reg reg,
+                                         GLuint delta )
+{   
+   reg.subnr += delta * type_sz(reg.type);
+   return reg;
+}
+
+
+static INLINE struct brw_reg offset( struct brw_reg reg,
+                                      GLuint delta )
+{
+   reg.nr += delta;
+   return reg;
+}
+
+
+static INLINE struct brw_reg byte_offset( struct brw_reg reg,
+                                           GLuint bytes )
+{
+   GLuint newoffset = reg.nr * REG_SIZE + reg.subnr + bytes;
+   reg.nr = newoffset / REG_SIZE;
+   reg.subnr = newoffset % REG_SIZE;
+   return reg;
+}
+   
+
+/** Construct unsigned word[16] register */
+static INLINE struct brw_reg brw_uw16_reg( GLuint file,
+                                            GLuint nr,
+                                            GLuint subnr )
+{
+   return suboffset(retype(brw_vec16_reg(file, nr, 0), BRW_REGISTER_TYPE_UW), subnr);
+}
+
+/** Construct unsigned word[8] register */
+static INLINE struct brw_reg brw_uw8_reg( GLuint file,
+                                           GLuint nr,
+                                           GLuint subnr )
+{
+   return suboffset(retype(brw_vec8_reg(file, nr, 0), BRW_REGISTER_TYPE_UW), subnr);
+}
+
+/** Construct unsigned word[1] register */
+static INLINE struct brw_reg brw_uw1_reg( GLuint file,
+                                           GLuint nr,
+                                           GLuint subnr )
+{
+   return suboffset(retype(brw_vec1_reg(file, nr, 0), BRW_REGISTER_TYPE_UW), subnr);
+}
+
+static INLINE struct brw_reg brw_imm_reg( GLuint type )
+{
+   return brw_reg( BRW_IMMEDIATE_VALUE,
+                  0,
+                  0,
+                  type,
+                  BRW_VERTICAL_STRIDE_0,
+                  BRW_WIDTH_1,
+                  BRW_HORIZONTAL_STRIDE_0,
+                  0,
+                  0);      
+}
+
+/** Construct float immediate register */
+static INLINE struct brw_reg brw_imm_f( GLfloat f )
+{
+   struct brw_reg imm = brw_imm_reg(BRW_REGISTER_TYPE_F);
+   imm.dw1.f = f;
+   return imm;
+}
+
+/** Construct integer immediate register */
+static INLINE struct brw_reg brw_imm_d( GLint d )
+{
+   struct brw_reg imm = brw_imm_reg(BRW_REGISTER_TYPE_D);
+   imm.dw1.d = d;
+   return imm;
+}
+
+/** Construct uint immediate register */
+static INLINE struct brw_reg brw_imm_ud( GLuint ud )
+{
+   struct brw_reg imm = brw_imm_reg(BRW_REGISTER_TYPE_UD);
+   imm.dw1.ud = ud;
+   return imm;
+}
+
+/** Construct ushort immediate register */
+static INLINE struct brw_reg brw_imm_uw( GLushort uw )
+{
+   struct brw_reg imm = brw_imm_reg(BRW_REGISTER_TYPE_UW);
+   imm.dw1.ud = uw | (uw << 16);
+   return imm;
+}
+
+/** Construct short immediate register */
+static INLINE struct brw_reg brw_imm_w( GLshort w )
+{
+   struct brw_reg imm = brw_imm_reg(BRW_REGISTER_TYPE_W);
+   imm.dw1.d = w | (w << 16);
+   return imm;
+}
+
+/* brw_imm_b and brw_imm_ub aren't supported by hardware - the type
+ * numbers alias with _V and _VF below:
+ */
+
+/** Construct vector of eight signed half-byte values */
+static INLINE struct brw_reg brw_imm_v( GLuint v )
+{
+   struct brw_reg imm = brw_imm_reg(BRW_REGISTER_TYPE_V);
+   imm.vstride = BRW_VERTICAL_STRIDE_0;
+   imm.width = BRW_WIDTH_8;
+   imm.hstride = BRW_HORIZONTAL_STRIDE_1;
+   imm.dw1.ud = v;
+   return imm;
+}
+
+/** Construct vector of four 8-bit float values */
+static INLINE struct brw_reg brw_imm_vf( GLuint v )
+{
+   struct brw_reg imm = brw_imm_reg(BRW_REGISTER_TYPE_VF);
+   imm.vstride = BRW_VERTICAL_STRIDE_0;
+   imm.width = BRW_WIDTH_4;
+   imm.hstride = BRW_HORIZONTAL_STRIDE_1;
+   imm.dw1.ud = v;
+   return imm;
+}
+
+#define VF_ZERO 0x0
+#define VF_ONE  0x30
+#define VF_NEG  (1<<7)
+
+static INLINE struct brw_reg brw_imm_vf4( GLuint v0, 
+                                           GLuint v1, 
+                                           GLuint v2,
+                                           GLuint v3)
+{
+   struct brw_reg imm = brw_imm_reg(BRW_REGISTER_TYPE_VF);
+   imm.vstride = BRW_VERTICAL_STRIDE_0;
+   imm.width = BRW_WIDTH_4;
+   imm.hstride = BRW_HORIZONTAL_STRIDE_1;
+   imm.dw1.ud = ((v0 << 0) |
+                (v1 << 8) |
+                (v2 << 16) |
+                (v3 << 24));
+   return imm;
+}
+
+
+static INLINE struct brw_reg brw_address( struct brw_reg reg )
+{
+   return brw_imm_uw(reg.nr * REG_SIZE + reg.subnr);
+}
+
+/** Construct float[1] general-purpose register */
+static INLINE struct brw_reg brw_vec1_grf( GLuint nr, GLuint subnr )
+{
+   return brw_vec1_reg(BRW_GENERAL_REGISTER_FILE, nr, subnr);
+}
+
+/** Construct float[2] general-purpose register */
+static INLINE struct brw_reg brw_vec2_grf( GLuint nr, GLuint subnr )
+{
+   return brw_vec2_reg(BRW_GENERAL_REGISTER_FILE, nr, subnr);
+}
+
+/** Construct float[4] general-purpose register */
+static INLINE struct brw_reg brw_vec4_grf( GLuint nr, GLuint subnr )
+{
+   return brw_vec4_reg(BRW_GENERAL_REGISTER_FILE, nr, subnr);
+}
+
+/** Construct float[8] general-purpose register */
+static INLINE struct brw_reg brw_vec8_grf( GLuint nr, GLuint subnr )
+{
+   return brw_vec8_reg(BRW_GENERAL_REGISTER_FILE, nr, subnr);
+}
+
+
+static INLINE struct brw_reg brw_uw8_grf( GLuint nr, GLuint subnr )
+{
+   return brw_uw8_reg(BRW_GENERAL_REGISTER_FILE, nr, subnr);
+}
+
+static INLINE struct brw_reg brw_uw16_grf( GLuint nr, GLuint subnr )
+{
+   return brw_uw16_reg(BRW_GENERAL_REGISTER_FILE, nr, subnr);
+}
+
+
+/** Construct null register (usually used for setting condition codes) */
+static INLINE struct brw_reg brw_null_reg( void )
+{
+   return brw_vec8_reg(BRW_ARCHITECTURE_REGISTER_FILE, 
+                      BRW_ARF_NULL, 
+                      0);
+}
+
+static INLINE struct brw_reg brw_address_reg( GLuint subnr )
+{
+   return brw_uw1_reg(BRW_ARCHITECTURE_REGISTER_FILE, 
+                     BRW_ARF_ADDRESS, 
+                     subnr);
+}
+
+/* If/else instructions break in align16 mode if writemask & swizzle
+ * aren't xyzw.  This goes against the convention for other scalar
+ * regs:
+ */
+static INLINE struct brw_reg brw_ip_reg( void )
+{
+   return brw_reg(BRW_ARCHITECTURE_REGISTER_FILE, 
+                 BRW_ARF_IP, 
+                 0,
+                 BRW_REGISTER_TYPE_UD,
+                 BRW_VERTICAL_STRIDE_4, /* ? */
+                 BRW_WIDTH_1,
+                 BRW_HORIZONTAL_STRIDE_0,
+                 BRW_SWIZZLE_XYZW, /* NOTE! */
+                 BRW_WRITEMASK_XYZW); /* NOTE! */
+}
+
+static INLINE struct brw_reg brw_acc_reg( void )
+{
+   return brw_vec8_reg(BRW_ARCHITECTURE_REGISTER_FILE, 
+                      BRW_ARF_ACCUMULATOR, 
+                      0);
+}
+
+
+static INLINE struct brw_reg brw_flag_reg( void )
+{
+   return brw_uw1_reg(BRW_ARCHITECTURE_REGISTER_FILE,
+                     BRW_ARF_FLAG,
+                     0);
+}
+
+
+static INLINE struct brw_reg brw_mask_reg( GLuint subnr )
+{
+   return brw_uw1_reg(BRW_ARCHITECTURE_REGISTER_FILE,
+                     BRW_ARF_MASK,
+                     subnr);
+}
+
+static INLINE struct brw_reg brw_message_reg( GLuint nr )
+{
+   assert(nr < BRW_MAX_MRF);
+   return brw_vec8_reg(BRW_MESSAGE_REGISTER_FILE,
+                      nr,
+                      0);
+}
+
+
+
+
+/* This is almost always called with a numeric constant argument, so
+ * make things easy to evaluate at compile time:
+ */
+static INLINE GLuint cvt( GLuint val )
+{
+   switch (val) {
+   case 0: return 0;
+   case 1: return 1;
+   case 2: return 2;
+   case 4: return 3;
+   case 8: return 4;
+   case 16: return 5;
+   case 32: return 6;
+   }
+   return 0;
+}
+
+static INLINE struct brw_reg stride( struct brw_reg reg,
+                                      GLuint vstride,
+                                      GLuint width,
+                                      GLuint hstride )
+{
+   reg.vstride = cvt(vstride);
+   reg.width = cvt(width) - 1;
+   reg.hstride = cvt(hstride);
+   return reg;
+}
+
+
+static INLINE struct brw_reg vec16( struct brw_reg reg )
+{
+   return stride(reg, 16,16,1);
+}
+
+static INLINE struct brw_reg vec8( struct brw_reg reg )
+{
+   return stride(reg, 8,8,1);
+}
+
+static INLINE struct brw_reg vec4( struct brw_reg reg )
+{
+   return stride(reg, 4,4,1);
+}
+
+static INLINE struct brw_reg vec2( struct brw_reg reg )
+{
+   return stride(reg, 2,2,1);
+}
+
+static INLINE struct brw_reg vec1( struct brw_reg reg )
+{
+   return stride(reg, 0,1,0);
+}
+
+
+static INLINE struct brw_reg get_element( struct brw_reg reg, GLuint elt )
+{
+   return vec1(suboffset(reg, elt));
+}
+
+static INLINE struct brw_reg get_element_ud( struct brw_reg reg, GLuint elt )
+{
+   return vec1(suboffset(retype(reg, BRW_REGISTER_TYPE_UD), elt));
+}
+
+
+static INLINE struct brw_reg brw_swizzle( struct brw_reg reg,
+                                           GLuint x,
+                                           GLuint y, 
+                                           GLuint z,
+                                           GLuint w)
+{
+   reg.dw1.bits.swizzle = BRW_SWIZZLE4(BRW_GET_SWZ(reg.dw1.bits.swizzle, x),
+                                      BRW_GET_SWZ(reg.dw1.bits.swizzle, y),
+                                      BRW_GET_SWZ(reg.dw1.bits.swizzle, z),
+                                      BRW_GET_SWZ(reg.dw1.bits.swizzle, w));
+   return reg;
+}
+
+
+static INLINE struct brw_reg brw_swizzle1( struct brw_reg reg,
+                                            GLuint x )
+{
+   return brw_swizzle(reg, x, x, x, x);
+}
+
+static INLINE struct brw_reg brw_writemask( struct brw_reg reg,
+                                             GLuint mask )
+{
+   reg.dw1.bits.writemask &= mask;
+   return reg;
+}
+
+static INLINE struct brw_reg brw_set_writemask( struct brw_reg reg,
+                                                 GLuint mask )
+{
+   reg.dw1.bits.writemask = mask;
+   return reg;
+}
+
+static INLINE struct brw_reg negate( struct brw_reg reg )
+{
+   reg.negate ^= 1;
+   return reg;
+}
+
+static INLINE struct brw_reg brw_abs( struct brw_reg reg )
+{
+   reg.abs = 1;
+   return reg;
+}
+
+/***********************************************************************
+ */
+static INLINE struct brw_reg brw_vec4_indirect( GLuint subnr,
+                                                 GLint offset )
+{
+   struct brw_reg reg =  brw_vec4_grf(0, 0);
+   reg.subnr = subnr;
+   reg.address_mode = BRW_ADDRESS_REGISTER_INDIRECT_REGISTER;
+   reg.dw1.bits.indirect_offset = offset;
+   return reg;
+}
+
+static INLINE struct brw_reg brw_vec1_indirect( GLuint subnr,
+                                                 GLint offset )
+{
+   struct brw_reg reg =  brw_vec1_grf(0, 0);
+   reg.subnr = subnr;
+   reg.address_mode = BRW_ADDRESS_REGISTER_INDIRECT_REGISTER;
+   reg.dw1.bits.indirect_offset = offset;
+   return reg;
+}
+
+static INLINE struct brw_reg deref_4f(struct brw_indirect ptr, GLint offset)
+{
+   return brw_vec4_indirect(ptr.addr_subnr, ptr.addr_offset + offset);
+}
+
+static INLINE struct brw_reg deref_1f(struct brw_indirect ptr, GLint offset)
+{
+   return brw_vec1_indirect(ptr.addr_subnr, ptr.addr_offset + offset);
+}
+
+static INLINE struct brw_reg deref_4b(struct brw_indirect ptr, GLint offset)
+{
+   return retype(deref_4f(ptr, offset), BRW_REGISTER_TYPE_B);
+}
+
+static INLINE struct brw_reg deref_1uw(struct brw_indirect ptr, GLint offset)
+{
+   return retype(deref_1f(ptr, offset), BRW_REGISTER_TYPE_UW);
+}
+
+static INLINE struct brw_reg deref_1d(struct brw_indirect ptr, GLint offset)
+{
+   return retype(deref_1f(ptr, offset), BRW_REGISTER_TYPE_D);
+}
+
+static INLINE struct brw_reg deref_1ud(struct brw_indirect ptr, GLint offset)
+{
+   return retype(deref_1f(ptr, offset), BRW_REGISTER_TYPE_UD);
+}
+
+static INLINE struct brw_reg get_addr_reg(struct brw_indirect ptr)
+{
+   return brw_address_reg(ptr.addr_subnr);
+}
+
+static INLINE struct brw_indirect brw_indirect_offset( struct brw_indirect ptr, GLint offset )
+{
+   ptr.addr_offset += offset;
+   return ptr;
+}
+
+static INLINE struct brw_indirect brw_indirect( GLuint addr_subnr, GLint offset )
+{
+   struct brw_indirect ptr;
+   ptr.addr_subnr = addr_subnr;
+   ptr.addr_offset = offset;
+   ptr.pad = 0;
+   return ptr;
+}
+
+/** Do two brw_regs refer to the same register? */
+static INLINE GLboolean
+brw_same_reg(struct brw_reg r1, struct brw_reg r2)
+{
+   return r1.file == r2.file && r1.nr == r2.nr;
+}
+
+static INLINE struct brw_instruction *current_insn( struct brw_compile *p)
+{
+   return &p->store[p->nr_insn];
+}
+
+void brw_pop_insn_state( struct brw_compile *p );
+void brw_push_insn_state( struct brw_compile *p );
+void brw_set_mask_control( struct brw_compile *p, GLuint value );
+void brw_set_saturate( struct brw_compile *p, GLuint value );
+void brw_set_access_mode( struct brw_compile *p, GLuint access_mode );
+void brw_set_compression_control( struct brw_compile *p, GLboolean control );
+void brw_set_predicate_control_flag_value( struct brw_compile *p, GLuint value );
+void brw_set_predicate_control( struct brw_compile *p, GLuint pc );
+void brw_set_conditionalmod( struct brw_compile *p, GLuint conditional );
+
+void brw_init_compile( struct brw_context *, struct brw_compile *p );
+
+enum pipe_error brw_get_program( struct brw_compile *p, 
+                                 const GLuint **program,
+                                 GLuint *sz );
+
+
+/* Helpers for regular instructions:
+ */
+#define ALU1(OP)                                       \
+struct brw_instruction *brw_##OP(struct brw_compile *p,        \
+             struct brw_reg dest,                      \
+             struct brw_reg src0);
+
+#define ALU2(OP)                                       \
+struct brw_instruction *brw_##OP(struct brw_compile *p,        \
+             struct brw_reg dest,                      \
+             struct brw_reg src0,                      \
+             struct brw_reg src1);
+
+ALU1(MOV)
+ALU2(SEL)
+ALU1(NOT)
+ALU2(AND)
+ALU2(OR)
+ALU2(XOR)
+ALU2(SHR)
+ALU2(SHL)
+ALU2(RSR)
+ALU2(RSL)
+ALU2(ASR)
+ALU2(JMPI)
+ALU2(ADD)
+ALU2(MUL)
+ALU1(FRC)
+ALU1(RNDD)
+ALU1(RNDZ)
+ALU2(MAC)
+ALU2(MACH)
+ALU1(LZD)
+ALU2(DP4)
+ALU2(DPH)
+ALU2(DP3)
+ALU2(DP2)
+ALU2(LINE)
+
+#undef ALU1
+#undef ALU2
+
+
+
+/* Helpers for SEND instruction:
+ */
+void brw_urb_WRITE(struct brw_compile *p,
+                  struct brw_reg dest,
+                  GLuint msg_reg_nr,
+                  struct brw_reg src0,
+                  GLboolean allocate,
+                  GLboolean used,
+                  GLuint msg_length,
+                  GLuint response_length,
+                  GLboolean eot,
+                  GLboolean writes_complete,
+                  GLuint offset,
+                  GLuint swizzle);
+
+void brw_ff_sync(struct brw_compile *p,
+                  struct brw_reg dest,
+                  GLuint msg_reg_nr,
+                  struct brw_reg src0,
+                  GLboolean allocate,
+                  GLboolean used,
+                  GLuint msg_length,
+                  GLuint response_length,
+                  GLboolean eot,
+                  GLboolean writes_complete,
+                  GLuint offset,
+                  GLuint swizzle);
+
+void brw_fb_WRITE(struct brw_compile *p,
+                  struct brw_reg dest,
+                  GLuint msg_reg_nr,
+                  struct brw_reg src0,
+                  GLuint binding_table_index,
+                  GLuint msg_length,
+                  GLuint response_length,
+                  GLboolean eot);
+
+void brw_SAMPLE(struct brw_compile *p,
+               struct brw_reg dest,
+               GLuint msg_reg_nr,
+               struct brw_reg src0,
+               GLuint binding_table_index,
+               GLuint sampler,
+               GLuint writemask,
+               GLuint msg_type,
+               GLuint response_length,
+               GLuint msg_length,
+               GLboolean eot,
+               GLuint header_present,
+               GLuint simd_mode);
+
+void brw_math_16( struct brw_compile *p,
+                 struct brw_reg dest,
+                 GLuint function,
+                 GLuint saturate,
+                 GLuint msg_reg_nr,
+                 struct brw_reg src,
+                 GLuint precision );
+
+void brw_math( struct brw_compile *p,
+              struct brw_reg dest,
+              GLuint function,
+              GLuint saturate,
+              GLuint msg_reg_nr,
+              struct brw_reg src,
+              GLuint data_type,
+              GLuint precision );
+
+void brw_dp_READ_16( struct brw_compile *p,
+                    struct brw_reg dest,
+                    GLuint scratch_offset );
+
+void brw_dp_READ_4( struct brw_compile *p,
+                    struct brw_reg dest,
+                    GLboolean relAddr,
+                    GLuint location,
+                    GLuint bind_table_index );
+
+void brw_dp_READ_4_vs( struct brw_compile *p,
+                       struct brw_reg dest,
+                       GLuint oword,
+                       GLboolean relAddr,
+                       struct brw_reg addrReg,
+                       GLuint location,
+                       GLuint bind_table_index );
+
+void brw_dp_WRITE_16( struct brw_compile *p,
+                     struct brw_reg src,
+                     GLuint scratch_offset );
+
+/* If/else/endif.  Works by manipulating the execution flags on each
+ * channel.
+ */
+struct brw_instruction *brw_IF(struct brw_compile *p, 
+                              GLuint execute_size);
+
+struct brw_instruction *brw_ELSE(struct brw_compile *p, 
+                                struct brw_instruction *if_insn);
+
+void brw_ENDIF(struct brw_compile *p, 
+              struct brw_instruction *if_or_else_insn);
+
+
+/* DO/WHILE loops:
+ */
+struct brw_instruction *brw_DO(struct brw_compile *p,
+                              GLuint execute_size);
+
+struct brw_instruction *brw_WHILE(struct brw_compile *p, 
+              struct brw_instruction *patch_insn);
+
+struct brw_instruction *brw_BREAK(struct brw_compile *p);
+struct brw_instruction *brw_CONT(struct brw_compile *p);
+/* Forward jumps:
+ */
+void brw_land_fwd_jump(struct brw_compile *p, 
+                      struct brw_instruction *jmp_insn);
+
+
+
+void brw_NOP(struct brw_compile *p);
+
+/* Special case: there is never a destination, execution size will be
+ * taken from src0:
+ */
+void brw_CMP(struct brw_compile *p,
+            struct brw_reg dest,
+            GLuint conditional,
+            struct brw_reg src0,
+            struct brw_reg src1);
+
+void brw_print_reg( struct brw_reg reg );
+
+
+/*********************************************************************** 
+ * brw_eu_util.c:
+ */
+
+void brw_copy_indirect_to_indirect(struct brw_compile *p,
+                                  struct brw_indirect dst_ptr,
+                                  struct brw_indirect src_ptr,
+                                  GLuint count);
+
+void brw_copy_from_indirect(struct brw_compile *p,
+                           struct brw_reg dst,
+                           struct brw_indirect ptr,
+                           GLuint count);
+
+void brw_copy4(struct brw_compile *p,
+              struct brw_reg dst,
+              struct brw_reg src,
+              GLuint count);
+
+void brw_copy8(struct brw_compile *p,
+              struct brw_reg dst,
+              struct brw_reg src,
+              GLuint count);
+
+void brw_math_invert( struct brw_compile *p, 
+                     struct brw_reg dst,
+                     struct brw_reg src);
+
+void brw_set_src1( struct brw_instruction *insn,
+                          struct brw_reg reg );
+#endif
diff --git a/src/gallium/drivers/i965/brw_eu_debug.c b/src/gallium/drivers/i965/brw_eu_debug.c
new file mode 100644 (file)
index 0000000..5989f5a
--- /dev/null
@@ -0,0 +1,94 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+#include "util/u_debug.h"
+
+#include "brw_eu.h"
+
+void brw_print_reg( struct brw_reg hwreg )
+{
+   static const char *file[] = {
+      "arf",
+      "grf",
+      "msg",
+      "imm"
+   };
+
+   static const char *type[] = {
+      "ud",
+      "d",
+      "uw",
+      "w",
+      "ub",
+      "vf",
+      "hf",
+      "f"
+   };
+
+   debug_printf("%s%s", 
+               hwreg.abs ? "abs/" : "",
+               hwreg.negate ? "-" : "");
+     
+   if (hwreg.file == BRW_GENERAL_REGISTER_FILE &&
+       hwreg.nr % 2 == 0 &&
+       hwreg.subnr == 0 &&
+       hwreg.vstride == BRW_VERTICAL_STRIDE_8 &&
+       hwreg.width == BRW_WIDTH_8 &&
+       hwreg.hstride == BRW_HORIZONTAL_STRIDE_1 &&
+       hwreg.type == BRW_REGISTER_TYPE_F) {
+      /* vector register */
+      debug_printf("vec%d", hwreg.nr);
+   }
+   else if (hwreg.file == BRW_GENERAL_REGISTER_FILE &&
+           hwreg.vstride == BRW_VERTICAL_STRIDE_0 &&
+           hwreg.width == BRW_WIDTH_1 &&
+           hwreg.hstride == BRW_HORIZONTAL_STRIDE_0 &&
+           hwreg.type == BRW_REGISTER_TYPE_F) {      
+      /* "scalar" register */
+      debug_printf("scl%d.%d", hwreg.nr, hwreg.subnr / 4);
+   }
+   else if (hwreg.file == BRW_IMMEDIATE_VALUE) {
+      debug_printf("imm %f", hwreg.dw1.f);
+   }
+   else {
+      debug_printf("%s%d.%d<%d;%d,%d>:%s", 
+                  file[hwreg.file],
+                  hwreg.nr,
+                  hwreg.subnr / type_sz(hwreg.type),
+                  hwreg.vstride ? (1<<(hwreg.vstride-1)) : 0,
+                  1<<hwreg.width,
+                  hwreg.hstride ? (1<<(hwreg.hstride-1)) : 0,          
+                  type[hwreg.type]);
+   }
+}
+
+
+
diff --git a/src/gallium/drivers/i965/brw_eu_emit.c b/src/gallium/drivers/i965/brw_eu_emit.c
new file mode 100644 (file)
index 0000000..4fe7b6a
--- /dev/null
@@ -0,0 +1,1433 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+     
+
+#include "brw_context.h"
+#include "brw_defines.h"
+#include "brw_eu.h"
+#include "brw_debug.h"
+#include "brw_disasm.h"
+
+
+
+
+/***********************************************************************
+ * Internal helper for constructing instructions
+ */
+
+static void guess_execution_size( struct brw_instruction *insn,
+                                 struct brw_reg reg )
+{
+   if (reg.width == BRW_WIDTH_8 && 
+       insn->header.compression_control == BRW_COMPRESSION_COMPRESSED) 
+      insn->header.execution_size = BRW_EXECUTE_16;
+   else
+      insn->header.execution_size = reg.width; /* note - definitions are compatible */
+}
+
+
+static void brw_set_dest( struct brw_instruction *insn,
+                         struct brw_reg dest )
+{
+   if (dest.type != BRW_ARCHITECTURE_REGISTER_FILE)
+      assert(dest.nr < 128);
+
+   insn->bits1.da1.dest_reg_file = dest.file;
+   insn->bits1.da1.dest_reg_type = dest.type;
+   insn->bits1.da1.dest_address_mode = dest.address_mode;
+
+   if (dest.address_mode == BRW_ADDRESS_DIRECT) {   
+      insn->bits1.da1.dest_reg_nr = dest.nr;
+
+      if (insn->header.access_mode == BRW_ALIGN_1) {
+        insn->bits1.da1.dest_subreg_nr = dest.subnr;
+        if (dest.hstride == BRW_HORIZONTAL_STRIDE_0)
+           dest.hstride = BRW_HORIZONTAL_STRIDE_1;
+        insn->bits1.da1.dest_horiz_stride = dest.hstride;
+      }
+      else {
+        insn->bits1.da16.dest_subreg_nr = dest.subnr / 16;
+        insn->bits1.da16.dest_writemask = dest.dw1.bits.writemask;
+      }
+   }
+   else {
+      insn->bits1.ia1.dest_subreg_nr = dest.subnr;
+
+      /* These are different sizes in align1 vs align16:
+       */
+      if (insn->header.access_mode == BRW_ALIGN_1) {
+        insn->bits1.ia1.dest_indirect_offset = dest.dw1.bits.indirect_offset;
+        if (dest.hstride == BRW_HORIZONTAL_STRIDE_0)
+           dest.hstride = BRW_HORIZONTAL_STRIDE_1;
+        insn->bits1.ia1.dest_horiz_stride = dest.hstride;
+      }
+      else {
+        insn->bits1.ia16.dest_indirect_offset = dest.dw1.bits.indirect_offset;
+      }
+   }
+
+   /* NEW: Set the execution size based on dest.width and
+    * insn->compression_control:
+    */
+   guess_execution_size(insn, dest);
+}
+
+static void brw_set_src0( struct brw_instruction *insn,
+                          struct brw_reg reg )
+{
+   assert(reg.file != BRW_MESSAGE_REGISTER_FILE);
+
+   if (reg.type != BRW_ARCHITECTURE_REGISTER_FILE)
+      assert(reg.nr < 128);
+
+   insn->bits1.da1.src0_reg_file = reg.file;
+   insn->bits1.da1.src0_reg_type = reg.type;
+   insn->bits2.da1.src0_abs = reg.abs;
+   insn->bits2.da1.src0_negate = reg.negate;
+   insn->bits2.da1.src0_address_mode = reg.address_mode;
+
+   if (reg.file == BRW_IMMEDIATE_VALUE) {
+      insn->bits3.ud = reg.dw1.ud;
+   
+      /* Required to set some fields in src1 as well:
+       */
+      insn->bits1.da1.src1_reg_file = 0; /* arf */
+      insn->bits1.da1.src1_reg_type = reg.type;
+   }
+   else 
+   {
+      if (reg.address_mode == BRW_ADDRESS_DIRECT) {
+        if (insn->header.access_mode == BRW_ALIGN_1) {
+           insn->bits2.da1.src0_subreg_nr = reg.subnr;
+           insn->bits2.da1.src0_reg_nr = reg.nr;
+        }
+        else {
+           insn->bits2.da16.src0_subreg_nr = reg.subnr / 16;
+           insn->bits2.da16.src0_reg_nr = reg.nr;
+        }
+      }
+      else {
+        insn->bits2.ia1.src0_subreg_nr = reg.subnr;
+
+        if (insn->header.access_mode == BRW_ALIGN_1) {
+           insn->bits2.ia1.src0_indirect_offset = reg.dw1.bits.indirect_offset; 
+        }
+        else {
+           insn->bits2.ia16.src0_subreg_nr = reg.dw1.bits.indirect_offset;
+        }
+      }
+
+      if (insn->header.access_mode == BRW_ALIGN_1) {
+        if (reg.width == BRW_WIDTH_1 && 
+            insn->header.execution_size == BRW_EXECUTE_1) {
+           insn->bits2.da1.src0_horiz_stride = BRW_HORIZONTAL_STRIDE_0;
+           insn->bits2.da1.src0_width = BRW_WIDTH_1;
+           insn->bits2.da1.src0_vert_stride = BRW_VERTICAL_STRIDE_0;
+        }
+        else {
+           insn->bits2.da1.src0_horiz_stride = reg.hstride;
+           insn->bits2.da1.src0_width = reg.width;
+           insn->bits2.da1.src0_vert_stride = reg.vstride;
+        }
+      }
+      else {
+        insn->bits2.da16.src0_swz_x = BRW_GET_SWZ(reg.dw1.bits.swizzle, BRW_CHANNEL_X);
+        insn->bits2.da16.src0_swz_y = BRW_GET_SWZ(reg.dw1.bits.swizzle, BRW_CHANNEL_Y);
+        insn->bits2.da16.src0_swz_z = BRW_GET_SWZ(reg.dw1.bits.swizzle, BRW_CHANNEL_Z);
+        insn->bits2.da16.src0_swz_w = BRW_GET_SWZ(reg.dw1.bits.swizzle, BRW_CHANNEL_W);
+
+        /* This is an oddity of the fact we're using the same
+         * descriptions for registers in align_16 as align_1:
+         */
+        if (reg.vstride == BRW_VERTICAL_STRIDE_8)
+           insn->bits2.da16.src0_vert_stride = BRW_VERTICAL_STRIDE_4;
+        else
+           insn->bits2.da16.src0_vert_stride = reg.vstride;
+      }
+   }
+}
+
+
+void brw_set_src1( struct brw_instruction *insn,
+                   struct brw_reg reg )
+{
+   assert(reg.file != BRW_MESSAGE_REGISTER_FILE);
+
+   assert(reg.nr < 128);
+
+   insn->bits1.da1.src1_reg_file = reg.file;
+   insn->bits1.da1.src1_reg_type = reg.type;
+   insn->bits3.da1.src1_abs = reg.abs;
+   insn->bits3.da1.src1_negate = reg.negate;
+
+   /* Only src1 can be immediate in two-argument instructions.
+    */
+   assert(insn->bits1.da1.src0_reg_file != BRW_IMMEDIATE_VALUE);
+
+   if (reg.file == BRW_IMMEDIATE_VALUE) {
+      insn->bits3.ud = reg.dw1.ud;
+   }
+   else {
+      /* This is a hardware restriction, which may or may not be lifted
+       * in the future:
+       */
+      assert (reg.address_mode == BRW_ADDRESS_DIRECT);
+      /*assert (reg.file == BRW_GENERAL_REGISTER_FILE); */
+
+      if (insn->header.access_mode == BRW_ALIGN_1) {
+        insn->bits3.da1.src1_subreg_nr = reg.subnr;
+        insn->bits3.da1.src1_reg_nr = reg.nr;
+      }
+      else {
+        insn->bits3.da16.src1_subreg_nr = reg.subnr / 16;
+        insn->bits3.da16.src1_reg_nr = reg.nr;
+      }
+
+      if (insn->header.access_mode == BRW_ALIGN_1) {
+        if (reg.width == BRW_WIDTH_1 && 
+            insn->header.execution_size == BRW_EXECUTE_1) {
+           insn->bits3.da1.src1_horiz_stride = BRW_HORIZONTAL_STRIDE_0;
+           insn->bits3.da1.src1_width = BRW_WIDTH_1;
+           insn->bits3.da1.src1_vert_stride = BRW_VERTICAL_STRIDE_0;
+        }
+        else {
+           insn->bits3.da1.src1_horiz_stride = reg.hstride;
+           insn->bits3.da1.src1_width = reg.width;
+           insn->bits3.da1.src1_vert_stride = reg.vstride;
+        }
+      }
+      else {
+        insn->bits3.da16.src1_swz_x = BRW_GET_SWZ(reg.dw1.bits.swizzle, BRW_CHANNEL_X);
+        insn->bits3.da16.src1_swz_y = BRW_GET_SWZ(reg.dw1.bits.swizzle, BRW_CHANNEL_Y);
+        insn->bits3.da16.src1_swz_z = BRW_GET_SWZ(reg.dw1.bits.swizzle, BRW_CHANNEL_Z);
+        insn->bits3.da16.src1_swz_w = BRW_GET_SWZ(reg.dw1.bits.swizzle, BRW_CHANNEL_W);
+
+        /* This is an oddity of the fact we're using the same
+         * descriptions for registers in align_16 as align_1:
+         */
+        if (reg.vstride == BRW_VERTICAL_STRIDE_8)
+           insn->bits3.da16.src1_vert_stride = BRW_VERTICAL_STRIDE_4;
+        else
+           insn->bits3.da16.src1_vert_stride = reg.vstride;
+      }
+   }
+}
+
+
+
+static void brw_set_math_message( struct brw_context *brw,
+                                 struct brw_instruction *insn,
+                                 GLuint msg_length,
+                                 GLuint response_length,
+                                 GLuint function,
+                                 GLuint integer_type,
+                                 GLboolean low_precision,
+                                 GLboolean saturate,
+                                 GLuint dataType )
+{
+   brw_set_src1(insn, brw_imm_d(0));
+
+   if (BRW_IS_IGDNG(brw)) {
+       insn->bits3.math_igdng.function = function;
+       insn->bits3.math_igdng.int_type = integer_type;
+       insn->bits3.math_igdng.precision = low_precision;
+       insn->bits3.math_igdng.saturate = saturate;
+       insn->bits3.math_igdng.data_type = dataType;
+       insn->bits3.math_igdng.snapshot = 0;
+       insn->bits3.math_igdng.header_present = 0;
+       insn->bits3.math_igdng.response_length = response_length;
+       insn->bits3.math_igdng.msg_length = msg_length;
+       insn->bits3.math_igdng.end_of_thread = 0;
+       insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_MATH;
+       insn->bits2.send_igdng.end_of_thread = 0;
+   } else {
+       insn->bits3.math.function = function;
+       insn->bits3.math.int_type = integer_type;
+       insn->bits3.math.precision = low_precision;
+       insn->bits3.math.saturate = saturate;
+       insn->bits3.math.data_type = dataType;
+       insn->bits3.math.response_length = response_length;
+       insn->bits3.math.msg_length = msg_length;
+       insn->bits3.math.msg_target = BRW_MESSAGE_TARGET_MATH;
+       insn->bits3.math.end_of_thread = 0;
+   }
+}
+
+
+static void brw_set_ff_sync_message( struct brw_context *brw,
+                                struct brw_instruction *insn,
+                                GLboolean allocate,
+                                GLboolean used,
+                                GLuint msg_length,
+                                GLuint response_length,
+                                GLboolean end_of_thread,
+                                GLboolean complete,
+                                GLuint offset,
+                                GLuint swizzle_control )
+{
+       brw_set_src1(insn, brw_imm_d(0));
+
+       insn->bits3.urb_igdng.opcode = 1;
+       insn->bits3.urb_igdng.offset = offset;
+       insn->bits3.urb_igdng.swizzle_control = swizzle_control;
+       insn->bits3.urb_igdng.allocate = allocate;
+       insn->bits3.urb_igdng.used = used;
+       insn->bits3.urb_igdng.complete = complete;
+       insn->bits3.urb_igdng.header_present = 1;
+       insn->bits3.urb_igdng.response_length = response_length;
+       insn->bits3.urb_igdng.msg_length = msg_length;
+       insn->bits3.urb_igdng.end_of_thread = end_of_thread;
+       insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_URB;
+       insn->bits2.send_igdng.end_of_thread = end_of_thread;
+}
+
+static void brw_set_urb_message( struct brw_context *brw,
+                                struct brw_instruction *insn,
+                                GLboolean allocate,
+                                GLboolean used,
+                                GLuint msg_length,
+                                GLuint response_length,
+                                GLboolean end_of_thread,
+                                GLboolean complete,
+                                GLuint offset,
+                                GLuint swizzle_control )
+{
+    brw_set_src1(insn, brw_imm_d(0));
+
+    if (BRW_IS_IGDNG(brw)) {
+        insn->bits3.urb_igdng.opcode = 0;      /* ? */
+        insn->bits3.urb_igdng.offset = offset;
+        insn->bits3.urb_igdng.swizzle_control = swizzle_control;
+        insn->bits3.urb_igdng.allocate = allocate;
+        insn->bits3.urb_igdng.used = used;     /* ? */
+        insn->bits3.urb_igdng.complete = complete;
+        insn->bits3.urb_igdng.header_present = 1;
+        insn->bits3.urb_igdng.response_length = response_length;
+        insn->bits3.urb_igdng.msg_length = msg_length;
+        insn->bits3.urb_igdng.end_of_thread = end_of_thread;
+        insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_URB;
+        insn->bits2.send_igdng.end_of_thread = end_of_thread;
+    } else {
+        insn->bits3.urb.opcode = 0;    /* ? */
+        insn->bits3.urb.offset = offset;
+        insn->bits3.urb.swizzle_control = swizzle_control;
+        insn->bits3.urb.allocate = allocate;
+        insn->bits3.urb.used = used;   /* ? */
+        insn->bits3.urb.complete = complete;
+        insn->bits3.urb.response_length = response_length;
+        insn->bits3.urb.msg_length = msg_length;
+        insn->bits3.urb.msg_target = BRW_MESSAGE_TARGET_URB;
+        insn->bits3.urb.end_of_thread = end_of_thread;
+    }
+}
+
+static void brw_set_dp_write_message( struct brw_context *brw,
+                                     struct brw_instruction *insn,
+                                     GLuint binding_table_index,
+                                     GLuint msg_control,
+                                     GLuint msg_type,
+                                     GLuint msg_length,
+                                     GLuint pixel_scoreboard_clear,
+                                     GLuint response_length,
+                                     GLuint end_of_thread )
+{
+   brw_set_src1(insn, brw_imm_d(0));
+
+   if (BRW_IS_IGDNG(brw)) {
+       insn->bits3.dp_write_igdng.binding_table_index = binding_table_index;
+       insn->bits3.dp_write_igdng.msg_control = msg_control;
+       insn->bits3.dp_write_igdng.pixel_scoreboard_clear = pixel_scoreboard_clear;
+       insn->bits3.dp_write_igdng.msg_type = msg_type;
+       insn->bits3.dp_write_igdng.send_commit_msg = 0;
+       insn->bits3.dp_write_igdng.header_present = 1;
+       insn->bits3.dp_write_igdng.response_length = response_length;
+       insn->bits3.dp_write_igdng.msg_length = msg_length;
+       insn->bits3.dp_write_igdng.end_of_thread = end_of_thread;
+       insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_DATAPORT_WRITE;
+       insn->bits2.send_igdng.end_of_thread = end_of_thread;
+   } else {
+       insn->bits3.dp_write.binding_table_index = binding_table_index;
+       insn->bits3.dp_write.msg_control = msg_control;
+       insn->bits3.dp_write.pixel_scoreboard_clear = pixel_scoreboard_clear;
+       insn->bits3.dp_write.msg_type = msg_type;
+       insn->bits3.dp_write.send_commit_msg = 0;
+       insn->bits3.dp_write.response_length = response_length;
+       insn->bits3.dp_write.msg_length = msg_length;
+       insn->bits3.dp_write.msg_target = BRW_MESSAGE_TARGET_DATAPORT_WRITE;
+       insn->bits3.dp_write.end_of_thread = end_of_thread;
+   }
+}
+
+static void brw_set_dp_read_message( struct brw_context *brw,
+                                     struct brw_instruction *insn,
+                                     GLuint binding_table_index,
+                                     GLuint msg_control,
+                                     GLuint msg_type,
+                                     GLuint target_cache,
+                                     GLuint msg_length,
+                                     GLuint response_length,
+                                     GLuint end_of_thread )
+{
+   brw_set_src1(insn, brw_imm_d(0));
+
+   if (BRW_IS_IGDNG(brw)) {
+       insn->bits3.dp_read_igdng.binding_table_index = binding_table_index;
+       insn->bits3.dp_read_igdng.msg_control = msg_control;
+       insn->bits3.dp_read_igdng.msg_type = msg_type;
+       insn->bits3.dp_read_igdng.target_cache = target_cache;
+       insn->bits3.dp_read_igdng.header_present = 1;
+       insn->bits3.dp_read_igdng.response_length = response_length;
+       insn->bits3.dp_read_igdng.msg_length = msg_length;
+       insn->bits3.dp_read_igdng.pad1 = 0;
+       insn->bits3.dp_read_igdng.end_of_thread = end_of_thread;
+       insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_DATAPORT_READ;
+       insn->bits2.send_igdng.end_of_thread = end_of_thread;
+   } else {
+       insn->bits3.dp_read.binding_table_index = binding_table_index; /*0:7*/
+       insn->bits3.dp_read.msg_control = msg_control;  /*8:11*/
+       insn->bits3.dp_read.msg_type = msg_type;  /*12:13*/
+       insn->bits3.dp_read.target_cache = target_cache;  /*14:15*/
+       insn->bits3.dp_read.response_length = response_length;  /*16:19*/
+       insn->bits3.dp_read.msg_length = msg_length;  /*20:23*/
+       insn->bits3.dp_read.msg_target = BRW_MESSAGE_TARGET_DATAPORT_READ; /*24:27*/
+       insn->bits3.dp_read.pad1 = 0;  /*28:30*/
+       insn->bits3.dp_read.end_of_thread = end_of_thread;  /*31*/
+   }
+}
+
+static void brw_set_sampler_message(struct brw_context *brw,
+                                    struct brw_instruction *insn,
+                                    GLuint binding_table_index,
+                                    GLuint sampler,
+                                    GLuint msg_type,
+                                    GLuint response_length,
+                                    GLuint msg_length,
+                                    GLboolean eot,
+                                    GLuint header_present,
+                                    GLuint simd_mode)
+{
+   assert(eot == 0);
+   brw_set_src1(insn, brw_imm_d(0));
+
+   if (BRW_IS_IGDNG(brw)) {
+      insn->bits3.sampler_igdng.binding_table_index = binding_table_index;
+      insn->bits3.sampler_igdng.sampler = sampler;
+      insn->bits3.sampler_igdng.msg_type = msg_type;
+      insn->bits3.sampler_igdng.simd_mode = simd_mode;
+      insn->bits3.sampler_igdng.header_present = header_present;
+      insn->bits3.sampler_igdng.response_length = response_length;
+      insn->bits3.sampler_igdng.msg_length = msg_length;
+      insn->bits3.sampler_igdng.end_of_thread = eot;
+      insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_SAMPLER;
+      insn->bits2.send_igdng.end_of_thread = eot;
+   } else if (BRW_IS_G4X(brw)) {
+      insn->bits3.sampler_g4x.binding_table_index = binding_table_index;
+      insn->bits3.sampler_g4x.sampler = sampler;
+      insn->bits3.sampler_g4x.msg_type = msg_type;
+      insn->bits3.sampler_g4x.response_length = response_length;
+      insn->bits3.sampler_g4x.msg_length = msg_length;
+      insn->bits3.sampler_g4x.end_of_thread = eot;
+      insn->bits3.sampler_g4x.msg_target = BRW_MESSAGE_TARGET_SAMPLER;
+   } else {
+      insn->bits3.sampler.binding_table_index = binding_table_index;
+      insn->bits3.sampler.sampler = sampler;
+      insn->bits3.sampler.msg_type = msg_type;
+      insn->bits3.sampler.return_format = BRW_SAMPLER_RETURN_FORMAT_FLOAT32;
+      insn->bits3.sampler.response_length = response_length;
+      insn->bits3.sampler.msg_length = msg_length;
+      insn->bits3.sampler.end_of_thread = eot;
+      insn->bits3.sampler.msg_target = BRW_MESSAGE_TARGET_SAMPLER;
+   }
+}
+
+
+
+static struct brw_instruction *next_insn( struct brw_compile *p, 
+                                         GLuint opcode )
+{
+   struct brw_instruction *insn;
+
+   if (0 && (BRW_DEBUG & DEBUG_DISASSEM))
+   {
+      if (p->nr_insn) 
+         brw_disasm_insn(stderr, &p->store[p->nr_insn-1]);
+   }
+
+   assert(p->nr_insn + 1 < BRW_EU_MAX_INSN);
+
+   insn = &p->store[p->nr_insn++];
+   memcpy(insn, p->current, sizeof(*insn));
+
+   /* Reset this one-shot flag: 
+    */
+
+   if (p->current->header.destreg__conditionalmod) {
+      p->current->header.destreg__conditionalmod = 0;
+      p->current->header.predicate_control = BRW_PREDICATE_NORMAL;
+   }
+
+   insn->header.opcode = opcode;
+   return insn;
+}
+
+
+static struct brw_instruction *brw_alu1( struct brw_compile *p,
+                                        GLuint opcode,
+                                        struct brw_reg dest,
+                                        struct brw_reg src )
+{
+   struct brw_instruction *insn = next_insn(p, opcode);
+   brw_set_dest(insn, dest);
+   brw_set_src0(insn, src);   
+   return insn;
+}
+
+static struct brw_instruction *brw_alu2(struct brw_compile *p,
+                                       GLuint opcode,
+                                       struct brw_reg dest,
+                                       struct brw_reg src0,
+                                       struct brw_reg src1 )
+{
+   struct brw_instruction *insn = next_insn(p, opcode);   
+   brw_set_dest(insn, dest);
+   brw_set_src0(insn, src0);
+   brw_set_src1(insn, src1);
+   return insn;
+}
+
+
+/***********************************************************************
+ * Convenience routines.
+ */
+#define ALU1(OP)                                       \
+struct brw_instruction *brw_##OP(struct brw_compile *p,        \
+             struct brw_reg dest,                      \
+             struct brw_reg src0)                      \
+{                                                      \
+   return brw_alu1(p, BRW_OPCODE_##OP, dest, src0);            \
+}
+
+#define ALU2(OP)                                       \
+struct brw_instruction *brw_##OP(struct brw_compile *p,        \
+             struct brw_reg dest,                      \
+             struct brw_reg src0,                      \
+             struct brw_reg src1)                      \
+{                                                      \
+   return brw_alu2(p, BRW_OPCODE_##OP, dest, src0, src1);      \
+}
+
+
+ALU1(MOV)
+ALU2(SEL)
+ALU1(NOT)
+ALU2(AND)
+ALU2(OR)
+ALU2(XOR)
+ALU2(SHR)
+ALU2(SHL)
+ALU2(RSR)
+ALU2(RSL)
+ALU2(ASR)
+ALU2(ADD)
+ALU2(MUL)
+ALU1(FRC)
+ALU1(RNDD)
+ALU1(RNDZ)
+ALU2(MAC)
+ALU2(MACH)
+ALU1(LZD)
+ALU2(DP4)
+ALU2(DPH)
+ALU2(DP3)
+ALU2(DP2)
+ALU2(LINE)
+
+
+
+
+void brw_NOP(struct brw_compile *p)
+{
+   struct brw_instruction *insn = next_insn(p, BRW_OPCODE_NOP);   
+   brw_set_dest(insn, retype(brw_vec4_grf(0,0), BRW_REGISTER_TYPE_UD));
+   brw_set_src0(insn, retype(brw_vec4_grf(0,0), BRW_REGISTER_TYPE_UD));
+   brw_set_src1(insn, brw_imm_ud(0x0));
+}
+
+
+
+
+
+/***********************************************************************
+ * Comparisons, if/else/endif
+ */
+
+struct brw_instruction *brw_JMPI(struct brw_compile *p, 
+                                 struct brw_reg dest,
+                                 struct brw_reg src0,
+                                 struct brw_reg src1)
+{
+   struct brw_instruction *insn = brw_alu2(p, BRW_OPCODE_JMPI, dest, src0, src1);
+
+   insn->header.execution_size = 1;
+   insn->header.compression_control = BRW_COMPRESSION_NONE;
+   insn->header.mask_control = BRW_MASK_DISABLE;
+
+   p->current->header.predicate_control = BRW_PREDICATE_NONE;
+
+   return insn;
+}
+
+/* EU takes the value from the flag register and pushes it onto some
+ * sort of a stack (presumably merging with any flag value already on
+ * the stack).  Within an if block, the flags at the top of the stack
+ * control execution on each channel of the unit, eg. on each of the
+ * 16 pixel values in our wm programs.
+ *
+ * When the matching 'else' instruction is reached (presumably by
+ * countdown of the instruction count patched in by our ELSE/ENDIF
+ * functions), the relevent flags are inverted.
+ *
+ * When the matching 'endif' instruction is reached, the flags are
+ * popped off.  If the stack is now empty, normal execution resumes.
+ *
+ * No attempt is made to deal with stack overflow (14 elements?).
+ */
+struct brw_instruction *brw_IF(struct brw_compile *p, GLuint execute_size)
+{
+   struct brw_instruction *insn;
+
+   if (p->single_program_flow) {
+      assert(execute_size == BRW_EXECUTE_1);
+
+      insn = next_insn(p, BRW_OPCODE_ADD);
+      insn->header.predicate_inverse = 1;
+   } else {
+      insn = next_insn(p, BRW_OPCODE_IF);
+   }
+
+   /* Override the defaults for this instruction:
+    */
+   brw_set_dest(insn, brw_ip_reg());
+   brw_set_src0(insn, brw_ip_reg());
+   brw_set_src1(insn, brw_imm_d(0x0));
+
+   insn->header.execution_size = execute_size;
+   insn->header.compression_control = BRW_COMPRESSION_NONE;
+   insn->header.predicate_control = BRW_PREDICATE_NORMAL;
+   insn->header.mask_control = BRW_MASK_ENABLE;
+   if (!p->single_program_flow)
+       insn->header.thread_control = BRW_THREAD_SWITCH;
+
+   p->current->header.predicate_control = BRW_PREDICATE_NONE;
+
+   return insn;
+}
+
+
+struct brw_instruction *brw_ELSE(struct brw_compile *p, 
+                                struct brw_instruction *if_insn)
+{
+   struct brw_instruction *insn;
+   GLuint br = 1;
+
+   if (BRW_IS_IGDNG(p->brw))
+      br = 2;
+
+   if (p->single_program_flow) {
+      insn = next_insn(p, BRW_OPCODE_ADD);
+   } else {
+      insn = next_insn(p, BRW_OPCODE_ELSE);
+   }
+
+   brw_set_dest(insn, brw_ip_reg());
+   brw_set_src0(insn, brw_ip_reg());
+   brw_set_src1(insn, brw_imm_d(0x0));
+
+   insn->header.compression_control = BRW_COMPRESSION_NONE;
+   insn->header.execution_size = if_insn->header.execution_size;
+   insn->header.mask_control = BRW_MASK_ENABLE;
+   if (!p->single_program_flow)
+       insn->header.thread_control = BRW_THREAD_SWITCH;
+
+   /* Patch the if instruction to point at this instruction.
+    */
+   if (p->single_program_flow) {
+      assert(if_insn->header.opcode == BRW_OPCODE_ADD);
+
+      if_insn->bits3.ud = (insn - if_insn + 1) * 16;
+   } else {
+      assert(if_insn->header.opcode == BRW_OPCODE_IF);
+
+      if_insn->bits3.if_else.jump_count = br * (insn - if_insn);
+      if_insn->bits3.if_else.pop_count = 0;
+      if_insn->bits3.if_else.pad0 = 0;
+   }
+
+   return insn;
+}
+
+void brw_ENDIF(struct brw_compile *p, 
+              struct brw_instruction *patch_insn)
+{
+   GLuint br = 1;
+
+   if (BRW_IS_IGDNG(p->brw))
+      br = 2; 
+   if (p->single_program_flow) {
+      /* In single program flow mode, there's no need to execute an ENDIF,
+       * since we don't need to do any stack operations, and if we're executing
+       * currently, we want to just continue executing.
+       */
+      struct brw_instruction *next = &p->store[p->nr_insn];
+
+      assert(patch_insn->header.opcode == BRW_OPCODE_ADD);
+
+      patch_insn->bits3.ud = (next - patch_insn) * 16;
+   } else {
+      struct brw_instruction *insn = next_insn(p, BRW_OPCODE_ENDIF);
+
+      brw_set_dest(insn, retype(brw_vec4_grf(0,0), BRW_REGISTER_TYPE_UD));
+      brw_set_src0(insn, retype(brw_vec4_grf(0,0), BRW_REGISTER_TYPE_UD));
+      brw_set_src1(insn, brw_imm_d(0x0));
+
+      insn->header.compression_control = BRW_COMPRESSION_NONE;
+      insn->header.execution_size = patch_insn->header.execution_size;
+      insn->header.mask_control = BRW_MASK_ENABLE;
+      insn->header.thread_control = BRW_THREAD_SWITCH;
+
+      assert(patch_insn->bits3.if_else.jump_count == 0);
+
+      /* Patch the if or else instructions to point at this or the next
+       * instruction respectively.
+       */
+      if (patch_insn->header.opcode == BRW_OPCODE_IF) {
+        /* Automagically turn it into an IFF:
+         */
+        patch_insn->header.opcode = BRW_OPCODE_IFF;
+        patch_insn->bits3.if_else.jump_count = br * (insn - patch_insn + 1);
+        patch_insn->bits3.if_else.pop_count = 0;
+        patch_insn->bits3.if_else.pad0 = 0;
+      } else if (patch_insn->header.opcode == BRW_OPCODE_ELSE) {
+        patch_insn->bits3.if_else.jump_count = br * (insn - patch_insn + 1);
+        patch_insn->bits3.if_else.pop_count = 1;
+        patch_insn->bits3.if_else.pad0 = 0;
+      } else {
+        assert(0);
+      }
+
+      /* Also pop item off the stack in the endif instruction:
+       */
+      insn->bits3.if_else.jump_count = 0;
+      insn->bits3.if_else.pop_count = 1;
+      insn->bits3.if_else.pad0 = 0;
+   }
+}
+
+struct brw_instruction *brw_BREAK(struct brw_compile *p)
+{
+   struct brw_instruction *insn;
+   insn = next_insn(p, BRW_OPCODE_BREAK);
+   brw_set_dest(insn, brw_ip_reg());
+   brw_set_src0(insn, brw_ip_reg());
+   brw_set_src1(insn, brw_imm_d(0x0));
+   insn->header.compression_control = BRW_COMPRESSION_NONE;
+   insn->header.execution_size = BRW_EXECUTE_8;
+   /* insn->header.mask_control = BRW_MASK_DISABLE; */
+   insn->bits3.if_else.pad0 = 0;
+   return insn;
+}
+
+struct brw_instruction *brw_CONT(struct brw_compile *p)
+{
+   struct brw_instruction *insn;
+   insn = next_insn(p, BRW_OPCODE_CONTINUE);
+   brw_set_dest(insn, brw_ip_reg());
+   brw_set_src0(insn, brw_ip_reg());
+   brw_set_src1(insn, brw_imm_d(0x0));
+   insn->header.compression_control = BRW_COMPRESSION_NONE;
+   insn->header.execution_size = BRW_EXECUTE_8;
+   /* insn->header.mask_control = BRW_MASK_DISABLE; */
+   insn->bits3.if_else.pad0 = 0;
+   return insn;
+}
+
+/* DO/WHILE loop:
+ */
+struct brw_instruction *brw_DO(struct brw_compile *p, GLuint execute_size)
+{
+   if (p->single_program_flow) {
+      return &p->store[p->nr_insn];
+   } else {
+      struct brw_instruction *insn = next_insn(p, BRW_OPCODE_DO);
+
+      /* Override the defaults for this instruction:
+       */
+      brw_set_dest(insn, brw_null_reg());
+      brw_set_src0(insn, brw_null_reg());
+      brw_set_src1(insn, brw_null_reg());
+
+      insn->header.compression_control = BRW_COMPRESSION_NONE;
+      insn->header.execution_size = execute_size;
+      insn->header.predicate_control = BRW_PREDICATE_NONE;
+      /* insn->header.mask_control = BRW_MASK_ENABLE; */
+      /* insn->header.mask_control = BRW_MASK_DISABLE; */
+
+      return insn;
+   }
+}
+
+
+
+struct brw_instruction *brw_WHILE(struct brw_compile *p, 
+                                  struct brw_instruction *do_insn)
+{
+   struct brw_instruction *insn;
+   GLuint br = 1;
+
+   if (BRW_IS_IGDNG(p->brw))
+      br = 2;
+
+   if (p->single_program_flow)
+      insn = next_insn(p, BRW_OPCODE_ADD);
+   else
+      insn = next_insn(p, BRW_OPCODE_WHILE);
+
+   brw_set_dest(insn, brw_ip_reg());
+   brw_set_src0(insn, brw_ip_reg());
+   brw_set_src1(insn, brw_imm_d(0x0));
+
+   insn->header.compression_control = BRW_COMPRESSION_NONE;
+
+   if (p->single_program_flow) {
+      insn->header.execution_size = BRW_EXECUTE_1;
+
+      insn->bits3.d = (do_insn - insn) * 16;
+   } else {
+      insn->header.execution_size = do_insn->header.execution_size;
+
+      assert(do_insn->header.opcode == BRW_OPCODE_DO);
+      insn->bits3.if_else.jump_count = br * (do_insn - insn + 1);
+      insn->bits3.if_else.pop_count = 0;
+      insn->bits3.if_else.pad0 = 0;
+   }
+
+/*    insn->header.mask_control = BRW_MASK_ENABLE; */
+
+   /* insn->header.mask_control = BRW_MASK_DISABLE; */
+   p->current->header.predicate_control = BRW_PREDICATE_NONE;   
+   return insn;
+}
+
+
+/* FORWARD JUMPS:
+ */
+void brw_land_fwd_jump(struct brw_compile *p, 
+                      struct brw_instruction *jmp_insn)
+{
+   struct brw_instruction *landing = &p->store[p->nr_insn];
+   GLuint jmpi = 1;
+
+   if (BRW_IS_IGDNG(p->brw))
+       jmpi = 2;
+
+   assert(jmp_insn->header.opcode == BRW_OPCODE_JMPI);
+   assert(jmp_insn->bits1.da1.src1_reg_file = BRW_IMMEDIATE_VALUE);
+
+   jmp_insn->bits3.ud = jmpi * ((landing - jmp_insn) - 1);
+}
+
+
+
+/* To integrate with the above, it makes sense that the comparison
+ * instruction should populate the flag register.  It might be simpler
+ * just to use the flag reg for most WM tasks?
+ */
+void brw_CMP(struct brw_compile *p,
+            struct brw_reg dest,
+            GLuint conditional,
+            struct brw_reg src0,
+            struct brw_reg src1)
+{
+   struct brw_instruction *insn = next_insn(p, BRW_OPCODE_CMP);
+
+   insn->header.destreg__conditionalmod = conditional;
+   brw_set_dest(insn, dest);
+   brw_set_src0(insn, src0);
+   brw_set_src1(insn, src1);
+
+/*    guess_execution_size(insn, src0); */
+
+
+   /* Make it so that future instructions will use the computed flag
+    * value until brw_set_predicate_control_flag_value() is called
+    * again.  
+    */
+   if (dest.file == BRW_ARCHITECTURE_REGISTER_FILE &&
+       dest.nr == 0) {
+      p->current->header.predicate_control = BRW_PREDICATE_NORMAL;
+      p->flag_value = 0xff;
+   }
+}
+
+
+
+/***********************************************************************
+ * Helpers for the various SEND message types:
+ */
+
+/** Extended math function, float[8].
+ */
+void brw_math( struct brw_compile *p,
+              struct brw_reg dest,
+              GLuint function,
+              GLuint saturate,
+              GLuint msg_reg_nr,
+              struct brw_reg src,
+              GLuint data_type,
+              GLuint precision )
+{
+   struct brw_instruction *insn = next_insn(p, BRW_OPCODE_SEND);
+   GLuint msg_length = (function == BRW_MATH_FUNCTION_POW) ? 2 : 1; 
+   GLuint response_length = (function == BRW_MATH_FUNCTION_SINCOS) ? 2 : 1; 
+
+   /* Example code doesn't set predicate_control for send
+    * instructions.
+    */
+   insn->header.predicate_control = 0; 
+   insn->header.destreg__conditionalmod = msg_reg_nr;
+
+   brw_set_dest(insn, dest);
+   brw_set_src0(insn, src);
+   brw_set_math_message(p->brw,
+                       insn, 
+                       msg_length, response_length, 
+                       function,
+                       BRW_MATH_INTEGER_UNSIGNED,
+                       precision,
+                       saturate,
+                       data_type);
+}
+
+/**
+ * Extended math function, float[16].
+ * Use 2 send instructions.
+ */
+void brw_math_16( struct brw_compile *p,
+                 struct brw_reg dest,
+                 GLuint function,
+                 GLuint saturate,
+                 GLuint msg_reg_nr,
+                 struct brw_reg src,
+                 GLuint precision )
+{
+   struct brw_instruction *insn;
+   GLuint msg_length = (function == BRW_MATH_FUNCTION_POW) ? 2 : 1; 
+   GLuint response_length = (function == BRW_MATH_FUNCTION_SINCOS) ? 2 : 1; 
+
+   /* First instruction:
+    */
+   brw_push_insn_state(p);
+   brw_set_predicate_control_flag_value(p, 0xff);
+   brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+
+   insn = next_insn(p, BRW_OPCODE_SEND);
+   insn->header.destreg__conditionalmod = msg_reg_nr;
+
+   brw_set_dest(insn, dest);
+   brw_set_src0(insn, src);
+   brw_set_math_message(p->brw,
+                       insn, 
+                       msg_length, response_length, 
+                       function,
+                       BRW_MATH_INTEGER_UNSIGNED,
+                       precision,
+                       saturate,
+                       BRW_MATH_DATA_VECTOR);
+
+   /* Second instruction:
+    */
+   insn = next_insn(p, BRW_OPCODE_SEND);
+   insn->header.compression_control = BRW_COMPRESSION_2NDHALF;
+   insn->header.destreg__conditionalmod = msg_reg_nr+1;
+
+   brw_set_dest(insn, offset(dest,1));
+   brw_set_src0(insn, src);
+   brw_set_math_message(p->brw, 
+                       insn, 
+                       msg_length, response_length, 
+                       function,
+                       BRW_MATH_INTEGER_UNSIGNED,
+                       precision,
+                       saturate,
+                       BRW_MATH_DATA_VECTOR);
+
+   brw_pop_insn_state(p);
+}
+
+
+/**
+ * Write block of 16 dwords/floats to the data port Render Cache scratch buffer.
+ * Scratch offset should be a multiple of 64.
+ * Used for register spilling.
+ */
+void brw_dp_WRITE_16( struct brw_compile *p,
+                     struct brw_reg src,
+                     GLuint scratch_offset )
+{
+   GLuint msg_reg_nr = 1;
+   {
+      brw_push_insn_state(p);
+      brw_set_mask_control(p, BRW_MASK_DISABLE);
+      brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+
+      /* set message header global offset field (reg 0, element 2) */
+      brw_MOV(p,
+             retype(brw_vec1_grf(0, 2), BRW_REGISTER_TYPE_D),
+             brw_imm_d(scratch_offset));
+
+      brw_pop_insn_state(p);
+   }
+
+   {
+      GLuint msg_length = 3;
+      struct brw_reg dest = retype(brw_null_reg(), BRW_REGISTER_TYPE_UW);
+      struct brw_instruction *insn = next_insn(p, BRW_OPCODE_SEND);
+   
+      insn->header.predicate_control = 0; /* XXX */
+      insn->header.compression_control = BRW_COMPRESSION_NONE; 
+      insn->header.destreg__conditionalmod = msg_reg_nr;
+  
+      brw_set_dest(insn, dest);
+      brw_set_src0(insn, src);
+
+      brw_set_dp_write_message(p->brw,
+                              insn,
+                              255, /* binding table index (255=stateless) */
+                              BRW_DATAPORT_OWORD_BLOCK_4_OWORDS, /* msg_control */
+                              BRW_DATAPORT_WRITE_MESSAGE_OWORD_BLOCK_WRITE, /* msg_type */
+                              msg_length,
+                              0, /* pixel scoreboard */
+                              0, /* response_length */
+                              0); /* eot */
+   }
+}
+
+
+/**
+ * Read block of 16 dwords/floats from the data port Render Cache scratch buffer.
+ * Scratch offset should be a multiple of 64.
+ * Used for register spilling.
+ */
+void brw_dp_READ_16( struct brw_compile *p,
+                     struct brw_reg dest,
+                     GLuint scratch_offset )
+{
+   GLuint msg_reg_nr = 1;
+   {
+      brw_push_insn_state(p);
+      brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+      brw_set_mask_control(p, BRW_MASK_DISABLE);
+
+      /* set message header global offset field (reg 0, element 2) */
+      brw_MOV(p,
+             retype(brw_vec1_grf(0, 2), BRW_REGISTER_TYPE_D),
+             brw_imm_d(scratch_offset));
+
+      brw_pop_insn_state(p);
+   }
+
+   {
+      struct brw_instruction *insn = next_insn(p, BRW_OPCODE_SEND);
+   
+      insn->header.predicate_control = 0; /* XXX */
+      insn->header.compression_control = BRW_COMPRESSION_NONE; 
+      insn->header.destreg__conditionalmod = msg_reg_nr;
+  
+      brw_set_dest(insn, dest);        /* UW? */
+      brw_set_src0(insn, retype(brw_vec8_grf(0, 0), BRW_REGISTER_TYPE_UW));
+
+      brw_set_dp_read_message(p->brw,
+                             insn,
+                             255, /* binding table index (255=stateless) */
+                             3,  /* msg_control (3 means 4 Owords) */
+                             BRW_DATAPORT_READ_MESSAGE_OWORD_BLOCK_READ, /* msg_type */
+                             1, /* target cache (render/scratch) */
+                             1, /* msg_length */
+                             2, /* response_length */
+                             0); /* eot */
+   }
+}
+
+
+/**
+ * Read a float[4] vector from the data port Data Cache (const buffer).
+ * Location (in buffer) should be a multiple of 16.
+ * Used for fetching shader constants.
+ * If relAddr is true, we'll do an indirect fetch using the address register.
+ */
+void brw_dp_READ_4( struct brw_compile *p,
+                    struct brw_reg dest,
+                    GLboolean relAddr,
+                    GLuint location,
+                    GLuint bind_table_index )
+{
+   /* XXX: relAddr not implemented */
+   GLuint msg_reg_nr = 1;
+   {
+      struct brw_reg b;
+      brw_push_insn_state(p);
+      brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+      brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+      brw_set_mask_control(p, BRW_MASK_DISABLE);
+
+   /* Setup MRF[1] with location/offset into const buffer */
+      b = brw_message_reg(msg_reg_nr);
+      b = retype(b, BRW_REGISTER_TYPE_UD);
+      /* XXX I think we're setting all the dwords of MRF[1] to 'location'.
+       * when the docs say only dword[2] should be set.  Hmmm.  But it works.
+       */
+      brw_MOV(p, b, brw_imm_ud(location));
+      brw_pop_insn_state(p);
+   }
+
+   {
+      struct brw_instruction *insn = next_insn(p, BRW_OPCODE_SEND);
+   
+      insn->header.predicate_control = BRW_PREDICATE_NONE;
+      insn->header.compression_control = BRW_COMPRESSION_NONE; 
+      insn->header.destreg__conditionalmod = msg_reg_nr;
+      insn->header.mask_control = BRW_MASK_DISABLE;
+  
+      /* cast dest to a uword[8] vector */
+      dest = retype(vec8(dest), BRW_REGISTER_TYPE_UW);
+
+      brw_set_dest(insn, dest);
+      brw_set_src0(insn, brw_null_reg());
+
+      brw_set_dp_read_message(p->brw,
+                             insn,
+                             bind_table_index,
+                             0,  /* msg_control (0 means 1 Oword) */
+                             BRW_DATAPORT_READ_MESSAGE_OWORD_BLOCK_READ, /* msg_type */
+                             0, /* source cache = data cache */
+                             1, /* msg_length */
+                             1, /* response_length (1 Oword) */
+                             0); /* eot */
+   }
+}
+
+
+/**
+ * Read float[4] constant(s) from VS constant buffer.
+ * For relative addressing, two float[4] constants will be read into 'dest'.
+ * Otherwise, one float[4] constant will be read into the lower half of 'dest'.
+ */
+void brw_dp_READ_4_vs(struct brw_compile *p,
+                      struct brw_reg dest,
+                      GLuint oword,
+                      GLboolean relAddr,
+                      struct brw_reg addrReg,
+                      GLuint location,
+                      GLuint bind_table_index)
+{
+   GLuint msg_reg_nr = 1;
+
+   assert(oword < 2);
+   /*
+   printf("vs const read msg, location %u, msg_reg_nr %d\n",
+          location, msg_reg_nr);
+   */
+
+   /* Setup MRF[1] with location/offset into const buffer */
+   {
+      struct brw_reg b;
+
+      brw_push_insn_state(p);
+      brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+      brw_set_mask_control(p, BRW_MASK_DISABLE);
+      brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+      /*brw_set_access_mode(p, BRW_ALIGN_16);*/
+
+      /* XXX I think we're setting all the dwords of MRF[1] to 'location'.
+       * when the docs say only dword[2] should be set.  Hmmm.  But it works.
+       */
+      b = brw_message_reg(msg_reg_nr);
+      b = retype(b, BRW_REGISTER_TYPE_UD);
+      /*b = get_element_ud(b, 2);*/
+      if (relAddr) {
+         brw_ADD(p, b, addrReg, brw_imm_ud(location));
+      }
+      else {
+         brw_MOV(p, b, brw_imm_ud(location));
+      }
+
+      brw_pop_insn_state(p);
+   }
+
+   {
+      struct brw_instruction *insn = next_insn(p, BRW_OPCODE_SEND);
+   
+      insn->header.predicate_control = BRW_PREDICATE_NONE;
+      insn->header.compression_control = BRW_COMPRESSION_NONE; 
+      insn->header.destreg__conditionalmod = msg_reg_nr;
+      insn->header.mask_control = BRW_MASK_DISABLE;
+      /*insn->header.access_mode = BRW_ALIGN_16;*/
+  
+      brw_set_dest(insn, dest);
+      brw_set_src0(insn, brw_null_reg());
+
+      brw_set_dp_read_message(p->brw,
+                             insn,
+                             bind_table_index,
+                             oword,  /* 0 = lower Oword, 1 = upper Oword */
+                             BRW_DATAPORT_READ_MESSAGE_OWORD_BLOCK_READ, /* msg_type */
+                             0, /* source cache = data cache */
+                             1, /* msg_length */
+                             1, /* response_length (1 Oword) */
+                             0); /* eot */
+   }
+}
+
+
+
+void brw_fb_WRITE(struct brw_compile *p,
+                  struct brw_reg dest,
+                  GLuint msg_reg_nr,
+                  struct brw_reg src0,
+                  GLuint binding_table_index,
+                  GLuint msg_length,
+                  GLuint response_length,
+                  GLboolean eot)
+{
+   struct brw_instruction *insn = next_insn(p, BRW_OPCODE_SEND);
+   
+   insn->header.predicate_control = 0; /* XXX */
+   insn->header.compression_control = BRW_COMPRESSION_NONE; 
+   insn->header.destreg__conditionalmod = msg_reg_nr;
+  
+   brw_set_dest(insn, dest);
+   brw_set_src0(insn, src0);
+   brw_set_dp_write_message(p->brw,
+                           insn,
+                           binding_table_index,
+                           BRW_DATAPORT_RENDER_TARGET_WRITE_SIMD16_SINGLE_SOURCE, /* msg_control */
+                           BRW_DATAPORT_WRITE_MESSAGE_RENDER_TARGET_WRITE, /* msg_type */
+                           msg_length,
+                           1,  /* pixel scoreboard */
+                           response_length, 
+                           eot);
+}
+
+
+/**
+ * Texture sample instruction.
+ * Note: the msg_type plus msg_length values determine exactly what kind
+ * of sampling operation is performed.  See volume 4, page 161 of docs.
+ */
+void brw_SAMPLE(struct brw_compile *p,
+               struct brw_reg dest,
+               GLuint msg_reg_nr,
+               struct brw_reg src0,
+               GLuint binding_table_index,
+               GLuint sampler,
+               GLuint writemask,
+               GLuint msg_type,
+               GLuint response_length,
+               GLuint msg_length,
+               GLboolean eot,
+               GLuint header_present,
+               GLuint simd_mode)
+{
+   GLboolean need_stall = 0;
+   
+   if (writemask == 0) {
+      /*debug_printf("%s: zero writemask??\n", __FUNCTION__); */
+      return;
+   }
+   
+   /* Hardware doesn't do destination dependency checking on send
+    * instructions properly.  Add a workaround which generates the
+    * dependency by other means.  In practice it seems like this bug
+    * only crops up for texture samples, and only where registers are
+    * written by the send and then written again later without being
+    * read in between.  Luckily for us, we already track that
+    * information and use it to modify the writemask for the
+    * instruction, so that is a guide for whether a workaround is
+    * needed.
+    */
+   if (writemask != BRW_WRITEMASK_XYZW) {
+      GLuint dst_offset = 0;
+      GLuint i, newmask = 0, len = 0;
+
+      for (i = 0; i < 4; i++) {
+        if (writemask & (1<<i))
+           break;
+        dst_offset += 2;
+      }
+      for (; i < 4; i++) {
+        if (!(writemask & (1<<i)))
+           break;
+        newmask |= 1<<i;
+        len++;
+      }
+
+      if (newmask != writemask) {
+        need_stall = 1;
+         /* debug_printf("need stall %x %x\n", newmask , writemask); */
+      }
+      else {
+        struct brw_reg m1 = brw_message_reg(msg_reg_nr);
+        
+        newmask = ~newmask & BRW_WRITEMASK_XYZW;
+
+        brw_push_insn_state(p);
+
+        brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+        brw_set_mask_control(p, BRW_MASK_DISABLE);
+
+        brw_MOV(p, m1, brw_vec8_grf(0,0));      
+        brw_MOV(p, get_element_ud(m1, 2), brw_imm_ud(newmask << 12)); 
+
+        brw_pop_insn_state(p);
+
+        src0 = retype(brw_null_reg(), BRW_REGISTER_TYPE_UW); 
+        dest = offset(dest, dst_offset);
+        response_length = len * 2;
+      }
+   }
+
+   {
+      struct brw_instruction *insn = next_insn(p, BRW_OPCODE_SEND);
+   
+      insn->header.predicate_control = 0; /* XXX */
+      insn->header.compression_control = BRW_COMPRESSION_NONE;
+      insn->header.destreg__conditionalmod = msg_reg_nr;
+
+      brw_set_dest(insn, dest);
+      brw_set_src0(insn, src0);
+      brw_set_sampler_message(p->brw, insn,
+                             binding_table_index,
+                             sampler,
+                             msg_type,
+                             response_length, 
+                             msg_length,
+                             eot,
+                             header_present,
+                             simd_mode);
+   }
+
+   if (need_stall) {
+      struct brw_reg reg = vec8(offset(dest, response_length-1));
+
+      /*  mov (8) r9.0<1>:f    r9.0<8;8,1>:f    { Align1 }
+       */
+      brw_push_insn_state(p);
+      brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+      brw_MOV(p, reg, reg);          
+      brw_pop_insn_state(p);
+   }
+
+}
+
+/* All these variables are pretty confusing - we might be better off
+ * using bitmasks and macros for this, in the old style.  Or perhaps
+ * just having the caller instantiate the fields in dword3 itself.
+ */
+void brw_urb_WRITE(struct brw_compile *p,
+                  struct brw_reg dest,
+                  GLuint msg_reg_nr,
+                  struct brw_reg src0,
+                  GLboolean allocate,
+                  GLboolean used,
+                  GLuint msg_length,
+                  GLuint response_length,
+                  GLboolean eot,
+                  GLboolean writes_complete,
+                  GLuint offset,
+                  GLuint swizzle)
+{
+   struct brw_instruction *insn = next_insn(p, BRW_OPCODE_SEND);
+
+   assert(msg_length < BRW_MAX_MRF);
+
+   brw_set_dest(insn, dest);
+   brw_set_src0(insn, src0);
+   brw_set_src1(insn, brw_imm_d(0));
+
+   insn->header.destreg__conditionalmod = msg_reg_nr;
+
+   brw_set_urb_message(p->brw,
+                      insn,
+                      allocate,
+                      used,
+                      msg_length,
+                      response_length, 
+                      eot, 
+                      writes_complete, 
+                      offset,
+                      swizzle);
+}
+
+void brw_ff_sync(struct brw_compile *p,
+                  struct brw_reg dest,
+                  GLuint msg_reg_nr,
+                  struct brw_reg src0,
+                  GLboolean allocate,
+                  GLboolean used,
+                  GLuint msg_length,
+                  GLuint response_length,
+                  GLboolean eot,
+                  GLboolean writes_complete,
+                  GLuint offset,
+                  GLuint swizzle)
+{
+   struct brw_instruction *insn = next_insn(p, BRW_OPCODE_SEND);
+
+   assert(msg_length < 16);
+
+   brw_set_dest(insn, dest);
+   brw_set_src0(insn, src0);
+   brw_set_src1(insn, brw_imm_d(0));
+
+   insn->header.destreg__conditionalmod = msg_reg_nr;
+
+   brw_set_ff_sync_message(p->brw,
+                      insn,
+                      allocate,
+                      used,
+                      msg_length,
+                      response_length, 
+                      eot, 
+                      writes_complete, 
+                      offset,
+                      swizzle);
+}
diff --git a/src/gallium/drivers/i965/brw_eu_util.c b/src/gallium/drivers/i965/brw_eu_util.c
new file mode 100644 (file)
index 0000000..5405cf1
--- /dev/null
@@ -0,0 +1,126 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+      
+
+#include "brw_context.h"
+#include "brw_defines.h"
+#include "brw_eu.h"
+
+
+void brw_math_invert( struct brw_compile *p, 
+                            struct brw_reg dst,
+                            struct brw_reg src)
+{
+   brw_math( p, 
+            dst,
+            BRW_MATH_FUNCTION_INV, 
+            BRW_MATH_SATURATE_NONE,
+            0,
+            src,
+            BRW_MATH_PRECISION_FULL, 
+            BRW_MATH_DATA_VECTOR );
+}
+
+
+
+void brw_copy4(struct brw_compile *p,
+              struct brw_reg dst,
+              struct brw_reg src,
+              GLuint count)
+{
+   GLuint i;
+
+   dst = vec4(dst);
+   src = vec4(src);
+
+   for (i = 0; i < count; i++)
+   {
+      GLuint delta = i*32;
+      brw_MOV(p, byte_offset(dst, delta),    byte_offset(src, delta));
+      brw_MOV(p, byte_offset(dst, delta+16), byte_offset(src, delta+16));
+   }
+}
+
+
+void brw_copy8(struct brw_compile *p,
+              struct brw_reg dst,
+              struct brw_reg src,
+              GLuint count)
+{
+   GLuint i;
+
+   dst = vec8(dst);
+   src = vec8(src);
+
+   for (i = 0; i < count; i++)
+   {
+      GLuint delta = i*32;
+      brw_MOV(p, byte_offset(dst, delta),    byte_offset(src, delta));
+   }
+}
+
+
+void brw_copy_indirect_to_indirect(struct brw_compile *p,
+                                  struct brw_indirect dst_ptr,
+                                  struct brw_indirect src_ptr,
+                                  GLuint count)
+{
+   GLuint i;
+
+   for (i = 0; i < count; i++)
+   {
+      GLuint delta = i*32;
+      brw_MOV(p, deref_4f(dst_ptr, delta),    deref_4f(src_ptr, delta));
+      brw_MOV(p, deref_4f(dst_ptr, delta+16), deref_4f(src_ptr, delta+16));
+   }
+}
+
+
+void brw_copy_from_indirect(struct brw_compile *p,
+                           struct brw_reg dst,
+                           struct brw_indirect ptr,
+                           GLuint count)
+{
+   GLuint i;
+
+   dst = vec4(dst);
+
+   for (i = 0; i < count; i++)
+   {
+      GLuint delta = i*32;
+      brw_MOV(p, byte_offset(dst, delta),    deref_4f(ptr, delta));
+      brw_MOV(p, byte_offset(dst, delta+16), deref_4f(ptr, delta+16));
+   }
+}
+
+
+
+
diff --git a/src/gallium/drivers/i965/brw_gs.c b/src/gallium/drivers/i965/brw_gs.c
new file mode 100644 (file)
index 0000000..921b201
--- /dev/null
@@ -0,0 +1,216 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+      
+#include "brw_batchbuffer.h"
+
+#include "brw_defines.h"
+#include "brw_context.h"
+#include "brw_eu.h"
+#include "brw_util.h"
+#include "brw_state.h"
+#include "brw_gs.h"
+
+
+
+static enum pipe_error compile_gs_prog( struct brw_context *brw,
+                                        struct brw_gs_prog_key *key,
+                                        struct brw_winsys_buffer **bo_out )
+{
+   struct brw_gs_compile c;
+   enum pipe_error ret;
+   const GLuint *program;
+   GLuint program_size;
+
+   memset(&c, 0, sizeof(c));
+   
+   c.key = *key;
+   c.need_ff_sync = BRW_IS_IGDNG(brw);
+   /* Need to locate the two positions present in vertex + header.
+    * These are currently hardcoded:
+    */
+   c.nr_attrs = c.key.nr_attrs;
+
+   if (BRW_IS_IGDNG(brw))
+      c.nr_regs = (c.nr_attrs + 1) / 2 + 3;  /* are vertices packed, or reg-aligned? */
+   else
+      c.nr_regs = (c.nr_attrs + 1) / 2 + 1;  /* are vertices packed, or reg-aligned? */
+
+   c.nr_bytes = c.nr_regs * REG_SIZE;
+
+   
+   /* Begin the compilation:
+    */
+   brw_init_compile(brw, &c.func);
+
+   c.func.single_program_flow = 1;
+
+   /* For some reason the thread is spawned with only 4 channels
+    * unmasked.  
+    */
+   brw_set_mask_control(&c.func, BRW_MASK_DISABLE);
+
+
+   /* Note that primitives which don't require a GS program have
+    * already been weeded out by this stage:
+    */
+   switch (key->primitive) {
+   case PIPE_PRIM_QUADS:
+      brw_gs_quads( &c ); 
+      break;
+   case PIPE_PRIM_QUAD_STRIP:
+      brw_gs_quad_strip( &c );
+      break;
+   case PIPE_PRIM_LINE_LOOP:
+      brw_gs_lines( &c );
+      break;
+   case PIPE_PRIM_LINES:
+      if (key->hint_gs_always)
+        brw_gs_lines( &c );
+      else {
+        return PIPE_OK;
+      }
+      break;
+   case PIPE_PRIM_TRIANGLES:
+      if (key->hint_gs_always)
+        brw_gs_tris( &c );
+      else {
+        return PIPE_OK;
+      }
+      break;
+   case PIPE_PRIM_POINTS:
+      if (key->hint_gs_always)
+        brw_gs_points( &c );
+      else {
+        return PIPE_OK;
+      }
+      break;
+   default:
+      assert(0);
+      return PIPE_ERROR_BAD_INPUT;
+   }
+
+   /* get the program
+    */
+   ret = brw_get_program(&c.func, &program, &program_size);
+   if (ret)
+      return ret;
+
+   /* Upload
+    */
+   ret = brw_upload_cache( &brw->cache, BRW_GS_PROG,
+                           &c.key, sizeof(c.key),
+                           NULL, 0,
+                           program, program_size,
+                           &c.prog_data,
+                           &brw->gs.prog_data,
+                           bo_out );
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+static const unsigned gs_prim[PIPE_PRIM_MAX] = {  
+   PIPE_PRIM_POINTS,
+   PIPE_PRIM_LINES,
+   PIPE_PRIM_LINE_LOOP,
+   PIPE_PRIM_LINES,
+   PIPE_PRIM_TRIANGLES,
+   PIPE_PRIM_TRIANGLES,
+   PIPE_PRIM_TRIANGLES,
+   PIPE_PRIM_QUADS,
+   PIPE_PRIM_QUAD_STRIP,
+   PIPE_PRIM_TRIANGLES
+};
+
+static void populate_key( struct brw_context *brw,
+                         struct brw_gs_prog_key *key )
+{
+   const struct brw_fs_signature *sig = &brw->curr.fragment_shader->signature;
+
+   memset(key, 0, sizeof(*key));
+
+   /* PIPE_NEW_FRAGMENT_SIGNATURE */
+   key->nr_attrs = sig->nr_inputs + 1;
+
+   /* BRW_NEW_PRIMITIVE */
+   key->primitive = gs_prim[brw->primitive];
+
+   key->hint_gs_always = 0;    /* debug code? */
+
+   key->need_gs_prog = (key->hint_gs_always ||
+                       brw->primitive == PIPE_PRIM_QUADS ||
+                       brw->primitive == PIPE_PRIM_QUAD_STRIP ||
+                       brw->primitive == PIPE_PRIM_LINE_LOOP);
+}
+
+/* Calculate interpolants for triangle and line rasterization.
+ */
+static int prepare_gs_prog(struct brw_context *brw)
+{
+   struct brw_gs_prog_key key;
+   enum pipe_error ret;
+
+   /* Populate the key:
+    */
+   populate_key(brw, &key);
+
+   if (brw->gs.prog_active != key.need_gs_prog) {
+      brw->state.dirty.cache |= CACHE_NEW_GS_PROG;
+      brw->gs.prog_active = key.need_gs_prog;
+   }
+
+   if (!brw->gs.prog_active)
+      return PIPE_OK;
+
+   if (brw_search_cache(&brw->cache, BRW_GS_PROG,
+                        &key, sizeof(key),
+                        NULL, 0,
+                        &brw->gs.prog_data,
+                        &brw->gs.prog_bo))
+      return PIPE_OK;
+
+   ret = compile_gs_prog( brw, &key, &brw->gs.prog_bo );
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+const struct brw_tracked_state brw_gs_prog = {
+   .dirty = {
+      .mesa  = PIPE_NEW_FRAGMENT_SIGNATURE,
+      .brw   = BRW_NEW_PRIMITIVE,
+      .cache = 0,
+   },
+   .prepare = prepare_gs_prog
+};
diff --git a/src/gallium/drivers/i965/brw_gs.h b/src/gallium/drivers/i965/brw_gs.h
new file mode 100644 (file)
index 0000000..6e616dc
--- /dev/null
@@ -0,0 +1,76 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#ifndef BRW_GS_H
+#define BRW_GS_H
+
+
+#include "brw_context.h"
+#include "brw_eu.h"
+
+#define MAX_GS_VERTS (4)            
+
+struct brw_gs_prog_key {
+   GLuint nr_attrs:8;
+   GLuint primitive:4;
+   GLuint hint_gs_always:1;
+   GLuint need_gs_prog:1;
+   GLuint pad:18;
+};
+
+struct brw_gs_compile {
+   struct brw_compile func;
+   struct brw_gs_prog_key key;
+   struct brw_gs_prog_data prog_data;
+   
+   struct {
+      struct brw_reg R0;
+      struct brw_reg vertex[MAX_GS_VERTS];
+   } reg;
+
+   /* 3 different ways of expressing vertex size:
+    */
+   GLuint nr_attrs;
+   GLuint nr_regs;
+   GLuint nr_bytes;
+   GLboolean need_ff_sync;
+};
+
+#define ATTR_SIZE  (4*4)
+
+void brw_gs_quads( struct brw_gs_compile *c );
+void brw_gs_quad_strip( struct brw_gs_compile *c );
+void brw_gs_tris( struct brw_gs_compile *c );
+void brw_gs_lines( struct brw_gs_compile *c );
+void brw_gs_points( struct brw_gs_compile *c );
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_gs_emit.c b/src/gallium/drivers/i965/brw_gs_emit.c
new file mode 100644 (file)
index 0000000..fd8e2ac
--- /dev/null
@@ -0,0 +1,181 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "brw_batchbuffer.h"
+
+#include "brw_defines.h"
+#include "brw_context.h"
+#include "brw_eu.h"
+#include "brw_util.h"
+#include "brw_gs.h"
+
+static void brw_gs_alloc_regs( struct brw_gs_compile *c,
+                              GLuint nr_verts )
+{
+   GLuint i = 0,j;
+
+   /* Register usage is static, precompute here:
+    */
+   c->reg.R0 = retype(brw_vec8_grf(i, 0), BRW_REGISTER_TYPE_UD); i++;
+
+   /* Payload vertices plus space for more generated vertices:
+    */
+   for (j = 0; j < nr_verts; j++) {
+      c->reg.vertex[j] = brw_vec4_grf(i, 0);
+      i += c->nr_regs;
+   }
+
+   c->prog_data.urb_read_length = c->nr_regs; 
+   c->prog_data.total_grf = i;
+}
+
+
+static void brw_gs_emit_vue(struct brw_gs_compile *c, 
+                           struct brw_reg vert,
+                           GLboolean last,
+                           GLuint header)
+{
+   struct brw_compile *p = &c->func;
+   GLboolean allocate = !last;
+
+   /* Overwrite PrimType and PrimStart in the message header, for
+    * each vertex in turn:
+    */
+   brw_MOV(p, get_element_ud(c->reg.R0, 2), brw_imm_ud(header));
+
+   /* Copy the vertex from vertn into m1..mN+1:
+    */
+   brw_copy8(p, brw_message_reg(1), vert, c->nr_regs);
+
+   /* Send each vertex as a seperate write to the urb.  This is
+    * different to the concept in brw_sf_emit.c, where subsequent
+    * writes are used to build up a single urb entry.  Each of these
+    * writes instantiates a seperate urb entry, and a new one must be
+    * allocated each time.
+    */
+   brw_urb_WRITE(p, 
+                allocate ? c->reg.R0 : retype(brw_null_reg(), BRW_REGISTER_TYPE_UD),
+                0,
+                c->reg.R0,
+                allocate,
+                1,             /* used */
+                c->nr_regs + 1, /* msg length */
+                allocate ? 1 : 0, /* response length */
+                allocate ? 0 : 1, /* eot */
+                1,             /* writes_complete */
+                0,             /* urb offset */
+                BRW_URB_SWIZZLE_NONE);
+}
+
+static void brw_gs_ff_sync(struct brw_gs_compile *c, int num_prim)
+{
+       struct brw_compile *p = &c->func;
+       brw_MOV(p, get_element_ud(c->reg.R0, 1), brw_imm_ud(num_prim));
+       brw_ff_sync(p, 
+                               c->reg.R0,
+                               0,
+                               c->reg.R0,
+                               1,      
+                               1,              /* used */
+                               1,      /* msg length */
+                               1,              /* response length */
+                               0,              /* eot */
+                               1,              /* write compelete */
+                               0,              /* urb offset */
+                               BRW_URB_SWIZZLE_NONE);
+}
+
+
+void brw_gs_quads( struct brw_gs_compile *c )
+{
+   brw_gs_alloc_regs(c, 4);
+   
+   /* Use polygons for correct edgeflag behaviour. Note that vertex 3
+    * is the PV for quads, but vertex 0 for polygons:
+    */
+   if (c->need_ff_sync)
+          brw_gs_ff_sync(c, 1);    
+   brw_gs_emit_vue(c, c->reg.vertex[3], 0, ((_3DPRIM_POLYGON << 2) | R02_PRIM_START));
+   brw_gs_emit_vue(c, c->reg.vertex[0], 0, (_3DPRIM_POLYGON << 2));
+   brw_gs_emit_vue(c, c->reg.vertex[1], 0, (_3DPRIM_POLYGON << 2)); 
+   brw_gs_emit_vue(c, c->reg.vertex[2], 1, ((_3DPRIM_POLYGON << 2) | R02_PRIM_END));
+}
+
+void brw_gs_quad_strip( struct brw_gs_compile *c )
+{
+   brw_gs_alloc_regs(c, 4);
+   
+   if (c->need_ff_sync)
+          brw_gs_ff_sync(c, 1);      
+   brw_gs_emit_vue(c, c->reg.vertex[2], 0, ((_3DPRIM_POLYGON << 2) | R02_PRIM_START));
+   brw_gs_emit_vue(c, c->reg.vertex[3], 0, (_3DPRIM_POLYGON << 2));
+   brw_gs_emit_vue(c, c->reg.vertex[0], 0, (_3DPRIM_POLYGON << 2)); 
+   brw_gs_emit_vue(c, c->reg.vertex[1], 1, ((_3DPRIM_POLYGON << 2) | R02_PRIM_END));
+}
+
+void brw_gs_tris( struct brw_gs_compile *c )
+{
+   brw_gs_alloc_regs(c, 3);
+
+   if (c->need_ff_sync)
+          brw_gs_ff_sync(c, 1);      
+   brw_gs_emit_vue(c, c->reg.vertex[0], 0, ((_3DPRIM_TRILIST << 2) | R02_PRIM_START));
+   brw_gs_emit_vue(c, c->reg.vertex[1], 0, (_3DPRIM_TRILIST << 2));
+   brw_gs_emit_vue(c, c->reg.vertex[2], 1, ((_3DPRIM_TRILIST << 2) | R02_PRIM_END));
+}
+
+void brw_gs_lines( struct brw_gs_compile *c )
+{
+   brw_gs_alloc_regs(c, 2);
+
+   if (c->need_ff_sync)
+          brw_gs_ff_sync(c, 1);      
+   brw_gs_emit_vue(c, c->reg.vertex[0], 0, ((_3DPRIM_LINESTRIP << 2) | R02_PRIM_START));
+   brw_gs_emit_vue(c, c->reg.vertex[1], 1, ((_3DPRIM_LINESTRIP << 2) | R02_PRIM_END));
+}
+
+void brw_gs_points( struct brw_gs_compile *c )
+{
+   brw_gs_alloc_regs(c, 1);
+
+   if (c->need_ff_sync)
+          brw_gs_ff_sync(c, 1);      
+   brw_gs_emit_vue(c, c->reg.vertex[0], 1, ((_3DPRIM_POINTLIST << 2) | R02_PRIM_START | R02_PRIM_END));
+}
+
+
+
+
+
+
+
+
diff --git a/src/gallium/drivers/i965/brw_gs_state.c b/src/gallium/drivers/i965/brw_gs_state.c
new file mode 100644 (file)
index 0000000..b64ec28
--- /dev/null
@@ -0,0 +1,169 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+#include "util/u_math.h"
+
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+#include "brw_debug.h"
+
+struct brw_gs_unit_key {
+   unsigned int total_grf;
+   unsigned int urb_entry_read_length;
+
+   unsigned int curbe_offset;
+
+   unsigned int nr_urb_entries, urb_size;
+   GLboolean prog_active;
+};
+
+static void
+gs_unit_populate_key(struct brw_context *brw, struct brw_gs_unit_key *key)
+{
+   memset(key, 0, sizeof(*key));
+
+   /* CACHE_NEW_GS_PROG */
+   key->prog_active = brw->gs.prog_active;
+   if (key->prog_active) {
+      key->total_grf = brw->gs.prog_data->total_grf;
+      key->urb_entry_read_length = brw->gs.prog_data->urb_read_length;
+   } else {
+      key->total_grf = 1;
+      key->urb_entry_read_length = 1;
+   }
+
+   /* BRW_NEW_CURBE_OFFSETS */
+   key->curbe_offset = brw->curbe.clip_start;
+
+   /* BRW_NEW_URB_FENCE */
+   key->nr_urb_entries = brw->urb.nr_gs_entries;
+   key->urb_size = brw->urb.vsize;
+}
+
+static enum pipe_error
+gs_unit_create_from_key(struct brw_context *brw, 
+                        struct brw_gs_unit_key *key,
+                        struct brw_winsys_reloc *reloc,
+                        unsigned nr_reloc,
+                        struct brw_winsys_buffer **bo_out)
+{
+   struct brw_gs_unit_state gs;
+   enum pipe_error ret;
+
+
+   memset(&gs, 0, sizeof(gs));
+
+   /* reloc */
+   gs.thread0.grf_reg_count = align(key->total_grf, 16) / 16 - 1;
+   gs.thread0.kernel_start_pointer = 0;
+
+   gs.thread1.floating_point_mode = BRW_FLOATING_POINT_NON_IEEE_754;
+   gs.thread1.single_program_flow = 1;
+
+   gs.thread3.dispatch_grf_start_reg = 1;
+   gs.thread3.const_urb_entry_read_offset = 0;
+   gs.thread3.const_urb_entry_read_length = 0;
+   gs.thread3.urb_entry_read_offset = 0;
+   gs.thread3.urb_entry_read_length = key->urb_entry_read_length;
+
+   gs.thread4.nr_urb_entries = key->nr_urb_entries;
+   gs.thread4.urb_entry_allocation_size = key->urb_size - 1;
+
+   if (key->nr_urb_entries >= 8)
+      gs.thread4.max_threads = 1;
+   else
+      gs.thread4.max_threads = 0;
+
+   if (BRW_IS_IGDNG(brw))
+      gs.thread4.rendering_enable = 1;
+
+   if (BRW_DEBUG & DEBUG_STATS)
+      gs.thread4.stats_enable = 1;
+
+   ret = brw_upload_cache(&brw->cache, BRW_GS_UNIT,
+                          key, sizeof(*key),
+                          reloc, nr_reloc,
+                          &gs, sizeof(gs),
+                          NULL, NULL,
+                          bo_out);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+static enum pipe_error prepare_gs_unit(struct brw_context *brw)
+{
+   struct brw_gs_unit_key key;
+   enum pipe_error ret;
+   struct brw_winsys_reloc reloc[1];
+   unsigned nr_reloc = 0;
+   unsigned grf_reg_count;
+
+   gs_unit_populate_key(brw, &key);
+
+   grf_reg_count = (align(key.total_grf, 16) / 16 - 1);
+
+   /* GS program relocation */
+   if (key.prog_active) {
+      make_reloc(&reloc[nr_reloc++],
+                 BRW_USAGE_STATE,
+                 grf_reg_count << 1,
+                 offsetof(struct brw_gs_unit_state, thread0),
+                 brw->gs.prog_bo);
+   }
+
+   if (brw_search_cache(&brw->cache, BRW_GS_UNIT,
+                        &key, sizeof(key),
+                        reloc, nr_reloc,
+                        NULL,
+                        &brw->gs.state_bo))
+      return PIPE_OK;
+
+   ret = gs_unit_create_from_key(brw, &key,
+                                 reloc, nr_reloc,
+                                 &brw->gs.state_bo);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+const struct brw_tracked_state brw_gs_unit = {
+   .dirty = {
+      .mesa  = 0,
+      .brw   = (BRW_NEW_CURBE_OFFSETS |
+               BRW_NEW_URB_FENCE),
+      .cache = CACHE_NEW_GS_PROG
+   },
+   .prepare = prepare_gs_unit,
+};
diff --git a/src/gallium/drivers/i965/brw_misc_state.c b/src/gallium/drivers/i965/brw_misc_state.c
new file mode 100644 (file)
index 0000000..e4b2422
--- /dev/null
@@ -0,0 +1,513 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+
+#include "brw_debug.h"
+#include "brw_batchbuffer.h"
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+#include "brw_screen.h"
+#include "brw_pipe_rast.h"
+
+
+
+
+
+/***********************************************************************
+ * Blend color
+ */
+
+static int upload_blend_constant_color(struct brw_context *brw)
+{
+   BRW_CACHED_BATCH_STRUCT(brw, &brw->curr.bcc);
+   return 0;
+}
+
+
+const struct brw_tracked_state brw_blend_constant_color = {
+   .dirty = {
+      .mesa = PIPE_NEW_BLEND_COLOR,
+      .brw = 0,
+      .cache = 0
+   },
+   .emit = upload_blend_constant_color
+};
+
+/***********************************************************************
+ * Drawing rectangle - framebuffer dimensions
+ */
+static int upload_drawing_rect(struct brw_context *brw)
+{
+   BEGIN_BATCH(4, NO_LOOP_CLIPRECTS);
+   OUT_BATCH(_3DSTATE_DRAWRECT_INFO_I965);
+   OUT_BATCH(0);
+   OUT_BATCH(((brw->curr.fb.width - 1) & 0xffff) |
+           ((brw->curr.fb.height - 1) << 16));
+   OUT_BATCH(0);
+   ADVANCE_BATCH();
+   return 0;
+}
+
+const struct brw_tracked_state brw_drawing_rect = {
+   .dirty = {
+      .mesa = PIPE_NEW_FRAMEBUFFER_DIMENSIONS,
+      .brw = 0,
+      .cache = 0
+   },
+   .emit = upload_drawing_rect
+};
+
+
+/***********************************************************************
+ * Binding table pointers
+ */
+
+static int prepare_binding_table_pointers(struct brw_context *brw)
+{
+   brw_add_validated_bo(brw, brw->vs.bind_bo);
+   brw_add_validated_bo(brw, brw->wm.bind_bo);
+   return 0;
+}
+
+/**
+ * Upload the binding table pointers, which point each stage's array of surface
+ * state pointers.
+ *
+ * The binding table pointers are relative to the surface state base address,
+ * which is 0.
+ */
+static int upload_binding_table_pointers(struct brw_context *brw)
+{
+   BEGIN_BATCH(6, IGNORE_CLIPRECTS);
+   OUT_BATCH(CMD_BINDING_TABLE_PTRS << 16 | (6 - 2));
+   if (brw->vs.bind_bo != NULL)
+      OUT_RELOC(brw->vs.bind_bo, 
+               BRW_USAGE_SAMPLER,
+               0); /* vs */
+   else
+      OUT_BATCH(0);
+   OUT_BATCH(0); /* gs */
+   OUT_BATCH(0); /* clip */
+   OUT_BATCH(0); /* sf */
+   OUT_RELOC(brw->wm.bind_bo,
+            BRW_USAGE_SAMPLER,
+            0); /* wm/ps */
+   ADVANCE_BATCH();
+   return 0;
+}
+
+const struct brw_tracked_state brw_binding_table_pointers = {
+   .dirty = {
+      .mesa = 0,
+      .brw = BRW_NEW_BATCH,
+      .cache = CACHE_NEW_SURF_BIND,
+   },
+   .prepare = prepare_binding_table_pointers,
+   .emit = upload_binding_table_pointers,
+};
+
+
+/**********************************************************************
+ * Upload pointers to the per-stage state.
+ *
+ * The state pointers in this packet are all relative to the general state
+ * base address set by CMD_STATE_BASE_ADDRESS, which is 0.
+ */
+static int upload_pipelined_state_pointers(struct brw_context *brw )
+{
+   BEGIN_BATCH(7, IGNORE_CLIPRECTS);
+   OUT_BATCH(CMD_PIPELINED_STATE_POINTERS << 16 | (7 - 2));
+   OUT_RELOC(brw->vs.state_bo, 
+            BRW_USAGE_STATE,
+            0);
+   if (brw->gs.prog_active)
+      OUT_RELOC(brw->gs.state_bo, 
+               BRW_USAGE_STATE,
+               1);
+   else
+      OUT_BATCH(0);
+   OUT_RELOC(brw->clip.state_bo, 
+            BRW_USAGE_STATE,
+            1);
+   OUT_RELOC(brw->sf.state_bo,
+            BRW_USAGE_STATE,
+            0);
+   OUT_RELOC(brw->wm.state_bo,
+            BRW_USAGE_STATE,
+            0);
+   OUT_RELOC(brw->cc.state_bo,
+            BRW_USAGE_STATE,
+            0);
+   ADVANCE_BATCH();
+
+   brw->state.dirty.brw |= BRW_NEW_PSP;
+   return 0;
+}
+
+
+static int prepare_psp_urb_cbs(struct brw_context *brw)
+{
+   brw_add_validated_bo(brw, brw->vs.state_bo);
+   brw_add_validated_bo(brw, brw->gs.state_bo);
+   brw_add_validated_bo(brw, brw->clip.state_bo);
+   brw_add_validated_bo(brw, brw->sf.state_bo);
+   brw_add_validated_bo(brw, brw->wm.state_bo);
+   brw_add_validated_bo(brw, brw->cc.state_bo);
+   return 0;
+}
+
+static int upload_psp_urb_cbs(struct brw_context *brw )
+{
+   int ret;
+   
+   ret = upload_pipelined_state_pointers(brw);
+   if (ret)
+      return ret;
+
+   ret = brw_upload_urb_fence(brw);
+   if (ret)
+      return ret;
+
+   ret = brw_upload_cs_urb_state(brw);
+   if (ret)
+      return ret;
+
+   return 0;
+}
+
+const struct brw_tracked_state brw_psp_urb_cbs = {
+   .dirty = {
+      .mesa = 0,
+      .brw = BRW_NEW_URB_FENCE | BRW_NEW_BATCH,
+      .cache = (CACHE_NEW_VS_UNIT | 
+               CACHE_NEW_GS_UNIT | 
+               CACHE_NEW_GS_PROG | 
+               CACHE_NEW_CLIP_UNIT | 
+               CACHE_NEW_SF_UNIT | 
+               CACHE_NEW_WM_UNIT | 
+               CACHE_NEW_CC_UNIT)
+   },
+   .prepare = prepare_psp_urb_cbs,
+   .emit = upload_psp_urb_cbs,
+};
+
+
+/***********************************************************************
+ * Depth buffer 
+ */
+
+static int prepare_depthbuffer(struct brw_context *brw)
+{
+   struct pipe_surface *zsbuf = brw->curr.fb.zsbuf;
+
+   if (zsbuf)
+      brw_add_validated_bo(brw, brw_surface(zsbuf)->bo);
+
+   return 0;
+}
+
+static int emit_depthbuffer(struct brw_context *brw)
+{
+   struct pipe_surface *surface = brw->curr.fb.zsbuf;
+   unsigned int len = (BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw)) ? 6 : 5;
+
+   if (surface == NULL) {
+      BEGIN_BATCH(len, IGNORE_CLIPRECTS);
+      OUT_BATCH(CMD_DEPTH_BUFFER << 16 | (len - 2));
+      OUT_BATCH((BRW_DEPTHFORMAT_D32_FLOAT << 18) |
+               (BRW_SURFACE_NULL << 29));
+      OUT_BATCH(0);
+      OUT_BATCH(0);
+      OUT_BATCH(0);
+
+      if (BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw))
+         OUT_BATCH(0);
+
+      ADVANCE_BATCH();
+   } else {
+      struct brw_winsys_buffer *bo;
+      unsigned int format;
+      unsigned int pitch;
+      unsigned int cpp;
+
+      switch (surface->format) {
+      case PIPE_FORMAT_Z16_UNORM:
+        format = BRW_DEPTHFORMAT_D16_UNORM;
+        cpp = 2;
+        break;
+      case PIPE_FORMAT_X8Z24_UNORM:
+      case PIPE_FORMAT_S8Z24_UNORM:
+        format = BRW_DEPTHFORMAT_D24_UNORM_S8_UINT;
+        cpp = 4;
+        break;
+      case PIPE_FORMAT_Z32_FLOAT:
+        format = BRW_DEPTHFORMAT_D32_FLOAT;
+        cpp = 4;
+        break;
+      default:
+        assert(0);
+        return PIPE_ERROR_BAD_INPUT;
+      }
+
+      bo = brw_surface(surface)->bo;
+      pitch = brw_surface(surface)->pitch;
+
+      BEGIN_BATCH(len, IGNORE_CLIPRECTS);
+      OUT_BATCH(CMD_DEPTH_BUFFER << 16 | (len - 2));
+      OUT_BATCH(((pitch * cpp) - 1) |
+               (format << 18) |
+               (BRW_TILEWALK_YMAJOR << 26) |
+               ((surface->layout != PIPE_SURFACE_LAYOUT_LINEAR) << 27) |
+               (BRW_SURFACE_2D << 29));
+      OUT_RELOC(bo,
+               BRW_USAGE_DEPTH_BUFFER,
+               surface->offset);
+      OUT_BATCH((BRW_SURFACE_MIPMAPLAYOUT_BELOW << 1) |
+               ((pitch - 1) << 6) |
+               ((surface->height - 1) << 19));
+      OUT_BATCH(0);
+
+      if (BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw))
+         OUT_BATCH(0);
+
+      ADVANCE_BATCH();
+   }
+
+   return 0;
+}
+
+const struct brw_tracked_state brw_depthbuffer = {
+   .dirty = {
+      .mesa = PIPE_NEW_DEPTH_BUFFER,
+      .brw = BRW_NEW_BATCH,
+      .cache = 0,
+   },
+   .prepare = prepare_depthbuffer,
+   .emit = emit_depthbuffer,
+};
+
+
+
+/***********************************************************************
+ * Polygon stipple packet
+ */
+
+static int upload_polygon_stipple(struct brw_context *brw)
+{
+   BRW_CACHED_BATCH_STRUCT(brw, &brw->curr.bps);
+   return 0;
+}
+
+const struct brw_tracked_state brw_polygon_stipple = {
+   .dirty = {
+      .mesa = PIPE_NEW_POLYGON_STIPPLE,
+      .brw = 0,
+      .cache = 0
+   },
+   .emit = upload_polygon_stipple
+};
+
+
+/***********************************************************************
+ * Line stipple packet
+ */
+
+static int upload_line_stipple(struct brw_context *brw)
+{
+   const struct brw_line_stipple *bls = &brw->curr.rast->bls;
+   if (bls->header.opcode) {
+      BRW_CACHED_BATCH_STRUCT(brw, bls);
+   }
+   return 0;
+}
+
+const struct brw_tracked_state brw_line_stipple = {
+   .dirty = {
+      .mesa = PIPE_NEW_RAST,
+      .brw = 0,
+      .cache = 0
+   },
+   .emit = upload_line_stipple
+};
+
+
+/***********************************************************************
+ * Misc invarient state packets
+ */
+
+static int upload_invarient_state( struct brw_context *brw )
+{
+   {
+      /* 0x61040000  Pipeline Select */
+      /*     PipelineSelect            : 0 */
+      struct brw_pipeline_select ps;
+
+      memset(&ps, 0, sizeof(ps));
+      if (BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw))
+        ps.header.opcode = CMD_PIPELINE_SELECT_GM45;
+      else
+        ps.header.opcode = CMD_PIPELINE_SELECT_965;
+      ps.header.pipeline_select = 0;
+      BRW_BATCH_STRUCT(brw, &ps);
+   }
+
+   {
+      struct brw_global_depth_offset_clamp gdo;
+      memset(&gdo, 0, sizeof(gdo));
+
+      /* Disable depth offset clamping. 
+       */
+      gdo.header.opcode = CMD_GLOBAL_DEPTH_OFFSET_CLAMP;
+      gdo.header.length = sizeof(gdo)/4 - 2;
+      gdo.depth_offset_clamp = 0.0;
+
+      BRW_BATCH_STRUCT(brw, &gdo);
+   }
+
+
+   /* 0x61020000  State Instruction Pointer */
+   {
+      struct brw_system_instruction_pointer sip;
+      memset(&sip, 0, sizeof(sip));
+
+      sip.header.opcode = CMD_STATE_INSN_POINTER;
+      sip.header.length = 0;
+      sip.bits0.pad = 0;
+      sip.bits0.system_instruction_pointer = 0;
+      BRW_BATCH_STRUCT(brw, &sip);
+   }
+
+   /* VF Statistics */
+   {
+      struct brw_vf_statistics vfs;
+      memset(&vfs, 0, sizeof(vfs));
+
+      if (BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw)) 
+        vfs.opcode = CMD_VF_STATISTICS_GM45;
+      else 
+        vfs.opcode = CMD_VF_STATISTICS_965;
+
+      if (BRW_DEBUG & DEBUG_STATS)
+        vfs.statistics_enable = 1; 
+
+      BRW_BATCH_STRUCT(brw, &vfs);
+   }
+   
+   if (!BRW_IS_965(brw))
+   {
+      struct brw_aa_line_parameters balp;
+
+      /* use legacy aa line coverage computation */
+      memset(&balp, 0, sizeof(balp));
+      balp.header.opcode = CMD_AA_LINE_PARAMETERS;
+      balp.header.length = sizeof(balp) / 4 - 2;
+   
+      BRW_BATCH_STRUCT(brw, &balp);
+   }
+
+   {
+      struct brw_polygon_stipple_offset bpso;
+      
+      /* This is invarient state in gallium:
+       */
+      memset(&bpso, 0, sizeof(bpso));
+      bpso.header.opcode = CMD_POLY_STIPPLE_OFFSET;
+      bpso.header.length = sizeof(bpso)/4-2;
+      bpso.bits0.y_offset = 0;
+      bpso.bits0.x_offset = 0;
+
+      BRW_BATCH_STRUCT(brw, &bpso);
+   }
+   
+   return 0;
+}
+
+const struct brw_tracked_state brw_invarient_state = {
+   .dirty = {
+      .mesa = 0,
+      .brw = BRW_NEW_CONTEXT,
+      .cache = 0
+   },
+   .emit = upload_invarient_state
+};
+
+
+/***********************************************************************
+ * State base address 
+ */
+
+/**
+ * Define the base addresses which some state is referenced from.
+ *
+ * This allows us to avoid having to emit relocations in many places for
+ * cached state, and instead emit pointers inside of large, mostly-static
+ * state pools.  This comes at the expense of memory, and more expensive cache
+ * misses.
+ */
+static int upload_state_base_address( struct brw_context *brw )
+{
+   /* Output the structure (brw_state_base_address) directly to the
+    * batchbuffer, so we can emit relocations inline.
+    */
+   if (BRW_IS_IGDNG(brw)) {
+       BEGIN_BATCH(8, IGNORE_CLIPRECTS);
+       OUT_BATCH(CMD_STATE_BASE_ADDRESS << 16 | (8 - 2));
+       OUT_BATCH(1); /* General state base address */
+       OUT_BATCH(1); /* Surface state base address */
+       OUT_BATCH(1); /* Indirect object base address */
+       OUT_BATCH(1); /* Instruction base address */
+       OUT_BATCH(1); /* General state upper bound */
+       OUT_BATCH(1); /* Indirect object upper bound */
+       OUT_BATCH(1); /* Instruction access upper bound */
+       ADVANCE_BATCH();
+   } else {
+       BEGIN_BATCH(6, IGNORE_CLIPRECTS);
+       OUT_BATCH(CMD_STATE_BASE_ADDRESS << 16 | (6 - 2));
+       OUT_BATCH(1); /* General state base address */
+       OUT_BATCH(1); /* Surface state base address */
+       OUT_BATCH(1); /* Indirect object base address */
+       OUT_BATCH(1); /* General state upper bound */
+       OUT_BATCH(1); /* Indirect object upper bound */
+       ADVANCE_BATCH();
+   }
+   return 0;
+}
+
+const struct brw_tracked_state brw_state_base_address = {
+   .dirty = {
+      .mesa = 0,
+      .brw = BRW_NEW_CONTEXT,
+      .cache = 0,
+   },
+   .emit = upload_state_base_address
+};
diff --git a/src/gallium/drivers/i965/brw_pipe_blend.c b/src/gallium/drivers/i965/brw_pipe_blend.c
new file mode 100644 (file)
index 0000000..b759a91
--- /dev/null
@@ -0,0 +1,208 @@
+
+#include "util/u_memory.h"
+#include "pipe/p_context.h"
+#include "pipe/p_state.h"
+
+#include "brw_context.h"
+#include "brw_defines.h"
+#include "brw_debug.h"
+
+static int translate_logicop(unsigned logicop)
+{
+   switch (logicop) {
+   case PIPE_LOGICOP_CLEAR:
+      return BRW_LOGICOPFUNCTION_CLEAR;
+   case PIPE_LOGICOP_AND:
+      return BRW_LOGICOPFUNCTION_AND;
+   case PIPE_LOGICOP_AND_REVERSE:
+      return BRW_LOGICOPFUNCTION_AND_REVERSE;
+   case PIPE_LOGICOP_COPY:
+      return BRW_LOGICOPFUNCTION_COPY;
+   case PIPE_LOGICOP_COPY_INVERTED:
+      return BRW_LOGICOPFUNCTION_COPY_INVERTED;
+   case PIPE_LOGICOP_AND_INVERTED:
+      return BRW_LOGICOPFUNCTION_AND_INVERTED;
+   case PIPE_LOGICOP_NOOP:
+      return BRW_LOGICOPFUNCTION_NOOP;
+   case PIPE_LOGICOP_XOR:
+      return BRW_LOGICOPFUNCTION_XOR;
+   case PIPE_LOGICOP_OR:
+      return BRW_LOGICOPFUNCTION_OR;
+   case PIPE_LOGICOP_OR_INVERTED:
+      return BRW_LOGICOPFUNCTION_OR_INVERTED;
+   case PIPE_LOGICOP_NOR:
+      return BRW_LOGICOPFUNCTION_NOR;
+   case PIPE_LOGICOP_EQUIV:
+      return BRW_LOGICOPFUNCTION_EQUIV;
+   case PIPE_LOGICOP_INVERT:
+      return BRW_LOGICOPFUNCTION_INVERT;
+   case PIPE_LOGICOP_OR_REVERSE:
+      return BRW_LOGICOPFUNCTION_OR_REVERSE;
+   case PIPE_LOGICOP_NAND:
+      return BRW_LOGICOPFUNCTION_NAND;
+   case PIPE_LOGICOP_SET:
+      return BRW_LOGICOPFUNCTION_SET;
+   default:
+      assert(0);
+      return BRW_LOGICOPFUNCTION_SET;
+   }
+}
+
+
+static unsigned translate_blend_equation( unsigned mode )
+{
+   switch (mode) {
+   case PIPE_BLEND_ADD: 
+      return BRW_BLENDFUNCTION_ADD; 
+   case PIPE_BLEND_MIN: 
+      return BRW_BLENDFUNCTION_MIN; 
+   case PIPE_BLEND_MAX: 
+      return BRW_BLENDFUNCTION_MAX; 
+   case PIPE_BLEND_SUBTRACT: 
+      return BRW_BLENDFUNCTION_SUBTRACT; 
+   case PIPE_BLEND_REVERSE_SUBTRACT: 
+      return BRW_BLENDFUNCTION_REVERSE_SUBTRACT; 
+   default: 
+      assert(0);
+      return BRW_BLENDFUNCTION_ADD;
+   }
+}
+
+static unsigned translate_blend_factor( unsigned factor )
+{
+   switch(factor) {
+   case PIPE_BLENDFACTOR_ZERO: 
+      return BRW_BLENDFACTOR_ZERO; 
+   case PIPE_BLENDFACTOR_SRC_ALPHA: 
+      return BRW_BLENDFACTOR_SRC_ALPHA; 
+   case PIPE_BLENDFACTOR_ONE: 
+      return BRW_BLENDFACTOR_ONE; 
+   case PIPE_BLENDFACTOR_SRC_COLOR: 
+      return BRW_BLENDFACTOR_SRC_COLOR; 
+   case PIPE_BLENDFACTOR_INV_SRC_COLOR: 
+      return BRW_BLENDFACTOR_INV_SRC_COLOR; 
+   case PIPE_BLENDFACTOR_DST_COLOR: 
+      return BRW_BLENDFACTOR_DST_COLOR; 
+   case PIPE_BLENDFACTOR_INV_DST_COLOR: 
+      return BRW_BLENDFACTOR_INV_DST_COLOR; 
+   case PIPE_BLENDFACTOR_INV_SRC_ALPHA:
+      return BRW_BLENDFACTOR_INV_SRC_ALPHA; 
+   case PIPE_BLENDFACTOR_DST_ALPHA: 
+      return BRW_BLENDFACTOR_DST_ALPHA; 
+   case PIPE_BLENDFACTOR_INV_DST_ALPHA:
+      return BRW_BLENDFACTOR_INV_DST_ALPHA; 
+   case PIPE_BLENDFACTOR_SRC_ALPHA_SATURATE: 
+      return BRW_BLENDFACTOR_SRC_ALPHA_SATURATE;
+   case PIPE_BLENDFACTOR_CONST_COLOR:
+      return BRW_BLENDFACTOR_CONST_COLOR; 
+   case PIPE_BLENDFACTOR_INV_CONST_COLOR:
+      return BRW_BLENDFACTOR_INV_CONST_COLOR;
+   case PIPE_BLENDFACTOR_CONST_ALPHA:
+      return BRW_BLENDFACTOR_CONST_ALPHA; 
+   case PIPE_BLENDFACTOR_INV_CONST_ALPHA:
+      return BRW_BLENDFACTOR_INV_CONST_ALPHA;
+   default:
+      assert(0);
+      return BRW_BLENDFACTOR_ZERO;
+   }   
+}
+
+static void *brw_create_blend_state( struct pipe_context *pipe,
+                                    const struct pipe_blend_state *templ )
+{
+   struct brw_blend_state *blend = CALLOC_STRUCT(brw_blend_state);
+   if (blend == NULL)
+      return NULL;
+
+   if (templ->logicop_enable) {
+      blend->cc2.logicop_enable = 1;
+      blend->cc5.logicop_func = translate_logicop(templ->logicop_func);
+   } 
+   else if (templ->blend_enable) {
+      blend->cc6.dest_blend_factor = translate_blend_factor(templ->rgb_dst_factor);
+      blend->cc6.src_blend_factor = translate_blend_factor(templ->rgb_src_factor);
+      blend->cc6.blend_function = translate_blend_equation(templ->rgb_func);
+
+      blend->cc5.ia_dest_blend_factor = translate_blend_factor(templ->alpha_dst_factor);
+      blend->cc5.ia_src_blend_factor = translate_blend_factor(templ->alpha_src_factor);
+      blend->cc5.ia_blend_function = translate_blend_equation(templ->alpha_func);
+
+      blend->cc3.blend_enable = 1;
+      blend->cc3.ia_blend_enable = 
+        (blend->cc6.dest_blend_factor != blend->cc5.ia_dest_blend_factor ||
+         blend->cc6.src_blend_factor != blend->cc5.ia_src_blend_factor ||
+         blend->cc6.blend_function != blend->cc5.ia_blend_function);
+
+      /* Per-surface blend enables, currently just follow global
+       * state:
+       */
+      blend->ss0.color_blend = 1;
+   }
+
+   blend->cc5.dither_enable = templ->dither;
+
+   if (BRW_DEBUG & DEBUG_STATS)
+      blend->cc5.statistics_enable = 1;
+
+   /* Per-surface color mask -- just follow global state:
+    */
+   blend->ss0.writedisable_red   = (templ->colormask & PIPE_MASK_R) ? 0 : 1;
+   blend->ss0.writedisable_green = (templ->colormask & PIPE_MASK_G) ? 0 : 1;
+   blend->ss0.writedisable_blue  = (templ->colormask & PIPE_MASK_B) ? 0 : 1;
+   blend->ss0.writedisable_alpha = (templ->colormask & PIPE_MASK_A) ? 0 : 1;
+
+   return (void *)blend;
+}
+
+static void brw_bind_blend_state(struct pipe_context *pipe,
+                                void *cso)
+{
+   struct brw_context *brw = brw_context(pipe);
+   brw->curr.blend = (const struct brw_blend_state *)cso;
+   brw->state.dirty.mesa |= PIPE_NEW_BLEND;
+}
+
+static void brw_delete_blend_state(struct pipe_context *pipe,
+                                 void *cso)
+{
+   struct brw_context *brw = brw_context(pipe);
+   assert((const void *)cso != (const void *)brw->curr.blend);
+   FREE(cso);
+}
+
+
+static void brw_set_blend_color(struct pipe_context *pipe,
+                               const struct pipe_blend_color *blend_color)
+{
+   struct brw_context *brw = brw_context(pipe);
+   struct brw_blend_constant_color *bcc = &brw->curr.bcc;
+
+   bcc->blend_constant_color[0] = blend_color->color[0];
+   bcc->blend_constant_color[1] = blend_color->color[1];
+   bcc->blend_constant_color[2] = blend_color->color[2];
+   bcc->blend_constant_color[3] = blend_color->color[3];
+
+   brw->state.dirty.mesa |= PIPE_NEW_BLEND_COLOR;
+}
+
+
+void brw_pipe_blend_init( struct brw_context *brw )
+{
+   brw->base.set_blend_color = brw_set_blend_color;
+   brw->base.create_blend_state = brw_create_blend_state;
+   brw->base.bind_blend_state = brw_bind_blend_state;
+   brw->base.delete_blend_state = brw_delete_blend_state;
+
+   {
+      struct brw_blend_constant_color *bcc = &brw->curr.bcc;
+
+      memset(bcc, 0, sizeof(*bcc));      
+      bcc->header.opcode = CMD_BLEND_CONSTANT_COLOR;
+      bcc->header.length = sizeof(*bcc)/4-2;
+   }
+
+}
+
+void brw_pipe_blend_cleanup( struct brw_context *brw )
+{
+}
diff --git a/src/gallium/drivers/i965/brw_pipe_clear.c b/src/gallium/drivers/i965/brw_pipe_clear.c
new file mode 100644 (file)
index 0000000..452e1e8
--- /dev/null
@@ -0,0 +1,218 @@
+/**************************************************************************
+ * 
+ * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include "util/u_pack_color.h"
+
+#include "pipe/p_state.h"
+
+#include "brw_batchbuffer.h"
+#include "brw_screen.h"
+#include "brw_context.h"
+
+#define MASK16 0xffff
+#define MASK24 0xffffff
+
+
+/**
+ * Use blitting to clear the renderbuffers named by 'flags'.
+ * Note: we can't use the ctx->DrawBuffer->_ColorDrawBufferIndexes field
+ * since that might include software renderbuffers or renderbuffers
+ * which we're clearing with triangles.
+ * \param mask  bitmask of BUFFER_BIT_* values indicating buffers to clear
+ */
+static enum pipe_error
+try_clear( struct brw_context *brw,
+           struct brw_surface *surface,
+           unsigned value )
+{
+   uint32_t BR13, CMD;
+   int x1 = 0;
+   int y1 = 0;
+   int x2 = surface->base.width;
+   int y2 = surface->base.height;
+   int pitch = surface->pitch;
+   int cpp = surface->cpp;
+
+   if (x2 == 0 || y2 == 0)
+      return 0;
+
+   debug_printf("%s dst:buf(%p)/%d+%d %d,%d sz:%dx%d\n",
+                __FUNCTION__,
+                (void *)surface->bo, pitch * cpp,
+                surface->base.offset,
+                x1, y1, x2 - x1, y2 - y1);
+
+   BR13 = 0xf0 << 16;
+   CMD = XY_COLOR_BLT_CMD | XY_BLT_WRITE_RGB | XY_BLT_WRITE_ALPHA;
+
+   /* Setup the blit command */
+   if (cpp == 4) {
+      BR13 |= BR13_8888;
+      CMD |= XY_BLT_WRITE_ALPHA | XY_BLT_WRITE_RGB;
+   }
+   else {
+      assert(cpp == 2);
+      BR13 |= BR13_565;
+   }
+
+   /* XXX: nasty hack for clearing depth buffers
+    */
+   if (surface->tiling == BRW_TILING_Y) {
+      x2 = pitch;
+   }
+
+   if (surface->tiling == BRW_TILING_X) {
+      CMD |= XY_DST_TILED;
+      pitch /= 4;
+   }
+
+   BR13 |= (pitch * cpp);
+
+   BEGIN_BATCH(6, 0);
+   OUT_BATCH(CMD);
+   OUT_BATCH(BR13);
+   OUT_BATCH((y1 << 16) | x1);
+   OUT_BATCH((y2 << 16) | x2);
+   OUT_RELOC(surface->bo,
+             BRW_USAGE_BLIT_DEST,
+             surface->base.offset);
+   OUT_BATCH(value);
+   ADVANCE_BATCH();
+
+   return 0;
+}
+
+
+
+
+static void color_clear(struct brw_context *brw, 
+                        struct brw_surface *bsurface,
+                        const float *rgba )
+{
+   enum pipe_error ret;
+   union util_color value;
+
+   util_pack_color( rgba, bsurface->base.format, &value );
+
+   if (bsurface->cpp == 2)
+      value.ui |= value.ui << 16;
+
+   ret = try_clear( brw, bsurface, value.ui );
+
+   if (ret != 0) {
+      brw_context_flush( brw );
+      ret = try_clear( brw, bsurface, value.ui );
+      assert( ret == 0 );
+   }
+}
+
+static void zstencil_clear(struct brw_context *brw, 
+                           struct brw_surface *bsurface,
+                           double depth,
+                           unsigned stencil )
+{
+   enum pipe_error ret;
+   unsigned value;
+
+   switch (bsurface->base.format) {
+   case PIPE_FORMAT_X8Z24_UNORM:
+   case PIPE_FORMAT_S8Z24_UNORM:
+      value = ((unsigned)(depth * MASK24) & MASK24);
+      break;
+   case PIPE_FORMAT_Z16_UNORM:
+      value = ((unsigned)(depth * MASK16) & MASK16);
+      break;
+   default:
+      assert(0);
+      return;
+   }
+
+   switch (bsurface->base.format) {
+   case PIPE_FORMAT_X8Z24_UNORM:
+   case PIPE_FORMAT_S8Z24_UNORM:
+      value = value | (stencil << 24);
+      break;
+
+   case PIPE_FORMAT_Z16_UNORM:
+      value = value | (value << 16);
+      break;
+
+   default:
+      break;
+   }
+
+   ret = try_clear( brw, bsurface, value );
+
+   if (ret != 0) {
+      brw_context_flush( brw );
+      ret = try_clear( brw, bsurface, value );
+      assert( ret == 0 );
+   }
+}
+
+
+
+/**
+ * Clear the given surface to the specified value.
+ * No masking, no scissor (clear entire buffer).
+ */
+static void brw_clear(struct pipe_context *pipe, 
+                      unsigned buffers,
+                      const float *rgba,
+                      double depth,
+                      unsigned stencil)
+{
+   struct brw_context *brw = brw_context( pipe );
+   int i;
+
+   if (buffers & PIPE_CLEAR_COLOR) {
+      for (i = 0; i < brw->curr.fb.nr_cbufs; i++) {
+         color_clear( brw, 
+                      brw_surface(brw->curr.fb.cbufs[i]),
+                      rgba );
+      }
+   }
+
+   if (buffers & PIPE_CLEAR_DEPTHSTENCIL) {
+      if (brw->curr.fb.zsbuf) {
+         zstencil_clear( brw,
+                         brw_surface(brw->curr.fb.zsbuf),
+                         depth, stencil );
+      }
+   }
+}
+
+
+void brw_pipe_clear_init( struct brw_context *brw )
+{
+   brw->base.clear = brw_clear;
+}
+
+
+void brw_pipe_clear_cleanup( struct brw_context *brw )
+{
+}
diff --git a/src/gallium/drivers/i965/brw_pipe_depth.c b/src/gallium/drivers/i965/brw_pipe_depth.c
new file mode 100644 (file)
index 0000000..e010d76
--- /dev/null
@@ -0,0 +1,172 @@
+
+#include "util/u_math.h"
+#include "util/u_memory.h"
+
+#include "brw_context.h"
+#include "brw_defines.h"
+
+/* XXX: Fixme - include this to get IZ_ defines
+ */
+#include "brw_wm.h"
+
+static unsigned brw_translate_compare_func(unsigned func)
+{
+   switch (func) {
+   case PIPE_FUNC_NEVER:
+      return BRW_COMPAREFUNCTION_NEVER;
+   case PIPE_FUNC_LESS:
+      return BRW_COMPAREFUNCTION_LESS;
+   case PIPE_FUNC_LEQUAL:
+      return BRW_COMPAREFUNCTION_LEQUAL;
+   case PIPE_FUNC_GREATER:
+      return BRW_COMPAREFUNCTION_GREATER;
+   case PIPE_FUNC_GEQUAL:
+      return BRW_COMPAREFUNCTION_GEQUAL;
+   case PIPE_FUNC_NOTEQUAL:
+      return BRW_COMPAREFUNCTION_NOTEQUAL;
+   case PIPE_FUNC_EQUAL:
+      return BRW_COMPAREFUNCTION_EQUAL;
+   case PIPE_FUNC_ALWAYS:
+      return BRW_COMPAREFUNCTION_ALWAYS;
+   default:
+      assert(0);
+      return BRW_COMPAREFUNCTION_ALWAYS;
+   }
+}
+
+static unsigned translate_stencil_op(unsigned op)
+{
+   switch (op) {
+   case PIPE_STENCIL_OP_KEEP:
+      return BRW_STENCILOP_KEEP;
+   case PIPE_STENCIL_OP_ZERO:
+      return BRW_STENCILOP_ZERO;
+   case PIPE_STENCIL_OP_REPLACE:
+      return BRW_STENCILOP_REPLACE;
+   case PIPE_STENCIL_OP_INCR:
+      return BRW_STENCILOP_INCRSAT;
+   case PIPE_STENCIL_OP_DECR:
+      return BRW_STENCILOP_DECRSAT;
+   case PIPE_STENCIL_OP_INCR_WRAP:
+      return BRW_STENCILOP_INCR;
+   case PIPE_STENCIL_OP_DECR_WRAP:
+      return BRW_STENCILOP_DECR;
+   case PIPE_STENCIL_OP_INVERT:
+      return BRW_STENCILOP_INVERT;
+   default:
+      assert(0);
+      return BRW_STENCILOP_ZERO;
+   }
+}
+
+static void create_bcc_state( struct brw_depth_stencil_state *zstencil,
+                             const struct pipe_depth_stencil_alpha_state *templ )
+{
+   if (templ->stencil[0].enabled) {
+      zstencil->cc0.stencil_enable = 1;
+      zstencil->cc0.stencil_func =
+        brw_translate_compare_func(templ->stencil[0].func);
+      zstencil->cc0.stencil_fail_op =
+        translate_stencil_op(templ->stencil[0].fail_op);
+      zstencil->cc0.stencil_pass_depth_fail_op =
+        translate_stencil_op(templ->stencil[0].zfail_op);
+      zstencil->cc0.stencil_pass_depth_pass_op =
+        translate_stencil_op(templ->stencil[0].zpass_op);
+      zstencil->cc1.stencil_ref = templ->stencil[0].ref_value;
+      zstencil->cc1.stencil_write_mask = templ->stencil[0].writemask;
+      zstencil->cc1.stencil_test_mask = templ->stencil[0].valuemask;
+
+      if (templ->stencil[1].enabled) {
+        zstencil->cc0.bf_stencil_enable = 1;
+        zstencil->cc0.bf_stencil_func =
+           brw_translate_compare_func(templ->stencil[1].func);
+        zstencil->cc0.bf_stencil_fail_op =
+           translate_stencil_op(templ->stencil[1].fail_op);
+        zstencil->cc0.bf_stencil_pass_depth_fail_op =
+           translate_stencil_op(templ->stencil[1].zfail_op);
+        zstencil->cc0.bf_stencil_pass_depth_pass_op =
+           translate_stencil_op(templ->stencil[1].zpass_op);
+        zstencil->cc1.bf_stencil_ref = templ->stencil[1].ref_value;
+        zstencil->cc2.bf_stencil_write_mask = templ->stencil[1].writemask;
+        zstencil->cc2.bf_stencil_test_mask = templ->stencil[1].valuemask;
+      }
+
+      zstencil->cc0.stencil_write_enable = (zstencil->cc1.stencil_write_mask ||
+                                           zstencil->cc2.bf_stencil_write_mask);
+   }
+
+
+   if (templ->alpha.enabled) {
+      zstencil->cc3.alpha_test = 1;
+      zstencil->cc3.alpha_test_func = brw_translate_compare_func(templ->alpha.func);
+      zstencil->cc3.alpha_test_format = BRW_ALPHATEST_FORMAT_UNORM8;
+      zstencil->cc7.alpha_ref.ub[0] = float_to_ubyte(templ->alpha.ref_value);
+   }
+
+   if (templ->depth.enabled) {
+      zstencil->cc2.depth_test = 1;
+      zstencil->cc2.depth_test_function = brw_translate_compare_func(templ->depth.func);
+      zstencil->cc2.depth_write_enable = templ->depth.writemask;
+   }
+}
+
+static void create_wm_iz_state( struct brw_depth_stencil_state *zstencil )
+{
+   if (zstencil->cc3.alpha_test)
+      zstencil->iz_lookup |= IZ_PS_KILL_ALPHATEST_BIT;
+
+   if (zstencil->cc2.depth_test)
+      zstencil->iz_lookup |= IZ_DEPTH_TEST_ENABLE_BIT;
+
+   if (zstencil->cc2.depth_write_enable)
+      zstencil->iz_lookup |= IZ_DEPTH_WRITE_ENABLE_BIT;
+
+   if (zstencil->cc0.stencil_enable)
+      zstencil->iz_lookup |= IZ_STENCIL_TEST_ENABLE_BIT;
+
+   if (zstencil->cc0.stencil_write_enable)
+      zstencil->iz_lookup |= IZ_STENCIL_WRITE_ENABLE_BIT;
+
+}
+
+
+static void *
+brw_create_depth_stencil_state( struct pipe_context *pipe,
+                               const struct pipe_depth_stencil_alpha_state *templ )
+{
+   struct brw_depth_stencil_state *zstencil = CALLOC_STRUCT(brw_depth_stencil_state);
+
+   create_bcc_state( zstencil, templ );
+   create_wm_iz_state( zstencil );
+
+   return (void *)zstencil;
+}
+
+
+static void brw_bind_depth_stencil_state(struct pipe_context *pipe,
+                                        void *cso)
+{
+   struct brw_context *brw = brw_context(pipe);
+   brw->curr.zstencil = (const struct brw_depth_stencil_state *)cso;
+   brw->state.dirty.mesa |= PIPE_NEW_DEPTH_STENCIL_ALPHA;
+}
+
+static void brw_delete_depth_stencil_state(struct pipe_context *pipe,
+                                          void *cso)
+{
+   struct brw_context *brw = brw_context(pipe);
+   assert((const void *)cso != (const void *)brw->curr.zstencil);
+   FREE(cso);
+}
+
+
+void brw_pipe_depth_stencil_init( struct brw_context *brw )
+{
+   brw->base.create_depth_stencil_alpha_state = brw_create_depth_stencil_state;
+   brw->base.bind_depth_stencil_alpha_state = brw_bind_depth_stencil_state;
+   brw->base.delete_depth_stencil_alpha_state = brw_delete_depth_stencil_state;
+}
+
+void brw_pipe_depth_stencil_cleanup( struct brw_context *brw )
+{
+}
diff --git a/src/gallium/drivers/i965/brw_pipe_fb.c b/src/gallium/drivers/i965/brw_pipe_fb.c
new file mode 100644 (file)
index 0000000..5d4e502
--- /dev/null
@@ -0,0 +1,84 @@
+#include "util/u_math.h"
+#include "pipe/p_context.h"
+#include "pipe/p_state.h"
+
+#include "brw_context.h"
+#include "brw_debug.h"
+
+/**
+ * called from intelDrawBuffer()
+ */
+static void brw_set_framebuffer_state( struct pipe_context *pipe, 
+                                      const struct pipe_framebuffer_state *fb )
+{
+   struct brw_context *brw = brw_context(pipe);
+   unsigned i;
+
+   /* Dimensions:
+    */
+   if (brw->curr.fb.width != fb->width ||
+       brw->curr.fb.height != fb->height) {
+      brw->curr.fb.width = fb->width;
+      brw->curr.fb.height = fb->height;
+      brw->state.dirty.mesa |= PIPE_NEW_FRAMEBUFFER_DIMENSIONS;
+   }
+   
+   /* Z/Stencil
+    */
+   if (brw->curr.fb.zsbuf != fb->zsbuf) {
+      pipe_surface_reference(&brw->curr.fb.zsbuf, fb->zsbuf);
+      brw->state.dirty.mesa |= PIPE_NEW_DEPTH_BUFFER;
+   }
+
+   /* Color buffers:
+    */
+   for (i = 0; i < PIPE_MAX_COLOR_BUFS; i++) {
+      if (brw->curr.fb.cbufs[i] != fb->cbufs[i]) {
+        brw->state.dirty.mesa |= PIPE_NEW_COLOR_BUFFERS;
+        pipe_surface_reference(&brw->curr.fb.cbufs[i], fb->cbufs[i]);
+      }
+   }
+   
+   if (brw->curr.fb.nr_cbufs != fb->nr_cbufs) {
+      brw->curr.fb.nr_cbufs = MIN2(BRW_MAX_DRAW_BUFFERS, fb->nr_cbufs);
+      brw->state.dirty.mesa |= PIPE_NEW_NR_CBUFS;
+   }
+}
+
+
+static void brw_set_viewport_state( struct pipe_context *pipe,
+                                   const struct pipe_viewport_state *viewport )
+{
+   struct brw_context *brw = brw_context(pipe);
+
+   brw->curr.viewport = *viewport;
+   brw->curr.ccv.min_depth = viewport->scale[2] * -1.0 + viewport->translate[2];
+   brw->curr.ccv.max_depth = viewport->scale[2] *  1.0 + viewport->translate[2];
+
+   if (0)
+      debug_printf("%s depth range %f .. %f\n",
+                   __FUNCTION__,
+                   brw->curr.ccv.min_depth,
+                   brw->curr.ccv.max_depth);
+
+   brw->state.dirty.mesa |= PIPE_NEW_VIEWPORT;
+}
+
+
+void brw_pipe_framebuffer_init( struct brw_context *brw )
+{
+   brw->base.set_framebuffer_state = brw_set_framebuffer_state;
+   brw->base.set_viewport_state = brw_set_viewport_state;
+}
+
+void brw_pipe_framebuffer_cleanup( struct brw_context *brw )
+{
+   struct pipe_framebuffer_state *fb = &brw->curr.fb;
+   int i;
+
+   for (i = 0; i < PIPE_MAX_COLOR_BUFS; i++) {
+      pipe_surface_reference(&fb->cbufs[i], NULL);
+   }
+
+   pipe_surface_reference(&fb->zsbuf, NULL);
+}
diff --git a/src/gallium/drivers/i965/brw_pipe_flush.c b/src/gallium/drivers/i965/brw_pipe_flush.c
new file mode 100644 (file)
index 0000000..fdc4814
--- /dev/null
@@ -0,0 +1,83 @@
+
+#include "util/u_upload_mgr.h"
+
+#include "brw_context.h"
+#include "brw_screen.h"
+#include "brw_batchbuffer.h"
+
+
+
+/* All batchbuffer flushes must go through this function.
+ */
+void brw_context_flush( struct brw_context *brw )
+{
+   /*
+    * 
+    */
+   brw_emit_query_end(brw);
+
+   /* Move to the end of the current upload buffer so that we'll force choosing
+    * a new buffer next time.
+    */
+   u_upload_flush( brw->vb.upload_vertex );
+   u_upload_flush( brw->vb.upload_index );
+
+   _brw_batchbuffer_flush( brw->batch, __FILE__, __LINE__ );
+
+   /* Mark all context state as needing to be re-emitted.
+    * This is probably not as severe as on 915, since almost all of our state
+    * is just in referenced buffers.
+    */
+   brw->state.dirty.brw |= BRW_NEW_CONTEXT;
+   brw->state.dirty.mesa |= ~0;
+   brw->state.dirty.brw |= ~0;
+   brw->state.dirty.cache |= ~0;
+
+   brw->curbe.need_new_bo = GL_TRUE;
+}
+
+static void
+brw_flush( struct pipe_context *pipe,
+           unsigned flags, 
+           struct pipe_fence_handle **fence )
+{
+   brw_context_flush( brw_context( pipe ) );
+   if (fence)
+      *fence = NULL;
+}
+
+static unsigned brw_is_buffer_referenced(struct pipe_context *pipe,
+                                  struct pipe_buffer *buffer)
+{
+   struct brw_context *brw = brw_context(pipe);
+   struct brw_screen *bscreen = brw_screen(brw->base.screen);
+
+   return brw_is_buffer_referenced_by_bo( bscreen,
+                                          buffer,
+                                          brw->batch->buf );
+}
+
+static unsigned brw_is_texture_referenced(struct pipe_context *pipe,
+                                   struct pipe_texture *texture,
+                                   unsigned face,
+                                   unsigned level)
+{
+   struct brw_context *brw = brw_context(pipe);
+   struct brw_screen *bscreen = brw_screen(brw->base.screen);
+
+   return brw_is_texture_referenced_by_bo( bscreen,
+                                           texture, face, level,
+                                           brw->batch->buf );
+}
+
+void brw_pipe_flush_init( struct brw_context *brw )
+{
+   brw->base.flush = brw_flush;
+   brw->base.is_buffer_referenced = brw_is_buffer_referenced;
+   brw->base.is_texture_referenced = brw_is_texture_referenced;
+}
+
+
+void brw_pipe_flush_cleanup( struct brw_context *brw )
+{
+}
diff --git a/src/gallium/drivers/i965/brw_pipe_misc.c b/src/gallium/drivers/i965/brw_pipe_misc.c
new file mode 100644 (file)
index 0000000..3035907
--- /dev/null
@@ -0,0 +1,54 @@
+
+#include "brw_context.h"
+#include "brw_structs.h"
+#include "brw_defines.h"
+
+static void brw_set_polygon_stipple( struct pipe_context *pipe,
+                                    const struct pipe_poly_stipple *stip )
+{
+   struct brw_context *brw = brw_context(pipe);
+   struct brw_polygon_stipple *bps = &brw->curr.bps;
+   GLuint i;
+
+   memset(bps, 0, sizeof *bps);
+   bps->header.opcode = CMD_POLY_STIPPLE_PATTERN;
+   bps->header.length = sizeof *bps/4-2;
+
+   for (i = 0; i < 32; i++)
+      bps->stipple[i] = stip->stipple[i]; /* don't invert */
+
+   brw->state.dirty.mesa |= PIPE_NEW_POLYGON_STIPPLE;
+}
+
+
+static void brw_set_scissor_state( struct pipe_context *pipe,
+                                   const struct pipe_scissor_state *scissor )
+{
+   struct brw_context *brw = brw_context(pipe);
+
+   brw->curr.scissor =  *scissor;
+   brw->state.dirty.mesa |= PIPE_NEW_SCISSOR;
+}
+
+
+static void brw_set_clip_state( struct pipe_context *pipe,
+                                const struct pipe_clip_state *clip )
+{
+   struct brw_context *brw = brw_context(pipe);
+
+   brw->curr.ucp = *clip;
+   brw->state.dirty.mesa |= PIPE_NEW_CLIP;
+}
+
+
+void brw_pipe_misc_init( struct brw_context *brw )
+{
+   brw->base.set_polygon_stipple = brw_set_polygon_stipple;
+   brw->base.set_scissor_state = brw_set_scissor_state;
+   brw->base.set_clip_state = brw_set_clip_state;
+}
+
+
+void brw_pipe_misc_cleanup( struct brw_context *brw )
+{
+}
diff --git a/src/gallium/drivers/i965/brw_pipe_query.c b/src/gallium/drivers/i965/brw_pipe_query.c
new file mode 100644 (file)
index 0000000..2eb8626
--- /dev/null
@@ -0,0 +1,263 @@
+/*
+ * Copyright Â© 2008 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *
+ */
+
+/** @file support for ARB_query_object
+ *
+ * ARB_query_object is implemented by using the PIPE_CONTROL command to stall
+ * execution on the completion of previous depth tests, and write the
+ * current PS_DEPTH_COUNT to a buffer object.
+ *
+ * We use before and after counts when drawing during a query so that
+ * we don't pick up other clients' query data in ours.  To reduce overhead,
+ * a single BO is used to record the query data for all active queries at
+ * once.  This also gives us a simple bound on how much batchbuffer space is
+ * required for handling queries, so that we can be sure that we won't
+ * have to emit a batchbuffer without getting the ending PS_DEPTH_COUNT.
+ */
+#include "util/u_simple_list.h"
+
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_batchbuffer.h"
+#include "brw_reg.h"
+
+/** Waits on the query object's BO and totals the results for this query */
+static boolean
+brw_query_get_result(struct pipe_context *pipe,
+                    struct pipe_query *q,
+                    boolean wait,
+                    uint64_t *result)
+{
+   struct brw_context *brw = brw_context(pipe);
+   struct brw_query_object *query = (struct brw_query_object *)q;
+
+   /* Map and count the pixels from the current query BO */
+   if (query->bo) {
+      int i;
+      uint64_t *map;
+      
+      if (brw->sws->bo_is_busy(query->bo) && !wait)
+        return FALSE;
+      
+      map = bo_map_read(brw->sws, query->bo);
+      if (map == NULL)
+        return FALSE;
+      
+      for (i = query->first_index; i <= query->last_index; i++) {
+        query->result += map[i * 2 + 1] - map[i * 2];
+      }
+
+      brw->sws->bo_unmap(query->bo);
+      bo_reference(&query->bo, NULL);
+   }
+
+   *result = query->result;
+   return TRUE;
+}
+
+static struct pipe_query *
+brw_query_create(struct pipe_context *pipe, unsigned type )
+{
+   struct brw_query_object *query;
+
+   switch (type) {
+   case PIPE_QUERY_OCCLUSION_COUNTER:
+      query = CALLOC_STRUCT( brw_query_object );
+      if (query == NULL)
+        return NULL;
+      return (struct pipe_query *)query;
+      
+   default:
+      return NULL;
+   }
+}
+
+static void
+brw_query_destroy(struct pipe_context *pipe, struct pipe_query *q)
+{
+   struct brw_query_object *query = (struct brw_query_object *)q;
+
+   bo_reference(&query->bo, NULL);
+   FREE(query);
+}
+
+static void
+brw_query_begin(struct pipe_context *pipe, struct pipe_query *q)
+{
+   struct brw_context *brw = brw_context(pipe);
+   struct brw_query_object *query = (struct brw_query_object *)q;
+
+   /* Reset our driver's tracking of query state. */
+   bo_reference(&query->bo, NULL);
+   query->result = 0;
+   query->first_index = -1;
+   query->last_index = -1;
+
+   insert_at_head(&brw->query.active_head, query);
+   brw->query.stats_wm++;
+   brw->state.dirty.mesa |= PIPE_NEW_QUERY;
+}
+
+static void
+brw_query_end(struct pipe_context *pipe, struct pipe_query *q)
+{
+   struct brw_context *brw = brw_context(pipe);
+   struct brw_query_object *query = (struct brw_query_object *)q;
+
+   /* Flush the batchbuffer in case it has writes to our query BO.
+    * Have later queries write to a new query BO so that further rendering
+    * doesn't delay the collection of our results.
+    */
+   if (query->bo) {
+      brw_emit_query_end(brw);
+      brw_context_flush( brw );
+
+      bo_reference(&brw->query.bo, NULL);
+   }
+
+   remove_from_list(query);
+   brw->query.stats_wm--;
+   brw->state.dirty.mesa |= PIPE_NEW_QUERY;
+}
+
+/***********************************************************************
+ * Internal functions and callbacks to implement queries 
+ */
+
+/** Called to set up the query BO and account for its aperture space */
+enum pipe_error
+brw_prepare_query_begin(struct brw_context *brw)
+{
+   enum pipe_error ret;
+
+   /* Skip if we're not doing any queries. */
+   if (is_empty_list(&brw->query.active_head))
+      return PIPE_OK;
+
+   /* Get a new query BO if we're going to need it. */
+   if (brw->query.bo == NULL ||
+       brw->query.index * 2 + 1 >= 4096 / sizeof(uint64_t)) {
+
+      ret = brw->sws->bo_alloc(brw->sws, BRW_BUFFER_TYPE_QUERY, 4096, 1,
+                               &brw->query.bo);
+      if (ret)
+         return ret;
+
+      brw->query.index = 0;
+   }
+
+   brw_add_validated_bo(brw, brw->query.bo);
+
+   return PIPE_OK;
+}
+
+/** Called just before primitive drawing to get a beginning PS_DEPTH_COUNT. */
+void
+brw_emit_query_begin(struct brw_context *brw)
+{
+   struct brw_query_object *query;
+
+   /* Skip if we're not doing any queries, or we've emitted the start. */
+   if (brw->query.active || is_empty_list(&brw->query.active_head))
+      return;
+
+   BEGIN_BATCH(4, IGNORE_CLIPRECTS);
+   OUT_BATCH(_3DSTATE_PIPE_CONTROL |
+            PIPE_CONTROL_DEPTH_STALL |
+            PIPE_CONTROL_WRITE_DEPTH_COUNT);
+   /* This object could be mapped cacheable, but we don't have an exposed
+    * mechanism to support that.  Since it's going uncached, tell GEM that
+    * we're writing to it.  The usual clflush should be all that's required
+    * to pick up the results.
+    */
+   OUT_RELOC(brw->query.bo,
+            BRW_USAGE_QUERY_RESULT,
+            PIPE_CONTROL_GLOBAL_GTT_WRITE |
+            ((brw->query.index * 2) * sizeof(uint64_t)));
+   OUT_BATCH(0);
+   OUT_BATCH(0);
+   ADVANCE_BATCH();
+
+   foreach(query, &brw->query.active_head) {
+      if (query->bo != brw->query.bo) {
+        uint64_t tmp;
+        
+        /* Propogate the results from this buffer to all of the
+         * active queries, as the bo is going away.
+         */
+        if (query->bo != NULL)
+           brw_query_get_result( &brw->base, 
+                                 (struct pipe_query *)query,
+                                 FALSE,
+                                 &tmp );
+
+        bo_reference( &query->bo, brw->query.bo );
+        query->first_index = brw->query.index;
+      }
+      query->last_index = brw->query.index;
+   }
+   brw->query.active = GL_TRUE;
+}
+
+/** Called at batchbuffer flush to get an ending PS_DEPTH_COUNT */
+void
+brw_emit_query_end(struct brw_context *brw)
+{
+   if (!brw->query.active)
+      return;
+
+   BEGIN_BATCH(4, IGNORE_CLIPRECTS);
+   OUT_BATCH(_3DSTATE_PIPE_CONTROL |
+            PIPE_CONTROL_DEPTH_STALL |
+            PIPE_CONTROL_WRITE_DEPTH_COUNT);
+   OUT_RELOC(brw->query.bo,
+            BRW_USAGE_QUERY_RESULT,
+            PIPE_CONTROL_GLOBAL_GTT_WRITE |
+            ((brw->query.index * 2 + 1) * sizeof(uint64_t)));
+   OUT_BATCH(0);
+   OUT_BATCH(0);
+   ADVANCE_BATCH();
+
+   brw->query.active = GL_FALSE;
+   brw->query.index++;
+}
+
+void brw_pipe_query_init( struct brw_context *brw )
+{
+   brw->base.create_query = brw_query_create;
+   brw->base.destroy_query = brw_query_destroy;
+   brw->base.begin_query = brw_query_begin;
+   brw->base.end_query = brw_query_end;
+   brw->base.get_query_result = brw_query_get_result;
+}
+
+
+void brw_pipe_query_cleanup( struct brw_context *brw )
+{
+   /* Unreference brw->query.bo ??
+    */
+}
diff --git a/src/gallium/drivers/i965/brw_pipe_rast.c b/src/gallium/drivers/i965/brw_pipe_rast.c
new file mode 100644 (file)
index 0000000..2117e91
--- /dev/null
@@ -0,0 +1,161 @@
+
+#include "util/u_memory.h"
+#include "pipe/p_defines.h"
+#include "brw_context.h"
+#include "brw_defines.h"
+#include "brw_pipe_rast.h"
+#include "brw_wm.h"
+
+
+static unsigned translate_fill( unsigned fill )
+{
+   switch (fill) {
+   case PIPE_POLYGON_MODE_FILL:
+      return CLIP_FILL;
+   case PIPE_POLYGON_MODE_LINE:
+      return CLIP_LINE;
+   case PIPE_POLYGON_MODE_POINT:
+      return CLIP_POINT;
+   default:
+      assert(0);
+      return CLIP_FILL;
+   }
+}
+
+
+/* Calculates the key for triangle-mode clipping.  Non-triangle
+ * clipping keys use much less information and are computed on the
+ * fly.
+ */
+static void
+calculate_clip_key_rast( const struct brw_context *brw,
+                        const struct pipe_rasterizer_state *templ,
+                        const struct brw_rasterizer_state *rast,
+                        struct brw_clip_prog_key *key)
+{
+   memset(key, 0, sizeof *key);
+
+   if (brw->chipset.is_igdng)
+       key->clip_mode = BRW_CLIPMODE_KERNEL_CLIP;
+   else
+       key->clip_mode = BRW_CLIPMODE_NORMAL;
+
+   key->do_flat_shading = templ->flatshade;
+
+   if (templ->cull_mode == PIPE_WINDING_BOTH) {
+      key->clip_mode = BRW_CLIPMODE_REJECT_ALL;
+      return;
+   }
+
+   key->fill_ccw = CLIP_CULL;
+   key->fill_cw = CLIP_CULL;
+
+   if (!(templ->cull_mode & PIPE_WINDING_CCW)) {
+      key->fill_ccw = translate_fill(templ->fill_ccw);
+   }
+
+   if (!(templ->cull_mode & PIPE_WINDING_CW)) {
+      key->fill_cw = translate_fill(templ->fill_cw);
+   }
+
+   if (key->fill_cw == CLIP_LINE ||
+       key->fill_ccw == CLIP_LINE ||
+       key->fill_cw == CLIP_POINT ||
+       key->fill_ccw == CLIP_POINT) {
+      key->do_unfilled = 1;
+      key->clip_mode = BRW_CLIPMODE_CLIP_NON_REJECTED;
+   }
+
+   key->offset_ccw = templ->offset_ccw;
+   key->offset_cw = templ->offset_cw;
+
+   if (templ->light_twoside && key->fill_cw != CLIP_CULL) 
+      key->copy_bfc_cw = 1;
+   
+   if (templ->light_twoside && key->fill_ccw != CLIP_CULL) 
+      key->copy_bfc_ccw = 1;
+}
+
+
+static void
+calculate_line_stipple_rast( const struct pipe_rasterizer_state *templ,
+                            struct brw_line_stipple *bls )
+{
+   GLfloat tmp = 1.0f / (templ->line_stipple_factor + 1);
+   GLint tmpi = tmp * (1<<13);
+
+   bls->header.opcode = CMD_LINE_STIPPLE_PATTERN;
+   bls->header.length = sizeof(*bls)/4 - 2;
+   bls->bits0.pattern = templ->line_stipple_pattern;
+   bls->bits1.repeat_count = templ->line_stipple_factor + 1;
+   bls->bits1.inverse_repeat_count = tmpi;
+}
+
+static void *brw_create_rasterizer_state( struct pipe_context *pipe,
+                                         const struct pipe_rasterizer_state *templ )
+{
+   struct brw_context *brw = brw_context(pipe);
+   struct brw_rasterizer_state *rast;
+
+   rast = CALLOC_STRUCT(brw_rasterizer_state);
+   if (rast == NULL)
+      return NULL;
+
+   rast->templ = *templ;
+
+   calculate_clip_key_rast( brw, templ, rast, &rast->clip_key );
+   
+   if (templ->line_stipple_enable)
+      calculate_line_stipple_rast( templ, &rast->bls );
+
+   /* Caclculate lookup value for WM IZ table.
+    */
+   if (templ->line_smooth) {
+      if (templ->fill_cw == PIPE_POLYGON_MODE_LINE &&
+         templ->fill_ccw == PIPE_POLYGON_MODE_LINE) {
+        rast->unfilled_aa_line = AA_ALWAYS;
+      }
+      else if (templ->fill_cw == PIPE_POLYGON_MODE_LINE ||
+              templ->fill_ccw == PIPE_POLYGON_MODE_LINE) {
+        rast->unfilled_aa_line = AA_SOMETIMES;
+      }
+      else {
+        rast->unfilled_aa_line = AA_NEVER;
+      }
+   }
+   else {
+      rast->unfilled_aa_line = AA_NEVER;
+   }
+
+   return (void *)rast;
+}
+
+
+static void brw_bind_rasterizer_state(struct pipe_context *pipe,
+                                void *cso)
+{
+   struct brw_context *brw = brw_context(pipe);
+   brw->curr.rast = (const struct brw_rasterizer_state *)cso;
+   brw->state.dirty.mesa |= PIPE_NEW_RAST;
+}
+
+static void brw_delete_rasterizer_state(struct pipe_context *pipe,
+                                 void *cso)
+{
+   struct brw_context *brw = brw_context(pipe);
+   assert((const void *)cso != (const void *)brw->curr.rast);
+   FREE(cso);
+}
+
+
+
+void brw_pipe_rast_init( struct brw_context *brw )
+{
+   brw->base.create_rasterizer_state = brw_create_rasterizer_state;
+   brw->base.bind_rasterizer_state = brw_bind_rasterizer_state;
+   brw->base.delete_rasterizer_state = brw_delete_rasterizer_state;
+}
+
+void brw_pipe_rast_cleanup( struct brw_context *brw )
+{
+}
diff --git a/src/gallium/drivers/i965/brw_pipe_rast.h b/src/gallium/drivers/i965/brw_pipe_rast.h
new file mode 100644 (file)
index 0000000..9354f01
--- /dev/null
@@ -0,0 +1,16 @@
+#ifndef BRW_PIPE_RAST_H
+#define BRW_PIPE_RAST_H
+
+#include "brw_clip.h"
+
+struct brw_rasterizer_state {
+   struct pipe_rasterizer_state templ; /* for draw module */
+
+   /* Precalculated hardware state:
+    */
+   struct brw_clip_prog_key clip_key;
+   struct brw_line_stipple bls;
+   unsigned unfilled_aa_line;
+};
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_pipe_sampler.c b/src/gallium/drivers/i965/brw_pipe_sampler.c
new file mode 100644 (file)
index 0000000..5ddc63f
--- /dev/null
@@ -0,0 +1,233 @@
+
+#include "util/u_memory.h"
+#include "util/u_math.h"
+
+#include "pipe/p_context.h"
+#include "pipe/p_state.h"
+
+#include "brw_context.h"
+#include "brw_defines.h"
+#include "brw_debug.h"
+
+
+
+/* The brw (and related graphics cores) do not support GL_CLAMP.  The
+ * Intel drivers for "other operating systems" implement GL_CLAMP as
+ * GL_CLAMP_TO_EDGE, so the same is done here.
+ */
+static GLuint translate_wrap_mode( unsigned wrap )
+{
+   switch( wrap ) {
+   case PIPE_TEX_WRAP_REPEAT: 
+      return BRW_TEXCOORDMODE_WRAP;
+
+   case PIPE_TEX_WRAP_CLAMP:
+   case PIPE_TEX_WRAP_CLAMP_TO_EDGE:
+      return BRW_TEXCOORDMODE_CLAMP;
+      
+   case PIPE_TEX_WRAP_CLAMP_TO_BORDER:
+      return BRW_TEXCOORDMODE_CLAMP_BORDER;
+
+   case PIPE_TEX_WRAP_MIRROR_REPEAT: 
+      return BRW_TEXCOORDMODE_MIRROR;
+
+   case PIPE_TEX_WRAP_MIRROR_CLAMP: 
+   case PIPE_TEX_WRAP_MIRROR_CLAMP_TO_EDGE: 
+   case PIPE_TEX_WRAP_MIRROR_CLAMP_TO_BORDER: 
+      return BRW_TEXCOORDMODE_MIRROR_ONCE;
+
+   default: 
+      return BRW_TEXCOORDMODE_WRAP;
+   }
+}
+
+static GLuint translate_img_filter( unsigned filter )
+{
+   switch (filter) {
+   case PIPE_TEX_FILTER_NEAREST:
+      return BRW_MAPFILTER_NEAREST;
+   case PIPE_TEX_FILTER_LINEAR:
+      return BRW_MAPFILTER_LINEAR;
+   case PIPE_TEX_FILTER_ANISO:
+      return BRW_MAPFILTER_ANISOTROPIC;
+   default:
+      assert(0);
+      return BRW_MAPFILTER_NEAREST;
+   }
+}
+
+static GLuint translate_mip_filter( unsigned filter )
+{
+   switch (filter) {
+   case PIPE_TEX_MIPFILTER_NONE: 
+      return BRW_MIPFILTER_NONE;
+   case PIPE_TEX_MIPFILTER_NEAREST:
+      return BRW_MIPFILTER_NEAREST;
+   case PIPE_TEX_MIPFILTER_LINEAR:
+      return BRW_MIPFILTER_LINEAR;
+   default:
+      assert(0);
+      return BRW_MIPFILTER_NONE;
+   }
+}
+
+/* XXX: not sure why there are special translations for the shadow tex
+ * compare functions.  In particular ALWAYS is translated to NEVER.
+ * Is this a hardware issue?  Does i965 really suffer from this?
+ */
+static GLuint translate_shadow_compare_func( unsigned func )
+{
+   switch (func) {
+   case PIPE_FUNC_NEVER: 
+       return BRW_COMPAREFUNCTION_ALWAYS;
+   case PIPE_FUNC_LESS: 
+       return BRW_COMPAREFUNCTION_LEQUAL;
+   case PIPE_FUNC_LEQUAL: 
+       return BRW_COMPAREFUNCTION_LESS;
+   case PIPE_FUNC_GREATER: 
+       return BRW_COMPAREFUNCTION_GEQUAL;
+   case PIPE_FUNC_GEQUAL: 
+      return BRW_COMPAREFUNCTION_GREATER;
+   case PIPE_FUNC_NOTEQUAL: 
+      return BRW_COMPAREFUNCTION_EQUAL;
+   case PIPE_FUNC_EQUAL: 
+      return BRW_COMPAREFUNCTION_NOTEQUAL;
+   case PIPE_FUNC_ALWAYS: 
+       return BRW_COMPAREFUNCTION_NEVER;
+   default:
+      assert(0);
+      return BRW_COMPAREFUNCTION_NEVER;
+   }
+}
+
+
+
+
+static void *
+brw_create_sampler_state( struct pipe_context *pipe,
+                          const struct pipe_sampler_state *template )
+{
+   struct brw_sampler *sampler = CALLOC_STRUCT(brw_sampler);
+
+   sampler->ss0.min_filter = translate_img_filter( template->min_img_filter );
+   sampler->ss0.mag_filter = translate_img_filter( template->mag_img_filter );
+   sampler->ss0.mip_filter = translate_mip_filter( template->min_mip_filter );
+
+
+   /* XXX: anisotropy logic slightly changed: 
+    */
+   if (template->max_anisotropy > 1.0) {
+      sampler->ss0.min_filter = BRW_MAPFILTER_ANISOTROPIC; 
+      sampler->ss0.mag_filter = BRW_MAPFILTER_ANISOTROPIC;
+
+      if (template->max_anisotropy > 2.0) {
+        sampler->ss3.max_aniso = MIN2((template->max_anisotropy - 2) / 2,
+                                      BRW_ANISORATIO_16);
+      }
+   }
+
+   sampler->ss1.r_wrap_mode = translate_wrap_mode(template->wrap_r);
+   sampler->ss1.s_wrap_mode = translate_wrap_mode(template->wrap_s);
+   sampler->ss1.t_wrap_mode = translate_wrap_mode(template->wrap_t);
+
+   /* Set LOD bias: 
+    */
+   sampler->ss0.lod_bias = 
+      util_signed_fixed(CLAMP(template->lod_bias, -16, 15), 6);
+
+
+   sampler->ss0.lod_preclamp = 1; /* OpenGL mode */
+   sampler->ss0.default_color_mode = 0; /* OpenGL/DX10 mode */
+
+   /* Set shadow function: 
+    */
+   if (template->compare_mode == PIPE_TEX_COMPARE_R_TO_TEXTURE) {
+
+      /* Shadowing is "enabled" by emitting a particular sampler
+       * message (sample_c).  So need to recompile WM program when
+       * shadow comparison is enabled on each/any texture unit.
+       */
+      sampler->ss0.shadow_function =
+        translate_shadow_compare_func(template->compare_func);
+   }
+
+   /* Set BaseMipLevel, MaxLOD, MinLOD: 
+    */
+   sampler->ss0.base_level = 
+      util_unsigned_fixed(0, 1);
+
+   sampler->ss1.max_lod = 
+      util_unsigned_fixed(CLAMP(template->max_lod, 0, 13), 6);
+
+   sampler->ss1.min_lod = 
+      util_unsigned_fixed(CLAMP(template->min_lod, 0, 13), 6);
+
+   return (void *)sampler;
+}
+
+static void brw_bind_sampler_state(struct pipe_context *pipe,
+                                   unsigned num, void **sampler)
+{
+   struct brw_context *brw = brw_context(pipe);
+   int i;
+
+   for (i = 0; i < num; i++)
+      brw->curr.sampler[i] = sampler[i];
+
+   for (i = num; i < brw->curr.num_samplers; i++)
+      brw->curr.sampler[i] = NULL;
+
+   brw->curr.num_samplers = num;
+   brw->state.dirty.mesa |= PIPE_NEW_SAMPLERS;
+}
+
+static void brw_delete_sampler_state(struct pipe_context *pipe,
+                                 void *cso)
+{
+   FREE(cso);
+}
+
+static void brw_set_sampler_textures(struct pipe_context *pipe,
+                                    unsigned num,
+                                    struct pipe_texture **texture)
+{
+   struct brw_context *brw = brw_context(pipe);
+   int i;
+
+   for (i = 0; i < num; i++)
+      pipe_texture_reference(&brw->curr.texture[i], texture[i]);
+
+   for (i = num; i < brw->curr.num_textures; i++)
+      pipe_texture_reference(&brw->curr.texture[i], NULL);
+
+   brw->curr.num_textures = num;
+   brw->state.dirty.mesa |= PIPE_NEW_BOUND_TEXTURES;
+}
+
+static void brw_set_vertex_sampler_textures(struct pipe_context *pipe,
+                                            unsigned num,
+                                            struct pipe_texture **texture)
+{
+}
+
+static void brw_bind_vertex_sampler_state(struct pipe_context *pipe,
+                                          unsigned num, void **sampler)
+{
+}
+
+
+void brw_pipe_sampler_init( struct brw_context *brw )
+{
+   brw->base.create_sampler_state = brw_create_sampler_state;
+   brw->base.delete_sampler_state = brw_delete_sampler_state;
+
+   brw->base.set_fragment_sampler_textures = brw_set_sampler_textures;
+   brw->base.bind_fragment_sampler_states = brw_bind_sampler_state;
+
+   brw->base.set_vertex_sampler_textures = brw_set_vertex_sampler_textures;
+   brw->base.bind_vertex_sampler_states = brw_bind_vertex_sampler_state;
+
+}
+void brw_pipe_sampler_cleanup( struct brw_context *brw )
+{
+}
diff --git a/src/gallium/drivers/i965/brw_pipe_shader.c b/src/gallium/drivers/i965/brw_pipe_shader.c
new file mode 100644 (file)
index 0000000..bb32d90
--- /dev/null
@@ -0,0 +1,303 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "util/u_memory.h"
+  
+#include "tgsi/tgsi_parse.h"
+#include "tgsi/tgsi_scan.h"
+
+#include "brw_context.h"
+#include "brw_util.h"
+#include "brw_wm.h"
+
+
+/**
+ * Determine if the given shader uses complex features such as flow
+ * conditionals, loops, subroutines.
+ */
+static GLboolean has_flow_control(const struct tgsi_shader_info *info)
+{
+    return (info->opcode_count[TGSI_OPCODE_ARL] > 0 ||
+           info->opcode_count[TGSI_OPCODE_IF] > 0 ||
+           info->opcode_count[TGSI_OPCODE_ENDIF] > 0 || /* redundant - IF */
+           info->opcode_count[TGSI_OPCODE_CAL] > 0 ||
+           info->opcode_count[TGSI_OPCODE_BRK] > 0 ||   /* redundant - BGNLOOP */
+           info->opcode_count[TGSI_OPCODE_RET] > 0 ||   /* redundant - CAL */
+           info->opcode_count[TGSI_OPCODE_BGNLOOP] > 0);
+}
+
+
+static void scan_immediates(const struct tgsi_token *tokens,
+                            const struct tgsi_shader_info *info,
+                            struct brw_immediate_data *imm)
+{
+   struct tgsi_parse_context parse;
+   boolean done = FALSE;
+
+   imm->nr = 0;
+   imm->data = MALLOC(info->immediate_count * 4 * sizeof(float));
+
+   tgsi_parse_init( &parse, tokens );
+   while (!tgsi_parse_end_of_tokens( &parse ) && !done) {
+      tgsi_parse_token( &parse );
+
+      switch (parse.FullToken.Token.Type) {
+      case TGSI_TOKEN_TYPE_DECLARATION:
+         break;
+
+      case TGSI_TOKEN_TYPE_IMMEDIATE: {
+        static const float id[4] = {0,0,0,1};
+        const float *value = &parse.FullToken.FullImmediate.u[0].Float;
+        unsigned size = parse.FullToken.FullImmediate.Immediate.NrTokens - 1;
+         unsigned i;
+
+        for (i = 0; i < size; i++)
+           imm->data[imm->nr][i] = value[i];
+
+        for (; i < 4; i++)
+           imm->data[imm->nr][i] = id[i];
+         
+         imm->nr++;
+        break;
+      }
+
+      case TGSI_TOKEN_TYPE_INSTRUCTION:
+        done = 1;
+        break;
+      }
+   }
+}
+
+
+static void brw_bind_fs_state( struct pipe_context *pipe, void *prog )
+{
+   struct brw_fragment_shader *fs = (struct brw_fragment_shader *)prog;
+   struct brw_context *brw = brw_context(pipe);
+   
+   if (brw->curr.fragment_shader == fs)
+      return;
+
+   if (brw->curr.fragment_shader == NULL ||
+       fs == NULL ||
+       memcmp(&brw->curr.fragment_shader->signature, &fs->signature,
+              brw_fs_signature_size(&fs->signature)) != 0) {
+      brw->state.dirty.mesa |= PIPE_NEW_FRAGMENT_SIGNATURE;
+   }
+
+   brw->curr.fragment_shader = fs;
+   brw->state.dirty.mesa |= PIPE_NEW_FRAGMENT_SHADER;
+}
+
+static void brw_bind_vs_state( struct pipe_context *pipe, void *prog )
+{
+   struct brw_context *brw = brw_context(pipe);
+
+   brw->curr.vertex_shader = (struct brw_vertex_shader *)prog;
+   brw->state.dirty.mesa |= PIPE_NEW_VERTEX_SHADER;
+}
+
+
+
+static void *brw_create_fs_state( struct pipe_context *pipe,
+                                 const struct pipe_shader_state *shader )
+{
+   struct brw_context *brw = brw_context(pipe);
+   struct brw_fragment_shader *fs;
+   int i;
+
+   fs = CALLOC_STRUCT(brw_fragment_shader);
+   if (fs == NULL)
+      return NULL;
+
+   /* Duplicate tokens, scan shader
+    */
+   fs->id = brw->program_id++;
+   fs->has_flow_control = has_flow_control(&fs->info);
+
+   fs->tokens = tgsi_dup_tokens(shader->tokens);
+   if (fs->tokens == NULL)
+      goto fail;
+
+   tgsi_scan_shader(fs->tokens, &fs->info);
+   scan_immediates(fs->tokens, &fs->info, &fs->immediates);
+
+   fs->signature.nr_inputs = fs->info.num_inputs;
+   for (i = 0; i < fs->info.num_inputs; i++) {
+      fs->signature.input[i].interp = fs->info.input_interpolate[i];
+      fs->signature.input[i].semantic = fs->info.input_semantic_name[i];
+      fs->signature.input[i].semantic_index = fs->info.input_semantic_index[i];
+   }
+
+   for (i = 0; i < fs->info.num_inputs; i++)
+      if (fs->info.input_semantic_name[i] == TGSI_SEMANTIC_POSITION)
+        fs->uses_depth = 1;
+
+   if (fs->info.uses_kill)
+      fs->iz_lookup |= IZ_PS_KILL_ALPHATEST_BIT;
+
+   if (fs->info.writes_z)
+      fs->iz_lookup |= IZ_PS_COMPUTES_DEPTH_BIT;
+
+   return (void *)fs;
+
+fail:
+   FREE(fs);
+   return NULL;
+}
+
+
+static void *brw_create_vs_state( struct pipe_context *pipe,
+                                 const struct pipe_shader_state *shader )
+{
+   struct brw_context *brw = brw_context(pipe);
+   struct brw_vertex_shader *vs;
+   unsigned i;
+
+   vs = CALLOC_STRUCT(brw_vertex_shader);
+   if (vs == NULL)
+      return NULL;
+
+   /* Duplicate tokens, scan shader
+    */
+   vs->tokens = tgsi_dup_tokens(shader->tokens);
+   if (vs->tokens == NULL)
+      goto fail;
+
+   tgsi_scan_shader(vs->tokens, &vs->info);
+   scan_immediates(vs->tokens, &vs->info, &vs->immediates);
+
+   vs->id = brw->program_id++;
+   vs->has_flow_control = has_flow_control(&vs->info);
+
+   vs->output_hpos = BRW_OUTPUT_NOT_PRESENT;
+   vs->output_color0 = BRW_OUTPUT_NOT_PRESENT;
+   vs->output_color1 = BRW_OUTPUT_NOT_PRESENT;
+   vs->output_bfc0 = BRW_OUTPUT_NOT_PRESENT;
+   vs->output_bfc1 = BRW_OUTPUT_NOT_PRESENT;
+   vs->output_edgeflag = BRW_OUTPUT_NOT_PRESENT;
+
+   for (i = 0; i < vs->info.num_outputs; i++) {
+      int index = vs->info.output_semantic_index[i];
+      switch (vs->info.output_semantic_name[i]) {
+      case TGSI_SEMANTIC_POSITION:
+         vs->output_hpos = i;
+         break;
+      case TGSI_SEMANTIC_COLOR:
+         if (index == 0)
+            vs->output_color0 = i;
+         else
+            vs->output_color1 = i;
+         break;
+      case TGSI_SEMANTIC_BCOLOR:
+         if (index == 0)
+            vs->output_bfc0 = i;
+         else
+            vs->output_bfc1 = i;
+         break;
+      case TGSI_SEMANTIC_EDGEFLAG:
+         vs->output_edgeflag = i;
+         break;
+      }
+   }
+
+   
+   /* Done:
+    */
+   return (void *)vs;
+
+fail:
+   FREE(vs);
+   return NULL;
+}
+
+
+static void brw_delete_fs_state( struct pipe_context *pipe, void *prog )
+{
+   struct brw_fragment_shader *fs = (struct brw_fragment_shader *)prog;
+
+   bo_reference(&fs->const_buffer, NULL);
+   FREE( (void *)fs->tokens );
+   FREE( fs );
+}
+
+
+static void brw_delete_vs_state( struct pipe_context *pipe, void *prog )
+{
+   struct brw_fragment_shader *vs = (struct brw_fragment_shader *)prog;
+
+   /* Delete draw shader
+    */
+   FREE( (void *)vs->tokens );
+   FREE( vs );
+}
+
+
+static void brw_set_constant_buffer(struct pipe_context *pipe,
+                                     uint shader, uint index,
+                                     const struct pipe_constant_buffer *buf)
+{
+   struct brw_context *brw = brw_context(pipe);
+
+   assert(index == 0);
+
+   if (shader == PIPE_SHADER_FRAGMENT) {
+      pipe_buffer_reference( &brw->curr.fragment_constants,
+                             buf->buffer );
+
+      brw->state.dirty.mesa |= PIPE_NEW_FRAGMENT_CONSTANTS;
+   }
+   else {
+      pipe_buffer_reference( &brw->curr.vertex_constants,
+                             buf->buffer );
+
+      brw->state.dirty.mesa |= PIPE_NEW_VERTEX_CONSTANTS;
+   }
+}
+
+
+void brw_pipe_shader_init( struct brw_context *brw )
+{
+   brw->base.set_constant_buffer = brw_set_constant_buffer;
+
+   brw->base.create_vs_state = brw_create_vs_state;
+   brw->base.bind_vs_state = brw_bind_vs_state;
+   brw->base.delete_vs_state = brw_delete_vs_state;
+
+   brw->base.create_fs_state = brw_create_fs_state;
+   brw->base.bind_fs_state = brw_bind_fs_state;
+   brw->base.delete_fs_state = brw_delete_fs_state;
+}
+
+void brw_pipe_shader_cleanup( struct brw_context *brw )
+{
+   pipe_buffer_reference( &brw->curr.fragment_constants, NULL );
+   pipe_buffer_reference( &brw->curr.vertex_constants, NULL );
+}
diff --git a/src/gallium/drivers/i965/brw_pipe_vertex.c b/src/gallium/drivers/i965/brw_pipe_vertex.c
new file mode 100644 (file)
index 0000000..e3c48e3
--- /dev/null
@@ -0,0 +1,71 @@
+#include "brw_context.h"
+
+
+static void brw_set_vertex_elements( struct pipe_context *pipe,
+                                    unsigned count,
+                                    const struct pipe_vertex_element *elements )
+{
+   struct brw_context *brw = brw_context(pipe);
+
+   memcpy(brw->curr.vertex_element, elements, count * sizeof(elements[0]));
+   brw->curr.num_vertex_elements = count;
+
+   brw->state.dirty.mesa |= PIPE_NEW_VERTEX_ELEMENT;
+}
+
+
+static void brw_set_vertex_buffers(struct pipe_context *pipe,
+                                  unsigned count,
+                                  const struct pipe_vertex_buffer *buffers)
+{
+   struct brw_context *brw = brw_context(pipe);
+   unsigned i;
+
+   /* Check for no change */
+   if (count == brw->curr.num_vertex_buffers &&
+       memcmp(brw->curr.vertex_buffer,
+              buffers,
+              count * sizeof buffers[0]) == 0)
+      return;
+
+   /* Adjust refcounts */
+   for (i = 0; i < count; i++) 
+      pipe_buffer_reference(&brw->curr.vertex_buffer[i].buffer, 
+                            buffers[i].buffer);
+
+   for ( ; i < brw->curr.num_vertex_buffers; i++)
+      pipe_buffer_reference(&brw->curr.vertex_buffer[i].buffer,
+                            NULL);
+
+   /* Copy remaining data */
+   memcpy(brw->curr.vertex_buffer, buffers, count * sizeof buffers[0]);
+   brw->curr.num_vertex_buffers = count;
+
+   brw->state.dirty.mesa |= PIPE_NEW_VERTEX_BUFFER;
+}
+
+
+void 
+brw_pipe_vertex_init( struct brw_context *brw )
+{
+   brw->base.set_vertex_buffers = brw_set_vertex_buffers;
+   brw->base.set_vertex_elements = brw_set_vertex_elements;
+}
+
+
+void 
+brw_pipe_vertex_cleanup( struct brw_context *brw )
+{
+
+   /* Release bound pipe vertex_buffers
+    */
+
+   /* Release some other stuff
+    */
+#if 0
+   for (i = 0; i < PIPE_MAX_ATTRIBS; i++) {
+      bo_reference(&brw->vb.inputs[i].bo, NULL);
+      brw->vb.inputs[i].bo = NULL;
+   }
+#endif
+}
diff --git a/src/gallium/drivers/i965/brw_reg.h b/src/gallium/drivers/i965/brw_reg.h
new file mode 100644 (file)
index 0000000..a63403b
--- /dev/null
@@ -0,0 +1,115 @@
+/**************************************************************************
+ * 
+ * Copyright 2003 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef BRW_REG_H
+#define BRW_REG_H
+
+#define CMD_MI                         (0x0 << 29)
+#define CMD_2D                         (0x2 << 29)
+#define CMD_3D                         (0x3 << 29)
+
+#define MI_NOOP                                (CMD_MI | 0)
+#define MI_BATCH_BUFFER_END            (CMD_MI | 0xA << 23)
+#define MI_FLUSH                       (CMD_MI | (4 << 23))
+
+#define _3DSTATE_DRAWRECT_INFO_I965    (CMD_3D | (3 << 27) | (1 << 24) | 0x2)
+
+/** @{
+ *
+ * PIPE_CONTROL operation, a combination MI_FLUSH and register write with
+ * additional flushing control.
+ */
+#define _3DSTATE_PIPE_CONTROL          (CMD_3D | (3 << 27) | (2 << 24) | 2)
+#define PIPE_CONTROL_NO_WRITE          (0 << 14)
+#define PIPE_CONTROL_WRITE_IMMEDIATE   (1 << 14)
+#define PIPE_CONTROL_WRITE_DEPTH_COUNT (2 << 14)
+#define PIPE_CONTROL_WRITE_TIMESTAMP   (3 << 14)
+#define PIPE_CONTROL_DEPTH_STALL       (1 << 13)
+#define PIPE_CONTROL_WRITE_FLUSH       (1 << 12)
+#define PIPE_CONTROL_INSTRUCTION_FLUSH (1 << 11)
+#define PIPE_CONTROL_INTERRUPT_ENABLE  (1 << 8)
+#define PIPE_CONTROL_PPGTT_WRITE       (0 << 2)
+#define PIPE_CONTROL_GLOBAL_GTT_WRITE  (1 << 2)
+
+/** @} */
+
+#define XY_SETUP_BLT_CMD               (CMD_2D | (0x01 << 22) | 6)
+#define XY_COLOR_BLT_CMD               (CMD_2D | (0x50 << 22) | 4)
+#define XY_SRC_COPY_BLT_CMD             (CMD_2D | (0x53 << 22) | 6)
+
+/* BR00 */
+#define XY_BLT_WRITE_ALPHA     (1 << 21)
+#define XY_BLT_WRITE_RGB       (1 << 20)
+#define XY_SRC_TILED           (1 << 15)
+#define XY_DST_TILED           (1 << 11)
+
+/* BR13 */
+#define BR13_565               (0x1 << 24)
+#define BR13_8888              (0x3 << 24)
+
+#define FENCE_LINEAR 0
+#define FENCE_XMAJOR 1
+#define FENCE_YMAJOR 2
+
+
+
+/* PCI IDs
+ */
+#define PCI_CHIP_I965_G                        0x29A2
+#define PCI_CHIP_I965_Q                        0x2992
+#define PCI_CHIP_I965_G_1              0x2982
+#define PCI_CHIP_I946_GZ               0x2972
+#define PCI_CHIP_I965_GM                0x2A02
+#define PCI_CHIP_I965_GME               0x2A12
+
+#define PCI_CHIP_GM45_GM                0x2A42
+
+#define PCI_CHIP_IGD_E_G                0x2E02
+#define PCI_CHIP_Q45_G                  0x2E12
+#define PCI_CHIP_G45_G                  0x2E22
+#define PCI_CHIP_G41_G                  0x2E32
+#define PCI_CHIP_B43_G                  0x2E42
+
+#define PCI_CHIP_ILD_G                  0x0042
+#define PCI_CHIP_ILM_G                  0x0046
+
+struct brw_chipset {
+   unsigned pci_id:16;
+   unsigned is_965:1;
+   unsigned is_igdng:1;
+   unsigned is_g4x:1;
+   unsigned pad:13;
+};
+
+
+/* XXX: hacks
+ */
+#define VERT_RESULT_HPOS 0     /* not always true */
+#define VERT_RESULT_PSIZ 10000 /* disabled */
+
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_screen.c b/src/gallium/drivers/i965/brw_screen.c
new file mode 100644 (file)
index 0000000..0ecacac
--- /dev/null
@@ -0,0 +1,403 @@
+/**************************************************************************
+ * 
+ * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+
+#include "pipe/p_inlines.h"
+#include "util/u_memory.h"
+#include "util/u_string.h"
+
+#include "brw_reg.h"
+#include "brw_context.h"
+#include "brw_screen.h"
+#include "brw_winsys.h"
+#include "brw_debug.h"
+
+#ifdef DEBUG
+static const struct debug_named_value debug_names[] = {
+   { "tex",   DEBUG_TEXTURE},
+   { "state", DEBUG_STATE},
+   { "ioctl", DEBUG_IOCTL},
+   { "blit",  DEBUG_BLIT},
+   { "curbe", DEBUG_CURBE},
+   { "fall",  DEBUG_FALLBACKS},
+   { "verb",  DEBUG_VERBOSE},
+   { "bat",   DEBUG_BATCH},
+   { "pix",   DEBUG_PIXEL},
+   { "wins",  DEBUG_WINSYS},
+   { "min",   DEBUG_MIN_URB},
+   { "dis",   DEBUG_DISASSEM},
+   { "sync",  DEBUG_SYNC},
+   { "prim",  DEBUG_PRIMS },
+   { "vert",  DEBUG_VERTS },
+   { "dma",   DEBUG_DMA },
+   { "san",   DEBUG_SANITY },
+   { "sleep", DEBUG_SLEEP },
+   { "stats", DEBUG_STATS },
+   { "sing",  DEBUG_SINGLE_THREAD },
+   { "thre",  DEBUG_SINGLE_THREAD },
+   { "wm",    DEBUG_WM },
+   { "urb",   DEBUG_URB },
+   { "vs",    DEBUG_VS },
+   { NULL,    0 }
+};
+
+static const struct debug_named_value dump_names[] = {
+   { "asm",   DUMP_ASM},
+   { "state", DUMP_STATE},
+   { "batch", DUMP_BATCH},
+   { NULL, 0 }
+};
+
+int BRW_DEBUG = 0;
+int BRW_DUMP = 0;
+
+#endif
+
+
+/*
+ * Probe functions
+ */
+
+
+static const char *
+brw_get_vendor(struct pipe_screen *screen)
+{
+   return "VMware, Inc.";
+}
+
+static const char *
+brw_get_name(struct pipe_screen *screen)
+{
+   static char buffer[128];
+   const char *chipset;
+
+   switch (brw_screen(screen)->chipset.pci_id) {
+   case PCI_CHIP_I965_G:
+      chipset = "I965_G";
+      break;
+   case PCI_CHIP_I965_Q:
+      chipset = "I965_Q";
+      break;
+   case PCI_CHIP_I965_G_1:
+      chipset = "I965_G_1";
+      break;
+   case PCI_CHIP_I946_GZ:
+      chipset = "I946_GZ";
+      break;
+   case PCI_CHIP_I965_GM:
+      chipset = "I965_GM";
+      break;
+   case PCI_CHIP_I965_GME:
+      chipset = "I965_GME";
+      break;
+   case PCI_CHIP_GM45_GM:
+      chipset = "GM45_GM";
+      break;
+   case PCI_CHIP_IGD_E_G:
+      chipset = "IGD_E_G";
+      break;
+   case PCI_CHIP_Q45_G:
+      chipset = "Q45_G";
+      break;
+   case PCI_CHIP_G45_G:
+      chipset = "G45_G";
+      break;
+   case PCI_CHIP_G41_G:
+      chipset = "G41_G";
+      break;
+   case PCI_CHIP_B43_G:
+      chipset = "B43_G";
+      break;
+   case PCI_CHIP_ILD_G:
+      chipset = "ILD_G";
+      break;
+   case PCI_CHIP_ILM_G:
+      chipset = "ILM_G";
+      break;
+   }
+
+   util_snprintf(buffer, sizeof(buffer), "i965 (chipset: %s)", chipset);
+   return buffer;
+}
+
+static int
+brw_get_param(struct pipe_screen *screen, int param)
+{
+   switch (param) {
+   case PIPE_CAP_MAX_TEXTURE_IMAGE_UNITS:
+      return 8;
+   case PIPE_CAP_NPOT_TEXTURES:
+      return 1;
+   case PIPE_CAP_TWO_SIDED_STENCIL:
+      return 1;
+   case PIPE_CAP_GLSL:
+      return 0;
+   case PIPE_CAP_ANISOTROPIC_FILTER:
+      return 0;
+   case PIPE_CAP_POINT_SPRITE:
+      return 0;
+   case PIPE_CAP_MAX_RENDER_TARGETS:
+      return 1;
+   case PIPE_CAP_OCCLUSION_QUERY:
+      return 0;
+   case PIPE_CAP_TEXTURE_SHADOW_MAP:
+      return 1;
+   case PIPE_CAP_MAX_TEXTURE_2D_LEVELS:
+      return 11; /* max 1024x1024 */
+   case PIPE_CAP_MAX_TEXTURE_3D_LEVELS:
+      return 8;  /* max 128x128x128 */
+   case PIPE_CAP_MAX_TEXTURE_CUBE_LEVELS:
+      return 11; /* max 1024x1024 */
+   default:
+      return 0;
+   }
+}
+
+static float
+brw_get_paramf(struct pipe_screen *screen, int param)
+{
+   switch (param) {
+   case PIPE_CAP_MAX_LINE_WIDTH:
+      /* fall-through */
+   case PIPE_CAP_MAX_LINE_WIDTH_AA:
+      return 7.5;
+
+   case PIPE_CAP_MAX_POINT_WIDTH:
+      /* fall-through */
+   case PIPE_CAP_MAX_POINT_WIDTH_AA:
+      return 255.0;
+
+   case PIPE_CAP_MAX_TEXTURE_ANISOTROPY:
+      return 4.0;
+
+   case PIPE_CAP_MAX_TEXTURE_LOD_BIAS:
+      return 16.0;
+
+   default:
+      return 0;
+   }
+}
+
+static boolean
+brw_is_format_supported(struct pipe_screen *screen,
+                         enum pipe_format format, 
+                         enum pipe_texture_target target,
+                         unsigned tex_usage, 
+                         unsigned geom_flags)
+{
+   static const enum pipe_format tex_supported[] = {
+      PIPE_FORMAT_L8_UNORM,
+      PIPE_FORMAT_I8_UNORM,
+      PIPE_FORMAT_A8_UNORM,
+      PIPE_FORMAT_L16_UNORM,
+      /*PIPE_FORMAT_I16_UNORM,*/
+      /*PIPE_FORMAT_A16_UNORM,*/
+      PIPE_FORMAT_A8L8_UNORM,
+      PIPE_FORMAT_R5G6B5_UNORM,
+      PIPE_FORMAT_A1R5G5B5_UNORM,
+      PIPE_FORMAT_A4R4G4B4_UNORM,
+      PIPE_FORMAT_X8R8G8B8_UNORM,
+      PIPE_FORMAT_A8R8G8B8_UNORM,
+      /* video */
+      PIPE_FORMAT_YCBCR,
+      PIPE_FORMAT_YCBCR_REV,
+      /* compressed */
+      /*PIPE_FORMAT_FXT1_RGBA,*/
+      PIPE_FORMAT_DXT1_RGB,
+      PIPE_FORMAT_DXT1_RGBA,
+      PIPE_FORMAT_DXT3_RGBA,
+      PIPE_FORMAT_DXT5_RGBA,
+      /* sRGB */
+      PIPE_FORMAT_R8G8B8A8_SRGB,
+      PIPE_FORMAT_A8L8_SRGB,
+      PIPE_FORMAT_L8_SRGB,
+      PIPE_FORMAT_DXT1_SRGB,
+      /* depth */
+      PIPE_FORMAT_Z32_FLOAT,
+      PIPE_FORMAT_X8Z24_UNORM,
+      PIPE_FORMAT_S8Z24_UNORM,
+      PIPE_FORMAT_Z16_UNORM,
+      /* signed */
+      PIPE_FORMAT_R8G8_SNORM,
+      PIPE_FORMAT_R8G8B8A8_SNORM,
+      PIPE_FORMAT_NONE  /* list terminator */
+   };
+   static const enum pipe_format render_supported[] = {
+      PIPE_FORMAT_X8R8G8B8_UNORM,
+      PIPE_FORMAT_A8R8G8B8_UNORM,
+      PIPE_FORMAT_R5G6B5_UNORM,
+      PIPE_FORMAT_NONE  /* list terminator */
+   };
+   static const enum pipe_format depth_supported[] = {
+      PIPE_FORMAT_Z32_FLOAT,
+      PIPE_FORMAT_X8Z24_UNORM,
+      PIPE_FORMAT_S8Z24_UNORM,
+      PIPE_FORMAT_Z16_UNORM,
+      PIPE_FORMAT_NONE  /* list terminator */
+   };
+   const enum pipe_format *list;
+   uint i;
+
+   if (tex_usage & PIPE_TEXTURE_USAGE_DEPTH_STENCIL)
+      list = depth_supported;
+   else if (tex_usage & PIPE_TEXTURE_USAGE_RENDER_TARGET)
+      list = render_supported;
+   else
+      list = tex_supported;
+
+   for (i = 0; list[i] != PIPE_FORMAT_NONE; i++) {
+      if (list[i] == format)
+         return TRUE;
+   }
+
+   return FALSE;
+}
+
+
+/*
+ * Fence functions
+ */
+
+
+static void
+brw_fence_reference(struct pipe_screen *screen,
+                     struct pipe_fence_handle **ptr,
+                     struct pipe_fence_handle *fence)
+{
+}
+
+static int
+brw_fence_signalled(struct pipe_screen *screen,
+                     struct pipe_fence_handle *fence,
+                     unsigned flags)
+{
+   return 0;                    /* XXX shouldn't this be a boolean? */
+}
+
+static int
+brw_fence_finish(struct pipe_screen *screen,
+                 struct pipe_fence_handle *fence,
+                 unsigned flags)
+{
+   return 0;
+}
+
+
+/*
+ * Generic functions
+ */
+
+
+static void
+brw_destroy_screen(struct pipe_screen *screen)
+{
+   struct brw_screen *bscreen = brw_screen(screen);
+
+   if (bscreen->sws)
+      bscreen->sws->destroy(bscreen->sws);
+
+   FREE(bscreen);
+}
+
+/**
+ * Create a new brw_screen object
+ */
+struct pipe_screen *
+brw_create_screen(struct brw_winsys_screen *sws, uint pci_id)
+{
+   struct brw_screen *bscreen;
+   struct brw_chipset chipset;
+
+#ifdef DEBUG
+   BRW_DEBUG = debug_get_flags_option("BRW_DEBUG", debug_names, 0);
+   BRW_DEBUG |= debug_get_flags_option("INTEL_DEBUG", debug_names, 0);
+   BRW_DEBUG |= DEBUG_STATS | DEBUG_MIN_URB | DEBUG_WM;
+
+   BRW_DUMP = debug_get_flags_option("BRW_DUMP", dump_names, 0);
+#endif
+
+   memset(&chipset, 0, sizeof chipset);
+
+   chipset.pci_id = pci_id;
+
+   switch (pci_id) {
+   case PCI_CHIP_I965_G:
+   case PCI_CHIP_I965_Q:
+   case PCI_CHIP_I965_G_1:
+   case PCI_CHIP_I946_GZ:
+   case PCI_CHIP_I965_GM:
+   case PCI_CHIP_I965_GME:
+      chipset.is_965 = TRUE;
+      break;
+
+   case PCI_CHIP_GM45_GM:
+   case PCI_CHIP_IGD_E_G:
+   case PCI_CHIP_Q45_G:
+   case PCI_CHIP_G45_G:
+   case PCI_CHIP_G41_G:
+   case PCI_CHIP_B43_G:
+      chipset.is_g4x = TRUE;
+      break;
+
+   case PCI_CHIP_ILD_G:
+   case PCI_CHIP_ILM_G:
+      chipset.is_igdng = TRUE;
+      break;
+
+   default:
+      debug_printf("%s: unknown pci id 0x%x, cannot create screen\n", 
+                   __FUNCTION__, pci_id);
+      return NULL;
+   }
+
+
+   bscreen = CALLOC_STRUCT(brw_screen);
+   if (!bscreen)
+      return NULL;
+
+   bscreen->chipset = chipset;
+   bscreen->sws = sws;
+   bscreen->base.winsys = NULL;
+   bscreen->base.destroy = brw_destroy_screen;
+   bscreen->base.get_name = brw_get_name;
+   bscreen->base.get_vendor = brw_get_vendor;
+   bscreen->base.get_param = brw_get_param;
+   bscreen->base.get_paramf = brw_get_paramf;
+   bscreen->base.is_format_supported = brw_is_format_supported;
+   bscreen->base.fence_reference = brw_fence_reference;
+   bscreen->base.fence_signalled = brw_fence_signalled;
+   bscreen->base.fence_finish = brw_fence_finish;
+
+   brw_screen_tex_init(bscreen);
+   brw_screen_tex_surface_init(bscreen);
+   brw_screen_buffer_init(bscreen);
+
+   bscreen->no_tiling = debug_get_option("BRW_NO_TILING", FALSE) != NULL;
+   
+   
+   return &bscreen->base;
+}
diff --git a/src/gallium/drivers/i965/brw_screen.h b/src/gallium/drivers/i965/brw_screen.h
new file mode 100644 (file)
index 0000000..7226d92
--- /dev/null
@@ -0,0 +1,199 @@
+/**************************************************************************
+ * 
+ * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef BRW_SCREEN_H
+#define BRW_SCREEN_H
+
+#include "pipe/p_state.h"
+#include "pipe/p_screen.h"
+
+#include "brw_reg.h"
+#include "brw_structs.h"
+
+struct brw_winsys_screen;
+
+
+/**
+ * Subclass of pipe_screen
+ */
+struct brw_screen
+{
+   struct pipe_screen base;
+   struct brw_chipset chipset;
+   struct brw_winsys_screen *sws;
+   boolean no_tiling;
+};
+
+/**
+ * Subclass of pipe_transfer
+ */
+struct brw_transfer
+{
+   struct pipe_transfer base;
+
+   unsigned offset;
+};
+
+struct brw_buffer
+{
+   struct pipe_buffer base;
+
+   /* One of either bo or user_buffer will be non-null, depending on
+    * whether this is a hardware or user buffer.
+    */
+   struct brw_winsys_buffer *bo;
+   void *user_buffer;
+
+   /* Mapped pointer??
+    */
+   void *ptr;
+};
+
+
+union brw_surface_id {
+   struct {
+      unsigned face:3;
+      unsigned zslice:13;
+      unsigned level:16;
+   } bits;
+   unsigned value;
+};
+
+
+struct brw_surface
+{
+   struct pipe_surface base;
+   
+   union brw_surface_id id;
+   unsigned cpp;
+   unsigned pitch;
+   unsigned draw_offset;
+   unsigned tiling;
+
+   struct brw_surface_state ss;
+   struct brw_winsys_buffer *bo;
+   struct brw_surface *next, *prev;
+};
+
+
+
+struct brw_texture
+{
+   struct pipe_texture base;
+   struct brw_winsys_buffer *bo;
+   struct brw_surface_state ss;
+
+   unsigned *image_offset[PIPE_MAX_TEXTURE_LEVELS];
+   unsigned nr_images[PIPE_MAX_TEXTURE_LEVELS];
+   unsigned level_offset[PIPE_MAX_TEXTURE_LEVELS];
+
+   boolean compressed;
+   unsigned brw_target;
+   unsigned pitch;
+   unsigned tiling;
+   unsigned cpp;
+   unsigned total_height;
+
+   struct brw_surface views[2];
+};
+
+
+
+/*
+ * Cast wrappers
+ */
+static INLINE struct brw_screen *
+brw_screen(struct pipe_screen *pscreen)
+{
+   return (struct brw_screen *) pscreen;
+}
+
+static INLINE struct brw_transfer *
+brw_transfer(struct pipe_transfer *transfer)
+{
+   return (struct brw_transfer *)transfer;
+}
+
+static INLINE struct brw_surface *
+brw_surface(struct pipe_surface *surface)
+{
+   return (struct brw_surface *)surface;
+}
+
+static INLINE struct brw_buffer *
+brw_buffer(struct pipe_buffer *buffer)
+{
+   return (struct brw_buffer *)buffer;
+}
+
+static INLINE struct brw_texture *
+brw_texture(struct pipe_texture *texture)
+{
+   return (struct brw_texture *)texture;
+}
+
+
+/* Pipe buffer helpers
+ */
+static INLINE boolean
+brw_buffer_is_user_buffer( const struct pipe_buffer *buf )
+{
+   return ((const struct brw_buffer *)buf)->user_buffer != NULL;
+}
+
+unsigned
+brw_surface_pitch( const struct pipe_surface *surface );
+
+/***********************************************************************
+ * Internal functions 
+ */
+GLboolean brw_texture_layout(struct brw_screen *brw_screen,
+                            struct brw_texture *tex );
+
+void brw_update_texture( struct brw_screen *brw_screen,
+                        struct brw_texture *tex );
+
+
+void brw_screen_tex_init( struct brw_screen *brw_screen );
+void brw_screen_tex_surface_init( struct brw_screen *brw_screen );
+
+void brw_screen_buffer_init(struct brw_screen *brw_screen);
+
+
+boolean brw_is_texture_referenced_by_bo( struct brw_screen *brw_screen,
+                                         struct pipe_texture *texture,
+                                         unsigned face, 
+                                         unsigned level,
+                                         struct brw_winsys_buffer *bo );
+
+boolean brw_is_buffer_referenced_by_bo( struct brw_screen *brw_screen,
+                                        struct pipe_buffer *buffer,
+                                        struct brw_winsys_buffer *bo );
+
+
+
+#endif /* BRW_SCREEN_H */
diff --git a/src/gallium/drivers/i965/brw_screen_buffers.c b/src/gallium/drivers/i965/brw_screen_buffers.c
new file mode 100644 (file)
index 0000000..d8141a3
--- /dev/null
@@ -0,0 +1,202 @@
+
+#include "util/u_memory.h"
+#include "util/u_math.h"
+
+#include "pipe/p_state.h"
+#include "pipe/p_defines.h"
+#include "pipe/p_inlines.h"
+
+#include "brw_screen.h"
+#include "brw_winsys.h"
+
+
+
+static void *
+brw_buffer_map_range( struct pipe_screen *screen,
+                      struct pipe_buffer *buffer,
+                      unsigned offset,
+                      unsigned length,
+                      unsigned usage )
+{
+   struct brw_screen *bscreen = brw_screen(screen); 
+   struct brw_winsys_screen *sws = bscreen->sws;
+   struct brw_buffer *buf = brw_buffer( buffer );
+
+   if (buf->user_buffer)
+      return buf->user_buffer;
+
+   return sws->bo_map( buf->bo, 
+                       BRW_DATA_OTHER,
+                       offset,
+                       length,
+                       (usage & PIPE_BUFFER_USAGE_CPU_WRITE) ? TRUE : FALSE,
+                       (usage & PIPE_BUFFER_USAGE_DISCARD) ? TRUE : FALSE,
+                       (usage & PIPE_BUFFER_USAGE_FLUSH_EXPLICIT) ? TRUE : FALSE);
+}
+
+static void *
+brw_buffer_map( struct pipe_screen *screen,
+                struct pipe_buffer *buffer,
+                unsigned usage )
+{
+   struct brw_screen *bscreen = brw_screen(screen); 
+   struct brw_winsys_screen *sws = bscreen->sws;
+   struct brw_buffer *buf = brw_buffer( buffer );
+
+   if (buf->user_buffer)
+      return buf->user_buffer;
+
+   return sws->bo_map( buf->bo, 
+                       BRW_DATA_OTHER,
+                       0,
+                       buf->base.size,
+                       (usage & PIPE_BUFFER_USAGE_CPU_WRITE) ? TRUE : FALSE,
+                       FALSE,
+                       FALSE);
+}
+
+
+static void 
+brw_buffer_flush_mapped_range( struct pipe_screen *screen,
+                               struct pipe_buffer *buffer,
+                               unsigned offset,
+                               unsigned length )
+{
+   struct brw_screen *bscreen = brw_screen(screen); 
+   struct brw_winsys_screen *sws = bscreen->sws;
+   struct brw_buffer *buf = brw_buffer( buffer );
+
+   if (buf->user_buffer)
+      return;
+
+   sws->bo_flush_range( buf->bo, 
+                        offset,
+                        length );
+}
+
+
+static void 
+brw_buffer_unmap( struct pipe_screen *screen,
+                   struct pipe_buffer *buffer )
+{
+   struct brw_screen *bscreen = brw_screen(screen); 
+   struct brw_winsys_screen *sws = bscreen->sws;
+   struct brw_buffer *buf = brw_buffer( buffer );
+   
+   if (buf->bo)
+      sws->bo_unmap(buf->bo);
+}
+
+static void
+brw_buffer_destroy( struct pipe_buffer *buffer )
+{
+   struct brw_buffer *buf = brw_buffer( buffer );
+
+   assert(!p_atomic_read(&buffer->reference.count));
+
+   bo_reference(&buf->bo, NULL);
+   FREE(buf);
+}
+
+
+static struct pipe_buffer *
+brw_buffer_create(struct pipe_screen *screen,
+                   unsigned alignment,
+                   unsigned usage,
+                   unsigned size)
+{
+   struct brw_screen *bscreen = brw_screen(screen);
+   struct brw_winsys_screen *sws = bscreen->sws;
+   struct brw_buffer *buf;
+   unsigned buffer_type;
+   enum pipe_error ret;
+   
+   buf = CALLOC_STRUCT(brw_buffer);
+   if (!buf)
+      return NULL;
+      
+   pipe_reference_init(&buf->base.reference, 1);
+   buf->base.screen = screen;
+   buf->base.alignment = alignment;
+   buf->base.usage = usage;
+   buf->base.size = size;
+
+   switch (usage & (PIPE_BUFFER_USAGE_VERTEX |
+                    PIPE_BUFFER_USAGE_INDEX |
+                    PIPE_BUFFER_USAGE_PIXEL |
+                    PIPE_BUFFER_USAGE_CONSTANT))
+   {
+   case PIPE_BUFFER_USAGE_VERTEX:
+   case PIPE_BUFFER_USAGE_INDEX:
+   case (PIPE_BUFFER_USAGE_VERTEX|PIPE_BUFFER_USAGE_INDEX):
+      buffer_type = BRW_BUFFER_TYPE_VERTEX;
+      break;
+      
+   case PIPE_BUFFER_USAGE_PIXEL:
+      buffer_type = BRW_BUFFER_TYPE_PIXEL;
+      break;
+
+   case PIPE_BUFFER_USAGE_CONSTANT:
+      buffer_type = BRW_BUFFER_TYPE_SHADER_CONSTANTS;
+      break;
+
+   default:
+      buffer_type = BRW_BUFFER_TYPE_GENERIC;
+      break;
+   }
+   
+   ret = sws->bo_alloc( sws, buffer_type,
+                        size, alignment,
+                        &buf->bo );
+   if (ret != PIPE_OK)
+      return NULL;
+      
+   return &buf->base; 
+}
+
+
+static struct pipe_buffer *
+brw_user_buffer_create(struct pipe_screen *screen,
+                       void *ptr,
+                       unsigned bytes)
+{
+   struct brw_buffer *buf;
+   
+   buf = CALLOC_STRUCT(brw_buffer);
+   if (!buf)
+      return NULL;
+      
+   buf->user_buffer = ptr;
+   
+   pipe_reference_init(&buf->base.reference, 1);
+   buf->base.screen = screen;
+   buf->base.alignment = 1;
+   buf->base.usage = 0;
+   buf->base.size = bytes;
+   
+   return &buf->base; 
+}
+
+
+boolean brw_is_buffer_referenced_by_bo( struct brw_screen *brw_screen,
+                                     struct pipe_buffer *buffer,
+                                     struct brw_winsys_buffer *bo )
+{
+   struct brw_buffer *buf = brw_buffer(buffer);
+   if (buf->bo == NULL)
+      return FALSE;
+
+   return brw_screen->sws->bo_references( bo, buf->bo );
+}
+
+   
+void brw_screen_buffer_init(struct brw_screen *brw_screen)
+{
+   brw_screen->base.buffer_create = brw_buffer_create;
+   brw_screen->base.user_buffer_create = brw_user_buffer_create;
+   brw_screen->base.buffer_map = brw_buffer_map;
+   brw_screen->base.buffer_map_range = brw_buffer_map_range;
+   brw_screen->base.buffer_flush_mapped_range = brw_buffer_flush_mapped_range;
+   brw_screen->base.buffer_unmap = brw_buffer_unmap;
+   brw_screen->base.buffer_destroy = brw_buffer_destroy;
+}
diff --git a/src/gallium/drivers/i965/brw_screen_surface.c b/src/gallium/drivers/i965/brw_screen_surface.c
new file mode 100644 (file)
index 0000000..e2b9954
--- /dev/null
@@ -0,0 +1,262 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "util/u_memory.h"
+#include "util/u_simple_list.h"
+#include "util/u_math.h"
+
+#include "pipe/p_screen.h"
+#include "brw_screen.h"
+#include "brw_defines.h"
+#include "brw_winsys.h"
+
+enum {
+   BRW_VIEW_LINEAR,
+   BRW_VIEW_IN_PLACE
+};
+
+
+static boolean need_linear_view( struct brw_screen *brw_screen,
+                                struct brw_texture *brw_texture,
+                                union brw_surface_id id,
+                                unsigned usage )
+{
+#if 0
+   /* XXX: what about IDGNG?
+    */
+   if (!BRW_IS_G4X(brw->brw_screen->pci_id))
+   {
+      struct gl_renderbuffer *rb = ctx->DrawBuffer->_ColorDrawBuffers[i];
+      struct intel_renderbuffer *irb = intel_renderbuffer(rb);
+
+      /* The original gen4 hardware couldn't set up WM surfaces pointing
+       * at an offset within a tile, which can happen when rendering to
+       * anything but the base level of a texture or the +X face/0 depth.
+       * This was fixed with the 4 Series hardware.
+       *
+       * For these original chips, you would have to make the depth and
+       * color destination surfaces include information on the texture
+       * type, LOD, face, and various limits to use them as a destination.
+       *
+       * This is easy in Gallium as surfaces are all backed by
+       * textures, but there's also a nasty requirement that the depth
+       * and the color surfaces all be of the same LOD, which is
+       * harder to get around as we can't look at a surface in
+       * isolation and decide if it's legal.
+       *
+       * Instead, end up being pessimistic and say that for i965,
+       * ... ??
+       */
+      if (brw_tex->tiling != I915_TILING_NONE &&
+         (brw_tex_image_offset(brw_tex, face, level, zslize) & 4095)) {
+        if (BRW_DEBUG & DEBUG_VIEW)
+           debug_printf("%s: need surface view for non-aligned tex image\n",
+                        __FUNCTION__);
+        return GL_TRUE;
+      }
+   }
+#endif
+
+   /* Tiled 3d textures don't have subsets that look like 2d surfaces:
+    */
+   
+   /* Everything else should be fine to render to in-place:
+    */
+   return GL_FALSE;
+}
+
+/* Look at all texture views and figure out if any of them need to be
+ * back-copied into the texture for sampling
+ */
+void brw_update_texture( struct brw_screen *brw_screen,
+                        struct brw_texture *tex )
+{
+   /* currently nothing to do */
+}
+
+
+/* Create a new surface with linear layout to serve as a render-target
+ * where it would be illegal (perhaps due to tiling constraints) to do
+ * this in-place.
+ * 
+ * Currently not implmented, not sure if it's needed.
+ */
+static struct brw_surface *create_linear_view( struct brw_screen *brw_screen,
+                                              struct brw_texture *tex,
+                                              union brw_surface_id id,
+                                              unsigned usage )
+{
+   return NULL;
+}
+
+
+/* Create a pipe_surface that just points directly into the existing
+ * texture's storage.
+ */
+static struct brw_surface *create_in_place_view( struct brw_screen *brw_screen,
+                                                 struct brw_texture *tex,
+                                                 union brw_surface_id id,
+                                                 unsigned usage )
+{
+   struct brw_surface *surface;
+
+   surface = CALLOC_STRUCT(brw_surface);
+   if (surface == NULL)
+      return NULL;
+
+   pipe_reference_init(&surface->base.reference, 1);
+
+   /* XXX: ignoring render-to-slice-of-3d-texture
+    */
+   assert(id.bits.zslice == 0);
+
+   surface->base.format = tex->base.format;
+   surface->base.width = u_minify(tex->base.width0, id.bits.level);
+   surface->base.height = u_minify(tex->base.height0, id.bits.level);
+   surface->base.offset = tex->image_offset[id.bits.level][id.bits.face];
+   surface->base.usage = usage;
+   surface->base.zslice = id.bits.zslice;
+   surface->base.face = id.bits.face;
+   surface->base.level = id.bits.level;
+   surface->id = id;
+   surface->cpp = tex->cpp;
+   surface->pitch = tex->pitch;
+   surface->tiling = tex->tiling;
+
+   bo_reference( &surface->bo, tex->bo );
+   pipe_texture_reference( &surface->base.texture, &tex->base );
+
+   surface->ss.ss0.surface_format = tex->ss.ss0.surface_format;
+   surface->ss.ss0.surface_type = BRW_SURFACE_2D;
+
+   if (tex->tiling == BRW_TILING_NONE) {
+      surface->ss.ss1.base_addr = surface->base.offset;
+   } else {
+      uint32_t tile_offset = surface->base.offset % 4096;
+
+      surface->ss.ss1.base_addr = surface->base.offset - tile_offset;
+
+      if (brw_screen->chipset.is_g4x) {
+        if (tex->tiling == BRW_TILING_X) {
+           /* Note that the low bits of these fields are missing, so
+            * there's the possibility of getting in trouble.
+            */
+           surface->ss.ss5.x_offset = (tile_offset % 512) / tex->cpp / 4;
+           surface->ss.ss5.y_offset = tile_offset / 512 / 2;
+        } else {
+           surface->ss.ss5.x_offset = (tile_offset % 128) / tex->cpp / 4;
+           surface->ss.ss5.y_offset = tile_offset / 128 / 2;
+        }
+      }
+      else {
+        assert(tile_offset == 0);
+      }
+   }
+
+#if 0
+   if (region_bo != NULL)
+      surface->ss.ss1.base_addr += region_bo->offset; /* reloc */
+#endif
+
+   surface->ss.ss2.width = surface->base.width - 1;
+   surface->ss.ss2.height = surface->base.height - 1;
+   surface->ss.ss3.tiled_surface = tex->ss.ss3.tiled_surface;
+   surface->ss.ss3.tile_walk = tex->ss.ss3.tile_walk;
+   surface->ss.ss3.pitch = tex->ss.ss3.pitch;
+
+   return surface;
+}
+
+/* Get a surface which is view into a texture 
+ */
+static struct pipe_surface *brw_get_tex_surface(struct pipe_screen *screen,
+                                               struct pipe_texture *pt,
+                                               unsigned face, unsigned level,
+                                               unsigned zslice,
+                                               unsigned usage )
+{
+   struct brw_texture *tex = brw_texture(pt);
+   struct brw_screen *bscreen = brw_screen(screen);
+   struct brw_surface *surface;
+   union brw_surface_id id;
+   int type;
+
+   id.bits.face = face;
+   id.bits.level = level;
+   id.bits.zslice = zslice;
+
+   if (need_linear_view(bscreen, tex, id, usage)) 
+      type = BRW_VIEW_LINEAR;
+   else
+      type = BRW_VIEW_IN_PLACE;
+
+   
+   foreach (surface, &tex->views[type]) {
+      if (id.value == surface->id.value)
+        return &surface->base;
+   }
+
+   switch (type) {
+   case BRW_VIEW_LINEAR:
+      surface = create_linear_view( bscreen, tex, id, usage );
+      break;
+   case BRW_VIEW_IN_PLACE:
+      surface = create_in_place_view( bscreen, tex, id, usage );
+      break;
+   default:
+      return NULL;
+   }
+
+   insert_at_head( &tex->views[type], surface );
+   return &surface->base;
+}
+
+
+static void brw_tex_surface_destroy( struct pipe_surface *surf )
+{
+   struct brw_surface *surface = brw_surface(surf);
+
+   /* Unreference texture, shared buffer:
+    */
+   remove_from_list(surface);
+   bo_reference(&surface->bo, NULL);
+   pipe_texture_reference( &surface->base.texture, NULL );
+
+
+   FREE(surface);
+}
+
+
+void brw_screen_tex_surface_init( struct brw_screen *brw_screen )
+{
+   brw_screen->base.get_tex_surface = brw_get_tex_surface;
+   brw_screen->base.tex_surface_destroy = brw_tex_surface_destroy;
+}
diff --git a/src/gallium/drivers/i965/brw_screen_tex_layout.c b/src/gallium/drivers/i965/brw_screen_tex_layout.c
new file mode 100644 (file)
index 0000000..894f4be
--- /dev/null
@@ -0,0 +1,414 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+
+#include "pipe/p_format.h"
+
+#include "util/u_math.h"
+#include "util/u_memory.h"
+
+#include "brw_screen.h"
+#include "brw_debug.h"
+#include "brw_winsys.h"
+
+/* Code to layout images in a mipmap tree for i965.
+ */
+
+static int 
+brw_tex_pitch_align (struct brw_texture *tex,
+                    int pitch)
+{
+   if (!tex->compressed) {
+      int pitch_align;
+
+      switch (tex->tiling) {
+      case BRW_TILING_X:
+        pitch_align = 512;
+        break;
+      case BRW_TILING_Y:
+        pitch_align = 128;
+        break;
+      default:
+        /* XXX: Untiled pitch alignment of 64 bytes for now to allow
+         * render-to-texture to work in all cases. This should
+         * probably be replaced at some point by some scheme to only
+         * do this when really necessary, for example standalone
+         * render target views.
+         */
+        pitch_align = 64;
+        break;
+      }
+
+      pitch = align(pitch * tex->cpp, pitch_align);
+      pitch /= tex->cpp;
+   }
+
+   return pitch;
+}
+
+
+static void 
+brw_tex_alignment_unit(enum pipe_format pf, 
+                      GLuint *w, GLuint *h)
+{
+    switch (pf) {
+    case PIPE_FORMAT_DXT1_RGB:
+    case PIPE_FORMAT_DXT1_RGBA:
+    case PIPE_FORMAT_DXT3_RGBA:
+    case PIPE_FORMAT_DXT5_RGBA:
+    case PIPE_FORMAT_DXT1_SRGB:
+    case PIPE_FORMAT_DXT1_SRGBA:
+    case PIPE_FORMAT_DXT3_SRGBA:
+    case PIPE_FORMAT_DXT5_SRGBA:
+        *w = 4;
+        *h = 4;
+        break;
+
+    default:
+        *w = 4;
+        *h = 2;
+        break;
+    }
+}
+
+
+static void 
+brw_tex_set_level_info(struct brw_texture *tex,
+                      GLuint level,
+                      GLuint nr_images,
+                      GLuint x, GLuint y,
+                      GLuint w, GLuint h, GLuint d)
+{
+
+   if (BRW_DEBUG & DEBUG_TEXTURE)
+      debug_printf("%s level %d size: %d,%d,%d offset %d,%d (0x%x)\n", __FUNCTION__,
+                  level, w, h, d, x, y, tex->level_offset[level]);
+
+   assert(tex->image_offset[level] == NULL);
+   assert(nr_images >= 1);
+
+   tex->level_offset[level] = (x + y * tex->pitch) * tex->cpp;
+   tex->nr_images[level] = nr_images;
+
+   tex->image_offset[level] = MALLOC(nr_images * sizeof(GLuint));
+   tex->image_offset[level][0] = 0;
+}
+
+
+static void
+brw_tex_set_image_offset(struct brw_texture *tex,
+                        GLuint level, GLuint img,
+                        GLuint x, GLuint y, 
+                        GLuint offset)
+{
+   assert((x == 0 && y == 0) || img != 0 || level != 0);
+   assert(img < tex->nr_images[level]);
+
+   if (BRW_DEBUG & DEBUG_TEXTURE)
+      debug_printf("%s level %d img %d pos %d,%d image_offset %x\n",
+                  __FUNCTION__, level, img, x, y, 
+                  tex->image_offset[level][img]);
+
+   tex->image_offset[level][img] = (x + y * tex->pitch) * tex->cpp + offset;
+}
+
+
+
+static void brw_layout_2d( struct brw_texture *tex )
+{
+   GLuint align_h = 2, align_w = 4;
+   GLuint level;
+   GLuint x = 0;
+   GLuint y = 0;
+   GLuint width = tex->base.width0;
+   GLuint height = tex->base.height0;
+
+   tex->pitch = tex->base.width0;
+   brw_tex_alignment_unit(tex->base.format, &align_w, &align_h);
+
+   if (tex->compressed) {
+       tex->pitch = align(tex->base.width0, align_w);
+   }
+
+   /* May need to adjust pitch to accomodate the placement of
+    * the 2nd mipmap.  This occurs when the alignment
+    * constraints of mipmap placement push the right edge of the
+    * 2nd mipmap out past the width of its parent.
+    */
+   if (tex->base.last_level > 0) {
+       GLuint mip1_width;
+
+       if (tex->compressed) {
+          mip1_width = (align(u_minify(tex->base.width0, 1), align_w) + 
+                        align(u_minify(tex->base.width0, 2), align_w));
+       } else {
+          mip1_width = (align(u_minify(tex->base.width0, 1), align_w) + 
+                        u_minify(tex->base.width0, 2));
+       }
+
+       if (mip1_width > tex->pitch) {
+           tex->pitch = mip1_width;
+       }
+   }
+
+   /* Pitch must be a whole number of dwords, even though we
+    * express it in texels.
+    */
+   tex->pitch = brw_tex_pitch_align (tex, tex->pitch);
+   tex->total_height = 0;
+
+   for ( level = 0 ; level <= tex->base.last_level ; level++ ) {
+      GLuint img_height;
+
+      brw_tex_set_level_info(tex, level, 1, x, y, width, height, 1);
+
+      if (tex->compressed)
+        img_height = MAX2(1, height/4);
+      else
+        img_height = align(height, align_h);
+
+
+      /* Because the images are packed better, the final offset
+       * might not be the maximal one:
+       */
+      tex->total_height = MAX2(tex->total_height, y + img_height);
+
+      /* Layout_below: step right after second mipmap.
+       */
+      if (level == 1) {
+        x += align(width, align_w);
+      }
+      else {
+        y += img_height;
+      }
+
+      width  = u_minify(width, 1);
+      height = u_minify(height, 1);
+   }
+}
+
+
+static boolean 
+brw_layout_cubemap_idgng( struct brw_texture *tex )
+{
+   GLuint align_h = 2, align_w = 4;
+   GLuint level;
+   GLuint x = 0;
+   GLuint y = 0;
+   GLuint width = tex->base.width0;
+   GLuint height = tex->base.height0;
+   GLuint qpitch = 0;
+   GLuint y_pitch = 0;
+
+   tex->pitch = tex->base.width0;
+   brw_tex_alignment_unit(tex->base.format, &align_w, &align_h);
+   y_pitch = align(height, align_h);
+
+   if (tex->compressed) {
+      tex->pitch = align(tex->base.width0, align_w);
+   }
+
+   if (tex->base.last_level != 0) {
+      GLuint mip1_width;
+
+      if (tex->compressed) {
+        mip1_width = (align(u_minify(tex->base.width0, 1), align_w) +
+                      align(u_minify(tex->base.width0, 2), align_w));
+      } else {
+        mip1_width = (align(u_minify(tex->base.width0, 1), align_w) +
+                      u_minify(tex->base.width0, 2));
+      }
+
+      if (mip1_width > tex->pitch) {
+        tex->pitch = mip1_width;
+      }
+   }
+
+   tex->pitch = brw_tex_pitch_align(tex, tex->pitch);
+
+   if (tex->compressed) {
+      qpitch = ((y_pitch + 
+                align(u_minify(y_pitch, 1), align_h) +
+                11 * align_h) / 4) * tex->pitch * tex->cpp;
+
+      tex->total_height = ((y_pitch + 
+                           align(u_minify(y_pitch, 1), align_h) + 
+                           11 * align_h) / 4) * 6;
+   } else {
+      qpitch = (y_pitch + 
+               align(u_minify(y_pitch, 1), align_h) + 
+               11 * align_h) * tex->pitch * tex->cpp;
+
+      tex->total_height = (y_pitch +
+                          align(u_minify(y_pitch, 1), align_h) +
+                          11 * align_h) * 6;
+   }
+
+   for (level = 0; level <= tex->base.last_level; level++) {
+      GLuint img_height;
+      GLuint nr_images = 6;
+      GLuint q = 0;
+
+      brw_tex_set_level_info(tex, level, nr_images, x, y, width, height, 1);
+
+      for (q = 0; q < nr_images; q++)
+        brw_tex_set_image_offset(tex, level, q, x, y, q * qpitch);
+
+      if (tex->compressed)
+        img_height = MAX2(1, height/4);
+      else
+        img_height = align(height, align_h);
+
+      if (level == 1) {
+        x += align(width, align_w);
+      }
+      else {
+        y += img_height;
+      }
+
+      width  = u_minify(width, 1);
+      height = u_minify(height, 1);
+   }
+
+   return TRUE;
+}
+
+
+static boolean
+brw_layout_3d_cube( struct brw_texture *tex )
+{
+   GLuint width  = tex->base.width0;
+   GLuint height = tex->base.height0;
+   GLuint depth = tex->base.depth0;
+   GLuint pack_x_pitch, pack_x_nr;
+   GLuint pack_y_pitch;
+   GLuint level;
+   GLuint align_h = 2;
+   GLuint align_w = 4;
+
+   tex->total_height = 0;
+   brw_tex_alignment_unit(tex->base.format, &align_w, &align_h);
+
+   if (tex->compressed) {
+      tex->pitch = align(width, align_w);
+      pack_y_pitch = (height + 3) / 4;
+   } else {
+      tex->pitch = brw_tex_pitch_align(tex, tex->base.width0);
+      pack_y_pitch = align(tex->base.height0, align_h);
+   }
+
+   pack_x_pitch = width;
+   pack_x_nr = 1;
+
+   for (level = 0 ; level <= tex->base.last_level ; level++) {
+      GLuint nr_images = tex->base.target == PIPE_TEXTURE_3D ? depth : 6;
+      GLint x = 0;
+      GLint y = 0;
+      GLint q, j;
+
+      brw_tex_set_level_info(tex, level, nr_images,
+                                  0, tex->total_height,
+                                  width, height, depth);
+
+      for (q = 0; q < nr_images;) {
+        for (j = 0; j < pack_x_nr && q < nr_images; j++, q++) {
+           brw_tex_set_image_offset(tex, level, q, x, y, 0);
+           x += pack_x_pitch;
+        }
+
+        x = 0;
+        y += pack_y_pitch;
+      }
+
+
+      tex->total_height += y;
+      width  = u_minify(width, 1);
+      height = u_minify(height, 1);
+      depth  = u_minify(depth, 1);
+
+      if (tex->compressed) {
+        pack_y_pitch = (height + 3) / 4;
+
+        if (pack_x_pitch > align(width, align_w)) {
+           pack_x_pitch = align(width, align_w);
+           pack_x_nr <<= 1;
+        }
+      } else {
+        if (pack_x_pitch > 4) {
+           pack_x_pitch >>= 1;
+           pack_x_nr <<= 1;
+           assert(pack_x_pitch * pack_x_nr <= tex->pitch);
+        }
+
+        if (pack_y_pitch > 2) {
+           pack_y_pitch >>= 1;
+           pack_y_pitch = align(pack_y_pitch, align_h);
+        }
+      }
+   }
+
+   /* The 965's sampler lays cachelines out according to how accesses
+    * in the texture surfaces run, so they may be "vertical" through
+    * memory.  As a result, the docs say in Surface Padding Requirements:
+    * Sampling Engine Surfaces that two extra rows of padding are required.
+    */
+   if (tex->base.target == PIPE_TEXTURE_CUBE)
+      tex->total_height += 2;
+
+   return TRUE;
+}
+
+
+
+GLboolean brw_texture_layout(struct brw_screen *brw_screen,
+                            struct brw_texture *tex )
+{
+   switch (tex->base.target) {
+   case PIPE_TEXTURE_CUBE:
+      if (brw_screen->chipset.is_igdng)
+        brw_layout_cubemap_idgng( tex );
+      else
+        brw_layout_3d_cube( tex );
+      break;
+           
+   case PIPE_TEXTURE_3D:
+      brw_layout_3d_cube( tex );
+      break;
+
+   default:
+      brw_layout_2d( tex );
+      break;
+   }
+
+   if (BRW_DEBUG & DEBUG_TEXTURE)
+      debug_printf("%s: %dx%dx%d - sz 0x%x\n", __FUNCTION__,
+                  tex->pitch,
+                  tex->total_height,
+                  tex->cpp,
+                  tex->pitch * tex->total_height * tex->cpp );
+
+   return GL_TRUE;
+}
diff --git a/src/gallium/drivers/i965/brw_screen_texture.c b/src/gallium/drivers/i965/brw_screen_texture.c
new file mode 100644 (file)
index 0000000..feb9d5f
--- /dev/null
@@ -0,0 +1,573 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "util/u_memory.h"
+#include "util/u_simple_list.h"
+#include "util/u_format.h"
+
+#include "brw_screen.h"
+#include "brw_defines.h"
+#include "brw_structs.h"
+#include "brw_winsys.h"
+
+
+
+static GLuint translate_tex_target( unsigned target )
+{
+   switch (target) {
+   case PIPE_TEXTURE_1D: 
+      return BRW_SURFACE_1D;
+
+   case PIPE_TEXTURE_2D: 
+      return BRW_SURFACE_2D;
+
+   case PIPE_TEXTURE_3D: 
+      return BRW_SURFACE_3D;
+
+   case PIPE_TEXTURE_CUBE:
+      return BRW_SURFACE_CUBE;
+
+   default: 
+      assert(0); 
+      return BRW_SURFACE_1D;
+   }
+}
+
+
+static GLuint translate_tex_format( enum pipe_format pf )
+{
+   switch( pf ) {
+   case PIPE_FORMAT_L8_UNORM:
+      return BRW_SURFACEFORMAT_L8_UNORM;
+
+   case PIPE_FORMAT_I8_UNORM:
+      return BRW_SURFACEFORMAT_I8_UNORM;
+
+   case PIPE_FORMAT_A8_UNORM:
+      return BRW_SURFACEFORMAT_A8_UNORM; 
+
+   case PIPE_FORMAT_L16_UNORM:
+      return BRW_SURFACEFORMAT_L16_UNORM;
+
+      /* XXX: Add these to gallium
+   case PIPE_FORMAT_I16_UNORM:
+      return BRW_SURFACEFORMAT_I16_UNORM;
+
+   case PIPE_FORMAT_A16_UNORM:
+      return BRW_SURFACEFORMAT_A16_UNORM; 
+      */
+
+   case PIPE_FORMAT_A8L8_UNORM:
+      return BRW_SURFACEFORMAT_L8A8_UNORM;
+
+   case PIPE_FORMAT_R5G6B5_UNORM:
+      return BRW_SURFACEFORMAT_B5G6R5_UNORM;
+
+   case PIPE_FORMAT_A1R5G5B5_UNORM:
+      return BRW_SURFACEFORMAT_B5G5R5A1_UNORM;
+
+   case PIPE_FORMAT_A4R4G4B4_UNORM:
+      return BRW_SURFACEFORMAT_B4G4R4A4_UNORM;
+
+   case PIPE_FORMAT_X8R8G8B8_UNORM:
+      return BRW_SURFACEFORMAT_R8G8B8X8_UNORM;
+
+   case PIPE_FORMAT_A8R8G8B8_UNORM:
+      return BRW_SURFACEFORMAT_B8G8R8A8_UNORM;
+
+   /*
+    * Video formats
+    */
+
+   case PIPE_FORMAT_YCBCR_REV:
+      return BRW_SURFACEFORMAT_YCRCB_NORMAL;
+
+   case PIPE_FORMAT_YCBCR:
+      return BRW_SURFACEFORMAT_YCRCB_SWAPUVY;
+
+   /*
+    * Compressed formats.
+    */
+      /* XXX: Add FXT to gallium?
+   case PIPE_FORMAT_FXT1_RGBA:
+      return BRW_SURFACEFORMAT_FXT1;
+      */
+
+   case PIPE_FORMAT_DXT1_RGB:
+       return BRW_SURFACEFORMAT_DXT1_RGB;
+
+   case PIPE_FORMAT_DXT1_RGBA:
+       return BRW_SURFACEFORMAT_BC1_UNORM;
+       
+   case PIPE_FORMAT_DXT3_RGBA:
+       return BRW_SURFACEFORMAT_BC2_UNORM;
+       
+   case PIPE_FORMAT_DXT5_RGBA:
+       return BRW_SURFACEFORMAT_BC3_UNORM;
+
+   /*
+    * sRGB formats
+    */
+
+   case PIPE_FORMAT_R8G8B8A8_SRGB:
+      return BRW_SURFACEFORMAT_B8G8R8A8_UNORM_SRGB;
+
+   case PIPE_FORMAT_A8L8_SRGB:
+      return BRW_SURFACEFORMAT_L8A8_UNORM_SRGB;
+
+   case PIPE_FORMAT_L8_SRGB:
+      return BRW_SURFACEFORMAT_L8_UNORM_SRGB;
+
+   case PIPE_FORMAT_DXT1_SRGB:
+      return BRW_SURFACEFORMAT_BC1_UNORM_SRGB;
+
+   /*
+    * Depth formats
+    */
+
+   case PIPE_FORMAT_Z16_UNORM:
+         return BRW_SURFACEFORMAT_I16_UNORM;
+
+   case PIPE_FORMAT_S8Z24_UNORM:
+   case PIPE_FORMAT_X8Z24_UNORM:
+         return BRW_SURFACEFORMAT_I24X8_UNORM;
+
+   case PIPE_FORMAT_Z32_FLOAT:
+         return BRW_SURFACEFORMAT_I32_FLOAT;
+
+      /* XXX: presumably for bump mapping.  Add this to mesa state
+       * tracker?
+       *
+       * XXX: Add flipped versions of these formats to Gallium.
+       */
+   case PIPE_FORMAT_R8G8_SNORM:
+      return BRW_SURFACEFORMAT_R8G8_SNORM;
+
+   case PIPE_FORMAT_R8G8B8A8_SNORM:
+      return BRW_SURFACEFORMAT_R8G8B8A8_SNORM;
+
+   default:
+      return BRW_SURFACEFORMAT_INVALID;
+   }
+}
+
+
+
+
+
+static struct pipe_texture *brw_texture_create( struct pipe_screen *screen,
+                                               const struct pipe_texture *templ )
+
+{  
+   struct brw_screen *bscreen = brw_screen(screen);
+   struct brw_texture *tex;
+   enum brw_buffer_type buffer_type;
+   enum pipe_error ret;
+   
+   tex = CALLOC_STRUCT(brw_texture);
+   if (tex == NULL)
+      return NULL;
+
+   memcpy(&tex->base, templ, sizeof *templ);
+   pipe_reference_init(&tex->base.reference, 1);
+   tex->base.screen = screen;
+
+   /* XXX: compressed textures need special treatment here
+    */
+   tex->cpp = util_format_get_blocksize(tex->base.format);
+   tex->compressed = util_format_is_compressed(tex->base.format);
+
+   make_empty_list(&tex->views[0]);
+   make_empty_list(&tex->views[1]);
+
+   /* XXX: No tiling with compressed textures??
+    */
+   if (tex->compressed == 0 &&
+       !bscreen->no_tiling) 
+   {
+      if (bscreen->chipset.is_965 &&
+         util_format_is_depth_or_stencil(templ->format))
+        tex->tiling = BRW_TILING_Y;
+      else
+        tex->tiling = BRW_TILING_X;
+   } 
+   else {
+      tex->tiling = BRW_TILING_NONE;
+   }
+
+
+
+
+   if (!brw_texture_layout( bscreen, tex ))
+      goto fail;
+
+   
+   if (templ->tex_usage & (PIPE_TEXTURE_USAGE_DISPLAY_TARGET |
+                           PIPE_TEXTURE_USAGE_PRIMARY)) {
+      buffer_type = BRW_BUFFER_TYPE_SCANOUT;
+   }
+   else {
+      buffer_type = BRW_BUFFER_TYPE_TEXTURE;
+   }
+
+   ret = bscreen->sws->bo_alloc( bscreen->sws,
+                                 buffer_type,
+                                 tex->pitch * tex->total_height * tex->cpp,
+                                 64,
+                                 &tex->bo );
+   if (ret)
+      goto fail;
+
+   tex->ss.ss0.mipmap_layout_mode = BRW_SURFACE_MIPMAPLAYOUT_BELOW;
+   tex->ss.ss0.surface_type = translate_tex_target(tex->base.target);
+   tex->ss.ss0.surface_format = translate_tex_format(tex->base.format);
+   assert(tex->ss.ss0.surface_format != BRW_SURFACEFORMAT_INVALID);
+
+   /* This is ok for all textures with channel width 8bit or less:
+    */
+/*    tex->ss.ss0.data_return_format = BRW_SURFACERETURNFORMAT_S1; */
+
+
+   /* XXX: what happens when tex->bo->offset changes???
+    */
+   tex->ss.ss1.base_addr = 0; /* reloc */
+   tex->ss.ss2.mip_count = tex->base.last_level;
+   tex->ss.ss2.width = tex->base.width0 - 1;
+   tex->ss.ss2.height = tex->base.height0 - 1;
+
+   switch (tex->tiling) {
+   case BRW_TILING_NONE:
+      tex->ss.ss3.tiled_surface = 0;
+      tex->ss.ss3.tile_walk = 0;
+      break;
+   case BRW_TILING_X:
+      tex->ss.ss3.tiled_surface = 1;
+      tex->ss.ss3.tile_walk = BRW_TILEWALK_XMAJOR;
+      break;
+   case BRW_TILING_Y:
+      tex->ss.ss3.tiled_surface = 1;
+      tex->ss.ss3.tile_walk = BRW_TILEWALK_YMAJOR;
+      break;
+   }
+
+   tex->ss.ss3.pitch = (tex->pitch * tex->cpp) - 1;
+   tex->ss.ss3.depth = tex->base.depth0 - 1;
+
+   tex->ss.ss4.min_lod = 0;
+   if (tex->base.target == PIPE_TEXTURE_CUBE) {
+      tex->ss.ss0.cube_pos_x = 1;
+      tex->ss.ss0.cube_pos_y = 1;
+      tex->ss.ss0.cube_pos_z = 1;
+      tex->ss.ss0.cube_neg_x = 1;
+      tex->ss.ss0.cube_neg_y = 1;
+      tex->ss.ss0.cube_neg_z = 1;
+   }
+
+   return &tex->base;
+
+fail:
+   bo_reference(&tex->bo, NULL);
+   FREE(tex);
+   return NULL;
+}
+
+static struct pipe_texture *brw_texture_blanket(struct pipe_screen *screen,
+                                               const struct pipe_texture *templ,
+                                               const unsigned *stride,
+                                               struct pipe_buffer *buffer)
+{
+   return NULL;
+}
+
+static void brw_texture_destroy(struct pipe_texture *pt)
+{
+   struct brw_texture *tex = brw_texture(pt);
+   bo_reference(&tex->bo, NULL);
+   FREE(pt);
+}
+
+
+static boolean brw_is_format_supported( struct pipe_screen *screen,
+                                       enum pipe_format format,
+                                       enum pipe_texture_target target,
+                                       unsigned tex_usage, 
+                                       unsigned geom_flags )
+{
+   return translate_tex_format(format) != BRW_SURFACEFORMAT_INVALID;
+}
+
+
+boolean brw_is_texture_referenced_by_bo( struct brw_screen *brw_screen,
+                                      struct pipe_texture *texture,
+                                      unsigned face, 
+                                      unsigned level,
+                                      struct brw_winsys_buffer *bo )
+{
+   struct brw_texture *tex = brw_texture(texture);
+   struct brw_surface *surf;
+   int i;
+
+   /* XXX: this is subject to false positives if the underlying
+    * texture BO is referenced, we can't tell whether the sub-region
+    * we care about participates in that.
+    */
+   if (brw_screen->sws->bo_references( bo, tex->bo ))
+      return TRUE;
+
+   /* Find any view on this texture for this face/level and see if it
+    * is referenced:
+    */
+   for (i = 0; i < 2; i++) {
+      foreach (surf, &tex->views[i]) {
+         if (surf->bo == tex->bo)
+            continue;
+
+         if (surf->id.bits.face != face ||
+             surf->id.bits.level != level)
+            continue;
+         
+         if (brw_screen->sws->bo_references( bo, surf->bo))
+            return TRUE;
+      }
+   }
+
+   return FALSE;
+}
+
+
+/*
+ * Transfer functions
+ */
+
+static struct pipe_transfer*
+brw_get_tex_transfer(struct pipe_screen *screen,
+                     struct pipe_texture *texture,
+                     unsigned face, unsigned level, unsigned zslice,
+                     enum pipe_transfer_usage usage, unsigned x, unsigned y,
+                     unsigned w, unsigned h)
+{
+   struct brw_texture *tex = brw_texture(texture);
+   struct brw_transfer *trans;
+   unsigned offset;  /* in bytes */
+
+   if (texture->target == PIPE_TEXTURE_CUBE) {
+      offset = tex->image_offset[level][face];
+   } else if (texture->target == PIPE_TEXTURE_3D) {
+      offset = tex->image_offset[level][zslice];
+   } else {
+      offset = tex->image_offset[level][0];
+      assert(face == 0);
+      assert(zslice == 0);
+   }
+
+   trans = CALLOC_STRUCT(brw_transfer);
+   if (trans) {
+      pipe_texture_reference(&trans->base.texture, texture);
+      trans->base.x = x;
+      trans->base.y = y;
+      trans->base.width = w;
+      trans->base.height = h;
+      trans->base.stride = tex->pitch * tex->cpp;
+      trans->offset = offset;
+      trans->base.usage = usage;
+   }
+   return &trans->base;
+}
+
+static void *
+brw_transfer_map(struct pipe_screen *screen,
+                 struct pipe_transfer *transfer)
+{
+   struct brw_texture *tex = brw_texture(transfer->texture);
+   struct brw_winsys_screen *sws = brw_screen(screen)->sws;
+   char *map;
+   unsigned usage = transfer->usage;
+
+   map = sws->bo_map(tex->bo, 
+                     BRW_DATA_OTHER,
+                     0,
+                     tex->bo->size,
+                     (usage & PIPE_TRANSFER_WRITE) ? TRUE : FALSE,
+                     (usage & 0) ? TRUE : FALSE,
+                     (usage & 0) ? TRUE : FALSE);
+
+   if (!map)
+      return NULL;
+
+   /* XXX: blocksize and compressed textures
+    */
+   return map + brw_transfer(transfer)->offset +
+      transfer->y /* / transfer->block.height */ * transfer->stride +
+      transfer->x /* / transfer->block.width */ * brw_texture(transfer->texture)->cpp;
+}
+
+static void
+brw_transfer_unmap(struct pipe_screen *screen,
+                   struct pipe_transfer *transfer)
+{
+   struct brw_texture *tex = brw_texture(transfer->texture);
+   struct brw_winsys_screen *sws = brw_screen(screen)->sws;
+
+   sws->bo_unmap(tex->bo);
+}
+
+static void
+brw_tex_transfer_destroy(struct pipe_transfer *trans)
+{
+   pipe_texture_reference(&trans->texture, NULL);
+   FREE(trans);
+}
+
+
+/*
+ * Functions exported to the winsys
+ */
+
+boolean brw_texture_get_winsys_buffer(struct pipe_texture *texture,
+                                      struct brw_winsys_buffer **buffer,
+                                      unsigned *stride)
+{
+   struct brw_texture *tex = brw_texture(texture);
+
+   *buffer = tex->bo;
+   if (stride)
+      *stride = tex->pitch * tex->cpp;
+
+   return TRUE;
+}
+
+struct pipe_texture * 
+brw_texture_blanket_winsys_buffer(struct pipe_screen *screen,
+                                  const struct pipe_texture *templ,
+                                  unsigned pitch,
+                                 unsigned tiling,
+                                  struct brw_winsys_buffer *buffer)
+{
+   struct brw_screen *bscreen = brw_screen(screen);
+   struct brw_texture *tex;
+
+   if (templ->target != PIPE_TEXTURE_2D ||
+       templ->last_level != 0 ||
+       templ->depth0 != 1)
+      return NULL;
+
+   if (util_format_is_compressed(templ->format))
+      return NULL;
+
+   tex = CALLOC_STRUCT(brw_texture);
+   if (!tex)
+      return NULL;
+
+   memcpy(&tex->base, templ, sizeof *templ);
+   pipe_reference_init(&tex->base.reference, 1);
+   tex->base.screen = screen;
+
+   /* XXX: cpp vs. blocksize
+    */
+   tex->cpp = util_format_get_blocksize(tex->base.format);
+   tex->tiling = tiling;
+
+   make_empty_list(&tex->views[0]);
+   make_empty_list(&tex->views[1]);
+
+   if (!brw_texture_layout(bscreen, tex))
+      goto fail;
+
+   /* XXX Maybe some more checks? */
+   if ((pitch / tex->cpp) < tex->pitch)
+      goto fail;
+
+   tex->pitch = pitch / tex->cpp;
+
+   tex->bo = buffer;
+
+   /* fix this warning */
+#if 0
+   if (tex->size > buffer->size)
+      goto fail;
+#endif
+
+   tex->ss.ss0.mipmap_layout_mode = BRW_SURFACE_MIPMAPLAYOUT_BELOW;
+   tex->ss.ss0.surface_type = translate_tex_target(tex->base.target);
+   tex->ss.ss0.surface_format = translate_tex_format(tex->base.format);
+   assert(tex->ss.ss0.surface_format != BRW_SURFACEFORMAT_INVALID);
+
+   /* This is ok for all textures with channel width 8bit or less:
+    */
+/*    tex->ss.ss0.data_return_format = BRW_SURFACERETURNFORMAT_S1; */
+
+
+   /* XXX: what happens when tex->bo->offset changes???
+    */
+   tex->ss.ss1.base_addr = 0; /* reloc */
+   tex->ss.ss2.mip_count = tex->base.last_level;
+   tex->ss.ss2.width = tex->base.width0 - 1;
+   tex->ss.ss2.height = tex->base.height0 - 1;
+
+   switch (tex->tiling) {
+   case BRW_TILING_NONE:
+      tex->ss.ss3.tiled_surface = 0;
+      tex->ss.ss3.tile_walk = 0;
+      break;
+   case BRW_TILING_X:
+      tex->ss.ss3.tiled_surface = 1;
+      tex->ss.ss3.tile_walk = BRW_TILEWALK_XMAJOR;
+      break;
+   case BRW_TILING_Y:
+      tex->ss.ss3.tiled_surface = 1;
+      tex->ss.ss3.tile_walk = BRW_TILEWALK_YMAJOR;
+      break;
+   }
+
+   tex->ss.ss3.pitch = (tex->pitch * tex->cpp) - 1;
+   tex->ss.ss3.depth = tex->base.depth0 - 1;
+
+   tex->ss.ss4.min_lod = 0;
+
+   return &tex->base;
+
+fail:
+   FREE(tex);
+   return NULL;
+}
+
+void brw_screen_tex_init( struct brw_screen *brw_screen )
+{
+   brw_screen->base.is_format_supported = brw_is_format_supported;
+   brw_screen->base.texture_create = brw_texture_create;
+   brw_screen->base.texture_destroy = brw_texture_destroy;
+   brw_screen->base.texture_blanket = brw_texture_blanket;
+   brw_screen->base.get_tex_transfer = brw_get_tex_transfer;
+   brw_screen->base.transfer_map = brw_transfer_map;
+   brw_screen->base.transfer_unmap = brw_transfer_unmap;
+   brw_screen->base.tex_transfer_destroy = brw_tex_transfer_destroy;
+}
diff --git a/src/gallium/drivers/i965/brw_sf.c b/src/gallium/drivers/i965/brw_sf.c
new file mode 100644 (file)
index 0000000..e1986a9
--- /dev/null
@@ -0,0 +1,216 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+  
+#include "pipe/p_state.h"
+
+#include "brw_batchbuffer.h"
+#include "brw_defines.h"
+#include "brw_context.h"
+#include "brw_pipe_rast.h"
+#include "brw_eu.h"
+#include "brw_util.h"
+#include "brw_sf.h"
+#include "brw_state.h"
+
+static enum pipe_error compile_sf_prog( struct brw_context *brw,
+                                        struct brw_sf_prog_key *key,
+                                        struct brw_winsys_buffer **bo_out )
+{
+   enum pipe_error ret;
+   struct brw_sf_compile c;
+   const GLuint *program;
+   GLuint program_size;
+
+   memset(&c, 0, sizeof(c));
+
+   /* Begin the compilation:
+    */
+   brw_init_compile(brw, &c.func);
+
+   c.key = *key;
+   c.nr_attrs = c.key.nr_attrs;
+   c.nr_attr_regs = (c.nr_attrs+1)/2;
+   c.nr_setup_attrs = c.key.nr_attrs;
+   c.nr_setup_regs = (c.nr_setup_attrs+1)/2;
+
+   c.prog_data.urb_read_length = c.nr_attr_regs;
+   c.prog_data.urb_entry_size = c.nr_setup_regs * 2;
+
+   /* Special case when there are no attributes to setup.
+    *
+    * XXX: should be able to set nr_setup_attrs to nr_attrs-1 -- but
+    * breaks vp-tris.c
+    */
+   if (c.nr_attrs - 1 == 0) {
+      c.nr_verts = 0;
+      brw_emit_null_setup( &c );
+   }
+   else {
+      /* Which primitive?  Or all three? 
+       */
+      switch (key->primitive) {
+      case SF_TRIANGLES:
+         c.nr_verts = 3;
+         brw_emit_tri_setup( &c, GL_TRUE );
+         break;
+      case SF_LINES:
+         c.nr_verts = 2;
+         brw_emit_line_setup( &c, GL_TRUE );
+         break;
+      case SF_POINTS:
+         c.nr_verts = 1;
+         if (key->do_point_sprite)
+            brw_emit_point_sprite_setup( &c, GL_TRUE );
+         else
+            brw_emit_point_setup( &c, GL_TRUE );
+         break;
+      case SF_UNFILLED_TRIS:
+         c.nr_verts = 3;
+         brw_emit_anyprim_setup( &c );
+         break;
+      default:
+         assert(0);
+         return PIPE_ERROR_BAD_INPUT;
+      }
+   }
+
+   /* get the program
+    */
+   ret = brw_get_program(&c.func, &program, &program_size);
+   if (ret)
+      return ret;
+
+   /* Upload
+    */
+   ret = brw_upload_cache( &brw->cache, BRW_SF_PROG,
+                           &c.key, sizeof(c.key),
+                           NULL, 0,
+                           program, program_size,
+                           &c.prog_data,
+                           &brw->sf.prog_data,
+                           bo_out);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+/* Calculate interpolants for triangle and line rasterization.
+ */
+static enum pipe_error upload_sf_prog(struct brw_context *brw)
+{
+   const struct brw_fs_signature *sig = &brw->curr.fragment_shader->signature;
+   struct brw_sf_prog_key key;
+   enum pipe_error ret;
+   unsigned i;
+
+   memset(&key, 0, sizeof(key));
+
+   /* Populate the key, noting state dependencies:
+    */
+
+   /* XXX: Add one to account for the position input.
+    */
+   /* PIPE_NEW_FRAGMENT_SIGNATURE */
+   key.nr_attrs = sig->nr_inputs + 1;
+
+
+   /* XXX: why is position required to be linear?  why do we care
+    * about it at all?
+    */
+   key.linear_attrs = 1;        /* position -- but why? */
+
+   for (i = 0; i < sig->nr_inputs; i++) {
+      switch (sig->input[i].interp) {
+      case TGSI_INTERPOLATE_CONSTANT:
+         break;
+      case TGSI_INTERPOLATE_LINEAR:
+         key.linear_attrs |= 1 << (i+1);
+         break;
+      case TGSI_INTERPOLATE_PERSPECTIVE:
+         key.persp_attrs |= 1 << (i+1);
+         break;
+      }
+   }
+
+   /* BRW_NEW_REDUCED_PRIMITIVE */
+   switch (brw->reduced_primitive) {
+   case PIPE_PRIM_TRIANGLES: 
+      /* PIPE_NEW_RAST
+       */
+      if (brw->curr.rast->templ.fill_cw != PIPE_POLYGON_MODE_FILL ||
+         brw->curr.rast->templ.fill_ccw != PIPE_POLYGON_MODE_FILL)
+        key.primitive = SF_UNFILLED_TRIS;
+      else
+        key.primitive = SF_TRIANGLES;
+      break;
+   case PIPE_PRIM_LINES: 
+      key.primitive = SF_LINES; 
+      break;
+   case PIPE_PRIM_POINTS: 
+      key.primitive = SF_POINTS; 
+      break;
+   }
+
+   key.do_point_sprite = brw->curr.rast->templ.point_sprite;
+   key.sprite_origin_lower_left = 0; /* XXX: ctx->Point.SpriteOrigin - fix rast state */
+   key.do_flat_shading = brw->curr.rast->templ.flatshade;
+   key.do_twoside_color = brw->curr.rast->templ.light_twoside;
+
+   if (key.do_twoside_color) {
+      key.frontface_ccw = (brw->curr.rast->templ.front_winding == 
+                          PIPE_WINDING_CCW);
+   }
+
+   if (brw_search_cache(&brw->cache, BRW_SF_PROG,
+                        &key, sizeof(key),
+                        NULL, 0,
+                        &brw->sf.prog_data,
+                        &brw->sf.prog_bo))
+      return PIPE_OK;
+
+   ret = compile_sf_prog( brw, &key, &brw->sf.prog_bo );
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+const struct brw_tracked_state brw_sf_prog = {
+   .dirty = {
+      .mesa  = (PIPE_NEW_RAST | PIPE_NEW_FRAGMENT_SIGNATURE),
+      .brw   = (BRW_NEW_REDUCED_PRIMITIVE),
+      .cache = 0
+   },
+   .prepare = upload_sf_prog
+};
+
diff --git a/src/gallium/drivers/i965/brw_sf.h b/src/gallium/drivers/i965/brw_sf.h
new file mode 100644 (file)
index 0000000..a895c7d
--- /dev/null
@@ -0,0 +1,122 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+   
+
+#ifndef BRW_SF_H
+#define BRW_SF_H
+
+
+#include "brw_context.h"
+#include "brw_eu.h"
+
+
+#define SF_POINTS    0
+#define SF_LINES     1
+#define SF_TRIANGLES 2
+#define SF_UNFILLED_TRIS   3
+
+struct brw_sf_prog_key {
+
+   /* Bitmask of linear and perspective interpolated inputs, 0..nr
+    */
+   GLuint persp_attrs:32;
+   GLuint linear_attrs:32;
+   GLuint point_coord_replace_attrs:32;
+
+   GLuint nr_attrs:8;
+   GLuint primitive:2;
+   GLuint do_twoside_color:1;
+   GLuint do_flat_shading:1;
+   GLuint frontface_ccw:1;
+   GLuint do_point_sprite:1;
+   GLuint sprite_origin_lower_left:1;
+   GLuint pad:17;
+
+   GLuint attr_col0:8;
+   GLuint attr_col1:8;
+   GLuint attr_bfc0:8;
+   GLuint attr_bfc1:8;
+};
+
+struct brw_sf_point_tex {
+   GLboolean CoordReplace;     
+};
+
+struct brw_sf_compile {
+   struct brw_compile func;
+   struct brw_sf_prog_key key;
+   struct brw_sf_prog_data prog_data;
+   
+   struct brw_reg pv;
+   struct brw_reg det;
+   struct brw_reg dx0;
+   struct brw_reg dx2;
+   struct brw_reg dy0;
+   struct brw_reg dy2;
+
+   /* z and 1/w passed in seperately:
+    */
+   struct brw_reg z[3];
+   struct brw_reg inv_w[3];
+   
+   /* The vertices:
+    */
+   struct brw_reg vert[3];
+
+    /* Temporaries, allocated after last vertex reg.
+    */
+   struct brw_reg inv_det;
+   struct brw_reg a1_sub_a0;
+   struct brw_reg a2_sub_a0;
+   struct brw_reg tmp;
+
+   struct brw_reg m1Cx;
+   struct brw_reg m2Cy;
+   struct brw_reg m3C0;
+
+   GLuint nr_verts;
+   GLuint nr_attrs;
+   GLuint nr_attr_regs;
+   GLuint nr_setup_attrs;
+   GLuint nr_setup_regs;
+
+   GLuint point_coord_replace_mask;
+};
+
+void brw_emit_null_setup( struct brw_sf_compile *c );
+void brw_emit_tri_setup( struct brw_sf_compile *c, GLboolean allocate );
+void brw_emit_line_setup( struct brw_sf_compile *c, GLboolean allocate );
+void brw_emit_point_setup( struct brw_sf_compile *c, GLboolean allocate );
+void brw_emit_point_sprite_setup( struct brw_sf_compile *c, GLboolean allocate );
+void brw_emit_anyprim_setup( struct brw_sf_compile *c );
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_sf_emit.c b/src/gallium/drivers/i965/brw_sf_emit.c
new file mode 100644 (file)
index 0000000..3b85725
--- /dev/null
@@ -0,0 +1,765 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+   
+
+#include "brw_batchbuffer.h"
+
+#include "brw_defines.h"
+#include "brw_context.h"
+#include "brw_eu.h"
+#include "brw_util.h"
+#include "brw_sf.h"
+
+
+static struct brw_reg get_vert_attr(struct brw_sf_compile *c,
+                                   struct brw_reg vert,
+                                   GLuint attr)
+{
+   GLuint off = attr / 2;
+   GLuint sub = attr % 2;
+
+   return brw_vec4_grf(vert.nr + off, sub * 4);
+}
+
+
+/*********************************************************************** 
+ * Twoside lighting
+ */
+static void copy_bfc( struct brw_sf_compile *c,
+                     struct brw_reg vert )
+{
+   struct brw_compile *p = &c->func;
+
+   if (c->key.attr_col0 && c->key.attr_bfc0)
+      brw_MOV(p, 
+             get_vert_attr(c, vert, c->key.attr_col0), 
+             get_vert_attr(c, vert, c->key.attr_bfc0));
+
+   if (c->key.attr_col1 && c->key.attr_bfc1)
+      brw_MOV(p, 
+             get_vert_attr(c, vert, c->key.attr_col1), 
+             get_vert_attr(c, vert, c->key.attr_bfc1));
+}
+
+
+static void do_twoside_color( struct brw_sf_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *if_insn;
+   GLuint backface_conditional = c->key.frontface_ccw ? BRW_CONDITIONAL_G : BRW_CONDITIONAL_L;
+
+   /* Already done in clip program:
+    */
+   if (c->key.primitive == SF_UNFILLED_TRIS)
+      return;
+
+   /* XXX: What happens if BFC isn't present?  This could only happen
+    * for user-supplied vertex programs, as t_vp_build.c always does
+    * the right thing.
+    */
+   if (!(c->key.attr_col0 && c->key.attr_bfc0) &&
+       !(c->key.attr_col1 && c->key.attr_bfc1))
+      return;
+   
+   /* Need to use BRW_EXECUTE_4 and also do an 4-wide compare in order
+    * to get all channels active inside the IF.  In the clipping code
+    * we run with NoMask, so it's not an option and we can use
+    * BRW_EXECUTE_1 for all comparisions.
+    */
+   brw_push_insn_state(p);
+   brw_CMP(p, vec4(brw_null_reg()), backface_conditional, c->det, brw_imm_f(0));
+   if_insn = brw_IF(p, BRW_EXECUTE_4); 
+   {
+      switch (c->nr_verts) {
+      case 3: copy_bfc(c, c->vert[2]);
+      case 2: copy_bfc(c, c->vert[1]);
+      case 1: copy_bfc(c, c->vert[0]);
+      }
+   }
+   brw_ENDIF(p, if_insn);
+   brw_pop_insn_state(p);
+}
+
+
+
+/***********************************************************************
+ * Flat shading
+ */
+
+#define VERT_RESULT_COLOR_BITS ((1<<VERT_RESULT_COL0) | \
+                                 (1<<VERT_RESULT_COL1))
+
+static void copy_colors( struct brw_sf_compile *c,
+                    struct brw_reg dst,
+                    struct brw_reg src)
+{
+   struct brw_compile *p = &c->func;
+
+   if (c->key.attr_col0)
+      brw_MOV(p, 
+             get_vert_attr(c, dst, c->key.attr_col0), 
+             get_vert_attr(c, src, c->key.attr_col0));
+
+   if (c->key.attr_col1)
+      brw_MOV(p, 
+             get_vert_attr(c, dst, c->key.attr_col1), 
+             get_vert_attr(c, src, c->key.attr_col1));
+
+}
+
+
+
+/* Need to use a computed jump to copy flatshaded attributes as the
+ * vertices are ordered according to y-coordinate before reaching this
+ * point, so the PV could be anywhere.
+ */
+static void do_flatshade_triangle( struct brw_sf_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg ip = brw_ip_reg();
+   GLuint jmpi = 1;
+   GLuint nr = 0;
+
+   if (c->key.attr_col0)
+      nr++;
+
+   if (c->key.attr_col1)
+      nr++;
+
+   if (nr == 0)
+      return;
+
+   /* Already done in clip program:
+    */
+   if (c->key.primitive == SF_UNFILLED_TRIS)
+      return;
+
+   if (BRW_IS_IGDNG(p->brw))
+       jmpi = 2;
+
+   brw_push_insn_state(p);
+   
+   brw_MUL(p, c->pv, c->pv, brw_imm_d(jmpi*(nr*2+1)));
+   brw_JMPI(p, ip, ip, c->pv);
+
+   copy_colors(c, c->vert[1], c->vert[0]);
+   copy_colors(c, c->vert[2], c->vert[0]);
+   brw_JMPI(p, ip, ip, brw_imm_d(jmpi*(nr*4+1)));
+
+   copy_colors(c, c->vert[0], c->vert[1]);
+   copy_colors(c, c->vert[2], c->vert[1]);
+   brw_JMPI(p, ip, ip, brw_imm_d(jmpi*nr*2));
+
+   copy_colors(c, c->vert[0], c->vert[2]);
+   copy_colors(c, c->vert[1], c->vert[2]);
+
+   brw_pop_insn_state(p);
+}
+       
+
+static void do_flatshade_line( struct brw_sf_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg ip = brw_ip_reg();
+   GLuint jmpi = 1;
+   GLuint nr = 0;
+
+   if (c->key.attr_col0)
+      nr++;
+
+   if (c->key.attr_col1)
+      nr++;
+
+   if (nr == 0)
+      return;
+
+   /* Already done in clip program: 
+    */
+   if (c->key.primitive == SF_UNFILLED_TRIS)
+      return;
+
+   if (BRW_IS_IGDNG(p->brw))
+       jmpi = 2;
+
+   brw_push_insn_state(p);
+   
+   brw_MUL(p, c->pv, c->pv, brw_imm_d(jmpi*(nr+1)));
+   brw_JMPI(p, ip, ip, c->pv);
+   copy_colors(c, c->vert[1], c->vert[0]);
+
+   brw_JMPI(p, ip, ip, brw_imm_ud(jmpi*nr));
+   copy_colors(c, c->vert[0], c->vert[1]);
+
+   brw_pop_insn_state(p);
+}
+
+       
+
+/***********************************************************************
+ * Triangle setup.
+ */
+
+
+static void alloc_regs( struct brw_sf_compile *c )
+{
+   GLuint reg, i;
+
+   /* Values computed by fixed function unit:
+    */
+   c->pv  = retype(brw_vec1_grf(1, 1), BRW_REGISTER_TYPE_D);
+   c->det = brw_vec1_grf(1, 2);
+   c->dx0 = brw_vec1_grf(1, 3);
+   c->dx2 = brw_vec1_grf(1, 4);
+   c->dy0 = brw_vec1_grf(1, 5);
+   c->dy2 = brw_vec1_grf(1, 6);
+
+   /* z and 1/w passed in seperately:
+    */
+   c->z[0]     = brw_vec1_grf(2, 0);
+   c->inv_w[0] = brw_vec1_grf(2, 1);
+   c->z[1]     = brw_vec1_grf(2, 2);
+   c->inv_w[1] = brw_vec1_grf(2, 3);
+   c->z[2]     = brw_vec1_grf(2, 4);
+   c->inv_w[2] = brw_vec1_grf(2, 5);
+   
+   /* The vertices:
+    */
+   reg = 3;
+   for (i = 0; i < c->nr_verts; i++) {
+      c->vert[i] = brw_vec8_grf(reg, 0);
+      reg += c->nr_attr_regs;
+   }
+
+   /* Temporaries, allocated after last vertex reg.
+    */
+   c->inv_det = brw_vec1_grf(reg, 0);  reg++;
+   c->a1_sub_a0 = brw_vec8_grf(reg, 0);  reg++;
+   c->a2_sub_a0 = brw_vec8_grf(reg, 0);  reg++;
+   c->tmp = brw_vec8_grf(reg, 0);  reg++;
+
+   /* Note grf allocation:
+    */
+   c->prog_data.total_grf = reg;
+   
+
+   /* Outputs of this program - interpolation coefficients for
+    * rasterization:
+    */
+   c->m1Cx = brw_vec8_reg(BRW_MESSAGE_REGISTER_FILE, 1, 0);
+   c->m2Cy = brw_vec8_reg(BRW_MESSAGE_REGISTER_FILE, 2, 0);
+   c->m3C0 = brw_vec8_reg(BRW_MESSAGE_REGISTER_FILE, 3, 0);
+}
+
+
+static void copy_z_inv_w( struct brw_sf_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   GLuint i;
+
+   brw_push_insn_state(p);
+       
+   /* Copy both scalars with a single MOV:
+    */
+   for (i = 0; i < c->nr_verts; i++)
+      brw_MOV(p, vec2(suboffset(c->vert[i], 2)), vec2(c->z[i]));
+        
+   brw_pop_insn_state(p);
+}
+
+
+static void invert_det( struct brw_sf_compile *c)
+{
+   /* Looks like we invert all 8 elements just to get 1/det in
+    * position 2 !?!
+    */
+   brw_math(&c->func, 
+           c->inv_det, 
+           BRW_MATH_FUNCTION_INV,
+           BRW_MATH_SATURATE_NONE,
+           0, 
+           c->det,
+           BRW_MATH_DATA_SCALAR,
+           BRW_MATH_PRECISION_FULL);
+
+}
+
+
+/* Two attributes packed into a wide register.  Figure out if either
+ * or both of them need linear/perspective interpolation.  Constant
+ * regs are left as-is.
+ */
+static GLboolean calculate_masks( struct brw_sf_compile *c,
+                                 GLuint reg,
+                                 GLushort *pc,
+                                 GLushort *pc_persp,
+                                 GLushort *pc_linear)
+{
+   GLboolean is_last_attr = (reg == c->nr_setup_regs - 1);
+   GLuint persp_mask = c->key.persp_attrs;
+   GLuint linear_mask = (c->key.persp_attrs | c->key.linear_attrs);
+
+   *pc_persp = 0;
+   *pc_linear = 0;
+   *pc = 0xf;
+      
+   if (persp_mask & (1 << (reg*2))) 
+      *pc_persp = 0xf;
+
+   if (linear_mask & (1 << (reg*2))) 
+      *pc_linear = 0xf;
+
+   /* Maybe only processs one attribute on the final round:
+    */
+   if (reg*2+1 < c->nr_setup_attrs) {
+      *pc |= 0xf0;
+
+      if (persp_mask & (1 << (reg*2+1))) 
+        *pc_persp |= 0xf0;
+
+      if (linear_mask & (1 << (reg*2+1))) 
+        *pc_linear |= 0xf0;
+   }
+
+   return is_last_attr;
+}
+
+
+void brw_emit_null_setup( struct brw_sf_compile *c )
+{
+   struct brw_compile *p = &c->func;
+
+   /* m0 is implicitly copied from r0 in the send instruction:
+    */  
+   brw_urb_WRITE(p, 
+                 brw_null_reg(),
+                 0,
+                 brw_vec8_grf(0, 0), /* r0, will be copied to m0 */
+                 0,    /* allocate */
+                 1,    /* used */
+                 1,    /* msg len */
+                 0,    /* response len */
+                 1,    /* eot */
+                 1,    /* writes complete */
+                 0,    /* offset */
+                 BRW_URB_SWIZZLE_TRANSPOSE); 
+}
+
+void brw_emit_tri_setup( struct brw_sf_compile *c, GLboolean allocate)
+{
+   struct brw_compile *p = &c->func;
+   GLuint i;
+
+   c->nr_verts = 3;
+
+   if (allocate)
+      alloc_regs(c);
+
+   invert_det(c);
+   copy_z_inv_w(c);
+
+   if (c->key.do_twoside_color) 
+      do_twoside_color(c);
+
+   if (c->key.do_flat_shading)
+      do_flatshade_triangle(c);
+      
+   
+   for (i = 0; i < c->nr_setup_regs; i++)
+   {
+      /* Pair of incoming attributes:
+       */
+      struct brw_reg a0 = offset(c->vert[0], i);
+      struct brw_reg a1 = offset(c->vert[1], i);
+      struct brw_reg a2 = offset(c->vert[2], i);
+      GLushort pc, pc_persp, pc_linear;
+      GLboolean last = calculate_masks(c, i, &pc, &pc_persp, &pc_linear);
+
+      if (pc_persp)
+      {
+        brw_set_predicate_control_flag_value(p, pc_persp);
+        brw_MUL(p, a0, a0, c->inv_w[0]);
+        brw_MUL(p, a1, a1, c->inv_w[1]);
+        brw_MUL(p, a2, a2, c->inv_w[2]);
+      }
+      
+      
+      /* Calculate coefficients for interpolated values:
+       */      
+      if (pc_linear)
+      {
+        brw_set_predicate_control_flag_value(p, pc_linear);
+
+        brw_ADD(p, c->a1_sub_a0, a1, negate(a0));
+        brw_ADD(p, c->a2_sub_a0, a2, negate(a0));
+
+        /* calculate dA/dx
+         */
+        brw_MUL(p, brw_null_reg(), c->a1_sub_a0, c->dy2);
+        brw_MAC(p, c->tmp, c->a2_sub_a0, negate(c->dy0));
+        brw_MUL(p, c->m1Cx, c->tmp, c->inv_det);
+               
+        /* calculate dA/dy
+         */
+        brw_MUL(p, brw_null_reg(), c->a2_sub_a0, c->dx0);
+        brw_MAC(p, c->tmp, c->a1_sub_a0, negate(c->dx2));
+        brw_MUL(p, c->m2Cy, c->tmp, c->inv_det);
+      }
+
+      {
+        brw_set_predicate_control_flag_value(p, pc); 
+        /* start point for interpolation
+         */
+        brw_MOV(p, c->m3C0, a0);
+      
+        /* Copy m0..m3 to URB.  m0 is implicitly copied from r0 in
+         * the send instruction:
+         */     
+        brw_urb_WRITE(p, 
+                      brw_null_reg(),
+                      0,
+                      brw_vec8_grf(0, 0), /* r0, will be copied to m0 */
+                      0,       /* allocate */
+                      1,       /* used */
+                      4,       /* msg len */
+                      0,       /* response len */
+                      last,    /* eot */
+                      last,    /* writes complete */
+                      i*4,     /* offset */
+                      BRW_URB_SWIZZLE_TRANSPOSE); /* XXX: Swizzle control "SF to windower" */
+      }
+   }
+}
+
+
+
+void brw_emit_line_setup( struct brw_sf_compile *c, GLboolean allocate)
+{
+   struct brw_compile *p = &c->func;
+   GLuint i;
+
+
+   c->nr_verts = 2;
+
+   if (allocate)
+      alloc_regs(c);
+
+   invert_det(c);
+   copy_z_inv_w(c);
+
+   if (c->key.do_flat_shading)
+      do_flatshade_line(c);
+
+   for (i = 0; i < c->nr_setup_regs; i++)
+   {
+      /* Pair of incoming attributes:
+       */
+      struct brw_reg a0 = offset(c->vert[0], i);
+      struct brw_reg a1 = offset(c->vert[1], i);
+      GLushort pc, pc_persp, pc_linear;
+      GLboolean last = calculate_masks(c, i, &pc, &pc_persp, &pc_linear);
+
+      if (pc_persp)
+      {
+        brw_set_predicate_control_flag_value(p, pc_persp);
+        brw_MUL(p, a0, a0, c->inv_w[0]);
+        brw_MUL(p, a1, a1, c->inv_w[1]);
+      }
+
+      /* Calculate coefficients for position, color:
+       */
+      if (pc_linear) {
+        brw_set_predicate_control_flag_value(p, pc_linear); 
+
+        brw_ADD(p, c->a1_sub_a0, a1, negate(a0));
+
+        brw_MUL(p, c->tmp, c->a1_sub_a0, c->dx0); 
+        brw_MUL(p, c->m1Cx, c->tmp, c->inv_det);
+               
+        brw_MUL(p, c->tmp, c->a1_sub_a0, c->dy0);
+        brw_MUL(p, c->m2Cy, c->tmp, c->inv_det);
+      }
+
+      {
+        brw_set_predicate_control_flag_value(p, pc); 
+
+        /* start point for interpolation
+         */
+        brw_MOV(p, c->m3C0, a0);
+
+        /* Copy m0..m3 to URB. 
+         */
+        brw_urb_WRITE(p, 
+                      brw_null_reg(),
+                      0,
+                      brw_vec8_grf(0, 0),
+                      0,       /* allocate */
+                      1,       /* used */
+                      4,       /* msg len */
+                      0,       /* response len */
+                      last,    /* eot */
+                      last,    /* writes complete */
+                      i*4,     /* urb destination offset */
+                      BRW_URB_SWIZZLE_TRANSPOSE); 
+      }
+   } 
+}
+
+void brw_emit_point_sprite_setup( struct brw_sf_compile *c, GLboolean allocate)
+{
+   struct brw_compile *p = &c->func;
+   GLuint i;
+
+   c->nr_verts = 1;
+
+   if (allocate)
+      alloc_regs(c);
+
+   copy_z_inv_w(c);
+
+   for (i = 0; i < c->nr_setup_regs; i++)
+   {
+      /* XXX: only seems to check point_coord_replace_attrs for every
+       * second attribute?!?
+       */
+      boolean coord_replace = !!(c->key.point_coord_replace_attrs & (1<<(2*i)));
+      struct brw_reg a0 = offset(c->vert[0], i);
+      GLushort pc, pc_persp, pc_linear;
+      GLboolean last = calculate_masks(c, i, &pc, &pc_persp, &pc_linear);
+            
+      if (pc_persp)
+      {                                
+        if (coord_replace) {
+           brw_set_predicate_control_flag_value(p, pc_persp);
+           brw_MUL(p, a0, a0, c->inv_w[0]);
+        }
+      }
+
+      if (coord_replace) {
+        /* Caculate 1.0/PointWidth */
+        brw_math(&c->func,
+                 c->tmp,
+                 BRW_MATH_FUNCTION_INV,
+                 BRW_MATH_SATURATE_NONE,
+                 0,
+                 c->dx0,
+                 BRW_MATH_DATA_SCALAR,
+                 BRW_MATH_PRECISION_FULL);
+
+        if (c->key.sprite_origin_lower_left) {
+           brw_MUL(p, c->m1Cx, c->tmp, c->inv_w[0]);
+           brw_MOV(p, vec1(suboffset(c->m1Cx, 1)), brw_imm_f(0.0));
+           brw_MUL(p, c->m2Cy, c->tmp, negate(c->inv_w[0]));
+           brw_MOV(p, vec1(suboffset(c->m2Cy, 0)), brw_imm_f(0.0));
+        } 
+        else {
+           brw_MUL(p, c->m1Cx, c->tmp, c->inv_w[0]);
+           brw_MOV(p, vec1(suboffset(c->m1Cx, 1)), brw_imm_f(0.0));
+           brw_MUL(p, c->m2Cy, c->tmp, c->inv_w[0]);
+           brw_MOV(p, vec1(suboffset(c->m2Cy, 0)), brw_imm_f(0.0));
+        }
+      } 
+      else {
+        brw_MOV(p, c->m1Cx, brw_imm_ud(0));
+        brw_MOV(p, c->m2Cy, brw_imm_ud(0));
+      }
+
+      {
+        brw_set_predicate_control_flag_value(p, pc); 
+        if (coord_replace) {
+           if (c->key.sprite_origin_lower_left) {
+              brw_MUL(p, c->m3C0, c->inv_w[0], brw_imm_f(1.0));
+              brw_MOV(p, vec1(suboffset(c->m3C0, 0)), brw_imm_f(0.0));
+           }
+           else {
+              brw_MOV(p, c->m3C0, brw_imm_f(0.0));
+           }
+        } 
+        else {
+           brw_MOV(p, c->m3C0, a0); /* constant value */
+        }
+
+        /* Copy m0..m3 to URB. 
+         */
+        brw_urb_WRITE(p, 
+                      brw_null_reg(),
+                      0,
+                      brw_vec8_grf(0, 0),
+                      0,       /* allocate */
+                      1,       /* used */
+                      4,       /* msg len */
+                      0,       /* response len */
+                      last,    /* eot */
+                      last,    /* writes complete */
+                      i*4,     /* urb destination offset */
+                      BRW_URB_SWIZZLE_TRANSPOSE);
+      }
+   }
+}
+
+/* Points setup - several simplifications as all attributes are
+ * constant across the face of the point (point sprites excluded!)
+ */
+void brw_emit_point_setup( struct brw_sf_compile *c, GLboolean allocate)
+{
+   struct brw_compile *p = &c->func;
+   GLuint i;
+
+   c->nr_verts = 1;
+   
+   if (allocate)
+      alloc_regs(c);
+
+   copy_z_inv_w(c);
+
+   brw_MOV(p, c->m1Cx, brw_imm_ud(0)); /* zero - move out of loop */
+   brw_MOV(p, c->m2Cy, brw_imm_ud(0)); /* zero - move out of loop */
+
+   for (i = 0; i < c->nr_setup_regs; i++)
+   {
+      struct brw_reg a0 = offset(c->vert[0], i);
+      GLushort pc, pc_persp, pc_linear;
+      GLboolean last = calculate_masks(c, i, &pc, &pc_persp, &pc_linear);
+            
+      if (pc_persp)
+      {                                
+        /* This seems odd as the values are all constant, but the
+         * fragment shader will be expecting it:
+         */
+        brw_set_predicate_control_flag_value(p, pc_persp);
+        brw_MUL(p, a0, a0, c->inv_w[0]);
+      }
+
+
+      /* The delta values are always zero, just send the starting
+       * coordinate.  Again, this is to fit in with the interpolation
+       * code in the fragment shader.
+       */
+      {
+        brw_set_predicate_control_flag_value(p, pc); 
+
+        brw_MOV(p, c->m3C0, a0); /* constant value */
+
+        /* Copy m0..m3 to URB. 
+         */
+        brw_urb_WRITE(p, 
+                      brw_null_reg(),
+                      0,
+                      brw_vec8_grf(0, 0),
+                      0,       /* allocate */
+                      1,       /* used */
+                      4,       /* msg len */
+                      0,       /* response len */
+                      last,    /* eot */
+                      last,    /* writes complete */
+                      i*4,     /* urb destination offset */
+                      BRW_URB_SWIZZLE_TRANSPOSE);
+      }
+   }
+}
+
+void brw_emit_anyprim_setup( struct brw_sf_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg ip = brw_ip_reg();
+   struct brw_reg payload_prim = brw_uw1_reg(BRW_GENERAL_REGISTER_FILE, 1, 0);
+   struct brw_reg payload_attr = get_element_ud(brw_vec1_reg(BRW_GENERAL_REGISTER_FILE, 1, 0), 0); 
+   struct brw_reg primmask;
+   struct brw_instruction *jmp;
+   struct brw_reg v1_null_ud = vec1(retype(brw_null_reg(), BRW_REGISTER_TYPE_UD));
+   
+   GLuint saveflag;
+
+   c->nr_verts = 3;
+   alloc_regs(c);
+
+   primmask = retype(get_element(c->tmp, 0), BRW_REGISTER_TYPE_UD);
+
+   brw_MOV(p, primmask, brw_imm_ud(1));
+   brw_SHL(p, primmask, primmask, payload_prim);
+
+   brw_set_conditionalmod(p, BRW_CONDITIONAL_Z);
+   brw_AND(p, v1_null_ud, primmask, brw_imm_ud((1<<_3DPRIM_TRILIST) |
+                                              (1<<_3DPRIM_TRISTRIP) |
+                                              (1<<_3DPRIM_TRIFAN) |
+                                              (1<<_3DPRIM_TRISTRIP_REVERSE) |
+                                              (1<<_3DPRIM_POLYGON) |
+                                              (1<<_3DPRIM_RECTLIST) |
+                                              (1<<_3DPRIM_TRIFAN_NOSTIPPLE)));
+   jmp = brw_JMPI(p, ip, ip, brw_imm_d(0));
+   {
+      saveflag = p->flag_value;
+      brw_push_insn_state(p); 
+      brw_emit_tri_setup( c, GL_FALSE );
+      brw_pop_insn_state(p);
+      p->flag_value = saveflag;
+      /* note - thread killed in subroutine, so must
+       * restore the flag which is changed when building
+       * the subroutine. fix #13240
+       */
+   }
+   brw_land_fwd_jump(p, jmp);
+
+   brw_set_conditionalmod(p, BRW_CONDITIONAL_Z);
+   brw_AND(p, v1_null_ud, primmask, brw_imm_ud((1<<_3DPRIM_LINELIST) |
+                                              (1<<_3DPRIM_LINESTRIP) |
+                                              (1<<_3DPRIM_LINELOOP) |
+                                              (1<<_3DPRIM_LINESTRIP_CONT) |
+                                              (1<<_3DPRIM_LINESTRIP_BF) |
+                                              (1<<_3DPRIM_LINESTRIP_CONT_BF)));
+   jmp = brw_JMPI(p, ip, ip, brw_imm_d(0));
+   {
+      saveflag = p->flag_value;
+      brw_push_insn_state(p); 
+      brw_emit_line_setup( c, GL_FALSE );
+      brw_pop_insn_state(p);
+      p->flag_value = saveflag;
+      /* note - thread killed in subroutine */
+   }
+   brw_land_fwd_jump(p, jmp); 
+
+   brw_set_conditionalmod(p, BRW_CONDITIONAL_Z);
+   brw_AND(p, v1_null_ud, payload_attr, brw_imm_ud(1<<BRW_SPRITE_POINT_ENABLE));
+   jmp = brw_JMPI(p, ip, ip, brw_imm_d(0));
+   {
+      saveflag = p->flag_value;
+      brw_push_insn_state(p); 
+      brw_emit_point_sprite_setup( c, GL_FALSE );
+      brw_pop_insn_state(p);
+      p->flag_value = saveflag;
+   }
+   brw_land_fwd_jump(p, jmp); 
+
+   brw_emit_point_setup( c, GL_FALSE );
+}
+
+
+
+
diff --git a/src/gallium/drivers/i965/brw_sf_state.c b/src/gallium/drivers/i965/brw_sf_state.c
new file mode 100644 (file)
index 0000000..25dc2b5
--- /dev/null
@@ -0,0 +1,333 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+   
+#include "util/u_math.h"
+
+#include "pipe/p_state.h"
+
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+#include "brw_debug.h"
+#include "brw_pipe_rast.h"
+
+static enum pipe_error upload_sf_vp(struct brw_context *brw)
+{
+   const struct pipe_viewport_state *vp = &brw->curr.viewport;
+   const struct pipe_scissor_state *scissor = &brw->curr.scissor;
+   struct brw_sf_viewport sfv;
+   enum pipe_error ret;
+
+   memset(&sfv, 0, sizeof(sfv));
+
+   /* PIPE_NEW_VIEWPORT, PIPE_NEW_SCISSOR */
+
+   sfv.viewport.m00 = vp->scale[0];
+   sfv.viewport.m11 = vp->scale[1];
+   sfv.viewport.m22 = vp->scale[2];
+   sfv.viewport.m30 = vp->translate[0];
+   sfv.viewport.m31 = vp->translate[1];
+   sfv.viewport.m32 = vp->translate[2];
+
+   sfv.scissor.xmin = scissor->minx;
+   sfv.scissor.xmax = scissor->maxx - 1; /* ? */
+   sfv.scissor.ymin = scissor->miny;
+   sfv.scissor.ymax = scissor->maxy - 1; /* ? */
+
+   ret = brw_cache_data( &brw->cache, BRW_SF_VP, &sfv, NULL, 0,
+                         &brw->sf.vp_bo );
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+const struct brw_tracked_state brw_sf_vp = {
+   .dirty = {
+      .mesa  = (PIPE_NEW_VIEWPORT | 
+               PIPE_NEW_SCISSOR),
+      .brw   = 0,
+      .cache = 0
+   },
+   .prepare = upload_sf_vp
+};
+
+struct brw_sf_unit_key {
+   unsigned int total_grf;
+   unsigned int urb_entry_read_length;
+   unsigned int nr_urb_entries, urb_size, sfsize;
+   
+   unsigned scissor:1;
+   unsigned line_smooth:1;
+   unsigned point_sprite:1;
+   unsigned point_attenuated:1;
+   unsigned front_face:2;
+   unsigned cull_mode:2;
+   unsigned flatshade_first:1;
+   unsigned gl_rasterization_rules:1;
+   unsigned line_last_pixel_enable:1;
+   float line_width;
+   float point_size;
+};
+
+static void
+sf_unit_populate_key(struct brw_context *brw, struct brw_sf_unit_key *key)
+{
+   const struct pipe_rasterizer_state *rast = &brw->curr.rast->templ;
+   memset(key, 0, sizeof(*key));
+
+   /* CACHE_NEW_SF_PROG */
+   key->total_grf = brw->sf.prog_data->total_grf;
+   key->urb_entry_read_length = brw->sf.prog_data->urb_read_length;
+
+   /* BRW_NEW_URB_FENCE */
+   key->nr_urb_entries = brw->urb.nr_sf_entries;
+   key->urb_size = brw->urb.vsize;
+   key->sfsize = brw->urb.sfsize;
+
+   /* PIPE_NEW_RAST */
+   key->scissor = rast->scissor;
+   key->front_face = rast->front_winding;
+   key->cull_mode = rast->cull_mode;
+   key->line_smooth = rast->line_smooth;
+   key->line_width = rast->line_width;
+   key->flatshade_first = rast->flatshade_first;
+   key->line_last_pixel_enable = rast->line_last_pixel;
+   key->gl_rasterization_rules = rast->gl_rasterization_rules;
+
+   key->point_sprite = rast->point_sprite;
+   key->point_attenuated = rast->point_size_per_vertex;
+
+   key->point_size = CLAMP(rast->point_size, 
+                          rast->point_size_min, 
+                          rast->point_size_max);
+}
+
+static enum pipe_error
+sf_unit_create_from_key(struct brw_context *brw,
+                        struct brw_sf_unit_key *key,
+                        struct brw_winsys_reloc *reloc,
+                        struct brw_winsys_buffer **bo_out)
+{
+   struct brw_sf_unit_state sf;
+   enum pipe_error ret;
+   int chipset_max_threads;
+   memset(&sf, 0, sizeof(sf));
+
+   sf.thread0.grf_reg_count = align(key->total_grf, 16) / 16 - 1;
+   /* reloc */
+   sf.thread0.kernel_start_pointer = 0;
+
+   sf.thread1.floating_point_mode = BRW_FLOATING_POINT_NON_IEEE_754;
+
+   sf.thread3.dispatch_grf_start_reg = 3;
+
+   if (BRW_IS_IGDNG(brw))
+       sf.thread3.urb_entry_read_offset = 3;
+   else
+       sf.thread3.urb_entry_read_offset = 1;
+
+   sf.thread3.urb_entry_read_length = key->urb_entry_read_length;
+
+   sf.thread4.nr_urb_entries = key->nr_urb_entries;
+   sf.thread4.urb_entry_allocation_size = key->sfsize - 1;
+
+   /* Each SF thread produces 1 PUE, and there can be up to 24(Pre-IGDNG) or 
+    * 48(IGDNG) threads 
+    */
+   if (BRW_IS_IGDNG(brw))
+      chipset_max_threads = 48;
+   else
+      chipset_max_threads = 24;
+
+   sf.thread4.max_threads = MIN2(chipset_max_threads, key->nr_urb_entries) - 1;
+
+   if (BRW_DEBUG & DEBUG_SINGLE_THREAD)
+      sf.thread4.max_threads = 0;
+
+   if (BRW_DEBUG & DEBUG_STATS)
+      sf.thread4.stats_enable = 1;
+
+   /* CACHE_NEW_SF_VP */
+   /* reloc */
+   sf.sf5.sf_viewport_state_offset = 0;
+
+   sf.sf5.viewport_transform = 1;
+
+   if (key->scissor)
+      sf.sf6.scissor = 1;
+
+   if (key->front_face == PIPE_WINDING_CCW)
+      sf.sf5.front_winding = BRW_FRONTWINDING_CCW;
+   else
+      sf.sf5.front_winding = BRW_FRONTWINDING_CW;
+
+   switch (key->cull_mode) {
+   case PIPE_WINDING_CCW:
+   case PIPE_WINDING_CW:
+      sf.sf6.cull_mode = (key->front_face == key->cull_mode ?
+                         BRW_CULLMODE_FRONT :
+                         BRW_CULLMODE_BACK);
+      break;
+   case PIPE_WINDING_BOTH:
+      sf.sf6.cull_mode = BRW_CULLMODE_BOTH;
+      break;
+   case PIPE_WINDING_NONE:
+      sf.sf6.cull_mode = BRW_CULLMODE_NONE;
+      break;
+   default:
+      assert(0);
+      sf.sf6.cull_mode = BRW_CULLMODE_NONE;
+      break;
+   }
+
+   /* _NEW_LINE */
+   /* XXX use ctx->Const.Min/MaxLineWidth here */
+   sf.sf6.line_width = CLAMP(key->line_width, 1.0, 5.0) * (1<<1);
+
+   sf.sf6.line_endcap_aa_region_width = 1;
+   if (key->line_smooth)
+      sf.sf6.aa_enable = 1;
+   else if (sf.sf6.line_width <= 0x2)
+       sf.sf6.line_width = 0;
+
+   /* XXX: gl_rasterization_rules?  something else?
+    */
+   sf.sf6.point_rast_rule = BRW_RASTRULE_UPPER_RIGHT;
+   sf.sf6.point_rast_rule = BRW_RASTRULE_LOWER_RIGHT;
+   sf.sf6.point_rast_rule = 1;
+
+   /* XXX clamp max depends on AA vs. non-AA */
+
+   /* _NEW_POINT */
+   sf.sf7.sprite_point = key->point_sprite;
+   sf.sf7.point_size = CLAMP(rint(key->point_size), 1, 255) * (1<<3);
+   sf.sf7.use_point_size_state = !key->point_attenuated;
+   sf.sf7.aa_line_distance_mode = 0;
+
+   /* might be BRW_NEW_PRIMITIVE if we have to adjust pv for polygons:
+    */
+   if (!key->flatshade_first) {
+      sf.sf7.trifan_pv = 2;
+      sf.sf7.linestrip_pv = 1;
+      sf.sf7.tristrip_pv = 2;
+   } else {
+      sf.sf7.trifan_pv = 1;
+      sf.sf7.linestrip_pv = 0;
+      sf.sf7.tristrip_pv = 0;
+   }
+
+   sf.sf7.line_last_pixel_enable = key->line_last_pixel_enable;
+
+   /* Set bias for OpenGL rasterization rules:
+    */
+   if (key->gl_rasterization_rules) {
+      sf.sf6.dest_org_vbias = 0x8;
+      sf.sf6.dest_org_hbias = 0x8;
+   }
+   else {
+      sf.sf6.dest_org_vbias = 0x0;
+      sf.sf6.dest_org_hbias = 0x0;
+   }
+
+   ret = brw_upload_cache(&brw->cache, BRW_SF_UNIT,
+                          key, sizeof(*key),
+                          reloc, 2,
+                          &sf, sizeof(sf),
+                          NULL, NULL,
+                          bo_out);
+   if (ret)
+      return ret;
+
+   
+   return PIPE_OK;
+}
+
+static enum pipe_error upload_sf_unit( struct brw_context *brw )
+{
+   struct brw_sf_unit_key key;
+   struct brw_winsys_reloc reloc[2];
+   unsigned total_grf;
+   unsigned viewport_transform;
+   unsigned front_winding;
+   enum pipe_error ret;
+
+   sf_unit_populate_key(brw, &key);
+   
+   /* XXX: cut this crap and pre calculate the key:
+    */
+   total_grf = (align(key.total_grf, 16) / 16 - 1);
+   viewport_transform = 1;
+   front_winding = (key.front_face == PIPE_WINDING_CCW ?
+                    BRW_FRONTWINDING_CCW :
+                    BRW_FRONTWINDING_CW);
+
+   /* Emit SF program relocation */
+   make_reloc(&reloc[0],
+              BRW_USAGE_STATE,
+              total_grf << 1,
+              offsetof(struct brw_sf_unit_state, thread0),
+              brw->sf.prog_bo);
+
+   /* Emit SF viewport relocation */
+   make_reloc(&reloc[1],
+              BRW_USAGE_STATE,
+              front_winding | (viewport_transform << 1),
+              offsetof(struct brw_sf_unit_state, sf5),
+              brw->sf.vp_bo);
+
+
+   if (brw_search_cache(&brw->cache, BRW_SF_UNIT,
+                        &key, sizeof(key),
+                        reloc, 2,
+                        NULL,
+                        &brw->sf.state_bo))
+      return PIPE_OK;
+
+
+   ret = sf_unit_create_from_key(brw, &key,
+                                 reloc,
+                                 &brw->sf.state_bo);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+const struct brw_tracked_state brw_sf_unit = {
+   .dirty = {
+      .mesa  = (PIPE_NEW_RAST),
+      .brw   = BRW_NEW_URB_FENCE,
+      .cache = (CACHE_NEW_SF_VP |
+               CACHE_NEW_SF_PROG)
+   },
+   .prepare = upload_sf_unit,
+};
diff --git a/src/gallium/drivers/i965/brw_state.h b/src/gallium/drivers/i965/brw_state.h
new file mode 100644 (file)
index 0000000..d2bbd01
--- /dev/null
@@ -0,0 +1,174 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+    
+
+#ifndef BRW_STATE_H
+#define BRW_STATE_H
+
+#include "pipe/p_defines.h"
+#include "util/u_memory.h"
+
+#include "brw_context.h"
+
+static INLINE void
+brw_add_validated_bo(struct brw_context *brw, struct brw_winsys_buffer *bo)
+{
+   assert(brw->state.validated_bo_count < Elements(brw->state.validated_bos));
+
+   if (bo != NULL) {
+      bo_reference( &brw->state.validated_bos[brw->state.validated_bo_count++],
+                    bo );
+   }
+}
+
+const struct brw_tracked_state brw_blend_constant_color;
+const struct brw_tracked_state brw_cc_unit;
+const struct brw_tracked_state brw_cc_vp;
+const struct brw_tracked_state brw_clip_prog;
+const struct brw_tracked_state brw_clip_unit;
+const struct brw_tracked_state brw_curbe_buffer;
+const struct brw_tracked_state brw_curbe_offsets;
+const struct brw_tracked_state brw_invarient_state;
+const struct brw_tracked_state brw_gs_prog;
+const struct brw_tracked_state brw_gs_unit;
+const struct brw_tracked_state brw_line_stipple;
+const struct brw_tracked_state brw_aa_line_parameters;
+const struct brw_tracked_state brw_pipelined_state_pointers;
+const struct brw_tracked_state brw_binding_table_pointers;
+const struct brw_tracked_state brw_depthbuffer;
+const struct brw_tracked_state brw_polygon_stipple;
+const struct brw_tracked_state brw_program_parameters;
+const struct brw_tracked_state brw_recalculate_urb_fence;
+const struct brw_tracked_state brw_sf_prog;
+const struct brw_tracked_state brw_sf_unit;
+const struct brw_tracked_state brw_sf_vp;
+const struct brw_tracked_state brw_state_base_address;
+const struct brw_tracked_state brw_urb_fence;
+const struct brw_tracked_state brw_vertex_state;
+const struct brw_tracked_state brw_vs_surfaces;
+const struct brw_tracked_state brw_vs_prog;
+const struct brw_tracked_state brw_vs_unit;
+const struct brw_tracked_state brw_wm_input_sizes;
+const struct brw_tracked_state brw_wm_prog;
+const struct brw_tracked_state brw_wm_samplers;
+const struct brw_tracked_state brw_wm_constant_surface;
+const struct brw_tracked_state brw_wm_surfaces;
+const struct brw_tracked_state brw_wm_unit;
+
+const struct brw_tracked_state brw_psp_urb_cbs;
+
+const struct brw_tracked_state brw_pipe_control;
+
+const struct brw_tracked_state brw_drawing_rect;
+const struct brw_tracked_state brw_indices;
+const struct brw_tracked_state brw_vertices;
+const struct brw_tracked_state brw_index_buffer;
+
+
+/***********************************************************************
+ * brw_state.c
+ */
+int brw_validate_state(struct brw_context *brw);
+int brw_upload_state(struct brw_context *brw);
+void brw_init_state(struct brw_context *brw);
+void brw_destroy_state(struct brw_context *brw);
+
+/***********************************************************************
+ * brw_state_cache.c
+ */
+enum pipe_error brw_cache_data(struct brw_cache *cache,
+                               enum brw_cache_id cache_id,
+                               const void *data,
+                               struct brw_winsys_reloc *relocs,
+                               GLuint nr_relocs,
+                               struct brw_winsys_buffer **bo_out );
+
+enum pipe_error brw_cache_data_sz(struct brw_cache *cache,
+                                  enum brw_cache_id cache_id,
+                                  const void *data,
+                                  GLuint data_size,
+                                  struct brw_winsys_reloc *relocs,
+                                  GLuint nr_relocs,
+                                  struct brw_winsys_buffer **bo_out);
+
+enum pipe_error brw_upload_cache( struct brw_cache *cache,
+                                  enum brw_cache_id cache_id,
+                                  const void *key,
+                                  GLuint key_sz,
+                                  struct brw_winsys_reloc *relocs,
+                                  GLuint nr_relocs,
+                                  const void *data,
+                                  GLuint data_sz,
+                                  const void *aux,
+                                  void *aux_return ,
+                                  struct brw_winsys_buffer **bo_out);
+
+boolean brw_search_cache( struct brw_cache *cache,
+                          enum brw_cache_id cache_id,
+                          const void *key,
+                          GLuint key_size,
+                          struct brw_winsys_reloc *relocs,
+                          GLuint nr_relocs,
+                          void *aux_return,
+                          struct brw_winsys_buffer **bo_out);
+
+void brw_state_cache_check_size( struct brw_context *brw );
+
+void brw_init_caches( struct brw_context *brw );
+void brw_destroy_caches( struct brw_context *brw );
+void brw_state_cache_bo_delete(struct brw_cache *cache, struct brw_winsys_buffer *bo);
+
+/***********************************************************************
+ * brw_state_batch.c
+ */
+#define BRW_BATCH_STRUCT(brw, s) brw_batchbuffer_data( brw->batch, (s), sizeof(*(s)), IGNORE_CLIPRECTS)
+#define BRW_CACHED_BATCH_STRUCT(brw, s) brw_cached_batch_struct( brw, (s), sizeof(*(s)) )
+
+GLboolean brw_cached_batch_struct( struct brw_context *brw,
+                                  const void *data,
+                                  GLuint sz );
+void brw_destroy_batch_cache( struct brw_context *brw );
+void brw_clear_batch_cache( struct brw_context *brw );
+
+/***********************************************************************
+ * brw_wm_surface_state.c 
+ */
+
+/***********************************************************************
+ * brw_state_debug.c
+ */
+void brw_update_dirty_counts( unsigned mesa,
+                             unsigned brw,
+                             unsigned cache );
+
+
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_state_batch.c b/src/gallium/drivers/i965/brw_state_batch.c
new file mode 100644 (file)
index 0000000..7d212e5
--- /dev/null
@@ -0,0 +1,98 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+     
+
+
+#include "brw_state.h"
+#include "brw_batchbuffer.h"
+
+
+
+/* A facility similar to the data caching code above, which aims to
+ * prevent identical commands being issued repeatedly.
+ */
+GLboolean brw_cached_batch_struct( struct brw_context *brw,
+                                  const void *data,
+                                  GLuint sz )
+{
+   struct brw_cached_batch_item *item = brw->cached_batch_items;
+   struct header *newheader = (struct header *)data;
+
+   if (brw->flags.always_emit_state) {
+      brw_batchbuffer_data(brw->batch, data, sz, IGNORE_CLIPRECTS);
+      return GL_TRUE;
+   }
+
+   while (item) {
+      if (item->header->opcode == newheader->opcode) {
+        if (item->sz == sz && memcmp(item->header, newheader, sz) == 0)
+           return GL_FALSE;
+        if (item->sz != sz) {
+           FREE(item->header);
+           item->header = MALLOC(sz);
+           item->sz = sz;
+        }
+        goto emit;
+      }
+      item = item->next;
+   }
+
+   assert(!item);
+   item = CALLOC_STRUCT(brw_cached_batch_item);
+   item->header = MALLOC(sz);
+   item->sz = sz;
+   item->next = brw->cached_batch_items;
+   brw->cached_batch_items = item;
+
+ emit:
+   memcpy(item->header, newheader, sz);
+   brw_batchbuffer_data(brw->batch, data, sz, IGNORE_CLIPRECTS);
+   return GL_TRUE;
+}
+
+void brw_clear_batch_cache( struct brw_context *brw )
+{
+   struct brw_cached_batch_item *item = brw->cached_batch_items;
+
+   while (item) {
+      struct brw_cached_batch_item *next = item->next;
+      free((void *)item->header);
+      free(item);
+      item = next;
+   }
+
+   brw->cached_batch_items = NULL;
+}
+
+void brw_destroy_batch_cache( struct brw_context *brw )
+{
+   brw_clear_batch_cache(brw);
+}
diff --git a/src/gallium/drivers/i965/brw_state_cache.c b/src/gallium/drivers/i965/brw_state_cache.c
new file mode 100644 (file)
index 0000000..16b643c
--- /dev/null
@@ -0,0 +1,617 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+/** @file brw_state_cache.c
+ *
+ * This file implements a simple static state cache for 965.  The consumers
+ * can query the hash table of state using a cache_id, opaque key data,
+ * and list of buffers that will be used in relocations, and receive the
+ * corresponding state buffer object of state (plus associated auxiliary
+ * data) in return.
+ *
+ * The inner workings are a simple hash table based on a CRC of the key data.
+ * The cache_id and relocation target buffers associated with the state
+ * buffer are included as auxiliary key data, but are not part of the hash
+ * value (this should be fixed, but will likely be fixed instead by making
+ * consumers use structured keys).
+ *
+ * Replacement is not implemented.  Instead, when the cache gets too big, at
+ * a safe point (unlock) we throw out all of the cache data and let it
+ * regenerate for the next rendering operation.
+ *
+ * The reloc structs need to be included as key data, otherwise the
+ * non-unique values stuffed in the offset in key data through
+ * brw_cache_data() may result in successful probe for state buffers
+ * even when the buffer being referenced doesn't match.  The result would be
+ * that the same state cache entry is used twice for different buffers,
+ * only one of the two buffers referenced gets put into the offset, and the
+ * incorrect program is run for the other instance.
+ */
+#include "util/u_memory.h"
+
+#include "brw_debug.h"
+#include "brw_state.h"
+#include "brw_batchbuffer.h"
+
+/* XXX: Fixme - have to include these to get the sizes of the prog_key
+ * structs:
+ */
+#include "brw_wm.h"
+#include "brw_vs.h"
+#include "brw_clip.h"
+#include "brw_sf.h"
+#include "brw_gs.h"
+
+
+static GLuint
+hash_key(const void *key, GLuint key_size,
+         struct brw_winsys_reloc *relocs, GLuint nr_relocs)
+{
+   GLuint *ikey = (GLuint *)key;
+   GLuint hash = 0, i;
+
+   assert(key_size % 4 == 0);
+
+   /* I'm sure this can be improved on:
+    */
+   for (i = 0; i < key_size/4; i++) {
+      hash ^= ikey[i];
+      hash = (hash << 5) | (hash >> 27);
+   }
+
+   /* Include the BO pointers as key data as well */
+   ikey = (GLuint *)relocs;
+   key_size = nr_relocs * sizeof(struct brw_winsys_reloc);
+   for (i = 0; i < key_size/4; i++) {
+      hash ^= ikey[i];
+      hash = (hash << 5) | (hash >> 27);
+   }
+
+   return hash;
+}
+
+
+/**
+ * Marks a new buffer as being chosen for the given cache id.
+ */
+static void
+update_cache_last(struct brw_cache *cache, enum brw_cache_id cache_id,
+                 struct brw_winsys_buffer *bo)
+{
+   if (bo == cache->last_bo[cache_id])
+      return; /* no change */
+
+   bo_reference( &cache->last_bo[cache_id],  bo );
+
+   cache->brw->state.dirty.cache |= 1 << cache_id;
+}
+
+
+static struct brw_cache_item *
+search_cache(struct brw_cache *cache, enum brw_cache_id cache_id,
+            GLuint hash, const void *key, GLuint key_size,
+            struct brw_winsys_reloc *relocs, GLuint nr_relocs)
+{
+   struct brw_cache_item *c;
+
+#if 0
+   int bucketcount = 0;
+
+   for (c = cache->items[hash % cache->size]; c; c = c->next)
+      bucketcount++;
+
+   debug_printf("bucket %d/%d = %d/%d items\n", hash % cache->size,
+          cache->size, bucketcount, cache->n_items);
+#endif
+
+   for (c = cache->items[hash % cache->size]; c; c = c->next) {
+      if (c->cache_id == cache_id &&
+         c->hash == hash &&
+         c->key_size == key_size &&
+         memcmp(c->key, key, key_size) == 0 &&
+         c->nr_relocs == nr_relocs &&
+         memcmp(c->relocs, relocs, nr_relocs * sizeof *relocs) == 0)
+        return c;
+   }
+
+   return NULL;
+}
+
+
+static void
+rehash(struct brw_cache *cache)
+{
+   struct brw_cache_item **items;
+   struct brw_cache_item *c, *next;
+   GLuint size, i;
+
+   size = cache->size * 3;
+   items = (struct brw_cache_item**) CALLOC(size, sizeof(*items));
+
+   for (i = 0; i < cache->size; i++)
+      for (c = cache->items[i]; c; c = next) {
+        next = c->next;
+        c->next = items[c->hash % size];
+        items[c->hash % size] = c;
+      }
+
+   FREE(cache->items);
+   cache->items = items;
+   cache->size = size;
+}
+
+
+/**
+ * Returns the buffer object matching cache_id and key, or NULL.
+ */
+boolean
+brw_search_cache(struct brw_cache *cache,
+                 enum brw_cache_id cache_id,
+                 const void *key,
+                 GLuint key_size,
+                 struct brw_winsys_reloc *relocs, 
+                GLuint nr_relocs,
+                 void *aux_return,
+                 struct brw_winsys_buffer **bo_out)
+{
+   struct brw_cache_item *item;
+   GLuint hash = hash_key(key, key_size, relocs, nr_relocs);
+
+   item = search_cache(cache, cache_id, hash, key, key_size,
+                      relocs, nr_relocs);
+
+   if (item) {
+      if (aux_return)
+         *(void **)aux_return = (void *)((char *)item->key + item->key_size);
+      
+      update_cache_last(cache, cache_id, item->bo);
+      bo_reference(bo_out, item->bo);
+      return TRUE;
+   }
+   
+   return FALSE;      
+}
+
+
+enum pipe_error
+brw_upload_cache( struct brw_cache *cache,
+                 enum brw_cache_id cache_id,
+                 const void *key,
+                 GLuint key_size,
+                 struct brw_winsys_reloc *relocs,
+                 GLuint nr_relocs,
+                 const void *data,
+                 GLuint data_size,
+                 const void *aux,
+                 void *aux_return,
+                  struct brw_winsys_buffer **bo_out)
+{
+   struct brw_cache_item *item = CALLOC_STRUCT(brw_cache_item);
+   GLuint hash = hash_key(key, key_size, relocs, nr_relocs);
+   GLuint relocs_size = nr_relocs * sizeof relocs[0];
+   GLuint aux_size = cache->aux_size[cache_id];
+   enum pipe_error ret;
+   void *tmp;
+   int i;
+
+   /* Create the buffer object to contain the data.  For now, use a
+    * single buffer type to describe all cached state atoms.  Later,
+    * may want to take advantage of hardware distinctions between
+    * these various entities.
+    */
+   ret = cache->sws->bo_alloc(cache->sws,
+                              cache->buffer_type,
+                              data_size, 1 << 6, 
+                              bo_out);
+   if (ret)
+      return ret;
+
+
+   /* Set up the memory containing the key, aux_data, and relocs */
+   tmp = MALLOC(key_size + aux_size + relocs_size);
+
+   memcpy(tmp, key, key_size);
+   memcpy((char *)tmp + key_size, aux, cache->aux_size[cache_id]);
+   memcpy((char *)tmp + key_size + aux_size, relocs, relocs_size);
+   for (i = 0; i < nr_relocs; i++) {
+      p_atomic_inc(&relocs[i].bo->reference.count);
+   }
+
+   item->cache_id = cache_id;
+   item->key = tmp;
+   item->hash = hash;
+   item->key_size = key_size;
+   item->relocs = (struct brw_winsys_reloc *)((char *)tmp + key_size + aux_size);
+   item->nr_relocs = nr_relocs;
+   bo_reference( &item->bo, *bo_out );
+   item->data_size = data_size;
+
+   if (cache->n_items > cache->size * 1.5)
+      rehash(cache);
+
+   hash %= cache->size;
+   item->next = cache->items[hash];
+   cache->items[hash] = item;
+   cache->n_items++;
+
+   if (aux_return) {
+      assert(cache->aux_size[cache_id]);
+      *(void **)aux_return = (void *)((char *)item->key + item->key_size);
+   }
+
+   if (BRW_DEBUG & DEBUG_STATE)
+      debug_printf("upload %s: %d bytes to cache id %d\n",
+                  cache->name[cache_id],
+                  data_size, cache_id);
+
+   /* Copy data to the buffer */
+   ret = cache->sws->bo_subdata(item->bo, 
+                                cache_id,
+                                0, data_size, data,
+                                relocs, nr_relocs);
+   if (ret)
+      return ret;
+
+   update_cache_last(cache, cache_id, item->bo);
+
+   return PIPE_OK;
+}
+
+
+/**
+ * This doesn't really work with aux data.  Use search/upload instead
+ */
+enum pipe_error
+brw_cache_data_sz(struct brw_cache *cache,
+                 enum brw_cache_id cache_id,
+                 const void *data,
+                 GLuint data_size,
+                 struct brw_winsys_reloc *relocs,
+                 GLuint nr_relocs,
+                  struct brw_winsys_buffer **bo_out)
+{
+   struct brw_cache_item *item;
+   GLuint hash = hash_key(data, data_size, relocs, nr_relocs);
+
+   item = search_cache(cache, cache_id, hash, data, data_size,
+                      relocs, nr_relocs);
+   if (item) {
+      update_cache_last(cache, cache_id, item->bo);
+
+      bo_reference(bo_out, item->bo);
+      return PIPE_OK;
+   }
+
+   return brw_upload_cache(cache, cache_id,
+                           data, data_size,
+                           relocs, nr_relocs,
+                           data, data_size,
+                           NULL, NULL,
+                           bo_out);
+}
+
+
+/**
+ * Wrapper around brw_cache_data_sz using the cache_id's canonical key size.
+ *
+ * If nr_relocs is nonzero, brw_search_cache()/brw_upload_cache() would be
+ * better to use, as the potentially changing offsets in the data-used-as-key
+ * will result in excessive cache misses.
+ * 
+ * XXX: above is no longer true -- can we remove some code?
+ */
+enum pipe_error
+brw_cache_data(struct brw_cache *cache,
+              enum brw_cache_id cache_id,
+              const void *data,
+              struct brw_winsys_reloc *relocs,
+              GLuint nr_relocs,
+               struct brw_winsys_buffer **bo_out)
+{
+   return brw_cache_data_sz(cache, cache_id, data, cache->key_size[cache_id],
+                           relocs, nr_relocs, bo_out);
+}
+
+
+static void
+brw_init_cache_id(struct brw_cache *cache,
+                  const char *name,
+                  enum brw_cache_id id,
+                  GLuint key_size,
+                  GLuint aux_size)
+{
+   cache->name[id] = strdup(name);
+   cache->key_size[id] = key_size;
+   cache->aux_size[id] = aux_size;
+}
+
+
+static void
+brw_init_general_state_cache(struct brw_context *brw)
+{
+   struct brw_cache *cache = &brw->cache;
+
+   cache->brw = brw;
+   cache->sws = brw->sws;
+
+   cache->buffer_type = BRW_BUFFER_TYPE_GENERAL_STATE;
+
+   cache->size = 7;
+   cache->n_items = 0;
+   cache->items = (struct brw_cache_item **)
+      CALLOC(cache->size, sizeof(struct brw_cache_item));
+
+   brw_init_cache_id(cache,
+                    "CC_VP",
+                    BRW_CC_VP,
+                    sizeof(struct brw_cc_viewport),
+                    0);
+
+   brw_init_cache_id(cache,
+                    "CC_UNIT",
+                    BRW_CC_UNIT,
+                    sizeof(struct brw_cc_unit_state),
+                    0);
+
+   brw_init_cache_id(cache,
+                    "WM_PROG",
+                    BRW_WM_PROG,
+                    sizeof(struct brw_wm_prog_key),
+                    sizeof(struct brw_wm_prog_data));
+
+   brw_init_cache_id(cache,
+                    "SAMPLER_DEFAULT_COLOR",
+                    BRW_SAMPLER_DEFAULT_COLOR,
+                    sizeof(struct brw_sampler_default_color),
+                    0);
+
+   brw_init_cache_id(cache,
+                    "SAMPLER",
+                    BRW_SAMPLER,
+                    0,         /* variable key/data size */
+                    0);
+
+   brw_init_cache_id(cache,
+                    "WM_UNIT",
+                    BRW_WM_UNIT,
+                    sizeof(struct brw_wm_unit_state),
+                    0);
+
+   brw_init_cache_id(cache,
+                    "SF_PROG",
+                    BRW_SF_PROG,
+                    sizeof(struct brw_sf_prog_key),
+                    sizeof(struct brw_sf_prog_data));
+
+   brw_init_cache_id(cache,
+                    "SF_VP",
+                    BRW_SF_VP,
+                    sizeof(struct brw_sf_viewport),
+                    0);
+
+   brw_init_cache_id(cache,
+                    "SF_UNIT",
+                    BRW_SF_UNIT,
+                    sizeof(struct brw_sf_unit_state),
+                    0);
+
+   brw_init_cache_id(cache,
+                    "VS_UNIT",
+                    BRW_VS_UNIT,
+                    sizeof(struct brw_vs_unit_state),
+                    0);
+
+   brw_init_cache_id(cache,
+                    "VS_PROG",
+                    BRW_VS_PROG,
+                    sizeof(struct brw_vs_prog_key),
+                    sizeof(struct brw_vs_prog_data));
+
+   brw_init_cache_id(cache,
+                    "CLIP_UNIT",
+                    BRW_CLIP_UNIT,
+                    sizeof(struct brw_clip_unit_state),
+                    0);
+
+   brw_init_cache_id(cache,
+                    "CLIP_PROG",
+                    BRW_CLIP_PROG,
+                    sizeof(struct brw_clip_prog_key),
+                    sizeof(struct brw_clip_prog_data));
+
+   brw_init_cache_id(cache,
+                    "GS_UNIT",
+                    BRW_GS_UNIT,
+                    sizeof(struct brw_gs_unit_state),
+                    0);
+
+   brw_init_cache_id(cache,
+                    "GS_PROG",
+                    BRW_GS_PROG,
+                    sizeof(struct brw_gs_prog_key),
+                    sizeof(struct brw_gs_prog_data));
+}
+
+
+static void
+brw_init_surface_state_cache(struct brw_context *brw)
+{
+   struct brw_cache *cache = &brw->surface_cache;
+
+   cache->brw = brw;
+   cache->sws = brw->sws;
+
+   cache->buffer_type = BRW_BUFFER_TYPE_SURFACE_STATE;
+
+   cache->size = 7;
+   cache->n_items = 0;
+   cache->items = (struct brw_cache_item **)
+      CALLOC(cache->size, sizeof(struct brw_cache_item));
+
+   brw_init_cache_id(cache,
+                    "SS_SURFACE",
+                    BRW_SS_SURFACE,
+                    sizeof(struct brw_surface_state),
+                    0);
+
+   brw_init_cache_id(cache,
+                    "SS_SURF_BIND",
+                    BRW_SS_SURF_BIND,
+                    0,
+                    0);
+}
+
+
+void
+brw_init_caches(struct brw_context *brw)
+{
+   brw_init_general_state_cache(brw);
+   brw_init_surface_state_cache(brw);
+}
+
+
+static void
+brw_clear_cache(struct brw_context *brw, struct brw_cache *cache)
+{
+   struct brw_cache_item *c, *next;
+   GLuint i;
+
+   if (BRW_DEBUG & DEBUG_STATE)
+      debug_printf("%s\n", __FUNCTION__);
+
+   for (i = 0; i < cache->size; i++) {
+      for (c = cache->items[i]; c; c = next) {
+        int j;
+
+        next = c->next;
+
+        for (j = 0; j < c->nr_relocs; j++)
+           bo_reference(&c->relocs[j].bo, NULL);
+
+        bo_reference(&c->bo, NULL);
+        FREE((void *)c->key);
+        FREE(c);
+      }
+      cache->items[i] = NULL;
+   }
+
+   cache->n_items = 0;
+
+   if (brw->curbe.last_buf) {
+      FREE(brw->curbe.last_buf);
+      brw->curbe.last_buf = NULL;
+   }
+
+   brw->state.dirty.mesa |= ~0;
+   brw->state.dirty.brw |= ~0;
+   brw->state.dirty.cache |= ~0;
+}
+
+/* Clear all entries from the cache that point to the given bo.
+ *
+ * This lets us release memory for reuse earlier for known-dead buffers,
+ * at the cost of walking the entire hash table.
+ */
+void
+brw_state_cache_bo_delete(struct brw_cache *cache, struct brw_winsys_buffer *bo)
+{
+   struct brw_cache_item **prev;
+   GLuint i;
+
+   if (BRW_DEBUG & DEBUG_STATE)
+      debug_printf("%s\n", __FUNCTION__);
+
+   for (i = 0; i < cache->size; i++) {
+      for (prev = &cache->items[i]; *prev;) {
+        struct brw_cache_item *c = *prev;
+
+        if (cache->sws->bo_references(c->bo, bo)) {
+           int j;
+
+           *prev = c->next;
+
+           for (j = 0; j < c->nr_relocs; j++)
+              bo_reference(&c->relocs[j].bo, NULL);
+
+           bo_reference(&c->bo, NULL);
+
+           FREE((void *)c->key);
+           FREE(c);
+           cache->n_items--;
+        } else {
+           prev = &c->next;
+        }
+      }
+   }
+}
+
+void
+brw_state_cache_check_size(struct brw_context *brw)
+{
+   if (BRW_DEBUG & DEBUG_STATE)
+      debug_printf("%s (n_items=%d)\n", __FUNCTION__, brw->cache.n_items);
+
+   /* un-tuned guess.  We've got around 20 state objects for a total of around
+    * 32k, so 1000 of them is around 1.5MB.
+    */
+   if (brw->cache.n_items > 1000)
+      brw_clear_cache(brw, &brw->cache);
+
+   if (brw->surface_cache.n_items > 1000)
+      brw_clear_cache(brw, &brw->surface_cache);
+}
+
+
+static void
+brw_destroy_cache(struct brw_context *brw, struct brw_cache *cache)
+{
+   GLuint i;
+
+   if (BRW_DEBUG & DEBUG_STATE)
+      debug_printf("%s\n", __FUNCTION__);
+
+   brw_clear_cache(brw, cache);
+   for (i = 0; i < BRW_MAX_CACHE; i++) {
+      bo_reference(&cache->last_bo[i], NULL);
+      FREE(cache->name[i]);
+   }
+   FREE(cache->items);
+   cache->items = NULL;
+   cache->size = 0;
+}
+
+
+void
+brw_destroy_caches(struct brw_context *brw)
+{
+   brw_destroy_cache(brw, &brw->cache);
+   brw_destroy_cache(brw, &brw->surface_cache);
+}
diff --git a/src/gallium/drivers/i965/brw_state_debug.c b/src/gallium/drivers/i965/brw_state_debug.c
new file mode 100644 (file)
index 0000000..049c278
--- /dev/null
@@ -0,0 +1,153 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+      
+
+
+#include "brw_context.h"
+#include "brw_state.h"
+
+
+struct dirty_bit_map {
+   uint32_t bit;
+   char *name;
+   uint32_t count;
+};
+
+#define DEFINE_BIT(name) {name, #name, 0}
+
+static struct dirty_bit_map mesa_bits[] = {
+   DEFINE_BIT(PIPE_NEW_DEPTH_STENCIL_ALPHA),
+   DEFINE_BIT(PIPE_NEW_RAST),
+   DEFINE_BIT(PIPE_NEW_BLEND),
+   DEFINE_BIT(PIPE_NEW_VIEWPORT),
+   DEFINE_BIT(PIPE_NEW_SAMPLERS),
+   DEFINE_BIT(PIPE_NEW_VERTEX_BUFFER),
+   DEFINE_BIT(PIPE_NEW_VERTEX_ELEMENT),
+   DEFINE_BIT(PIPE_NEW_FRAGMENT_SHADER),
+   DEFINE_BIT(PIPE_NEW_VERTEX_SHADER),
+   DEFINE_BIT(PIPE_NEW_FRAGMENT_CONSTANTS),
+   DEFINE_BIT(PIPE_NEW_VERTEX_CONSTANTS),
+   DEFINE_BIT(PIPE_NEW_CLIP),
+   DEFINE_BIT(PIPE_NEW_INDEX_BUFFER),
+   DEFINE_BIT(PIPE_NEW_INDEX_RANGE),
+   DEFINE_BIT(PIPE_NEW_BLEND_COLOR),
+   DEFINE_BIT(PIPE_NEW_POLYGON_STIPPLE),
+   DEFINE_BIT(PIPE_NEW_FRAMEBUFFER_DIMENSIONS),
+   DEFINE_BIT(PIPE_NEW_DEPTH_BUFFER),
+   DEFINE_BIT(PIPE_NEW_COLOR_BUFFERS),
+   DEFINE_BIT(PIPE_NEW_QUERY),
+   DEFINE_BIT(PIPE_NEW_SCISSOR),
+   DEFINE_BIT(PIPE_NEW_BOUND_TEXTURES),
+   DEFINE_BIT(PIPE_NEW_NR_CBUFS),
+   {0, 0, 0}
+};
+
+static struct dirty_bit_map brw_bits[] = {
+   DEFINE_BIT(BRW_NEW_URB_FENCE),
+   DEFINE_BIT(BRW_NEW_FRAGMENT_PROGRAM),
+   DEFINE_BIT(BRW_NEW_VERTEX_PROGRAM),
+   DEFINE_BIT(BRW_NEW_INPUT_DIMENSIONS),
+   DEFINE_BIT(BRW_NEW_CURBE_OFFSETS),
+   DEFINE_BIT(BRW_NEW_REDUCED_PRIMITIVE),
+   DEFINE_BIT(BRW_NEW_PRIMITIVE),
+   DEFINE_BIT(BRW_NEW_CONTEXT),
+   DEFINE_BIT(BRW_NEW_WM_INPUT_DIMENSIONS),
+   DEFINE_BIT(BRW_NEW_PSP),
+   DEFINE_BIT(BRW_NEW_WM_SURFACES),
+   DEFINE_BIT(BRW_NEW_xxx),
+   DEFINE_BIT(BRW_NEW_INDICES),
+   {0, 0, 0}
+};
+
+static struct dirty_bit_map cache_bits[] = {
+   DEFINE_BIT(CACHE_NEW_CC_VP),
+   DEFINE_BIT(CACHE_NEW_CC_UNIT),
+   DEFINE_BIT(CACHE_NEW_WM_PROG),
+   DEFINE_BIT(CACHE_NEW_SAMPLER_DEFAULT_COLOR),
+   DEFINE_BIT(CACHE_NEW_SAMPLER),
+   DEFINE_BIT(CACHE_NEW_WM_UNIT),
+   DEFINE_BIT(CACHE_NEW_SF_PROG),
+   DEFINE_BIT(CACHE_NEW_SF_VP),
+   DEFINE_BIT(CACHE_NEW_SF_UNIT),
+   DEFINE_BIT(CACHE_NEW_VS_UNIT),
+   DEFINE_BIT(CACHE_NEW_VS_PROG),
+   DEFINE_BIT(CACHE_NEW_GS_UNIT),
+   DEFINE_BIT(CACHE_NEW_GS_PROG),
+   DEFINE_BIT(CACHE_NEW_CLIP_VP),
+   DEFINE_BIT(CACHE_NEW_CLIP_UNIT),
+   DEFINE_BIT(CACHE_NEW_CLIP_PROG),
+   DEFINE_BIT(CACHE_NEW_SURFACE),
+   DEFINE_BIT(CACHE_NEW_SURF_BIND),
+   {0, 0, 0}
+};
+
+
+static void
+brw_update_dirty_count(struct dirty_bit_map *bit_map, int32_t bits)
+{
+   int i;
+
+   for (i = 0; i < 32; i++) {
+      if (bit_map[i].bit == 0)
+        return;
+
+      if (bit_map[i].bit & bits)
+        bit_map[i].count++;
+   }
+}
+
+static void
+brw_print_dirty_count(struct dirty_bit_map *bit_map, int32_t bits)
+{
+   int i;
+
+   for (i = 0; i < 32; i++) {
+      if (bit_map[i].bit == 0)
+        return;
+
+      debug_printf("0x%08x: %12d (%s)\n",
+             bit_map[i].bit, bit_map[i].count, bit_map[i].name);
+   }
+}
+
+void
+brw_update_dirty_counts( unsigned mesa,
+                        unsigned brw,
+                        unsigned cache )
+{
+   static int dirty_count = 0;
+
+   brw_update_dirty_count(mesa_bits, mesa);
+   brw_update_dirty_count(brw_bits, brw);
+   brw_update_dirty_count(cache_bits, cache);
+      if (dirty_count++ % 1000 == 0) {
+        brw_print_dirty_count(mesa_bits, mesa);
+        brw_print_dirty_count(brw_bits, brw);
+        brw_print_dirty_count(cache_bits, cache);
+        debug_printf("\n");
+      }
+}
diff --git a/src/gallium/drivers/i965/brw_state_upload.c b/src/gallium/drivers/i965/brw_state_upload.c
new file mode 100644 (file)
index 0000000..f8b91ef
--- /dev/null
@@ -0,0 +1,270 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+       
+
+
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_batchbuffer.h"
+#include "brw_debug.h"
+
+const struct brw_tracked_state *atoms[] =
+{
+/*   &brw_wm_input_sizes, */
+   &brw_vs_prog,
+   &brw_gs_prog, 
+   &brw_clip_prog, 
+   &brw_sf_prog,
+   &brw_wm_prog,
+
+   /* Once all the programs are done, we know how large urb entry
+    * sizes need to be and can decide if we need to change the urb
+    * layout.
+    */
+   &brw_curbe_offsets,
+   &brw_recalculate_urb_fence,
+
+   &brw_cc_vp,
+   &brw_cc_unit,
+
+   &brw_vs_surfaces,           /* must do before unit */
+   /*&brw_wm_constant_surface,*/       /* must do before wm surfaces/bind bo */
+   &brw_wm_surfaces,           /* must do before samplers and unit */
+   &brw_wm_samplers,
+
+   &brw_wm_unit,
+   &brw_sf_vp,
+   &brw_sf_unit,
+   &brw_vs_unit,               /* always required, enabled or not */
+   &brw_clip_unit,
+   &brw_gs_unit,  
+
+   /* Command packets:
+    */
+   &brw_invarient_state,
+   &brw_state_base_address,
+
+   &brw_binding_table_pointers,
+   &brw_blend_constant_color,
+
+   &brw_depthbuffer,
+   &brw_polygon_stipple,
+   &brw_line_stipple,
+
+   &brw_psp_urb_cbs,
+
+   &brw_drawing_rect,
+   &brw_indices,
+   &brw_index_buffer,
+   &brw_vertices,
+
+   &brw_curbe_buffer
+};
+
+
+void brw_init_state( struct brw_context *brw )
+{
+   brw_init_caches(brw);
+}
+
+
+void brw_destroy_state( struct brw_context *brw )
+{
+   brw_destroy_caches(brw);
+   brw_destroy_batch_cache(brw);
+}
+
+/***********************************************************************
+ */
+
+static GLboolean check_state( const struct brw_state_flags *a,
+                             const struct brw_state_flags *b )
+{
+   return ((a->mesa & b->mesa) ||
+          (a->brw & b->brw) ||
+          (a->cache & b->cache));
+}
+
+static void accumulate_state( struct brw_state_flags *a,
+                             const struct brw_state_flags *b )
+{
+   a->mesa |= b->mesa;
+   a->brw |= b->brw;
+   a->cache |= b->cache;
+}
+
+
+static void xor_states( struct brw_state_flags *result,
+                            const struct brw_state_flags *a,
+                             const struct brw_state_flags *b )
+{
+   result->mesa = a->mesa ^ b->mesa;
+   result->brw = a->brw ^ b->brw;
+   result->cache = a->cache ^ b->cache;
+}
+
+static void
+brw_clear_validated_bos(struct brw_context *brw)
+{
+   int i;
+
+   /* Clear the last round of validated bos */
+   for (i = 0; i < brw->state.validated_bo_count; i++) {
+      bo_reference(&brw->state.validated_bos[i], NULL);
+   }
+   brw->state.validated_bo_count = 0;
+}
+
+
+/***********************************************************************
+ * Emit all state:
+ */
+enum pipe_error brw_validate_state( struct brw_context *brw )
+{
+   struct brw_state_flags *state = &brw->state.dirty;
+   GLuint i;
+   int ret;
+
+   brw_clear_validated_bos(brw);
+   brw_add_validated_bo(brw, brw->batch->buf);
+
+   if (brw->flags.always_emit_state) {
+      state->mesa |= ~0;
+      state->brw |= ~0;
+      state->cache |= ~0;
+   }
+
+   if (state->mesa == 0 &&
+       state->cache == 0 &&
+       state->brw == 0)
+      return 0;
+
+   if (brw->state.dirty.brw & BRW_NEW_CONTEXT)
+      brw_clear_batch_cache(brw);
+
+   /* do prepare stage for all atoms */
+   for (i = 0; i < Elements(atoms); i++) {
+      const struct brw_tracked_state *atom = atoms[i];
+
+      if (check_state(state, &atom->dirty)) {
+         if (atom->prepare) {
+            ret = atom->prepare(brw);
+           if (ret)
+              return ret;
+        }
+      }
+   }
+
+   /* Make sure that the textures which are referenced by the current
+    * brw fragment program are actually present/valid.
+    * If this fails, we can experience GPU lock-ups.
+    */
+   {
+      const struct brw_fragment_shader *fp = brw->curr.fragment_shader;
+      if (fp) {
+         assert(fp->info.file_max[TGSI_FILE_SAMPLER] < (int)brw->curr.num_samplers);
+        /*assert(fp->info.texture_max <= brw->curr.num_textures);*/
+      }
+   }
+
+   return 0;
+}
+
+
+enum pipe_error brw_upload_state(struct brw_context *brw)
+{
+   struct brw_state_flags *state = &brw->state.dirty;
+   int ret;
+   int i;
+
+   brw_clear_validated_bos(brw);
+
+   if (BRW_DEBUG) {
+      /* Debug version which enforces various sanity checks on the
+       * state flags which are generated and checked to help ensure
+       * state atoms are ordered correctly in the list.
+       */
+      struct brw_state_flags examined, prev;      
+      memset(&examined, 0, sizeof(examined));
+      prev = *state;
+
+      for (i = 0; i < Elements(atoms); i++) {
+        const struct brw_tracked_state *atom = atoms[i];
+        struct brw_state_flags generated;
+
+        assert(atom->dirty.mesa ||
+               atom->dirty.brw ||
+               atom->dirty.cache);
+
+        if (check_state(state, &atom->dirty)) {
+           if (atom->emit) {
+              ret = atom->emit( brw );
+              if (ret)
+                 return ret;
+           }
+        }
+
+        accumulate_state(&examined, &atom->dirty);
+
+        /* generated = (prev ^ state)
+         * if (examined & generated)
+         *     fail;
+         */
+        xor_states(&generated, &prev, state);
+        assert(!check_state(&examined, &generated));
+        prev = *state;
+      }
+   }
+   else {
+      for (i = 0; i < Elements(atoms); i++) {   
+        const struct brw_tracked_state *atom = atoms[i];
+
+        if (check_state(state, &atom->dirty)) {
+           if (atom->emit) {
+              ret = atom->emit( brw );
+              if (ret)
+                 return ret;
+           }
+        }
+      }
+   }
+
+   if (BRW_DEBUG & DEBUG_STATE) {
+      brw_update_dirty_counts( state->mesa, 
+                              state->brw,
+                              state->cache );
+   }
+   
+   /* Clear dirty flags:
+    */
+   memset(state, 0, sizeof(*state));
+   return 0;
+}
diff --git a/src/gallium/drivers/i965/brw_structs.h b/src/gallium/drivers/i965/brw_structs.h
new file mode 100644 (file)
index 0000000..bf10bc0
--- /dev/null
@@ -0,0 +1,1576 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+        
+
+#ifndef BRW_STRUCTS_H
+#define BRW_STRUCTS_H
+
+#include "brw_types.h"
+
+/** Number of general purpose registers (VS, WM, etc) */
+#define BRW_MAX_GRF 128
+
+/** Number of message register file registers */
+#define BRW_MAX_MRF 16
+
+
+/* Command packets:
+ */
+struct header 
+{
+   GLuint length:16; 
+   GLuint opcode:16; 
+};
+
+
+union header_union
+{
+   struct header bits;
+   GLuint dword;
+};
+
+struct brw_3d_control
+{   
+   struct 
+   {
+      GLuint length:8;
+      GLuint notify_enable:1;
+      GLuint pad:3;
+      GLuint wc_flush_enable:1; 
+      GLuint depth_stall_enable:1; 
+      GLuint operation:2; 
+      GLuint opcode:16; 
+   } header;
+   
+   struct
+   {
+      GLuint pad:2;
+      GLuint dest_addr_type:1; 
+      GLuint dest_addr:29; 
+   } dest;
+   
+   GLuint dword2;   
+   GLuint dword3;   
+};
+
+
+struct brw_3d_primitive
+{
+   struct
+   {
+      GLuint length:8; 
+      GLuint pad:2;
+      GLuint topology:5; 
+      GLuint indexed:1; 
+      GLuint opcode:16; 
+   } header;
+
+   GLuint verts_per_instance;  
+   GLuint start_vert_location;  
+   GLuint instance_count;  
+   GLuint start_instance_location;  
+   GLuint base_vert_location;  
+};
+
+/* These seem to be passed around as function args, so it works out
+ * better to keep them as #defines:
+ */
+#define BRW_FLUSH_READ_CACHE           0x1
+#define BRW_FLUSH_STATE_CACHE          0x2
+#define BRW_INHIBIT_FLUSH_RENDER_CACHE 0x4
+#define BRW_FLUSH_SNAPSHOT_COUNTERS    0x8
+
+struct brw_mi_flush
+{
+   GLuint flags:4;
+   GLuint pad:12;
+   GLuint opcode:16;
+};
+
+struct brw_vf_statistics
+{
+   GLuint statistics_enable:1;
+   GLuint pad:15;
+   GLuint opcode:16;
+};
+
+
+
+struct brw_binding_table_pointers
+{
+   struct header header;
+   GLuint vs; 
+   GLuint gs; 
+   GLuint clp; 
+   GLuint sf; 
+   GLuint wm; 
+};
+
+
+struct brw_blend_constant_color
+{
+   struct header header;
+   GLfloat blend_constant_color[4];  
+};
+
+
+struct brw_depthbuffer
+{
+   union header_union header;
+   
+   union {
+      struct {
+        GLuint pitch:18; 
+        GLuint format:3; 
+        GLuint pad:2;
+        GLuint software_tiled_rendering_mode:2;
+        GLuint depth_offset_disable:1; 
+        GLuint tile_walk:1; 
+        GLuint tiled_surface:1; 
+        GLuint pad2:1;
+        GLuint surface_type:3; 
+      } bits;
+      GLuint dword;
+   } dword1;
+   
+   GLuint dword2_base_addr; 
+   union {
+      struct {
+        GLuint pad:1;
+        GLuint mipmap_layout:1; 
+        GLuint lod:4; 
+        GLuint width:13; 
+        GLuint height:13; 
+      } bits;
+      GLuint dword;
+   } dword3;
+
+   union {
+      struct {
+        GLuint pad:10;
+        GLuint min_array_element:11; 
+        GLuint depth:11; 
+      } bits;
+      GLuint dword;
+   } dword4;
+};
+
+struct brw_depthbuffer_g4x
+{
+   union header_union header;
+   
+   union {
+      struct {
+        GLuint pitch:18; 
+        GLuint format:3; 
+        GLuint pad:2;
+        GLuint software_tiled_rendering_mode:2;
+        GLuint depth_offset_disable:1; 
+        GLuint tile_walk:1; 
+        GLuint tiled_surface:1; 
+        GLuint pad2:1;
+        GLuint surface_type:3; 
+      } bits;
+      GLuint dword;
+   } dword1;
+   
+   GLuint dword2_base_addr; 
+   union {
+      struct {
+        GLuint pad:1;
+        GLuint mipmap_layout:1; 
+        GLuint lod:4; 
+        GLuint width:13; 
+        GLuint height:13; 
+      } bits;
+      GLuint dword;
+   } dword3;
+
+   union {
+      struct {
+        GLuint pad:10;
+        GLuint min_array_element:11; 
+        GLuint depth:11; 
+      } bits;
+      GLuint dword;
+   } dword4;
+
+   union {
+      struct {
+         GLuint xoffset:16;
+         GLuint yoffset:16;
+      } bits;
+      GLuint dword;
+   } dword5;   /* NEW in Integrated Graphics Device */
+};
+
+struct brw_drawrect
+{
+   struct header header;
+   GLuint xmin:16; 
+   GLuint ymin:16; 
+   GLuint xmax:16; 
+   GLuint ymax:16; 
+   GLuint xorg:16;  
+   GLuint yorg:16;  
+};
+
+
+
+
+struct brw_global_depth_offset_clamp
+{
+   struct header header;
+   GLfloat depth_offset_clamp;  
+};
+
+struct brw_indexbuffer
+{   
+   union {
+      struct
+      {
+        GLuint length:8; 
+        GLuint index_format:2; 
+        GLuint cut_index_enable:1; 
+        GLuint pad:5; 
+        GLuint opcode:16; 
+      } bits;
+      GLuint dword;
+
+   } header;
+
+   GLuint buffer_start; 
+   GLuint buffer_end; 
+};
+
+/* NEW in Integrated Graphics Device */
+struct brw_aa_line_parameters
+{
+   struct header header;
+
+   struct {
+      GLuint aa_coverage_scope:8;
+      GLuint pad0:8;
+      GLuint aa_coverage_bias:8;
+      GLuint pad1:8;
+   } bits0;
+
+   struct {
+      GLuint aa_coverage_endcap_slope:8;
+      GLuint pad0:8;
+      GLuint aa_coverage_endcap_bias:8;
+      GLuint pad1:8;
+   } bits1;
+};
+
+struct brw_line_stipple
+{   
+   struct header header;
+  
+   struct
+   {
+      GLuint pattern:16; 
+      GLuint pad:16;
+   } bits0;
+   
+   struct
+   {
+      GLuint repeat_count:9; 
+      GLuint pad:7;
+      GLuint inverse_repeat_count:16; 
+   } bits1;
+};
+
+
+struct brw_pipelined_state_pointers
+{
+   struct header header;
+   
+   struct {
+      GLuint pad:5;
+      GLuint offset:27; /* Offset from GENERAL_STATE_BASE */
+   } vs;
+   
+   struct
+   {
+      GLuint enable:1;
+      GLuint pad:4;
+      GLuint offset:27; /* Offset from GENERAL_STATE_BASE */
+   } gs;
+   
+   struct
+   {
+      GLuint enable:1;
+      GLuint pad:4;
+      GLuint offset:27; /* Offset from GENERAL_STATE_BASE */
+   } clp;
+   
+   struct
+   {
+      GLuint pad:5;
+      GLuint offset:27; /* Offset from GENERAL_STATE_BASE */
+   } sf;
+
+   struct
+   {
+      GLuint pad:5;
+      GLuint offset:27; /* Offset from GENERAL_STATE_BASE */
+   } wm;
+   
+   struct
+   {
+      GLuint pad:5;
+      GLuint offset:27; /* Offset from GENERAL_STATE_BASE. KW: check me! */
+   } cc;
+};
+
+
+struct brw_polygon_stipple_offset
+{
+   struct header header;
+
+   struct {
+      GLuint y_offset:5; 
+      GLuint pad:3;
+      GLuint x_offset:5; 
+      GLuint pad0:19;
+   } bits0;
+};
+
+
+
+struct brw_polygon_stipple
+{
+   struct header header;
+   GLuint stipple[32];
+};
+
+
+
+struct brw_pipeline_select
+{
+   struct
+   {
+      GLuint pipeline_select:1;   
+      GLuint pad:15;
+      GLuint opcode:16;   
+   } header;
+};
+
+
+struct brw_pipe_control
+{
+   struct
+   {
+      GLuint length:8;
+      GLuint notify_enable:1;
+      GLuint texture_cache_flush_enable:1;
+      GLuint indirect_state_pointers_disable:1;
+      GLuint instruction_state_cache_flush_enable:1;
+      GLuint write_cache_flush_enable:1;
+      GLuint depth_stall_enable:1;
+      GLuint post_sync_operation:2;
+
+      GLuint opcode:16;
+   } header;
+
+   struct
+   {
+      GLuint pad:2;
+      GLuint dest_addr_type:1;
+      GLuint dest_addr:29;
+   } bits1;
+
+   GLuint data0;
+   GLuint data1;
+};
+
+
+struct brw_urb_fence
+{
+   struct
+   {
+      GLuint length:8;   
+      GLuint vs_realloc:1;   
+      GLuint gs_realloc:1;   
+      GLuint clp_realloc:1;   
+      GLuint sf_realloc:1;   
+      GLuint vfe_realloc:1;   
+      GLuint cs_realloc:1;   
+      GLuint pad:2;
+      GLuint opcode:16;   
+   } header;
+
+   struct
+   {
+      GLuint vs_fence:10;  
+      GLuint gs_fence:10;  
+      GLuint clp_fence:10;  
+      GLuint pad:2;
+   } bits0;
+
+   struct
+   {
+      GLuint sf_fence:10;  
+      GLuint vf_fence:10;  
+      GLuint cs_fence:11;  
+      GLuint pad:1;
+   } bits1;
+};
+
+struct brw_cs_urb_state
+{
+   struct header header;
+
+   struct
+   {
+      GLuint nr_urb_entries:3;   
+      GLuint pad:1;
+      GLuint urb_entry_size:5;   
+      GLuint pad0:23;
+   } bits0;
+};
+
+struct brw_constant_buffer
+{
+   struct
+   {
+      GLuint length:8;   
+      GLuint valid:1;   
+      GLuint pad:7;
+      GLuint opcode:16;   
+   } header;
+
+   struct
+   {
+      GLuint buffer_length:6;   
+      GLuint buffer_address:26;  
+   } bits0;
+};
+
+struct brw_state_base_address
+{
+   struct header header;
+
+   struct
+   {
+      GLuint modify_enable:1;
+      GLuint pad:4;
+      GLuint general_state_address:27;  
+   } bits0;
+
+   struct
+   {
+      GLuint modify_enable:1;
+      GLuint pad:4;
+      GLuint surface_state_address:27;  
+   } bits1;
+
+   struct
+   {
+      GLuint modify_enable:1;
+      GLuint pad:4;
+      GLuint indirect_object_state_address:27;  
+   } bits2;
+
+   struct
+   {
+      GLuint modify_enable:1;
+      GLuint pad:11;
+      GLuint general_state_upper_bound:20;  
+   } bits3;
+
+   struct
+   {
+      GLuint modify_enable:1;
+      GLuint pad:11;
+      GLuint indirect_object_state_upper_bound:20;  
+   } bits4;
+};
+
+struct brw_state_prefetch
+{
+   struct header header;
+
+   struct
+   {
+      GLuint prefetch_count:3;   
+      GLuint pad:3;
+      GLuint prefetch_pointer:26;  
+   } bits0;
+};
+
+struct brw_system_instruction_pointer
+{
+   struct header header;
+
+   struct
+   {
+      GLuint pad:4;
+      GLuint system_instruction_pointer:28;  
+   } bits0;
+};
+
+
+
+
+/* State structs for the various fixed function units:
+ */
+
+
+struct thread0
+{
+   GLuint pad0:1;
+   GLuint grf_reg_count:3; 
+   GLuint pad1:2;
+   GLuint kernel_start_pointer:26; /* Offset from GENERAL_STATE_BASE */
+};
+
+struct thread1
+{
+   GLuint ext_halt_exception_enable:1; 
+   GLuint sw_exception_enable:1; 
+   GLuint mask_stack_exception_enable:1; 
+   GLuint timeout_exception_enable:1; 
+   GLuint illegal_op_exception_enable:1; 
+   GLuint pad0:3;
+   GLuint depth_coef_urb_read_offset:6;        /* WM only */
+   GLuint pad1:2;
+   GLuint floating_point_mode:1; 
+   GLuint thread_priority:1; 
+   GLuint binding_table_entry_count:8; 
+   GLuint pad3:5;
+   GLuint single_program_flow:1; 
+};
+
+struct thread2
+{
+   GLuint per_thread_scratch_space:4; 
+   GLuint pad0:6;
+   GLuint scratch_space_base_pointer:22; 
+};
+
+   
+struct thread3
+{
+   GLuint dispatch_grf_start_reg:4; 
+   GLuint urb_entry_read_offset:6; 
+   GLuint pad0:1;
+   GLuint urb_entry_read_length:6; 
+   GLuint pad1:1;
+   GLuint const_urb_entry_read_offset:6; 
+   GLuint pad2:1;
+   GLuint const_urb_entry_read_length:6; 
+   GLuint pad3:1;
+};
+
+
+
+struct brw_clip_unit_state
+{
+   struct thread0 thread0;
+   struct
+   {
+      GLuint pad0:7;
+      GLuint sw_exception_enable:1;
+      GLuint pad1:3;
+      GLuint mask_stack_exception_enable:1;
+      GLuint pad2:1;
+      GLuint illegal_op_exception_enable:1;
+      GLuint pad3:2;
+      GLuint floating_point_mode:1;
+      GLuint thread_priority:1;
+      GLuint binding_table_entry_count:8;
+      GLuint pad4:5;
+      GLuint single_program_flow:1;
+   } thread1;
+
+   struct thread2 thread2;
+   struct thread3 thread3;
+
+   struct
+   {
+      GLuint pad0:9;
+      GLuint gs_output_stats:1; /* not always */
+      GLuint stats_enable:1; 
+      GLuint nr_urb_entries:7; 
+      GLuint pad1:1;
+      GLuint urb_entry_allocation_size:5; 
+      GLuint pad2:1;
+      GLuint max_threads:5;    /* may be less */
+      GLuint pad3:2;
+   } thread4;   
+      
+   struct
+   {
+      GLuint pad0:13;
+      GLuint clip_mode:3; 
+      GLuint userclip_enable_flags:8; 
+      GLuint userclip_must_clip:1; 
+      GLuint negative_w_clip_test:1;
+      GLuint guard_band_enable:1; 
+      GLuint viewport_z_clip_enable:1; 
+      GLuint viewport_xy_clip_enable:1; 
+      GLuint vertex_position_space:1; 
+      GLuint api_mode:1; 
+      GLuint pad2:1;
+   } clip5;
+   
+   struct
+   {
+      GLuint pad0:5;
+      GLuint clipper_viewport_state_ptr:27; 
+   } clip6;
+
+   
+   GLfloat viewport_xmin;  
+   GLfloat viewport_xmax;  
+   GLfloat viewport_ymin;  
+   GLfloat viewport_ymax;  
+};
+
+
+
+struct brw_cc_unit_state
+{
+   struct brw_cc0
+   {
+      GLuint pad0:3;
+      GLuint bf_stencil_pass_depth_pass_op:3; 
+      GLuint bf_stencil_pass_depth_fail_op:3; 
+      GLuint bf_stencil_fail_op:3; 
+      GLuint bf_stencil_func:3; 
+      GLuint bf_stencil_enable:1; 
+      GLuint pad1:2;
+      GLuint stencil_write_enable:1; 
+      GLuint stencil_pass_depth_pass_op:3; 
+      GLuint stencil_pass_depth_fail_op:3; 
+      GLuint stencil_fail_op:3; 
+      GLuint stencil_func:3; 
+      GLuint stencil_enable:1; 
+   } cc0;
+
+   
+   struct brw_cc1
+   {
+      GLuint bf_stencil_ref:8; 
+      GLuint stencil_write_mask:8; 
+      GLuint stencil_test_mask:8; 
+      GLuint stencil_ref:8; 
+   } cc1;
+
+   
+   struct brw_cc2
+   {
+      GLuint logicop_enable:1; 
+      GLuint pad0:10;
+      GLuint depth_write_enable:1; 
+      GLuint depth_test_function:3; 
+      GLuint depth_test:1; 
+      GLuint bf_stencil_write_mask:8; 
+      GLuint bf_stencil_test_mask:8; 
+   } cc2;
+
+   
+   struct brw_cc3
+   {
+      GLuint pad0:8;
+      GLuint alpha_test_func:3; 
+      GLuint alpha_test:1; 
+      GLuint blend_enable:1; 
+      GLuint ia_blend_enable:1; 
+      GLuint pad1:1;
+      GLuint alpha_test_format:1;
+      GLuint pad2:16;
+   } cc3;
+   
+   struct brw_cc4
+   {
+      GLuint pad0:5; 
+      GLuint cc_viewport_state_offset:27; /* Offset from GENERAL_STATE_BASE */
+   } cc4;
+   
+   struct brw_cc5
+   {
+      GLuint pad0:2;
+      GLuint ia_dest_blend_factor:5; 
+      GLuint ia_src_blend_factor:5; 
+      GLuint ia_blend_function:3; 
+      GLuint statistics_enable:1; 
+      GLuint logicop_func:4; 
+      GLuint pad1:11;
+      GLuint dither_enable:1; 
+   } cc5;
+
+   struct brw_cc6
+   {
+      GLuint clamp_post_alpha_blend:1; 
+      GLuint clamp_pre_alpha_blend:1; 
+      GLuint clamp_range:2; 
+      GLuint pad0:11;
+      GLuint y_dither_offset:2; 
+      GLuint x_dither_offset:2; 
+      GLuint dest_blend_factor:5; 
+      GLuint src_blend_factor:5; 
+      GLuint blend_function:3; 
+   } cc6;
+
+   struct brw_cc7 {
+      union {
+        GLfloat f;  
+        GLubyte ub[4];
+      } alpha_ref;
+   } cc7;
+};
+
+
+
+struct brw_sf_unit_state
+{
+   struct thread0 thread0;
+   struct thread1 thread1;
+   struct thread2 thread2;
+   struct thread3 thread3;
+
+   struct
+   {
+      GLuint pad0:10;
+      GLuint stats_enable:1; 
+      GLuint nr_urb_entries:7; 
+      GLuint pad1:1;
+      GLuint urb_entry_allocation_size:5; 
+      GLuint pad2:1;
+      GLuint max_threads:6; 
+      GLuint pad3:1;
+   } thread4;   
+
+   struct
+   {
+      GLuint front_winding:1; 
+      GLuint viewport_transform:1; 
+      GLuint pad0:3;
+      GLuint sf_viewport_state_offset:27; /* Offset from GENERAL_STATE_BASE */
+   } sf5;
+   
+   struct
+   {
+      GLuint pad0:9;
+      GLuint dest_org_vbias:4; 
+      GLuint dest_org_hbias:4; 
+      GLuint scissor:1; 
+      GLuint disable_2x2_trifilter:1; 
+      GLuint disable_zero_pix_trifilter:1; 
+      GLuint point_rast_rule:2; 
+      GLuint line_endcap_aa_region_width:2; 
+      GLuint line_width:4; 
+      GLuint fast_scissor_disable:1; 
+      GLuint cull_mode:2; 
+      GLuint aa_enable:1; 
+   } sf6;
+
+   struct
+   {
+      GLuint point_size:11; 
+      GLuint use_point_size_state:1; 
+      GLuint subpixel_precision:1; 
+      GLuint sprite_point:1; 
+      GLuint pad0:10;
+      GLuint aa_line_distance_mode:1;
+      GLuint trifan_pv:2; 
+      GLuint linestrip_pv:2; 
+      GLuint tristrip_pv:2; 
+      GLuint line_last_pixel_enable:1; 
+   } sf7;
+
+};
+
+
+struct brw_gs_unit_state
+{
+   struct thread0 thread0;
+   struct thread1 thread1;
+   struct thread2 thread2;
+   struct thread3 thread3;
+
+   struct
+   {
+      GLuint pad0:8;
+      GLuint rendering_enable:1; /* for IGDNG */
+      GLuint pad4:1;
+      GLuint stats_enable:1; 
+      GLuint nr_urb_entries:7; 
+      GLuint pad1:1;
+      GLuint urb_entry_allocation_size:5; 
+      GLuint pad2:1;
+      GLuint max_threads:5; 
+      GLuint pad3:2;
+   } thread4;   
+      
+   struct
+   {
+      GLuint sampler_count:3; 
+      GLuint pad0:2;
+      GLuint sampler_state_pointer:27; 
+   } gs5;
+
+   
+   struct
+   {
+      GLuint max_vp_index:4; 
+      GLuint pad0:12;
+      GLuint svbi_post_inc_value:10;
+      GLuint pad1:1;
+      GLuint svbi_post_inc_enable:1;
+      GLuint svbi_payload:1;
+      GLuint discard_adjaceny:1;
+      GLuint reorder_enable:1; 
+      GLuint pad2:1;
+   } gs6;
+};
+
+
+struct brw_vs_unit_state
+{
+   struct thread0 thread0;
+   struct thread1 thread1;
+   struct thread2 thread2;
+   struct thread3 thread3;
+   
+   struct
+   {
+      GLuint pad0:10;
+      GLuint stats_enable:1; 
+      GLuint nr_urb_entries:7; 
+      GLuint pad1:1;
+      GLuint urb_entry_allocation_size:5; 
+      GLuint pad2:1;
+      GLuint max_threads:6; 
+      GLuint pad3:1;
+   } thread4;   
+
+   struct
+   {
+      GLuint sampler_count:3; 
+      GLuint pad0:2;
+      GLuint sampler_state_pointer:27; 
+   } vs5;
+
+   struct
+   {
+      GLuint vs_enable:1; 
+      GLuint vert_cache_disable:1; 
+      GLuint pad0:30;
+   } vs6;
+};
+
+
+struct brw_wm_unit_state
+{
+   struct thread0 thread0;
+   struct thread1 thread1;
+   struct thread2 thread2;
+   struct thread3 thread3;
+   
+   struct {
+      GLuint stats_enable:1; 
+      GLuint depth_buffer_clear:1;
+      GLuint sampler_count:3; 
+      GLuint sampler_state_pointer:27; 
+   } wm4;
+   
+   struct
+   {
+      GLuint enable_8_pix:1; 
+      GLuint enable_16_pix:1; 
+      GLuint enable_32_pix:1; 
+      GLuint enable_con_32_pix:1;
+      GLuint enable_con_64_pix:1;
+      GLuint pad0:5;
+      GLuint legacy_global_depth_bias:1; 
+      GLuint line_stipple:1; 
+      GLuint depth_offset:1; 
+      GLuint polygon_stipple:1; 
+      GLuint line_aa_region_width:2; 
+      GLuint line_endcap_aa_region_width:2; 
+      GLuint early_depth_test:1; 
+      GLuint thread_dispatch_enable:1; 
+      GLuint program_uses_depth:1; 
+      GLuint program_computes_depth:1; 
+      GLuint program_uses_killpixel:1; 
+      GLuint legacy_line_rast: 1; 
+      GLuint transposed_urb_read_enable:1; 
+      GLuint max_threads:7; 
+   } wm5;
+   
+   GLfloat global_depth_offset_constant;  
+   GLfloat global_depth_offset_scale;   
+   
+   /* for IGDNG only */
+   struct {
+      GLuint pad0:1;
+      GLuint grf_reg_count_1:3; 
+      GLuint pad1:2;
+      GLuint kernel_start_pointer_1:26;
+   } wm8;       
+
+   struct {
+      GLuint pad0:1;
+      GLuint grf_reg_count_2:3; 
+      GLuint pad1:2;
+      GLuint kernel_start_pointer_2:26;
+   } wm9;       
+
+   struct {
+      GLuint pad0:1;
+      GLuint grf_reg_count_3:3; 
+      GLuint pad1:2;
+      GLuint kernel_start_pointer_3:26;
+   } wm10;       
+};
+
+struct brw_sampler_default_color {
+   GLfloat color[4];
+};
+
+struct brw_sampler_state
+{
+   
+   struct brw_ss0
+   {
+      GLuint shadow_function:3; 
+      GLuint lod_bias:11; 
+      GLuint min_filter:3; 
+      GLuint mag_filter:3; 
+      GLuint mip_filter:2; 
+      GLuint base_level:5; 
+      GLuint pad:1;
+      GLuint lod_preclamp:1; 
+      GLuint default_color_mode:1; 
+      GLuint pad0:1;
+      GLuint disable:1; 
+   } ss0;
+
+   struct brw_ss1
+   {
+      GLuint r_wrap_mode:3; 
+      GLuint t_wrap_mode:3; 
+      GLuint s_wrap_mode:3; 
+      GLuint pad:3;
+      GLuint max_lod:10; 
+      GLuint min_lod:10; 
+   } ss1;
+
+   
+   struct brw_ss2
+   {
+      GLuint pad:5;
+      GLuint default_color_pointer:27; 
+   } ss2;
+   
+   struct brw_ss3
+   {
+      GLuint pad:19;
+      GLuint max_aniso:3; 
+      GLuint chroma_key_mode:1; 
+      GLuint chroma_key_index:2; 
+      GLuint chroma_key_enable:1; 
+      GLuint monochrome_filter_width:3; 
+      GLuint monochrome_filter_height:3; 
+   } ss3;
+};
+
+
+struct brw_clipper_viewport
+{
+   GLfloat xmin;  
+   GLfloat xmax;  
+   GLfloat ymin;  
+   GLfloat ymax;  
+};
+
+struct brw_cc_viewport
+{
+   GLfloat min_depth;  
+   GLfloat max_depth;  
+};
+
+struct brw_sf_viewport
+{
+   struct {
+      GLfloat m00;  
+      GLfloat m11;  
+      GLfloat m22;  
+      GLfloat m30;  
+      GLfloat m31;  
+      GLfloat m32;  
+   } viewport;
+
+   /* scissor coordinates are inclusive */
+   struct {
+      GLshort xmin;
+      GLshort ymin;
+      GLshort xmax;
+      GLshort ymax;
+   } scissor;
+};
+
+/* Documented in the subsystem/shared-functions/sampler chapter...
+ */
+struct brw_surface_state
+{
+   struct brw_surf_ss0 {
+      GLuint cube_pos_z:1; 
+      GLuint cube_neg_z:1; 
+      GLuint cube_pos_y:1; 
+      GLuint cube_neg_y:1; 
+      GLuint cube_pos_x:1; 
+      GLuint cube_neg_x:1; 
+      GLuint pad:4;
+      GLuint mipmap_layout_mode:1; 
+      GLuint vert_line_stride_ofs:1; 
+      GLuint vert_line_stride:1; 
+      GLuint color_blend:1; 
+      GLuint writedisable_blue:1; 
+      GLuint writedisable_green:1; 
+      GLuint writedisable_red:1; 
+      GLuint writedisable_alpha:1; 
+      GLuint surface_format:9;     /**< BRW_SURFACEFORMAT_x */
+      GLuint data_return_format:1; 
+      GLuint pad0:1;
+      GLuint surface_type:3;       /**< BRW_SURFACE_1D/2D/3D/CUBE */
+   } ss0;
+   
+   struct brw_surf_ss1 {
+      GLuint base_addr;  
+   } ss1;
+   
+   struct brw_surf_ss2 {
+      GLuint pad:2;
+      GLuint mip_count:4; 
+      GLuint width:13; 
+      GLuint height:13; 
+   } ss2;
+
+   struct brw_surf_ss3 {
+      GLuint tile_walk:1; 
+      GLuint tiled_surface:1; 
+      GLuint pad:1; 
+      GLuint pitch:18; 
+      GLuint depth:11; 
+   } ss3;
+   
+   struct brw_surf_ss4 {
+      GLuint multisample_position_palette_index:3;
+      GLuint pad1:1;
+      GLuint num_multisamples:3;
+      GLuint pad0:1;
+      GLuint render_target_view_extent:9;
+      GLuint min_array_elt:11;
+      GLuint min_lod:4; 
+   } ss4;
+
+   struct brw_surf_ss5 {
+      GLuint pad1:16;
+      GLuint llc_mapping:1;
+      GLuint mlc_mapping:1;
+      GLuint gfdt:1;
+      GLuint gfdt_src:1;
+      GLuint y_offset:4;
+      GLuint pad0:1;
+      GLuint x_offset:7;
+   } ss5;   /* New in G4X */
+
+};
+
+
+
+struct brw_vertex_buffer_state
+{
+   struct {
+      GLuint pitch:11; 
+      GLuint pad:15;
+      GLuint access_type:1; 
+      GLuint vb_index:5; 
+   } vb0;
+   
+   GLuint start_addr; 
+   GLuint max_index;   
+#if 1
+   GLuint instance_data_step_rate; /* not included for sequential/random vertices? */
+#endif
+};
+
+#define BRW_VBP_MAX 17
+
+struct brw_vb_array_state {
+   struct header header;
+   struct brw_vertex_buffer_state vb[BRW_VBP_MAX];
+};
+
+
+struct brw_vertex_element_state
+{
+   struct
+   {
+      GLuint src_offset:11; 
+      GLuint pad:5;
+      GLuint src_format:9; 
+      GLuint pad0:1;
+      GLuint valid:1; 
+      GLuint vertex_buffer_index:5; 
+   } ve0;
+   
+   struct
+   {
+      GLuint dst_offset:8; 
+      GLuint pad:8;
+      GLuint vfcomponent3:4; 
+      GLuint vfcomponent2:4; 
+      GLuint vfcomponent1:4; 
+      GLuint vfcomponent0:4; 
+   } ve1;
+};
+
+#define BRW_VEP_MAX 18
+
+struct brw_vertex_element_packet {
+   struct header header;
+   struct brw_vertex_element_state ve[BRW_VEP_MAX]; /* note: less than _TNL_ATTRIB_MAX */
+};
+
+
+struct brw_urb_immediate {
+   GLuint opcode:4;
+   GLuint offset:6;
+   GLuint swizzle_control:2; 
+   GLuint pad:1;
+   GLuint allocate:1;
+   GLuint used:1;
+   GLuint complete:1;
+   GLuint response_length:4;
+   GLuint msg_length:4;
+   GLuint msg_target:4;
+   GLuint pad1:3;
+   GLuint end_of_thread:1;
+};
+
+/* Instruction format for the execution units:
+ */
+struct brw_instruction
+{
+   struct 
+   {
+      GLuint opcode:7;
+      GLuint pad:1;
+      GLuint access_mode:1;
+      GLuint mask_control:1;
+      GLuint dependency_control:2;
+      GLuint compression_control:2;
+      GLuint thread_control:2;
+      GLuint predicate_control:4;
+      GLuint predicate_inverse:1;
+      GLuint execution_size:3;
+      GLuint destreg__conditionalmod:4; /* destreg - send, conditionalmod - others */
+      GLuint pad0:2;
+      GLuint debug_control:1;
+      GLuint saturate:1;
+   } header;
+
+   union {
+      struct
+      {
+        GLuint dest_reg_file:2;
+        GLuint dest_reg_type:3;
+        GLuint src0_reg_file:2;
+        GLuint src0_reg_type:3;
+        GLuint src1_reg_file:2;
+        GLuint src1_reg_type:3;
+        GLuint pad:1;
+        GLuint dest_subreg_nr:5;
+        GLuint dest_reg_nr:8;
+        GLuint dest_horiz_stride:2;
+        GLuint dest_address_mode:1;
+      } da1;
+
+      struct
+      {
+        GLuint dest_reg_file:2;
+        GLuint dest_reg_type:3;
+        GLuint src0_reg_file:2;
+        GLuint src0_reg_type:3;
+        GLuint src1_reg_file:2;        /* 0x00000c00 */
+        GLuint src1_reg_type:3;        /* 0x00007000 */
+        GLuint pad:1;
+        GLint dest_indirect_offset:10; /* offset against the deref'd address reg */
+        GLuint dest_subreg_nr:3; /* subnr for the address reg a0.x */
+        GLuint dest_horiz_stride:2;
+        GLuint dest_address_mode:1;
+      } ia1;
+
+      struct
+      {
+        GLuint dest_reg_file:2;
+        GLuint dest_reg_type:3;
+        GLuint src0_reg_file:2;
+        GLuint src0_reg_type:3;
+        GLuint src1_reg_file:2;
+        GLuint src1_reg_type:3;
+        GLuint pad:1;
+        GLuint dest_writemask:4;
+        GLuint dest_subreg_nr:1;
+        GLuint dest_reg_nr:8;
+        GLuint pad1:2;
+        GLuint dest_address_mode:1;
+      } da16;
+
+      struct
+      {
+        GLuint dest_reg_file:2;
+        GLuint dest_reg_type:3;
+        GLuint src0_reg_file:2;
+        GLuint src0_reg_type:3;
+        GLuint pad0:6;
+        GLuint dest_writemask:4;
+        GLint dest_indirect_offset:6;
+        GLuint dest_subreg_nr:3;
+        GLuint pad1:2;
+        GLuint dest_address_mode:1;
+      } ia16;
+   } bits1;
+
+
+   union {
+      struct
+      {
+        GLuint src0_subreg_nr:5;
+        GLuint src0_reg_nr:8;
+        GLuint src0_abs:1;
+        GLuint src0_negate:1;
+        GLuint src0_address_mode:1;
+        GLuint src0_horiz_stride:2;
+        GLuint src0_width:3;
+        GLuint src0_vert_stride:4;
+        GLuint flag_reg_nr:1;
+        GLuint pad:6;
+      } da1;
+
+      struct
+      {
+        GLint src0_indirect_offset:10;
+        GLuint src0_subreg_nr:3;
+        GLuint src0_abs:1;
+        GLuint src0_negate:1;
+        GLuint src0_address_mode:1;
+        GLuint src0_horiz_stride:2;
+        GLuint src0_width:3;
+        GLuint src0_vert_stride:4;
+        GLuint flag_reg_nr:1;
+        GLuint pad:6;  
+      } ia1;
+
+      struct
+      {
+        GLuint src0_swz_x:2;
+        GLuint src0_swz_y:2;
+        GLuint src0_subreg_nr:1;
+        GLuint src0_reg_nr:8;
+        GLuint src0_abs:1;
+        GLuint src0_negate:1;
+        GLuint src0_address_mode:1;
+        GLuint src0_swz_z:2;
+        GLuint src0_swz_w:2;
+        GLuint pad0:1;
+        GLuint src0_vert_stride:4;
+        GLuint flag_reg_nr:1;
+        GLuint pad1:6;
+      } da16;
+
+      struct
+      {
+        GLuint src0_swz_x:2;
+        GLuint src0_swz_y:2;
+        GLint src0_indirect_offset:6;
+        GLuint src0_subreg_nr:3;
+        GLuint src0_abs:1;
+        GLuint src0_negate:1;
+        GLuint src0_address_mode:1;
+        GLuint src0_swz_z:2;
+        GLuint src0_swz_w:2;
+        GLuint pad0:1;
+        GLuint src0_vert_stride:4;
+        GLuint flag_reg_nr:1;
+        GLuint pad1:6;
+      } ia16;
+
+       struct 
+       {
+           GLuint pad:26;
+           GLuint end_of_thread:1;
+           GLuint pad1:1;
+           GLuint sfid:4;
+       } send_igdng;  /* for IGDNG only */
+
+   } bits2;
+
+   union
+   {
+      struct
+      {
+        GLuint src1_subreg_nr:5;
+        GLuint src1_reg_nr:8;
+        GLuint src1_abs:1;
+        GLuint src1_negate:1;
+        GLuint src1_address_mode:1;
+        GLuint src1_horiz_stride:2;
+        GLuint src1_width:3;
+        GLuint src1_vert_stride:4;
+        GLuint pad0:7;
+      } da1;
+
+      struct
+      {
+        GLuint src1_swz_x:2;
+        GLuint src1_swz_y:2;
+        GLuint src1_subreg_nr:1;
+        GLuint src1_reg_nr:8;
+        GLuint src1_abs:1;
+        GLuint src1_negate:1;
+        GLuint src1_address_mode:1;
+        GLuint src1_swz_z:2;
+        GLuint src1_swz_w:2;
+        GLuint pad1:1;
+        GLuint src1_vert_stride:4;
+        GLuint pad2:7;
+      } da16;
+
+      struct
+      {
+        GLint  src1_indirect_offset:10;
+        GLuint src1_subreg_nr:3;
+        GLuint src1_abs:1;
+        GLuint src1_negate:1;
+        GLuint src1_address_mode:1;
+        GLuint src1_horiz_stride:2;
+        GLuint src1_width:3;
+        GLuint src1_vert_stride:4;
+        GLuint flag_reg_nr:1;
+        GLuint pad1:6; 
+      } ia1;
+
+      struct
+      {
+        GLuint src1_swz_x:2;
+        GLuint src1_swz_y:2;
+        GLint  src1_indirect_offset:6;
+        GLuint src1_subreg_nr:3;
+        GLuint src1_abs:1;
+        GLuint src1_negate:1;
+        GLuint pad0:1;
+        GLuint src1_swz_z:2;
+        GLuint src1_swz_w:2;
+        GLuint pad1:1;
+        GLuint src1_vert_stride:4;
+        GLuint flag_reg_nr:1;
+        GLuint pad2:6;
+      } ia16;
+
+
+      struct
+      {
+        GLint  jump_count:16;  /* note: signed */
+        GLuint  pop_count:4;
+        GLuint  pad0:12;
+      } if_else;
+
+      struct {
+        GLuint function:4;
+        GLuint int_type:1;
+        GLuint precision:1;
+        GLuint saturate:1;
+        GLuint data_type:1;
+        GLuint pad0:8;
+        GLuint response_length:4;
+        GLuint msg_length:4;
+        GLuint msg_target:4;
+        GLuint pad1:3;
+        GLuint end_of_thread:1;
+      } math;
+
+      struct {
+        GLuint function:4;
+        GLuint int_type:1;
+        GLuint precision:1;
+        GLuint saturate:1;
+        GLuint data_type:1;
+        GLuint snapshot:1;
+        GLuint pad0:10;
+        GLuint header_present:1;
+        GLuint response_length:5;
+        GLuint msg_length:4;
+        GLuint pad1:2;
+        GLuint end_of_thread:1;
+      } math_igdng;
+
+      struct {
+        GLuint binding_table_index:8;
+        GLuint sampler:4;
+        GLuint return_format:2; 
+        GLuint msg_type:2;   
+        GLuint response_length:4;
+        GLuint msg_length:4;
+        GLuint msg_target:4;
+        GLuint pad1:3;
+        GLuint end_of_thread:1;
+      } sampler;
+
+      struct {
+         GLuint binding_table_index:8;
+         GLuint sampler:4;
+         GLuint msg_type:4;
+         GLuint response_length:4;
+         GLuint msg_length:4;
+         GLuint msg_target:4;
+         GLuint pad1:3;
+         GLuint end_of_thread:1;
+      } sampler_g4x;
+
+      struct {
+        GLuint binding_table_index:8;
+        GLuint sampler:4;
+        GLuint msg_type:4;
+        GLuint simd_mode:2;
+        GLuint pad0:1;
+        GLuint header_present:1;
+        GLuint response_length:5;
+        GLuint msg_length:4;
+        GLuint pad1:2;
+        GLuint end_of_thread:1;
+      } sampler_igdng;
+
+      struct brw_urb_immediate urb;
+
+      struct {
+        GLuint opcode:4;
+        GLuint offset:6;
+        GLuint swizzle_control:2; 
+        GLuint pad:1;
+        GLuint allocate:1;
+        GLuint used:1;
+        GLuint complete:1;
+        GLuint pad0:3;
+        GLuint header_present:1;
+        GLuint response_length:5;
+        GLuint msg_length:4;
+        GLuint pad1:2;
+        GLuint end_of_thread:1;
+      } urb_igdng;
+
+      struct {
+        GLuint binding_table_index:8;
+        GLuint msg_control:4;  
+        GLuint msg_type:2;  
+        GLuint target_cache:2;    
+        GLuint response_length:4;
+        GLuint msg_length:4;
+        GLuint msg_target:4;
+        GLuint pad1:3;
+        GLuint end_of_thread:1;
+      } dp_read;
+
+      struct {
+        GLuint binding_table_index:8;
+        GLuint msg_control:3;  
+        GLuint msg_type:3;  
+        GLuint target_cache:2;    
+        GLuint pad0:3;
+        GLuint header_present:1;
+        GLuint response_length:5;
+        GLuint msg_length:4;
+        GLuint pad1:2;
+        GLuint end_of_thread:1;
+      } dp_read_igdng;
+
+      struct {
+        GLuint binding_table_index:8;
+        GLuint msg_control:3;
+        GLuint pixel_scoreboard_clear:1;
+        GLuint msg_type:3;    
+        GLuint send_commit_msg:1;
+        GLuint response_length:4;
+        GLuint msg_length:4;
+        GLuint msg_target:4;
+        GLuint pad1:3;
+        GLuint end_of_thread:1;
+      } dp_write;
+
+      struct {
+        GLuint binding_table_index:8;
+        GLuint msg_control:3;
+        GLuint pixel_scoreboard_clear:1;
+        GLuint msg_type:3;    
+        GLuint send_commit_msg:1;
+        GLuint pad0:3;
+        GLuint header_present:1;
+        GLuint response_length:5;
+        GLuint msg_length:4;
+        GLuint pad1:2;
+        GLuint end_of_thread:1;
+      } dp_write_igdng;
+
+      struct {
+        GLuint pad:16;
+        GLuint response_length:4;
+        GLuint msg_length:4;
+        GLuint msg_target:4;
+        GLuint pad1:3;
+        GLuint end_of_thread:1;
+      } generic;
+
+      struct {
+        GLuint pad:19;
+        GLuint header_present:1;
+        GLuint response_length:5;
+        GLuint msg_length:4;
+        GLuint pad1:2;
+        GLuint end_of_thread:1;
+      } generic_igdng;
+
+      GLint d;
+      GLuint ud;
+      float f;
+   } bits3;
+};
+
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_structs_dump.c b/src/gallium/drivers/i965/brw_structs_dump.c
new file mode 100644 (file)
index 0000000..cd40fc6
--- /dev/null
@@ -0,0 +1,1247 @@
+/**************************************************************************
+ *
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * Dump i965 data structures.
+ *
+ * Generated automatically from brw_structs.h by brw_structs_dump.py.
+ */
+
+#include "util/u_debug.h"
+
+#include "brw_types.h"
+#include "brw_structs.h"
+#include "brw_structs_dump.h"
+
+void
+brw_dump_3d_control(const struct brw_3d_control *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.notify_enable = 0x%x\n", (*ptr).header.notify_enable);
+   debug_printf("\t\t.header.wc_flush_enable = 0x%x\n", (*ptr).header.wc_flush_enable);
+   debug_printf("\t\t.header.depth_stall_enable = 0x%x\n", (*ptr).header.depth_stall_enable);
+   debug_printf("\t\t.header.operation = 0x%x\n", (*ptr).header.operation);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.dest.dest_addr_type = 0x%x\n", (*ptr).dest.dest_addr_type);
+   debug_printf("\t\t.dest.dest_addr = 0x%x\n", (*ptr).dest.dest_addr);
+   debug_printf("\t\t.dword2 = 0x%x\n", (*ptr).dword2);
+   debug_printf("\t\t.dword3 = 0x%x\n", (*ptr).dword3);
+}
+
+void
+brw_dump_3d_primitive(const struct brw_3d_primitive *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.topology = 0x%x\n", (*ptr).header.topology);
+   debug_printf("\t\t.header.indexed = 0x%x\n", (*ptr).header.indexed);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.verts_per_instance = 0x%x\n", (*ptr).verts_per_instance);
+   debug_printf("\t\t.start_vert_location = 0x%x\n", (*ptr).start_vert_location);
+   debug_printf("\t\t.instance_count = 0x%x\n", (*ptr).instance_count);
+   debug_printf("\t\t.start_instance_location = 0x%x\n", (*ptr).start_instance_location);
+   debug_printf("\t\t.base_vert_location = 0x%x\n", (*ptr).base_vert_location);
+}
+
+void
+brw_dump_aa_line_parameters(const struct brw_aa_line_parameters *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.bits0.aa_coverage_scope = 0x%x\n", (*ptr).bits0.aa_coverage_scope);
+   debug_printf("\t\t.bits0.aa_coverage_bias = 0x%x\n", (*ptr).bits0.aa_coverage_bias);
+   debug_printf("\t\t.bits1.aa_coverage_endcap_slope = 0x%x\n", (*ptr).bits1.aa_coverage_endcap_slope);
+   debug_printf("\t\t.bits1.aa_coverage_endcap_bias = 0x%x\n", (*ptr).bits1.aa_coverage_endcap_bias);
+}
+
+void
+brw_dump_binding_table_pointers(const struct brw_binding_table_pointers *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.vs = 0x%x\n", (*ptr).vs);
+   debug_printf("\t\t.gs = 0x%x\n", (*ptr).gs);
+   debug_printf("\t\t.clp = 0x%x\n", (*ptr).clp);
+   debug_printf("\t\t.sf = 0x%x\n", (*ptr).sf);
+   debug_printf("\t\t.wm = 0x%x\n", (*ptr).wm);
+}
+
+void
+brw_dump_blend_constant_color(const struct brw_blend_constant_color *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.blend_constant_color[0] = %f\n", (*ptr).blend_constant_color[0]);
+   debug_printf("\t\t.blend_constant_color[1] = %f\n", (*ptr).blend_constant_color[1]);
+   debug_printf("\t\t.blend_constant_color[2] = %f\n", (*ptr).blend_constant_color[2]);
+   debug_printf("\t\t.blend_constant_color[3] = %f\n", (*ptr).blend_constant_color[3]);
+}
+
+void
+brw_dump_cc0(const struct brw_cc0 *ptr)
+{
+   debug_printf("\t\t.bf_stencil_pass_depth_pass_op = 0x%x\n", (*ptr).bf_stencil_pass_depth_pass_op);
+   debug_printf("\t\t.bf_stencil_pass_depth_fail_op = 0x%x\n", (*ptr).bf_stencil_pass_depth_fail_op);
+   debug_printf("\t\t.bf_stencil_fail_op = 0x%x\n", (*ptr).bf_stencil_fail_op);
+   debug_printf("\t\t.bf_stencil_func = 0x%x\n", (*ptr).bf_stencil_func);
+   debug_printf("\t\t.bf_stencil_enable = 0x%x\n", (*ptr).bf_stencil_enable);
+   debug_printf("\t\t.stencil_write_enable = 0x%x\n", (*ptr).stencil_write_enable);
+   debug_printf("\t\t.stencil_pass_depth_pass_op = 0x%x\n", (*ptr).stencil_pass_depth_pass_op);
+   debug_printf("\t\t.stencil_pass_depth_fail_op = 0x%x\n", (*ptr).stencil_pass_depth_fail_op);
+   debug_printf("\t\t.stencil_fail_op = 0x%x\n", (*ptr).stencil_fail_op);
+   debug_printf("\t\t.stencil_func = 0x%x\n", (*ptr).stencil_func);
+   debug_printf("\t\t.stencil_enable = 0x%x\n", (*ptr).stencil_enable);
+}
+
+void
+brw_dump_cc1(const struct brw_cc1 *ptr)
+{
+   debug_printf("\t\t.bf_stencil_ref = 0x%x\n", (*ptr).bf_stencil_ref);
+   debug_printf("\t\t.stencil_write_mask = 0x%x\n", (*ptr).stencil_write_mask);
+   debug_printf("\t\t.stencil_test_mask = 0x%x\n", (*ptr).stencil_test_mask);
+   debug_printf("\t\t.stencil_ref = 0x%x\n", (*ptr).stencil_ref);
+}
+
+void
+brw_dump_cc2(const struct brw_cc2 *ptr)
+{
+   debug_printf("\t\t.logicop_enable = 0x%x\n", (*ptr).logicop_enable);
+   debug_printf("\t\t.depth_write_enable = 0x%x\n", (*ptr).depth_write_enable);
+   debug_printf("\t\t.depth_test_function = 0x%x\n", (*ptr).depth_test_function);
+   debug_printf("\t\t.depth_test = 0x%x\n", (*ptr).depth_test);
+   debug_printf("\t\t.bf_stencil_write_mask = 0x%x\n", (*ptr).bf_stencil_write_mask);
+   debug_printf("\t\t.bf_stencil_test_mask = 0x%x\n", (*ptr).bf_stencil_test_mask);
+}
+
+void
+brw_dump_cc3(const struct brw_cc3 *ptr)
+{
+   debug_printf("\t\t.alpha_test_func = 0x%x\n", (*ptr).alpha_test_func);
+   debug_printf("\t\t.alpha_test = 0x%x\n", (*ptr).alpha_test);
+   debug_printf("\t\t.blend_enable = 0x%x\n", (*ptr).blend_enable);
+   debug_printf("\t\t.ia_blend_enable = 0x%x\n", (*ptr).ia_blend_enable);
+   debug_printf("\t\t.alpha_test_format = 0x%x\n", (*ptr).alpha_test_format);
+}
+
+void
+brw_dump_cc4(const struct brw_cc4 *ptr)
+{
+   debug_printf("\t\t.cc_viewport_state_offset = 0x%x\n", (*ptr).cc_viewport_state_offset);
+}
+
+void
+brw_dump_cc5(const struct brw_cc5 *ptr)
+{
+   debug_printf("\t\t.ia_dest_blend_factor = 0x%x\n", (*ptr).ia_dest_blend_factor);
+   debug_printf("\t\t.ia_src_blend_factor = 0x%x\n", (*ptr).ia_src_blend_factor);
+   debug_printf("\t\t.ia_blend_function = 0x%x\n", (*ptr).ia_blend_function);
+   debug_printf("\t\t.statistics_enable = 0x%x\n", (*ptr).statistics_enable);
+   debug_printf("\t\t.logicop_func = 0x%x\n", (*ptr).logicop_func);
+   debug_printf("\t\t.dither_enable = 0x%x\n", (*ptr).dither_enable);
+}
+
+void
+brw_dump_cc6(const struct brw_cc6 *ptr)
+{
+   debug_printf("\t\t.clamp_post_alpha_blend = 0x%x\n", (*ptr).clamp_post_alpha_blend);
+   debug_printf("\t\t.clamp_pre_alpha_blend = 0x%x\n", (*ptr).clamp_pre_alpha_blend);
+   debug_printf("\t\t.clamp_range = 0x%x\n", (*ptr).clamp_range);
+   debug_printf("\t\t.y_dither_offset = 0x%x\n", (*ptr).y_dither_offset);
+   debug_printf("\t\t.x_dither_offset = 0x%x\n", (*ptr).x_dither_offset);
+   debug_printf("\t\t.dest_blend_factor = 0x%x\n", (*ptr).dest_blend_factor);
+   debug_printf("\t\t.src_blend_factor = 0x%x\n", (*ptr).src_blend_factor);
+   debug_printf("\t\t.blend_function = 0x%x\n", (*ptr).blend_function);
+}
+
+void
+brw_dump_cc7(const struct brw_cc7 *ptr)
+{
+   debug_printf("\t\t.alpha_ref.f = %f\n", (*ptr).alpha_ref.f);
+   debug_printf("\t\t.alpha_ref.ub[0] = 0x%x\n", (*ptr).alpha_ref.ub[0]);
+   debug_printf("\t\t.alpha_ref.ub[1] = 0x%x\n", (*ptr).alpha_ref.ub[1]);
+   debug_printf("\t\t.alpha_ref.ub[2] = 0x%x\n", (*ptr).alpha_ref.ub[2]);
+   debug_printf("\t\t.alpha_ref.ub[3] = 0x%x\n", (*ptr).alpha_ref.ub[3]);
+}
+
+void
+brw_dump_cc_unit_state(const struct brw_cc_unit_state *ptr)
+{
+   debug_printf("\t\t.cc0.bf_stencil_pass_depth_pass_op = 0x%x\n", (*ptr).cc0.bf_stencil_pass_depth_pass_op);
+   debug_printf("\t\t.cc0.bf_stencil_pass_depth_fail_op = 0x%x\n", (*ptr).cc0.bf_stencil_pass_depth_fail_op);
+   debug_printf("\t\t.cc0.bf_stencil_fail_op = 0x%x\n", (*ptr).cc0.bf_stencil_fail_op);
+   debug_printf("\t\t.cc0.bf_stencil_func = 0x%x\n", (*ptr).cc0.bf_stencil_func);
+   debug_printf("\t\t.cc0.bf_stencil_enable = 0x%x\n", (*ptr).cc0.bf_stencil_enable);
+   debug_printf("\t\t.cc0.stencil_write_enable = 0x%x\n", (*ptr).cc0.stencil_write_enable);
+   debug_printf("\t\t.cc0.stencil_pass_depth_pass_op = 0x%x\n", (*ptr).cc0.stencil_pass_depth_pass_op);
+   debug_printf("\t\t.cc0.stencil_pass_depth_fail_op = 0x%x\n", (*ptr).cc0.stencil_pass_depth_fail_op);
+   debug_printf("\t\t.cc0.stencil_fail_op = 0x%x\n", (*ptr).cc0.stencil_fail_op);
+   debug_printf("\t\t.cc0.stencil_func = 0x%x\n", (*ptr).cc0.stencil_func);
+   debug_printf("\t\t.cc0.stencil_enable = 0x%x\n", (*ptr).cc0.stencil_enable);
+   debug_printf("\t\t.cc1.bf_stencil_ref = 0x%x\n", (*ptr).cc1.bf_stencil_ref);
+   debug_printf("\t\t.cc1.stencil_write_mask = 0x%x\n", (*ptr).cc1.stencil_write_mask);
+   debug_printf("\t\t.cc1.stencil_test_mask = 0x%x\n", (*ptr).cc1.stencil_test_mask);
+   debug_printf("\t\t.cc1.stencil_ref = 0x%x\n", (*ptr).cc1.stencil_ref);
+   debug_printf("\t\t.cc2.logicop_enable = 0x%x\n", (*ptr).cc2.logicop_enable);
+   debug_printf("\t\t.cc2.depth_write_enable = 0x%x\n", (*ptr).cc2.depth_write_enable);
+   debug_printf("\t\t.cc2.depth_test_function = 0x%x\n", (*ptr).cc2.depth_test_function);
+   debug_printf("\t\t.cc2.depth_test = 0x%x\n", (*ptr).cc2.depth_test);
+   debug_printf("\t\t.cc2.bf_stencil_write_mask = 0x%x\n", (*ptr).cc2.bf_stencil_write_mask);
+   debug_printf("\t\t.cc2.bf_stencil_test_mask = 0x%x\n", (*ptr).cc2.bf_stencil_test_mask);
+   debug_printf("\t\t.cc3.alpha_test_func = 0x%x\n", (*ptr).cc3.alpha_test_func);
+   debug_printf("\t\t.cc3.alpha_test = 0x%x\n", (*ptr).cc3.alpha_test);
+   debug_printf("\t\t.cc3.blend_enable = 0x%x\n", (*ptr).cc3.blend_enable);
+   debug_printf("\t\t.cc3.ia_blend_enable = 0x%x\n", (*ptr).cc3.ia_blend_enable);
+   debug_printf("\t\t.cc3.alpha_test_format = 0x%x\n", (*ptr).cc3.alpha_test_format);
+   debug_printf("\t\t.cc4.cc_viewport_state_offset = 0x%x\n", (*ptr).cc4.cc_viewport_state_offset);
+   debug_printf("\t\t.cc5.ia_dest_blend_factor = 0x%x\n", (*ptr).cc5.ia_dest_blend_factor);
+   debug_printf("\t\t.cc5.ia_src_blend_factor = 0x%x\n", (*ptr).cc5.ia_src_blend_factor);
+   debug_printf("\t\t.cc5.ia_blend_function = 0x%x\n", (*ptr).cc5.ia_blend_function);
+   debug_printf("\t\t.cc5.statistics_enable = 0x%x\n", (*ptr).cc5.statistics_enable);
+   debug_printf("\t\t.cc5.logicop_func = 0x%x\n", (*ptr).cc5.logicop_func);
+   debug_printf("\t\t.cc5.dither_enable = 0x%x\n", (*ptr).cc5.dither_enable);
+   debug_printf("\t\t.cc6.clamp_post_alpha_blend = 0x%x\n", (*ptr).cc6.clamp_post_alpha_blend);
+   debug_printf("\t\t.cc6.clamp_pre_alpha_blend = 0x%x\n", (*ptr).cc6.clamp_pre_alpha_blend);
+   debug_printf("\t\t.cc6.clamp_range = 0x%x\n", (*ptr).cc6.clamp_range);
+   debug_printf("\t\t.cc6.y_dither_offset = 0x%x\n", (*ptr).cc6.y_dither_offset);
+   debug_printf("\t\t.cc6.x_dither_offset = 0x%x\n", (*ptr).cc6.x_dither_offset);
+   debug_printf("\t\t.cc6.dest_blend_factor = 0x%x\n", (*ptr).cc6.dest_blend_factor);
+   debug_printf("\t\t.cc6.src_blend_factor = 0x%x\n", (*ptr).cc6.src_blend_factor);
+   debug_printf("\t\t.cc6.blend_function = 0x%x\n", (*ptr).cc6.blend_function);
+   debug_printf("\t\t.cc7.alpha_ref.f = %f\n", (*ptr).cc7.alpha_ref.f);
+   debug_printf("\t\t.cc7.alpha_ref.ub[0] = 0x%x\n", (*ptr).cc7.alpha_ref.ub[0]);
+   debug_printf("\t\t.cc7.alpha_ref.ub[1] = 0x%x\n", (*ptr).cc7.alpha_ref.ub[1]);
+   debug_printf("\t\t.cc7.alpha_ref.ub[2] = 0x%x\n", (*ptr).cc7.alpha_ref.ub[2]);
+   debug_printf("\t\t.cc7.alpha_ref.ub[3] = 0x%x\n", (*ptr).cc7.alpha_ref.ub[3]);
+}
+
+void
+brw_dump_cc_viewport(const struct brw_cc_viewport *ptr)
+{
+   debug_printf("\t\t.min_depth = %f\n", (*ptr).min_depth);
+   debug_printf("\t\t.max_depth = %f\n", (*ptr).max_depth);
+}
+
+void
+brw_dump_clip_unit_state(const struct brw_clip_unit_state *ptr)
+{
+   debug_printf("\t\t.thread0.grf_reg_count = 0x%x\n", (*ptr).thread0.grf_reg_count);
+   debug_printf("\t\t.thread0.kernel_start_pointer = 0x%x\n", (*ptr).thread0.kernel_start_pointer);
+   debug_printf("\t\t.thread1.sw_exception_enable = 0x%x\n", (*ptr).thread1.sw_exception_enable);
+   debug_printf("\t\t.thread1.mask_stack_exception_enable = 0x%x\n", (*ptr).thread1.mask_stack_exception_enable);
+   debug_printf("\t\t.thread1.illegal_op_exception_enable = 0x%x\n", (*ptr).thread1.illegal_op_exception_enable);
+   debug_printf("\t\t.thread1.floating_point_mode = 0x%x\n", (*ptr).thread1.floating_point_mode);
+   debug_printf("\t\t.thread1.thread_priority = 0x%x\n", (*ptr).thread1.thread_priority);
+   debug_printf("\t\t.thread1.binding_table_entry_count = 0x%x\n", (*ptr).thread1.binding_table_entry_count);
+   debug_printf("\t\t.thread1.single_program_flow = 0x%x\n", (*ptr).thread1.single_program_flow);
+   debug_printf("\t\t.thread2.per_thread_scratch_space = 0x%x\n", (*ptr).thread2.per_thread_scratch_space);
+   debug_printf("\t\t.thread2.scratch_space_base_pointer = 0x%x\n", (*ptr).thread2.scratch_space_base_pointer);
+   debug_printf("\t\t.thread3.dispatch_grf_start_reg = 0x%x\n", (*ptr).thread3.dispatch_grf_start_reg);
+   debug_printf("\t\t.thread3.urb_entry_read_offset = 0x%x\n", (*ptr).thread3.urb_entry_read_offset);
+   debug_printf("\t\t.thread3.urb_entry_read_length = 0x%x\n", (*ptr).thread3.urb_entry_read_length);
+   debug_printf("\t\t.thread3.const_urb_entry_read_offset = 0x%x\n", (*ptr).thread3.const_urb_entry_read_offset);
+   debug_printf("\t\t.thread3.const_urb_entry_read_length = 0x%x\n", (*ptr).thread3.const_urb_entry_read_length);
+   debug_printf("\t\t.thread4.gs_output_stats = 0x%x\n", (*ptr).thread4.gs_output_stats);
+   debug_printf("\t\t.thread4.stats_enable = 0x%x\n", (*ptr).thread4.stats_enable);
+   debug_printf("\t\t.thread4.nr_urb_entries = 0x%x\n", (*ptr).thread4.nr_urb_entries);
+   debug_printf("\t\t.thread4.urb_entry_allocation_size = 0x%x\n", (*ptr).thread4.urb_entry_allocation_size);
+   debug_printf("\t\t.thread4.max_threads = 0x%x\n", (*ptr).thread4.max_threads);
+   debug_printf("\t\t.clip5.clip_mode = 0x%x\n", (*ptr).clip5.clip_mode);
+   debug_printf("\t\t.clip5.userclip_enable_flags = 0x%x\n", (*ptr).clip5.userclip_enable_flags);
+   debug_printf("\t\t.clip5.userclip_must_clip = 0x%x\n", (*ptr).clip5.userclip_must_clip);
+   debug_printf("\t\t.clip5.negative_w_clip_test = 0x%x\n", (*ptr).clip5.negative_w_clip_test);
+   debug_printf("\t\t.clip5.guard_band_enable = 0x%x\n", (*ptr).clip5.guard_band_enable);
+   debug_printf("\t\t.clip5.viewport_z_clip_enable = 0x%x\n", (*ptr).clip5.viewport_z_clip_enable);
+   debug_printf("\t\t.clip5.viewport_xy_clip_enable = 0x%x\n", (*ptr).clip5.viewport_xy_clip_enable);
+   debug_printf("\t\t.clip5.vertex_position_space = 0x%x\n", (*ptr).clip5.vertex_position_space);
+   debug_printf("\t\t.clip5.api_mode = 0x%x\n", (*ptr).clip5.api_mode);
+   debug_printf("\t\t.clip6.clipper_viewport_state_ptr = 0x%x\n", (*ptr).clip6.clipper_viewport_state_ptr);
+   debug_printf("\t\t.viewport_xmin = %f\n", (*ptr).viewport_xmin);
+   debug_printf("\t\t.viewport_xmax = %f\n", (*ptr).viewport_xmax);
+   debug_printf("\t\t.viewport_ymin = %f\n", (*ptr).viewport_ymin);
+   debug_printf("\t\t.viewport_ymax = %f\n", (*ptr).viewport_ymax);
+}
+
+void
+brw_dump_clipper_viewport(const struct brw_clipper_viewport *ptr)
+{
+   debug_printf("\t\t.xmin = %f\n", (*ptr).xmin);
+   debug_printf("\t\t.xmax = %f\n", (*ptr).xmax);
+   debug_printf("\t\t.ymin = %f\n", (*ptr).ymin);
+   debug_printf("\t\t.ymax = %f\n", (*ptr).ymax);
+}
+
+void
+brw_dump_constant_buffer(const struct brw_constant_buffer *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.valid = 0x%x\n", (*ptr).header.valid);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.bits0.buffer_length = 0x%x\n", (*ptr).bits0.buffer_length);
+   debug_printf("\t\t.bits0.buffer_address = 0x%x\n", (*ptr).bits0.buffer_address);
+}
+
+void
+brw_dump_cs_urb_state(const struct brw_cs_urb_state *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.bits0.nr_urb_entries = 0x%x\n", (*ptr).bits0.nr_urb_entries);
+   debug_printf("\t\t.bits0.urb_entry_size = 0x%x\n", (*ptr).bits0.urb_entry_size);
+}
+
+void
+brw_dump_depthbuffer(const struct brw_depthbuffer *ptr)
+{
+   debug_printf("\t\t.header.bits.length = 0x%x\n", (*ptr).header.bits.length);
+   debug_printf("\t\t.header.bits.opcode = 0x%x\n", (*ptr).header.bits.opcode);
+   debug_printf("\t\t.dword1.bits.pitch = 0x%x\n", (*ptr).dword1.bits.pitch);
+   debug_printf("\t\t.dword1.bits.format = 0x%x\n", (*ptr).dword1.bits.format);
+   debug_printf("\t\t.dword1.bits.software_tiled_rendering_mode = 0x%x\n", (*ptr).dword1.bits.software_tiled_rendering_mode);
+   debug_printf("\t\t.dword1.bits.depth_offset_disable = 0x%x\n", (*ptr).dword1.bits.depth_offset_disable);
+   debug_printf("\t\t.dword1.bits.tile_walk = 0x%x\n", (*ptr).dword1.bits.tile_walk);
+   debug_printf("\t\t.dword1.bits.tiled_surface = 0x%x\n", (*ptr).dword1.bits.tiled_surface);
+   debug_printf("\t\t.dword1.bits.surface_type = 0x%x\n", (*ptr).dword1.bits.surface_type);
+   debug_printf("\t\t.dword2_base_addr = 0x%x\n", (*ptr).dword2_base_addr);
+   debug_printf("\t\t.dword3.bits.mipmap_layout = 0x%x\n", (*ptr).dword3.bits.mipmap_layout);
+   debug_printf("\t\t.dword3.bits.lod = 0x%x\n", (*ptr).dword3.bits.lod);
+   debug_printf("\t\t.dword3.bits.width = 0x%x\n", (*ptr).dword3.bits.width);
+   debug_printf("\t\t.dword3.bits.height = 0x%x\n", (*ptr).dword3.bits.height);
+   debug_printf("\t\t.dword4.bits.min_array_element = 0x%x\n", (*ptr).dword4.bits.min_array_element);
+   debug_printf("\t\t.dword4.bits.depth = 0x%x\n", (*ptr).dword4.bits.depth);
+}
+
+void
+brw_dump_depthbuffer_g4x(const struct brw_depthbuffer_g4x *ptr)
+{
+   debug_printf("\t\t.header.bits.length = 0x%x\n", (*ptr).header.bits.length);
+   debug_printf("\t\t.header.bits.opcode = 0x%x\n", (*ptr).header.bits.opcode);
+   debug_printf("\t\t.dword1.bits.pitch = 0x%x\n", (*ptr).dword1.bits.pitch);
+   debug_printf("\t\t.dword1.bits.format = 0x%x\n", (*ptr).dword1.bits.format);
+   debug_printf("\t\t.dword1.bits.software_tiled_rendering_mode = 0x%x\n", (*ptr).dword1.bits.software_tiled_rendering_mode);
+   debug_printf("\t\t.dword1.bits.depth_offset_disable = 0x%x\n", (*ptr).dword1.bits.depth_offset_disable);
+   debug_printf("\t\t.dword1.bits.tile_walk = 0x%x\n", (*ptr).dword1.bits.tile_walk);
+   debug_printf("\t\t.dword1.bits.tiled_surface = 0x%x\n", (*ptr).dword1.bits.tiled_surface);
+   debug_printf("\t\t.dword1.bits.surface_type = 0x%x\n", (*ptr).dword1.bits.surface_type);
+   debug_printf("\t\t.dword2_base_addr = 0x%x\n", (*ptr).dword2_base_addr);
+   debug_printf("\t\t.dword3.bits.mipmap_layout = 0x%x\n", (*ptr).dword3.bits.mipmap_layout);
+   debug_printf("\t\t.dword3.bits.lod = 0x%x\n", (*ptr).dword3.bits.lod);
+   debug_printf("\t\t.dword3.bits.width = 0x%x\n", (*ptr).dword3.bits.width);
+   debug_printf("\t\t.dword3.bits.height = 0x%x\n", (*ptr).dword3.bits.height);
+   debug_printf("\t\t.dword4.bits.min_array_element = 0x%x\n", (*ptr).dword4.bits.min_array_element);
+   debug_printf("\t\t.dword4.bits.depth = 0x%x\n", (*ptr).dword4.bits.depth);
+   debug_printf("\t\t.dword5.bits.xoffset = 0x%x\n", (*ptr).dword5.bits.xoffset);
+   debug_printf("\t\t.dword5.bits.yoffset = 0x%x\n", (*ptr).dword5.bits.yoffset);
+}
+
+void
+brw_dump_drawrect(const struct brw_drawrect *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.xmin = 0x%x\n", (*ptr).xmin);
+   debug_printf("\t\t.ymin = 0x%x\n", (*ptr).ymin);
+   debug_printf("\t\t.xmax = 0x%x\n", (*ptr).xmax);
+   debug_printf("\t\t.ymax = 0x%x\n", (*ptr).ymax);
+   debug_printf("\t\t.xorg = 0x%x\n", (*ptr).xorg);
+   debug_printf("\t\t.yorg = 0x%x\n", (*ptr).yorg);
+}
+
+void
+brw_dump_global_depth_offset_clamp(const struct brw_global_depth_offset_clamp *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.depth_offset_clamp = %f\n", (*ptr).depth_offset_clamp);
+}
+
+void
+brw_dump_gs_unit_state(const struct brw_gs_unit_state *ptr)
+{
+   debug_printf("\t\t.thread0.grf_reg_count = 0x%x\n", (*ptr).thread0.grf_reg_count);
+   debug_printf("\t\t.thread0.kernel_start_pointer = 0x%x\n", (*ptr).thread0.kernel_start_pointer);
+   debug_printf("\t\t.thread1.ext_halt_exception_enable = 0x%x\n", (*ptr).thread1.ext_halt_exception_enable);
+   debug_printf("\t\t.thread1.sw_exception_enable = 0x%x\n", (*ptr).thread1.sw_exception_enable);
+   debug_printf("\t\t.thread1.mask_stack_exception_enable = 0x%x\n", (*ptr).thread1.mask_stack_exception_enable);
+   debug_printf("\t\t.thread1.timeout_exception_enable = 0x%x\n", (*ptr).thread1.timeout_exception_enable);
+   debug_printf("\t\t.thread1.illegal_op_exception_enable = 0x%x\n", (*ptr).thread1.illegal_op_exception_enable);
+   debug_printf("\t\t.thread1.depth_coef_urb_read_offset = 0x%x\n", (*ptr).thread1.depth_coef_urb_read_offset);
+   debug_printf("\t\t.thread1.floating_point_mode = 0x%x\n", (*ptr).thread1.floating_point_mode);
+   debug_printf("\t\t.thread1.thread_priority = 0x%x\n", (*ptr).thread1.thread_priority);
+   debug_printf("\t\t.thread1.binding_table_entry_count = 0x%x\n", (*ptr).thread1.binding_table_entry_count);
+   debug_printf("\t\t.thread1.single_program_flow = 0x%x\n", (*ptr).thread1.single_program_flow);
+   debug_printf("\t\t.thread2.per_thread_scratch_space = 0x%x\n", (*ptr).thread2.per_thread_scratch_space);
+   debug_printf("\t\t.thread2.scratch_space_base_pointer = 0x%x\n", (*ptr).thread2.scratch_space_base_pointer);
+   debug_printf("\t\t.thread3.dispatch_grf_start_reg = 0x%x\n", (*ptr).thread3.dispatch_grf_start_reg);
+   debug_printf("\t\t.thread3.urb_entry_read_offset = 0x%x\n", (*ptr).thread3.urb_entry_read_offset);
+   debug_printf("\t\t.thread3.urb_entry_read_length = 0x%x\n", (*ptr).thread3.urb_entry_read_length);
+   debug_printf("\t\t.thread3.const_urb_entry_read_offset = 0x%x\n", (*ptr).thread3.const_urb_entry_read_offset);
+   debug_printf("\t\t.thread3.const_urb_entry_read_length = 0x%x\n", (*ptr).thread3.const_urb_entry_read_length);
+   debug_printf("\t\t.thread4.rendering_enable = 0x%x\n", (*ptr).thread4.rendering_enable);
+   debug_printf("\t\t.thread4.stats_enable = 0x%x\n", (*ptr).thread4.stats_enable);
+   debug_printf("\t\t.thread4.nr_urb_entries = 0x%x\n", (*ptr).thread4.nr_urb_entries);
+   debug_printf("\t\t.thread4.urb_entry_allocation_size = 0x%x\n", (*ptr).thread4.urb_entry_allocation_size);
+   debug_printf("\t\t.thread4.max_threads = 0x%x\n", (*ptr).thread4.max_threads);
+   debug_printf("\t\t.gs5.sampler_count = 0x%x\n", (*ptr).gs5.sampler_count);
+   debug_printf("\t\t.gs5.sampler_state_pointer = 0x%x\n", (*ptr).gs5.sampler_state_pointer);
+   debug_printf("\t\t.gs6.max_vp_index = 0x%x\n", (*ptr).gs6.max_vp_index);
+   debug_printf("\t\t.gs6.svbi_post_inc_value = 0x%x\n", (*ptr).gs6.svbi_post_inc_value);
+   debug_printf("\t\t.gs6.svbi_post_inc_enable = 0x%x\n", (*ptr).gs6.svbi_post_inc_enable);
+   debug_printf("\t\t.gs6.svbi_payload = 0x%x\n", (*ptr).gs6.svbi_payload);
+   debug_printf("\t\t.gs6.discard_adjaceny = 0x%x\n", (*ptr).gs6.discard_adjaceny);
+   debug_printf("\t\t.gs6.reorder_enable = 0x%x\n", (*ptr).gs6.reorder_enable);
+}
+
+void
+brw_dump_indexbuffer(const struct brw_indexbuffer *ptr)
+{
+   debug_printf("\t\t.header.bits.length = 0x%x\n", (*ptr).header.bits.length);
+   debug_printf("\t\t.header.bits.index_format = 0x%x\n", (*ptr).header.bits.index_format);
+   debug_printf("\t\t.header.bits.cut_index_enable = 0x%x\n", (*ptr).header.bits.cut_index_enable);
+   debug_printf("\t\t.header.bits.opcode = 0x%x\n", (*ptr).header.bits.opcode);
+   debug_printf("\t\t.buffer_start = 0x%x\n", (*ptr).buffer_start);
+   debug_printf("\t\t.buffer_end = 0x%x\n", (*ptr).buffer_end);
+}
+
+void
+brw_dump_line_stipple(const struct brw_line_stipple *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.bits0.pattern = 0x%x\n", (*ptr).bits0.pattern);
+   debug_printf("\t\t.bits1.repeat_count = 0x%x\n", (*ptr).bits1.repeat_count);
+   debug_printf("\t\t.bits1.inverse_repeat_count = 0x%x\n", (*ptr).bits1.inverse_repeat_count);
+}
+
+void
+brw_dump_mi_flush(const struct brw_mi_flush *ptr)
+{
+   debug_printf("\t\t.flags = 0x%x\n", (*ptr).flags);
+   debug_printf("\t\t.opcode = 0x%x\n", (*ptr).opcode);
+}
+
+void
+brw_dump_pipe_control(const struct brw_pipe_control *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.notify_enable = 0x%x\n", (*ptr).header.notify_enable);
+   debug_printf("\t\t.header.texture_cache_flush_enable = 0x%x\n", (*ptr).header.texture_cache_flush_enable);
+   debug_printf("\t\t.header.indirect_state_pointers_disable = 0x%x\n", (*ptr).header.indirect_state_pointers_disable);
+   debug_printf("\t\t.header.instruction_state_cache_flush_enable = 0x%x\n", (*ptr).header.instruction_state_cache_flush_enable);
+   debug_printf("\t\t.header.write_cache_flush_enable = 0x%x\n", (*ptr).header.write_cache_flush_enable);
+   debug_printf("\t\t.header.depth_stall_enable = 0x%x\n", (*ptr).header.depth_stall_enable);
+   debug_printf("\t\t.header.post_sync_operation = 0x%x\n", (*ptr).header.post_sync_operation);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.bits1.dest_addr_type = 0x%x\n", (*ptr).bits1.dest_addr_type);
+   debug_printf("\t\t.bits1.dest_addr = 0x%x\n", (*ptr).bits1.dest_addr);
+   debug_printf("\t\t.data0 = 0x%x\n", (*ptr).data0);
+   debug_printf("\t\t.data1 = 0x%x\n", (*ptr).data1);
+}
+
+void
+brw_dump_pipeline_select(const struct brw_pipeline_select *ptr)
+{
+   debug_printf("\t\t.header.pipeline_select = 0x%x\n", (*ptr).header.pipeline_select);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+}
+
+void
+brw_dump_pipelined_state_pointers(const struct brw_pipelined_state_pointers *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.vs.offset = 0x%x\n", (*ptr).vs.offset);
+   debug_printf("\t\t.gs.enable = 0x%x\n", (*ptr).gs.enable);
+   debug_printf("\t\t.gs.offset = 0x%x\n", (*ptr).gs.offset);
+   debug_printf("\t\t.clp.enable = 0x%x\n", (*ptr).clp.enable);
+   debug_printf("\t\t.clp.offset = 0x%x\n", (*ptr).clp.offset);
+   debug_printf("\t\t.sf.offset = 0x%x\n", (*ptr).sf.offset);
+   debug_printf("\t\t.wm.offset = 0x%x\n", (*ptr).wm.offset);
+   debug_printf("\t\t.cc.offset = 0x%x\n", (*ptr).cc.offset);
+}
+
+void
+brw_dump_polygon_stipple(const struct brw_polygon_stipple *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.stipple[0] = 0x%x\n", (*ptr).stipple[0]);
+   debug_printf("\t\t.stipple[1] = 0x%x\n", (*ptr).stipple[1]);
+   debug_printf("\t\t.stipple[2] = 0x%x\n", (*ptr).stipple[2]);
+   debug_printf("\t\t.stipple[3] = 0x%x\n", (*ptr).stipple[3]);
+   debug_printf("\t\t.stipple[4] = 0x%x\n", (*ptr).stipple[4]);
+   debug_printf("\t\t.stipple[5] = 0x%x\n", (*ptr).stipple[5]);
+   debug_printf("\t\t.stipple[6] = 0x%x\n", (*ptr).stipple[6]);
+   debug_printf("\t\t.stipple[7] = 0x%x\n", (*ptr).stipple[7]);
+   debug_printf("\t\t.stipple[8] = 0x%x\n", (*ptr).stipple[8]);
+   debug_printf("\t\t.stipple[9] = 0x%x\n", (*ptr).stipple[9]);
+   debug_printf("\t\t.stipple[10] = 0x%x\n", (*ptr).stipple[10]);
+   debug_printf("\t\t.stipple[11] = 0x%x\n", (*ptr).stipple[11]);
+   debug_printf("\t\t.stipple[12] = 0x%x\n", (*ptr).stipple[12]);
+   debug_printf("\t\t.stipple[13] = 0x%x\n", (*ptr).stipple[13]);
+   debug_printf("\t\t.stipple[14] = 0x%x\n", (*ptr).stipple[14]);
+   debug_printf("\t\t.stipple[15] = 0x%x\n", (*ptr).stipple[15]);
+   debug_printf("\t\t.stipple[16] = 0x%x\n", (*ptr).stipple[16]);
+   debug_printf("\t\t.stipple[17] = 0x%x\n", (*ptr).stipple[17]);
+   debug_printf("\t\t.stipple[18] = 0x%x\n", (*ptr).stipple[18]);
+   debug_printf("\t\t.stipple[19] = 0x%x\n", (*ptr).stipple[19]);
+   debug_printf("\t\t.stipple[20] = 0x%x\n", (*ptr).stipple[20]);
+   debug_printf("\t\t.stipple[21] = 0x%x\n", (*ptr).stipple[21]);
+   debug_printf("\t\t.stipple[22] = 0x%x\n", (*ptr).stipple[22]);
+   debug_printf("\t\t.stipple[23] = 0x%x\n", (*ptr).stipple[23]);
+   debug_printf("\t\t.stipple[24] = 0x%x\n", (*ptr).stipple[24]);
+   debug_printf("\t\t.stipple[25] = 0x%x\n", (*ptr).stipple[25]);
+   debug_printf("\t\t.stipple[26] = 0x%x\n", (*ptr).stipple[26]);
+   debug_printf("\t\t.stipple[27] = 0x%x\n", (*ptr).stipple[27]);
+   debug_printf("\t\t.stipple[28] = 0x%x\n", (*ptr).stipple[28]);
+   debug_printf("\t\t.stipple[29] = 0x%x\n", (*ptr).stipple[29]);
+   debug_printf("\t\t.stipple[30] = 0x%x\n", (*ptr).stipple[30]);
+   debug_printf("\t\t.stipple[31] = 0x%x\n", (*ptr).stipple[31]);
+}
+
+void
+brw_dump_polygon_stipple_offset(const struct brw_polygon_stipple_offset *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.bits0.y_offset = 0x%x\n", (*ptr).bits0.y_offset);
+   debug_printf("\t\t.bits0.x_offset = 0x%x\n", (*ptr).bits0.x_offset);
+}
+
+void
+brw_dump_sampler_default_color(const struct brw_sampler_default_color *ptr)
+{
+   debug_printf("\t\t.color[0] = %f\n", (*ptr).color[0]);
+   debug_printf("\t\t.color[1] = %f\n", (*ptr).color[1]);
+   debug_printf("\t\t.color[2] = %f\n", (*ptr).color[2]);
+   debug_printf("\t\t.color[3] = %f\n", (*ptr).color[3]);
+}
+
+void
+brw_dump_sampler_state(const struct brw_sampler_state *ptr)
+{
+   debug_printf("\t\t.ss0.shadow_function = 0x%x\n", (*ptr).ss0.shadow_function);
+   debug_printf("\t\t.ss0.lod_bias = 0x%x\n", (*ptr).ss0.lod_bias);
+   debug_printf("\t\t.ss0.min_filter = 0x%x\n", (*ptr).ss0.min_filter);
+   debug_printf("\t\t.ss0.mag_filter = 0x%x\n", (*ptr).ss0.mag_filter);
+   debug_printf("\t\t.ss0.mip_filter = 0x%x\n", (*ptr).ss0.mip_filter);
+   debug_printf("\t\t.ss0.base_level = 0x%x\n", (*ptr).ss0.base_level);
+   debug_printf("\t\t.ss0.lod_preclamp = 0x%x\n", (*ptr).ss0.lod_preclamp);
+   debug_printf("\t\t.ss0.default_color_mode = 0x%x\n", (*ptr).ss0.default_color_mode);
+   debug_printf("\t\t.ss0.disable = 0x%x\n", (*ptr).ss0.disable);
+   debug_printf("\t\t.ss1.r_wrap_mode = 0x%x\n", (*ptr).ss1.r_wrap_mode);
+   debug_printf("\t\t.ss1.t_wrap_mode = 0x%x\n", (*ptr).ss1.t_wrap_mode);
+   debug_printf("\t\t.ss1.s_wrap_mode = 0x%x\n", (*ptr).ss1.s_wrap_mode);
+   debug_printf("\t\t.ss1.max_lod = 0x%x\n", (*ptr).ss1.max_lod);
+   debug_printf("\t\t.ss1.min_lod = 0x%x\n", (*ptr).ss1.min_lod);
+   debug_printf("\t\t.ss2.default_color_pointer = 0x%x\n", (*ptr).ss2.default_color_pointer);
+   debug_printf("\t\t.ss3.max_aniso = 0x%x\n", (*ptr).ss3.max_aniso);
+   debug_printf("\t\t.ss3.chroma_key_mode = 0x%x\n", (*ptr).ss3.chroma_key_mode);
+   debug_printf("\t\t.ss3.chroma_key_index = 0x%x\n", (*ptr).ss3.chroma_key_index);
+   debug_printf("\t\t.ss3.chroma_key_enable = 0x%x\n", (*ptr).ss3.chroma_key_enable);
+   debug_printf("\t\t.ss3.monochrome_filter_width = 0x%x\n", (*ptr).ss3.monochrome_filter_width);
+   debug_printf("\t\t.ss3.monochrome_filter_height = 0x%x\n", (*ptr).ss3.monochrome_filter_height);
+}
+
+void
+brw_dump_sf_unit_state(const struct brw_sf_unit_state *ptr)
+{
+   debug_printf("\t\t.thread0.grf_reg_count = 0x%x\n", (*ptr).thread0.grf_reg_count);
+   debug_printf("\t\t.thread0.kernel_start_pointer = 0x%x\n", (*ptr).thread0.kernel_start_pointer);
+   debug_printf("\t\t.thread1.ext_halt_exception_enable = 0x%x\n", (*ptr).thread1.ext_halt_exception_enable);
+   debug_printf("\t\t.thread1.sw_exception_enable = 0x%x\n", (*ptr).thread1.sw_exception_enable);
+   debug_printf("\t\t.thread1.mask_stack_exception_enable = 0x%x\n", (*ptr).thread1.mask_stack_exception_enable);
+   debug_printf("\t\t.thread1.timeout_exception_enable = 0x%x\n", (*ptr).thread1.timeout_exception_enable);
+   debug_printf("\t\t.thread1.illegal_op_exception_enable = 0x%x\n", (*ptr).thread1.illegal_op_exception_enable);
+   debug_printf("\t\t.thread1.depth_coef_urb_read_offset = 0x%x\n", (*ptr).thread1.depth_coef_urb_read_offset);
+   debug_printf("\t\t.thread1.floating_point_mode = 0x%x\n", (*ptr).thread1.floating_point_mode);
+   debug_printf("\t\t.thread1.thread_priority = 0x%x\n", (*ptr).thread1.thread_priority);
+   debug_printf("\t\t.thread1.binding_table_entry_count = 0x%x\n", (*ptr).thread1.binding_table_entry_count);
+   debug_printf("\t\t.thread1.single_program_flow = 0x%x\n", (*ptr).thread1.single_program_flow);
+   debug_printf("\t\t.thread2.per_thread_scratch_space = 0x%x\n", (*ptr).thread2.per_thread_scratch_space);
+   debug_printf("\t\t.thread2.scratch_space_base_pointer = 0x%x\n", (*ptr).thread2.scratch_space_base_pointer);
+   debug_printf("\t\t.thread3.dispatch_grf_start_reg = 0x%x\n", (*ptr).thread3.dispatch_grf_start_reg);
+   debug_printf("\t\t.thread3.urb_entry_read_offset = 0x%x\n", (*ptr).thread3.urb_entry_read_offset);
+   debug_printf("\t\t.thread3.urb_entry_read_length = 0x%x\n", (*ptr).thread3.urb_entry_read_length);
+   debug_printf("\t\t.thread3.const_urb_entry_read_offset = 0x%x\n", (*ptr).thread3.const_urb_entry_read_offset);
+   debug_printf("\t\t.thread3.const_urb_entry_read_length = 0x%x\n", (*ptr).thread3.const_urb_entry_read_length);
+   debug_printf("\t\t.thread4.stats_enable = 0x%x\n", (*ptr).thread4.stats_enable);
+   debug_printf("\t\t.thread4.nr_urb_entries = 0x%x\n", (*ptr).thread4.nr_urb_entries);
+   debug_printf("\t\t.thread4.urb_entry_allocation_size = 0x%x\n", (*ptr).thread4.urb_entry_allocation_size);
+   debug_printf("\t\t.thread4.max_threads = 0x%x\n", (*ptr).thread4.max_threads);
+   debug_printf("\t\t.sf5.front_winding = 0x%x\n", (*ptr).sf5.front_winding);
+   debug_printf("\t\t.sf5.viewport_transform = 0x%x\n", (*ptr).sf5.viewport_transform);
+   debug_printf("\t\t.sf5.sf_viewport_state_offset = 0x%x\n", (*ptr).sf5.sf_viewport_state_offset);
+   debug_printf("\t\t.sf6.dest_org_vbias = 0x%x\n", (*ptr).sf6.dest_org_vbias);
+   debug_printf("\t\t.sf6.dest_org_hbias = 0x%x\n", (*ptr).sf6.dest_org_hbias);
+   debug_printf("\t\t.sf6.scissor = 0x%x\n", (*ptr).sf6.scissor);
+   debug_printf("\t\t.sf6.disable_2x2_trifilter = 0x%x\n", (*ptr).sf6.disable_2x2_trifilter);
+   debug_printf("\t\t.sf6.disable_zero_pix_trifilter = 0x%x\n", (*ptr).sf6.disable_zero_pix_trifilter);
+   debug_printf("\t\t.sf6.point_rast_rule = 0x%x\n", (*ptr).sf6.point_rast_rule);
+   debug_printf("\t\t.sf6.line_endcap_aa_region_width = 0x%x\n", (*ptr).sf6.line_endcap_aa_region_width);
+   debug_printf("\t\t.sf6.line_width = 0x%x\n", (*ptr).sf6.line_width);
+   debug_printf("\t\t.sf6.fast_scissor_disable = 0x%x\n", (*ptr).sf6.fast_scissor_disable);
+   debug_printf("\t\t.sf6.cull_mode = 0x%x\n", (*ptr).sf6.cull_mode);
+   debug_printf("\t\t.sf6.aa_enable = 0x%x\n", (*ptr).sf6.aa_enable);
+   debug_printf("\t\t.sf7.point_size = 0x%x\n", (*ptr).sf7.point_size);
+   debug_printf("\t\t.sf7.use_point_size_state = 0x%x\n", (*ptr).sf7.use_point_size_state);
+   debug_printf("\t\t.sf7.subpixel_precision = 0x%x\n", (*ptr).sf7.subpixel_precision);
+   debug_printf("\t\t.sf7.sprite_point = 0x%x\n", (*ptr).sf7.sprite_point);
+   debug_printf("\t\t.sf7.aa_line_distance_mode = 0x%x\n", (*ptr).sf7.aa_line_distance_mode);
+   debug_printf("\t\t.sf7.trifan_pv = 0x%x\n", (*ptr).sf7.trifan_pv);
+   debug_printf("\t\t.sf7.linestrip_pv = 0x%x\n", (*ptr).sf7.linestrip_pv);
+   debug_printf("\t\t.sf7.tristrip_pv = 0x%x\n", (*ptr).sf7.tristrip_pv);
+   debug_printf("\t\t.sf7.line_last_pixel_enable = 0x%x\n", (*ptr).sf7.line_last_pixel_enable);
+}
+
+void
+brw_dump_sf_viewport(const struct brw_sf_viewport *ptr)
+{
+   debug_printf("\t\t.viewport.m00 = %f\n", (*ptr).viewport.m00);
+   debug_printf("\t\t.viewport.m11 = %f\n", (*ptr).viewport.m11);
+   debug_printf("\t\t.viewport.m22 = %f\n", (*ptr).viewport.m22);
+   debug_printf("\t\t.viewport.m30 = %f\n", (*ptr).viewport.m30);
+   debug_printf("\t\t.viewport.m31 = %f\n", (*ptr).viewport.m31);
+   debug_printf("\t\t.viewport.m32 = %f\n", (*ptr).viewport.m32);
+   debug_printf("\t\t.scissor.xmin = 0x%x\n", (*ptr).scissor.xmin);
+   debug_printf("\t\t.scissor.ymin = 0x%x\n", (*ptr).scissor.ymin);
+   debug_printf("\t\t.scissor.xmax = 0x%x\n", (*ptr).scissor.xmax);
+   debug_printf("\t\t.scissor.ymax = 0x%x\n", (*ptr).scissor.ymax);
+}
+
+void
+brw_dump_ss0(const struct brw_ss0 *ptr)
+{
+   debug_printf("\t\t.shadow_function = 0x%x\n", (*ptr).shadow_function);
+   debug_printf("\t\t.lod_bias = 0x%x\n", (*ptr).lod_bias);
+   debug_printf("\t\t.min_filter = 0x%x\n", (*ptr).min_filter);
+   debug_printf("\t\t.mag_filter = 0x%x\n", (*ptr).mag_filter);
+   debug_printf("\t\t.mip_filter = 0x%x\n", (*ptr).mip_filter);
+   debug_printf("\t\t.base_level = 0x%x\n", (*ptr).base_level);
+   debug_printf("\t\t.lod_preclamp = 0x%x\n", (*ptr).lod_preclamp);
+   debug_printf("\t\t.default_color_mode = 0x%x\n", (*ptr).default_color_mode);
+   debug_printf("\t\t.disable = 0x%x\n", (*ptr).disable);
+}
+
+void
+brw_dump_ss1(const struct brw_ss1 *ptr)
+{
+   debug_printf("\t\t.r_wrap_mode = 0x%x\n", (*ptr).r_wrap_mode);
+   debug_printf("\t\t.t_wrap_mode = 0x%x\n", (*ptr).t_wrap_mode);
+   debug_printf("\t\t.s_wrap_mode = 0x%x\n", (*ptr).s_wrap_mode);
+   debug_printf("\t\t.max_lod = 0x%x\n", (*ptr).max_lod);
+   debug_printf("\t\t.min_lod = 0x%x\n", (*ptr).min_lod);
+}
+
+void
+brw_dump_ss2(const struct brw_ss2 *ptr)
+{
+   debug_printf("\t\t.default_color_pointer = 0x%x\n", (*ptr).default_color_pointer);
+}
+
+void
+brw_dump_ss3(const struct brw_ss3 *ptr)
+{
+   debug_printf("\t\t.max_aniso = 0x%x\n", (*ptr).max_aniso);
+   debug_printf("\t\t.chroma_key_mode = 0x%x\n", (*ptr).chroma_key_mode);
+   debug_printf("\t\t.chroma_key_index = 0x%x\n", (*ptr).chroma_key_index);
+   debug_printf("\t\t.chroma_key_enable = 0x%x\n", (*ptr).chroma_key_enable);
+   debug_printf("\t\t.monochrome_filter_width = 0x%x\n", (*ptr).monochrome_filter_width);
+   debug_printf("\t\t.monochrome_filter_height = 0x%x\n", (*ptr).monochrome_filter_height);
+}
+
+void
+brw_dump_state_base_address(const struct brw_state_base_address *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.bits0.modify_enable = 0x%x\n", (*ptr).bits0.modify_enable);
+   debug_printf("\t\t.bits0.general_state_address = 0x%x\n", (*ptr).bits0.general_state_address);
+   debug_printf("\t\t.bits1.modify_enable = 0x%x\n", (*ptr).bits1.modify_enable);
+   debug_printf("\t\t.bits1.surface_state_address = 0x%x\n", (*ptr).bits1.surface_state_address);
+   debug_printf("\t\t.bits2.modify_enable = 0x%x\n", (*ptr).bits2.modify_enable);
+   debug_printf("\t\t.bits2.indirect_object_state_address = 0x%x\n", (*ptr).bits2.indirect_object_state_address);
+   debug_printf("\t\t.bits3.modify_enable = 0x%x\n", (*ptr).bits3.modify_enable);
+   debug_printf("\t\t.bits3.general_state_upper_bound = 0x%x\n", (*ptr).bits3.general_state_upper_bound);
+   debug_printf("\t\t.bits4.modify_enable = 0x%x\n", (*ptr).bits4.modify_enable);
+   debug_printf("\t\t.bits4.indirect_object_state_upper_bound = 0x%x\n", (*ptr).bits4.indirect_object_state_upper_bound);
+}
+
+void
+brw_dump_state_prefetch(const struct brw_state_prefetch *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.bits0.prefetch_count = 0x%x\n", (*ptr).bits0.prefetch_count);
+   debug_printf("\t\t.bits0.prefetch_pointer = 0x%x\n", (*ptr).bits0.prefetch_pointer);
+}
+
+void
+brw_dump_surf_ss0(const struct brw_surf_ss0 *ptr)
+{
+   debug_printf("\t\t.cube_pos_z = 0x%x\n", (*ptr).cube_pos_z);
+   debug_printf("\t\t.cube_neg_z = 0x%x\n", (*ptr).cube_neg_z);
+   debug_printf("\t\t.cube_pos_y = 0x%x\n", (*ptr).cube_pos_y);
+   debug_printf("\t\t.cube_neg_y = 0x%x\n", (*ptr).cube_neg_y);
+   debug_printf("\t\t.cube_pos_x = 0x%x\n", (*ptr).cube_pos_x);
+   debug_printf("\t\t.cube_neg_x = 0x%x\n", (*ptr).cube_neg_x);
+   debug_printf("\t\t.mipmap_layout_mode = 0x%x\n", (*ptr).mipmap_layout_mode);
+   debug_printf("\t\t.vert_line_stride_ofs = 0x%x\n", (*ptr).vert_line_stride_ofs);
+   debug_printf("\t\t.vert_line_stride = 0x%x\n", (*ptr).vert_line_stride);
+   debug_printf("\t\t.color_blend = 0x%x\n", (*ptr).color_blend);
+   debug_printf("\t\t.writedisable_blue = 0x%x\n", (*ptr).writedisable_blue);
+   debug_printf("\t\t.writedisable_green = 0x%x\n", (*ptr).writedisable_green);
+   debug_printf("\t\t.writedisable_red = 0x%x\n", (*ptr).writedisable_red);
+   debug_printf("\t\t.writedisable_alpha = 0x%x\n", (*ptr).writedisable_alpha);
+   debug_printf("\t\t.surface_format = 0x%x\n", (*ptr).surface_format);
+   debug_printf("\t\t.data_return_format = 0x%x\n", (*ptr).data_return_format);
+   debug_printf("\t\t.surface_type = 0x%x\n", (*ptr).surface_type);
+}
+
+void
+brw_dump_surf_ss1(const struct brw_surf_ss1 *ptr)
+{
+   debug_printf("\t\t.base_addr = 0x%x\n", (*ptr).base_addr);
+}
+
+void
+brw_dump_surf_ss2(const struct brw_surf_ss2 *ptr)
+{
+   debug_printf("\t\t.mip_count = 0x%x\n", (*ptr).mip_count);
+   debug_printf("\t\t.width = 0x%x\n", (*ptr).width);
+   debug_printf("\t\t.height = 0x%x\n", (*ptr).height);
+}
+
+void
+brw_dump_surf_ss3(const struct brw_surf_ss3 *ptr)
+{
+   debug_printf("\t\t.tile_walk = 0x%x\n", (*ptr).tile_walk);
+   debug_printf("\t\t.tiled_surface = 0x%x\n", (*ptr).tiled_surface);
+   debug_printf("\t\t.pitch = 0x%x\n", (*ptr).pitch);
+   debug_printf("\t\t.depth = 0x%x\n", (*ptr).depth);
+}
+
+void
+brw_dump_surf_ss4(const struct brw_surf_ss4 *ptr)
+{
+   debug_printf("\t\t.multisample_position_palette_index = 0x%x\n", (*ptr).multisample_position_palette_index);
+   debug_printf("\t\t.num_multisamples = 0x%x\n", (*ptr).num_multisamples);
+   debug_printf("\t\t.render_target_view_extent = 0x%x\n", (*ptr).render_target_view_extent);
+   debug_printf("\t\t.min_array_elt = 0x%x\n", (*ptr).min_array_elt);
+   debug_printf("\t\t.min_lod = 0x%x\n", (*ptr).min_lod);
+}
+
+void
+brw_dump_surf_ss5(const struct brw_surf_ss5 *ptr)
+{
+   debug_printf("\t\t.llc_mapping = 0x%x\n", (*ptr).llc_mapping);
+   debug_printf("\t\t.mlc_mapping = 0x%x\n", (*ptr).mlc_mapping);
+   debug_printf("\t\t.gfdt = 0x%x\n", (*ptr).gfdt);
+   debug_printf("\t\t.gfdt_src = 0x%x\n", (*ptr).gfdt_src);
+   debug_printf("\t\t.y_offset = 0x%x\n", (*ptr).y_offset);
+   debug_printf("\t\t.x_offset = 0x%x\n", (*ptr).x_offset);
+}
+
+void
+brw_dump_surface_state(const struct brw_surface_state *ptr)
+{
+   debug_printf("\t\t.ss0.cube_pos_z = 0x%x\n", (*ptr).ss0.cube_pos_z);
+   debug_printf("\t\t.ss0.cube_neg_z = 0x%x\n", (*ptr).ss0.cube_neg_z);
+   debug_printf("\t\t.ss0.cube_pos_y = 0x%x\n", (*ptr).ss0.cube_pos_y);
+   debug_printf("\t\t.ss0.cube_neg_y = 0x%x\n", (*ptr).ss0.cube_neg_y);
+   debug_printf("\t\t.ss0.cube_pos_x = 0x%x\n", (*ptr).ss0.cube_pos_x);
+   debug_printf("\t\t.ss0.cube_neg_x = 0x%x\n", (*ptr).ss0.cube_neg_x);
+   debug_printf("\t\t.ss0.mipmap_layout_mode = 0x%x\n", (*ptr).ss0.mipmap_layout_mode);
+   debug_printf("\t\t.ss0.vert_line_stride_ofs = 0x%x\n", (*ptr).ss0.vert_line_stride_ofs);
+   debug_printf("\t\t.ss0.vert_line_stride = 0x%x\n", (*ptr).ss0.vert_line_stride);
+   debug_printf("\t\t.ss0.color_blend = 0x%x\n", (*ptr).ss0.color_blend);
+   debug_printf("\t\t.ss0.writedisable_blue = 0x%x\n", (*ptr).ss0.writedisable_blue);
+   debug_printf("\t\t.ss0.writedisable_green = 0x%x\n", (*ptr).ss0.writedisable_green);
+   debug_printf("\t\t.ss0.writedisable_red = 0x%x\n", (*ptr).ss0.writedisable_red);
+   debug_printf("\t\t.ss0.writedisable_alpha = 0x%x\n", (*ptr).ss0.writedisable_alpha);
+   debug_printf("\t\t.ss0.surface_format = 0x%x\n", (*ptr).ss0.surface_format);
+   debug_printf("\t\t.ss0.data_return_format = 0x%x\n", (*ptr).ss0.data_return_format);
+   debug_printf("\t\t.ss0.surface_type = 0x%x\n", (*ptr).ss0.surface_type);
+   debug_printf("\t\t.ss1.base_addr = 0x%x\n", (*ptr).ss1.base_addr);
+   debug_printf("\t\t.ss2.mip_count = 0x%x\n", (*ptr).ss2.mip_count);
+   debug_printf("\t\t.ss2.width = 0x%x\n", (*ptr).ss2.width);
+   debug_printf("\t\t.ss2.height = 0x%x\n", (*ptr).ss2.height);
+   debug_printf("\t\t.ss3.tile_walk = 0x%x\n", (*ptr).ss3.tile_walk);
+   debug_printf("\t\t.ss3.tiled_surface = 0x%x\n", (*ptr).ss3.tiled_surface);
+   debug_printf("\t\t.ss3.pitch = 0x%x\n", (*ptr).ss3.pitch);
+   debug_printf("\t\t.ss3.depth = 0x%x\n", (*ptr).ss3.depth);
+   debug_printf("\t\t.ss4.multisample_position_palette_index = 0x%x\n", (*ptr).ss4.multisample_position_palette_index);
+   debug_printf("\t\t.ss4.num_multisamples = 0x%x\n", (*ptr).ss4.num_multisamples);
+   debug_printf("\t\t.ss4.render_target_view_extent = 0x%x\n", (*ptr).ss4.render_target_view_extent);
+   debug_printf("\t\t.ss4.min_array_elt = 0x%x\n", (*ptr).ss4.min_array_elt);
+   debug_printf("\t\t.ss4.min_lod = 0x%x\n", (*ptr).ss4.min_lod);
+   debug_printf("\t\t.ss5.llc_mapping = 0x%x\n", (*ptr).ss5.llc_mapping);
+   debug_printf("\t\t.ss5.mlc_mapping = 0x%x\n", (*ptr).ss5.mlc_mapping);
+   debug_printf("\t\t.ss5.gfdt = 0x%x\n", (*ptr).ss5.gfdt);
+   debug_printf("\t\t.ss5.gfdt_src = 0x%x\n", (*ptr).ss5.gfdt_src);
+   debug_printf("\t\t.ss5.y_offset = 0x%x\n", (*ptr).ss5.y_offset);
+   debug_printf("\t\t.ss5.x_offset = 0x%x\n", (*ptr).ss5.x_offset);
+}
+
+void
+brw_dump_system_instruction_pointer(const struct brw_system_instruction_pointer *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.bits0.system_instruction_pointer = 0x%x\n", (*ptr).bits0.system_instruction_pointer);
+}
+
+void
+brw_dump_urb_fence(const struct brw_urb_fence *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.vs_realloc = 0x%x\n", (*ptr).header.vs_realloc);
+   debug_printf("\t\t.header.gs_realloc = 0x%x\n", (*ptr).header.gs_realloc);
+   debug_printf("\t\t.header.clp_realloc = 0x%x\n", (*ptr).header.clp_realloc);
+   debug_printf("\t\t.header.sf_realloc = 0x%x\n", (*ptr).header.sf_realloc);
+   debug_printf("\t\t.header.vfe_realloc = 0x%x\n", (*ptr).header.vfe_realloc);
+   debug_printf("\t\t.header.cs_realloc = 0x%x\n", (*ptr).header.cs_realloc);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.bits0.vs_fence = 0x%x\n", (*ptr).bits0.vs_fence);
+   debug_printf("\t\t.bits0.gs_fence = 0x%x\n", (*ptr).bits0.gs_fence);
+   debug_printf("\t\t.bits0.clp_fence = 0x%x\n", (*ptr).bits0.clp_fence);
+   debug_printf("\t\t.bits1.sf_fence = 0x%x\n", (*ptr).bits1.sf_fence);
+   debug_printf("\t\t.bits1.vf_fence = 0x%x\n", (*ptr).bits1.vf_fence);
+   debug_printf("\t\t.bits1.cs_fence = 0x%x\n", (*ptr).bits1.cs_fence);
+}
+
+void
+brw_dump_urb_immediate(const struct brw_urb_immediate *ptr)
+{
+   debug_printf("\t\t.opcode = 0x%x\n", (*ptr).opcode);
+   debug_printf("\t\t.offset = 0x%x\n", (*ptr).offset);
+   debug_printf("\t\t.swizzle_control = 0x%x\n", (*ptr).swizzle_control);
+   debug_printf("\t\t.allocate = 0x%x\n", (*ptr).allocate);
+   debug_printf("\t\t.used = 0x%x\n", (*ptr).used);
+   debug_printf("\t\t.complete = 0x%x\n", (*ptr).complete);
+   debug_printf("\t\t.response_length = 0x%x\n", (*ptr).response_length);
+   debug_printf("\t\t.msg_length = 0x%x\n", (*ptr).msg_length);
+   debug_printf("\t\t.msg_target = 0x%x\n", (*ptr).msg_target);
+   debug_printf("\t\t.end_of_thread = 0x%x\n", (*ptr).end_of_thread);
+}
+
+void
+brw_dump_vb_array_state(const struct brw_vb_array_state *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.vb[0].vb0.pitch = 0x%x\n", (*ptr).vb[0].vb0.pitch);
+   debug_printf("\t\t.vb[0].vb0.access_type = 0x%x\n", (*ptr).vb[0].vb0.access_type);
+   debug_printf("\t\t.vb[0].vb0.vb_index = 0x%x\n", (*ptr).vb[0].vb0.vb_index);
+   debug_printf("\t\t.vb[0].start_addr = 0x%x\n", (*ptr).vb[0].start_addr);
+   debug_printf("\t\t.vb[0].max_index = 0x%x\n", (*ptr).vb[0].max_index);
+   debug_printf("\t\t.vb[0].instance_data_step_rate = 0x%x\n", (*ptr).vb[0].instance_data_step_rate);
+   debug_printf("\t\t.vb[1].vb0.pitch = 0x%x\n", (*ptr).vb[1].vb0.pitch);
+   debug_printf("\t\t.vb[1].vb0.access_type = 0x%x\n", (*ptr).vb[1].vb0.access_type);
+   debug_printf("\t\t.vb[1].vb0.vb_index = 0x%x\n", (*ptr).vb[1].vb0.vb_index);
+   debug_printf("\t\t.vb[1].start_addr = 0x%x\n", (*ptr).vb[1].start_addr);
+   debug_printf("\t\t.vb[1].max_index = 0x%x\n", (*ptr).vb[1].max_index);
+   debug_printf("\t\t.vb[1].instance_data_step_rate = 0x%x\n", (*ptr).vb[1].instance_data_step_rate);
+   debug_printf("\t\t.vb[2].vb0.pitch = 0x%x\n", (*ptr).vb[2].vb0.pitch);
+   debug_printf("\t\t.vb[2].vb0.access_type = 0x%x\n", (*ptr).vb[2].vb0.access_type);
+   debug_printf("\t\t.vb[2].vb0.vb_index = 0x%x\n", (*ptr).vb[2].vb0.vb_index);
+   debug_printf("\t\t.vb[2].start_addr = 0x%x\n", (*ptr).vb[2].start_addr);
+   debug_printf("\t\t.vb[2].max_index = 0x%x\n", (*ptr).vb[2].max_index);
+   debug_printf("\t\t.vb[2].instance_data_step_rate = 0x%x\n", (*ptr).vb[2].instance_data_step_rate);
+   debug_printf("\t\t.vb[3].vb0.pitch = 0x%x\n", (*ptr).vb[3].vb0.pitch);
+   debug_printf("\t\t.vb[3].vb0.access_type = 0x%x\n", (*ptr).vb[3].vb0.access_type);
+   debug_printf("\t\t.vb[3].vb0.vb_index = 0x%x\n", (*ptr).vb[3].vb0.vb_index);
+   debug_printf("\t\t.vb[3].start_addr = 0x%x\n", (*ptr).vb[3].start_addr);
+   debug_printf("\t\t.vb[3].max_index = 0x%x\n", (*ptr).vb[3].max_index);
+   debug_printf("\t\t.vb[3].instance_data_step_rate = 0x%x\n", (*ptr).vb[3].instance_data_step_rate);
+   debug_printf("\t\t.vb[4].vb0.pitch = 0x%x\n", (*ptr).vb[4].vb0.pitch);
+   debug_printf("\t\t.vb[4].vb0.access_type = 0x%x\n", (*ptr).vb[4].vb0.access_type);
+   debug_printf("\t\t.vb[4].vb0.vb_index = 0x%x\n", (*ptr).vb[4].vb0.vb_index);
+   debug_printf("\t\t.vb[4].start_addr = 0x%x\n", (*ptr).vb[4].start_addr);
+   debug_printf("\t\t.vb[4].max_index = 0x%x\n", (*ptr).vb[4].max_index);
+   debug_printf("\t\t.vb[4].instance_data_step_rate = 0x%x\n", (*ptr).vb[4].instance_data_step_rate);
+   debug_printf("\t\t.vb[5].vb0.pitch = 0x%x\n", (*ptr).vb[5].vb0.pitch);
+   debug_printf("\t\t.vb[5].vb0.access_type = 0x%x\n", (*ptr).vb[5].vb0.access_type);
+   debug_printf("\t\t.vb[5].vb0.vb_index = 0x%x\n", (*ptr).vb[5].vb0.vb_index);
+   debug_printf("\t\t.vb[5].start_addr = 0x%x\n", (*ptr).vb[5].start_addr);
+   debug_printf("\t\t.vb[5].max_index = 0x%x\n", (*ptr).vb[5].max_index);
+   debug_printf("\t\t.vb[5].instance_data_step_rate = 0x%x\n", (*ptr).vb[5].instance_data_step_rate);
+   debug_printf("\t\t.vb[6].vb0.pitch = 0x%x\n", (*ptr).vb[6].vb0.pitch);
+   debug_printf("\t\t.vb[6].vb0.access_type = 0x%x\n", (*ptr).vb[6].vb0.access_type);
+   debug_printf("\t\t.vb[6].vb0.vb_index = 0x%x\n", (*ptr).vb[6].vb0.vb_index);
+   debug_printf("\t\t.vb[6].start_addr = 0x%x\n", (*ptr).vb[6].start_addr);
+   debug_printf("\t\t.vb[6].max_index = 0x%x\n", (*ptr).vb[6].max_index);
+   debug_printf("\t\t.vb[6].instance_data_step_rate = 0x%x\n", (*ptr).vb[6].instance_data_step_rate);
+   debug_printf("\t\t.vb[7].vb0.pitch = 0x%x\n", (*ptr).vb[7].vb0.pitch);
+   debug_printf("\t\t.vb[7].vb0.access_type = 0x%x\n", (*ptr).vb[7].vb0.access_type);
+   debug_printf("\t\t.vb[7].vb0.vb_index = 0x%x\n", (*ptr).vb[7].vb0.vb_index);
+   debug_printf("\t\t.vb[7].start_addr = 0x%x\n", (*ptr).vb[7].start_addr);
+   debug_printf("\t\t.vb[7].max_index = 0x%x\n", (*ptr).vb[7].max_index);
+   debug_printf("\t\t.vb[7].instance_data_step_rate = 0x%x\n", (*ptr).vb[7].instance_data_step_rate);
+   debug_printf("\t\t.vb[8].vb0.pitch = 0x%x\n", (*ptr).vb[8].vb0.pitch);
+   debug_printf("\t\t.vb[8].vb0.access_type = 0x%x\n", (*ptr).vb[8].vb0.access_type);
+   debug_printf("\t\t.vb[8].vb0.vb_index = 0x%x\n", (*ptr).vb[8].vb0.vb_index);
+   debug_printf("\t\t.vb[8].start_addr = 0x%x\n", (*ptr).vb[8].start_addr);
+   debug_printf("\t\t.vb[8].max_index = 0x%x\n", (*ptr).vb[8].max_index);
+   debug_printf("\t\t.vb[8].instance_data_step_rate = 0x%x\n", (*ptr).vb[8].instance_data_step_rate);
+   debug_printf("\t\t.vb[9].vb0.pitch = 0x%x\n", (*ptr).vb[9].vb0.pitch);
+   debug_printf("\t\t.vb[9].vb0.access_type = 0x%x\n", (*ptr).vb[9].vb0.access_type);
+   debug_printf("\t\t.vb[9].vb0.vb_index = 0x%x\n", (*ptr).vb[9].vb0.vb_index);
+   debug_printf("\t\t.vb[9].start_addr = 0x%x\n", (*ptr).vb[9].start_addr);
+   debug_printf("\t\t.vb[9].max_index = 0x%x\n", (*ptr).vb[9].max_index);
+   debug_printf("\t\t.vb[9].instance_data_step_rate = 0x%x\n", (*ptr).vb[9].instance_data_step_rate);
+   debug_printf("\t\t.vb[10].vb0.pitch = 0x%x\n", (*ptr).vb[10].vb0.pitch);
+   debug_printf("\t\t.vb[10].vb0.access_type = 0x%x\n", (*ptr).vb[10].vb0.access_type);
+   debug_printf("\t\t.vb[10].vb0.vb_index = 0x%x\n", (*ptr).vb[10].vb0.vb_index);
+   debug_printf("\t\t.vb[10].start_addr = 0x%x\n", (*ptr).vb[10].start_addr);
+   debug_printf("\t\t.vb[10].max_index = 0x%x\n", (*ptr).vb[10].max_index);
+   debug_printf("\t\t.vb[10].instance_data_step_rate = 0x%x\n", (*ptr).vb[10].instance_data_step_rate);
+   debug_printf("\t\t.vb[11].vb0.pitch = 0x%x\n", (*ptr).vb[11].vb0.pitch);
+   debug_printf("\t\t.vb[11].vb0.access_type = 0x%x\n", (*ptr).vb[11].vb0.access_type);
+   debug_printf("\t\t.vb[11].vb0.vb_index = 0x%x\n", (*ptr).vb[11].vb0.vb_index);
+   debug_printf("\t\t.vb[11].start_addr = 0x%x\n", (*ptr).vb[11].start_addr);
+   debug_printf("\t\t.vb[11].max_index = 0x%x\n", (*ptr).vb[11].max_index);
+   debug_printf("\t\t.vb[11].instance_data_step_rate = 0x%x\n", (*ptr).vb[11].instance_data_step_rate);
+   debug_printf("\t\t.vb[12].vb0.pitch = 0x%x\n", (*ptr).vb[12].vb0.pitch);
+   debug_printf("\t\t.vb[12].vb0.access_type = 0x%x\n", (*ptr).vb[12].vb0.access_type);
+   debug_printf("\t\t.vb[12].vb0.vb_index = 0x%x\n", (*ptr).vb[12].vb0.vb_index);
+   debug_printf("\t\t.vb[12].start_addr = 0x%x\n", (*ptr).vb[12].start_addr);
+   debug_printf("\t\t.vb[12].max_index = 0x%x\n", (*ptr).vb[12].max_index);
+   debug_printf("\t\t.vb[12].instance_data_step_rate = 0x%x\n", (*ptr).vb[12].instance_data_step_rate);
+   debug_printf("\t\t.vb[13].vb0.pitch = 0x%x\n", (*ptr).vb[13].vb0.pitch);
+   debug_printf("\t\t.vb[13].vb0.access_type = 0x%x\n", (*ptr).vb[13].vb0.access_type);
+   debug_printf("\t\t.vb[13].vb0.vb_index = 0x%x\n", (*ptr).vb[13].vb0.vb_index);
+   debug_printf("\t\t.vb[13].start_addr = 0x%x\n", (*ptr).vb[13].start_addr);
+   debug_printf("\t\t.vb[13].max_index = 0x%x\n", (*ptr).vb[13].max_index);
+   debug_printf("\t\t.vb[13].instance_data_step_rate = 0x%x\n", (*ptr).vb[13].instance_data_step_rate);
+   debug_printf("\t\t.vb[14].vb0.pitch = 0x%x\n", (*ptr).vb[14].vb0.pitch);
+   debug_printf("\t\t.vb[14].vb0.access_type = 0x%x\n", (*ptr).vb[14].vb0.access_type);
+   debug_printf("\t\t.vb[14].vb0.vb_index = 0x%x\n", (*ptr).vb[14].vb0.vb_index);
+   debug_printf("\t\t.vb[14].start_addr = 0x%x\n", (*ptr).vb[14].start_addr);
+   debug_printf("\t\t.vb[14].max_index = 0x%x\n", (*ptr).vb[14].max_index);
+   debug_printf("\t\t.vb[14].instance_data_step_rate = 0x%x\n", (*ptr).vb[14].instance_data_step_rate);
+   debug_printf("\t\t.vb[15].vb0.pitch = 0x%x\n", (*ptr).vb[15].vb0.pitch);
+   debug_printf("\t\t.vb[15].vb0.access_type = 0x%x\n", (*ptr).vb[15].vb0.access_type);
+   debug_printf("\t\t.vb[15].vb0.vb_index = 0x%x\n", (*ptr).vb[15].vb0.vb_index);
+   debug_printf("\t\t.vb[15].start_addr = 0x%x\n", (*ptr).vb[15].start_addr);
+   debug_printf("\t\t.vb[15].max_index = 0x%x\n", (*ptr).vb[15].max_index);
+   debug_printf("\t\t.vb[15].instance_data_step_rate = 0x%x\n", (*ptr).vb[15].instance_data_step_rate);
+   debug_printf("\t\t.vb[16].vb0.pitch = 0x%x\n", (*ptr).vb[16].vb0.pitch);
+   debug_printf("\t\t.vb[16].vb0.access_type = 0x%x\n", (*ptr).vb[16].vb0.access_type);
+   debug_printf("\t\t.vb[16].vb0.vb_index = 0x%x\n", (*ptr).vb[16].vb0.vb_index);
+   debug_printf("\t\t.vb[16].start_addr = 0x%x\n", (*ptr).vb[16].start_addr);
+   debug_printf("\t\t.vb[16].max_index = 0x%x\n", (*ptr).vb[16].max_index);
+   debug_printf("\t\t.vb[16].instance_data_step_rate = 0x%x\n", (*ptr).vb[16].instance_data_step_rate);
+}
+
+void
+brw_dump_vertex_buffer_state(const struct brw_vertex_buffer_state *ptr)
+{
+   debug_printf("\t\t.vb0.pitch = 0x%x\n", (*ptr).vb0.pitch);
+   debug_printf("\t\t.vb0.access_type = 0x%x\n", (*ptr).vb0.access_type);
+   debug_printf("\t\t.vb0.vb_index = 0x%x\n", (*ptr).vb0.vb_index);
+   debug_printf("\t\t.start_addr = 0x%x\n", (*ptr).start_addr);
+   debug_printf("\t\t.max_index = 0x%x\n", (*ptr).max_index);
+   debug_printf("\t\t.instance_data_step_rate = 0x%x\n", (*ptr).instance_data_step_rate);
+}
+
+void
+brw_dump_vertex_element_packet(const struct brw_vertex_element_packet *ptr)
+{
+   debug_printf("\t\t.header.length = 0x%x\n", (*ptr).header.length);
+   debug_printf("\t\t.header.opcode = 0x%x\n", (*ptr).header.opcode);
+   debug_printf("\t\t.ve[0].ve0.src_offset = 0x%x\n", (*ptr).ve[0].ve0.src_offset);
+   debug_printf("\t\t.ve[0].ve0.src_format = 0x%x\n", (*ptr).ve[0].ve0.src_format);
+   debug_printf("\t\t.ve[0].ve0.valid = 0x%x\n", (*ptr).ve[0].ve0.valid);
+   debug_printf("\t\t.ve[0].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[0].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[0].ve1.dst_offset = 0x%x\n", (*ptr).ve[0].ve1.dst_offset);
+   debug_printf("\t\t.ve[0].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[0].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[0].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[0].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[0].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[0].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[0].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[0].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[1].ve0.src_offset = 0x%x\n", (*ptr).ve[1].ve0.src_offset);
+   debug_printf("\t\t.ve[1].ve0.src_format = 0x%x\n", (*ptr).ve[1].ve0.src_format);
+   debug_printf("\t\t.ve[1].ve0.valid = 0x%x\n", (*ptr).ve[1].ve0.valid);
+   debug_printf("\t\t.ve[1].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[1].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[1].ve1.dst_offset = 0x%x\n", (*ptr).ve[1].ve1.dst_offset);
+   debug_printf("\t\t.ve[1].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[1].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[1].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[1].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[1].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[1].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[1].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[1].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[2].ve0.src_offset = 0x%x\n", (*ptr).ve[2].ve0.src_offset);
+   debug_printf("\t\t.ve[2].ve0.src_format = 0x%x\n", (*ptr).ve[2].ve0.src_format);
+   debug_printf("\t\t.ve[2].ve0.valid = 0x%x\n", (*ptr).ve[2].ve0.valid);
+   debug_printf("\t\t.ve[2].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[2].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[2].ve1.dst_offset = 0x%x\n", (*ptr).ve[2].ve1.dst_offset);
+   debug_printf("\t\t.ve[2].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[2].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[2].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[2].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[2].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[2].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[2].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[2].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[3].ve0.src_offset = 0x%x\n", (*ptr).ve[3].ve0.src_offset);
+   debug_printf("\t\t.ve[3].ve0.src_format = 0x%x\n", (*ptr).ve[3].ve0.src_format);
+   debug_printf("\t\t.ve[3].ve0.valid = 0x%x\n", (*ptr).ve[3].ve0.valid);
+   debug_printf("\t\t.ve[3].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[3].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[3].ve1.dst_offset = 0x%x\n", (*ptr).ve[3].ve1.dst_offset);
+   debug_printf("\t\t.ve[3].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[3].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[3].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[3].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[3].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[3].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[3].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[3].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[4].ve0.src_offset = 0x%x\n", (*ptr).ve[4].ve0.src_offset);
+   debug_printf("\t\t.ve[4].ve0.src_format = 0x%x\n", (*ptr).ve[4].ve0.src_format);
+   debug_printf("\t\t.ve[4].ve0.valid = 0x%x\n", (*ptr).ve[4].ve0.valid);
+   debug_printf("\t\t.ve[4].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[4].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[4].ve1.dst_offset = 0x%x\n", (*ptr).ve[4].ve1.dst_offset);
+   debug_printf("\t\t.ve[4].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[4].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[4].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[4].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[4].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[4].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[4].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[4].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[5].ve0.src_offset = 0x%x\n", (*ptr).ve[5].ve0.src_offset);
+   debug_printf("\t\t.ve[5].ve0.src_format = 0x%x\n", (*ptr).ve[5].ve0.src_format);
+   debug_printf("\t\t.ve[5].ve0.valid = 0x%x\n", (*ptr).ve[5].ve0.valid);
+   debug_printf("\t\t.ve[5].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[5].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[5].ve1.dst_offset = 0x%x\n", (*ptr).ve[5].ve1.dst_offset);
+   debug_printf("\t\t.ve[5].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[5].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[5].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[5].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[5].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[5].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[5].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[5].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[6].ve0.src_offset = 0x%x\n", (*ptr).ve[6].ve0.src_offset);
+   debug_printf("\t\t.ve[6].ve0.src_format = 0x%x\n", (*ptr).ve[6].ve0.src_format);
+   debug_printf("\t\t.ve[6].ve0.valid = 0x%x\n", (*ptr).ve[6].ve0.valid);
+   debug_printf("\t\t.ve[6].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[6].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[6].ve1.dst_offset = 0x%x\n", (*ptr).ve[6].ve1.dst_offset);
+   debug_printf("\t\t.ve[6].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[6].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[6].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[6].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[6].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[6].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[6].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[6].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[7].ve0.src_offset = 0x%x\n", (*ptr).ve[7].ve0.src_offset);
+   debug_printf("\t\t.ve[7].ve0.src_format = 0x%x\n", (*ptr).ve[7].ve0.src_format);
+   debug_printf("\t\t.ve[7].ve0.valid = 0x%x\n", (*ptr).ve[7].ve0.valid);
+   debug_printf("\t\t.ve[7].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[7].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[7].ve1.dst_offset = 0x%x\n", (*ptr).ve[7].ve1.dst_offset);
+   debug_printf("\t\t.ve[7].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[7].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[7].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[7].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[7].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[7].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[7].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[7].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[8].ve0.src_offset = 0x%x\n", (*ptr).ve[8].ve0.src_offset);
+   debug_printf("\t\t.ve[8].ve0.src_format = 0x%x\n", (*ptr).ve[8].ve0.src_format);
+   debug_printf("\t\t.ve[8].ve0.valid = 0x%x\n", (*ptr).ve[8].ve0.valid);
+   debug_printf("\t\t.ve[8].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[8].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[8].ve1.dst_offset = 0x%x\n", (*ptr).ve[8].ve1.dst_offset);
+   debug_printf("\t\t.ve[8].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[8].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[8].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[8].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[8].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[8].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[8].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[8].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[9].ve0.src_offset = 0x%x\n", (*ptr).ve[9].ve0.src_offset);
+   debug_printf("\t\t.ve[9].ve0.src_format = 0x%x\n", (*ptr).ve[9].ve0.src_format);
+   debug_printf("\t\t.ve[9].ve0.valid = 0x%x\n", (*ptr).ve[9].ve0.valid);
+   debug_printf("\t\t.ve[9].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[9].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[9].ve1.dst_offset = 0x%x\n", (*ptr).ve[9].ve1.dst_offset);
+   debug_printf("\t\t.ve[9].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[9].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[9].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[9].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[9].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[9].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[9].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[9].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[10].ve0.src_offset = 0x%x\n", (*ptr).ve[10].ve0.src_offset);
+   debug_printf("\t\t.ve[10].ve0.src_format = 0x%x\n", (*ptr).ve[10].ve0.src_format);
+   debug_printf("\t\t.ve[10].ve0.valid = 0x%x\n", (*ptr).ve[10].ve0.valid);
+   debug_printf("\t\t.ve[10].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[10].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[10].ve1.dst_offset = 0x%x\n", (*ptr).ve[10].ve1.dst_offset);
+   debug_printf("\t\t.ve[10].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[10].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[10].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[10].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[10].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[10].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[10].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[10].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[11].ve0.src_offset = 0x%x\n", (*ptr).ve[11].ve0.src_offset);
+   debug_printf("\t\t.ve[11].ve0.src_format = 0x%x\n", (*ptr).ve[11].ve0.src_format);
+   debug_printf("\t\t.ve[11].ve0.valid = 0x%x\n", (*ptr).ve[11].ve0.valid);
+   debug_printf("\t\t.ve[11].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[11].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[11].ve1.dst_offset = 0x%x\n", (*ptr).ve[11].ve1.dst_offset);
+   debug_printf("\t\t.ve[11].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[11].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[11].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[11].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[11].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[11].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[11].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[11].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[12].ve0.src_offset = 0x%x\n", (*ptr).ve[12].ve0.src_offset);
+   debug_printf("\t\t.ve[12].ve0.src_format = 0x%x\n", (*ptr).ve[12].ve0.src_format);
+   debug_printf("\t\t.ve[12].ve0.valid = 0x%x\n", (*ptr).ve[12].ve0.valid);
+   debug_printf("\t\t.ve[12].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[12].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[12].ve1.dst_offset = 0x%x\n", (*ptr).ve[12].ve1.dst_offset);
+   debug_printf("\t\t.ve[12].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[12].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[12].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[12].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[12].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[12].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[12].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[12].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[13].ve0.src_offset = 0x%x\n", (*ptr).ve[13].ve0.src_offset);
+   debug_printf("\t\t.ve[13].ve0.src_format = 0x%x\n", (*ptr).ve[13].ve0.src_format);
+   debug_printf("\t\t.ve[13].ve0.valid = 0x%x\n", (*ptr).ve[13].ve0.valid);
+   debug_printf("\t\t.ve[13].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[13].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[13].ve1.dst_offset = 0x%x\n", (*ptr).ve[13].ve1.dst_offset);
+   debug_printf("\t\t.ve[13].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[13].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[13].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[13].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[13].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[13].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[13].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[13].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[14].ve0.src_offset = 0x%x\n", (*ptr).ve[14].ve0.src_offset);
+   debug_printf("\t\t.ve[14].ve0.src_format = 0x%x\n", (*ptr).ve[14].ve0.src_format);
+   debug_printf("\t\t.ve[14].ve0.valid = 0x%x\n", (*ptr).ve[14].ve0.valid);
+   debug_printf("\t\t.ve[14].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[14].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[14].ve1.dst_offset = 0x%x\n", (*ptr).ve[14].ve1.dst_offset);
+   debug_printf("\t\t.ve[14].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[14].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[14].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[14].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[14].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[14].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[14].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[14].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[15].ve0.src_offset = 0x%x\n", (*ptr).ve[15].ve0.src_offset);
+   debug_printf("\t\t.ve[15].ve0.src_format = 0x%x\n", (*ptr).ve[15].ve0.src_format);
+   debug_printf("\t\t.ve[15].ve0.valid = 0x%x\n", (*ptr).ve[15].ve0.valid);
+   debug_printf("\t\t.ve[15].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[15].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[15].ve1.dst_offset = 0x%x\n", (*ptr).ve[15].ve1.dst_offset);
+   debug_printf("\t\t.ve[15].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[15].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[15].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[15].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[15].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[15].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[15].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[15].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[16].ve0.src_offset = 0x%x\n", (*ptr).ve[16].ve0.src_offset);
+   debug_printf("\t\t.ve[16].ve0.src_format = 0x%x\n", (*ptr).ve[16].ve0.src_format);
+   debug_printf("\t\t.ve[16].ve0.valid = 0x%x\n", (*ptr).ve[16].ve0.valid);
+   debug_printf("\t\t.ve[16].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[16].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[16].ve1.dst_offset = 0x%x\n", (*ptr).ve[16].ve1.dst_offset);
+   debug_printf("\t\t.ve[16].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[16].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[16].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[16].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[16].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[16].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[16].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[16].ve1.vfcomponent0);
+   debug_printf("\t\t.ve[17].ve0.src_offset = 0x%x\n", (*ptr).ve[17].ve0.src_offset);
+   debug_printf("\t\t.ve[17].ve0.src_format = 0x%x\n", (*ptr).ve[17].ve0.src_format);
+   debug_printf("\t\t.ve[17].ve0.valid = 0x%x\n", (*ptr).ve[17].ve0.valid);
+   debug_printf("\t\t.ve[17].ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve[17].ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve[17].ve1.dst_offset = 0x%x\n", (*ptr).ve[17].ve1.dst_offset);
+   debug_printf("\t\t.ve[17].ve1.vfcomponent3 = 0x%x\n", (*ptr).ve[17].ve1.vfcomponent3);
+   debug_printf("\t\t.ve[17].ve1.vfcomponent2 = 0x%x\n", (*ptr).ve[17].ve1.vfcomponent2);
+   debug_printf("\t\t.ve[17].ve1.vfcomponent1 = 0x%x\n", (*ptr).ve[17].ve1.vfcomponent1);
+   debug_printf("\t\t.ve[17].ve1.vfcomponent0 = 0x%x\n", (*ptr).ve[17].ve1.vfcomponent0);
+}
+
+void
+brw_dump_vertex_element_state(const struct brw_vertex_element_state *ptr)
+{
+   debug_printf("\t\t.ve0.src_offset = 0x%x\n", (*ptr).ve0.src_offset);
+   debug_printf("\t\t.ve0.src_format = 0x%x\n", (*ptr).ve0.src_format);
+   debug_printf("\t\t.ve0.valid = 0x%x\n", (*ptr).ve0.valid);
+   debug_printf("\t\t.ve0.vertex_buffer_index = 0x%x\n", (*ptr).ve0.vertex_buffer_index);
+   debug_printf("\t\t.ve1.dst_offset = 0x%x\n", (*ptr).ve1.dst_offset);
+   debug_printf("\t\t.ve1.vfcomponent3 = 0x%x\n", (*ptr).ve1.vfcomponent3);
+   debug_printf("\t\t.ve1.vfcomponent2 = 0x%x\n", (*ptr).ve1.vfcomponent2);
+   debug_printf("\t\t.ve1.vfcomponent1 = 0x%x\n", (*ptr).ve1.vfcomponent1);
+   debug_printf("\t\t.ve1.vfcomponent0 = 0x%x\n", (*ptr).ve1.vfcomponent0);
+}
+
+void
+brw_dump_vf_statistics(const struct brw_vf_statistics *ptr)
+{
+   debug_printf("\t\t.statistics_enable = 0x%x\n", (*ptr).statistics_enable);
+   debug_printf("\t\t.opcode = 0x%x\n", (*ptr).opcode);
+}
+
+void
+brw_dump_vs_unit_state(const struct brw_vs_unit_state *ptr)
+{
+   debug_printf("\t\t.thread0.grf_reg_count = 0x%x\n", (*ptr).thread0.grf_reg_count);
+   debug_printf("\t\t.thread0.kernel_start_pointer = 0x%x\n", (*ptr).thread0.kernel_start_pointer);
+   debug_printf("\t\t.thread1.ext_halt_exception_enable = 0x%x\n", (*ptr).thread1.ext_halt_exception_enable);
+   debug_printf("\t\t.thread1.sw_exception_enable = 0x%x\n", (*ptr).thread1.sw_exception_enable);
+   debug_printf("\t\t.thread1.mask_stack_exception_enable = 0x%x\n", (*ptr).thread1.mask_stack_exception_enable);
+   debug_printf("\t\t.thread1.timeout_exception_enable = 0x%x\n", (*ptr).thread1.timeout_exception_enable);
+   debug_printf("\t\t.thread1.illegal_op_exception_enable = 0x%x\n", (*ptr).thread1.illegal_op_exception_enable);
+   debug_printf("\t\t.thread1.depth_coef_urb_read_offset = 0x%x\n", (*ptr).thread1.depth_coef_urb_read_offset);
+   debug_printf("\t\t.thread1.floating_point_mode = 0x%x\n", (*ptr).thread1.floating_point_mode);
+   debug_printf("\t\t.thread1.thread_priority = 0x%x\n", (*ptr).thread1.thread_priority);
+   debug_printf("\t\t.thread1.binding_table_entry_count = 0x%x\n", (*ptr).thread1.binding_table_entry_count);
+   debug_printf("\t\t.thread1.single_program_flow = 0x%x\n", (*ptr).thread1.single_program_flow);
+   debug_printf("\t\t.thread2.per_thread_scratch_space = 0x%x\n", (*ptr).thread2.per_thread_scratch_space);
+   debug_printf("\t\t.thread2.scratch_space_base_pointer = 0x%x\n", (*ptr).thread2.scratch_space_base_pointer);
+   debug_printf("\t\t.thread3.dispatch_grf_start_reg = 0x%x\n", (*ptr).thread3.dispatch_grf_start_reg);
+   debug_printf("\t\t.thread3.urb_entry_read_offset = 0x%x\n", (*ptr).thread3.urb_entry_read_offset);
+   debug_printf("\t\t.thread3.urb_entry_read_length = 0x%x\n", (*ptr).thread3.urb_entry_read_length);
+   debug_printf("\t\t.thread3.const_urb_entry_read_offset = 0x%x\n", (*ptr).thread3.const_urb_entry_read_offset);
+   debug_printf("\t\t.thread3.const_urb_entry_read_length = 0x%x\n", (*ptr).thread3.const_urb_entry_read_length);
+   debug_printf("\t\t.thread4.stats_enable = 0x%x\n", (*ptr).thread4.stats_enable);
+   debug_printf("\t\t.thread4.nr_urb_entries = 0x%x\n", (*ptr).thread4.nr_urb_entries);
+   debug_printf("\t\t.thread4.urb_entry_allocation_size = 0x%x\n", (*ptr).thread4.urb_entry_allocation_size);
+   debug_printf("\t\t.thread4.max_threads = 0x%x\n", (*ptr).thread4.max_threads);
+   debug_printf("\t\t.vs5.sampler_count = 0x%x\n", (*ptr).vs5.sampler_count);
+   debug_printf("\t\t.vs5.sampler_state_pointer = 0x%x\n", (*ptr).vs5.sampler_state_pointer);
+   debug_printf("\t\t.vs6.vs_enable = 0x%x\n", (*ptr).vs6.vs_enable);
+   debug_printf("\t\t.vs6.vert_cache_disable = 0x%x\n", (*ptr).vs6.vert_cache_disable);
+}
+
+void
+brw_dump_wm_unit_state(const struct brw_wm_unit_state *ptr)
+{
+   debug_printf("\t\t.thread0.grf_reg_count = 0x%x\n", (*ptr).thread0.grf_reg_count);
+   debug_printf("\t\t.thread0.kernel_start_pointer = 0x%x\n", (*ptr).thread0.kernel_start_pointer);
+   debug_printf("\t\t.thread1.ext_halt_exception_enable = 0x%x\n", (*ptr).thread1.ext_halt_exception_enable);
+   debug_printf("\t\t.thread1.sw_exception_enable = 0x%x\n", (*ptr).thread1.sw_exception_enable);
+   debug_printf("\t\t.thread1.mask_stack_exception_enable = 0x%x\n", (*ptr).thread1.mask_stack_exception_enable);
+   debug_printf("\t\t.thread1.timeout_exception_enable = 0x%x\n", (*ptr).thread1.timeout_exception_enable);
+   debug_printf("\t\t.thread1.illegal_op_exception_enable = 0x%x\n", (*ptr).thread1.illegal_op_exception_enable);
+   debug_printf("\t\t.thread1.depth_coef_urb_read_offset = 0x%x\n", (*ptr).thread1.depth_coef_urb_read_offset);
+   debug_printf("\t\t.thread1.floating_point_mode = 0x%x\n", (*ptr).thread1.floating_point_mode);
+   debug_printf("\t\t.thread1.thread_priority = 0x%x\n", (*ptr).thread1.thread_priority);
+   debug_printf("\t\t.thread1.binding_table_entry_count = 0x%x\n", (*ptr).thread1.binding_table_entry_count);
+   debug_printf("\t\t.thread1.single_program_flow = 0x%x\n", (*ptr).thread1.single_program_flow);
+   debug_printf("\t\t.thread2.per_thread_scratch_space = 0x%x\n", (*ptr).thread2.per_thread_scratch_space);
+   debug_printf("\t\t.thread2.scratch_space_base_pointer = 0x%x\n", (*ptr).thread2.scratch_space_base_pointer);
+   debug_printf("\t\t.thread3.dispatch_grf_start_reg = 0x%x\n", (*ptr).thread3.dispatch_grf_start_reg);
+   debug_printf("\t\t.thread3.urb_entry_read_offset = 0x%x\n", (*ptr).thread3.urb_entry_read_offset);
+   debug_printf("\t\t.thread3.urb_entry_read_length = 0x%x\n", (*ptr).thread3.urb_entry_read_length);
+   debug_printf("\t\t.thread3.const_urb_entry_read_offset = 0x%x\n", (*ptr).thread3.const_urb_entry_read_offset);
+   debug_printf("\t\t.thread3.const_urb_entry_read_length = 0x%x\n", (*ptr).thread3.const_urb_entry_read_length);
+   debug_printf("\t\t.wm4.stats_enable = 0x%x\n", (*ptr).wm4.stats_enable);
+   debug_printf("\t\t.wm4.depth_buffer_clear = 0x%x\n", (*ptr).wm4.depth_buffer_clear);
+   debug_printf("\t\t.wm4.sampler_count = 0x%x\n", (*ptr).wm4.sampler_count);
+   debug_printf("\t\t.wm4.sampler_state_pointer = 0x%x\n", (*ptr).wm4.sampler_state_pointer);
+   debug_printf("\t\t.wm5.enable_8_pix = 0x%x\n", (*ptr).wm5.enable_8_pix);
+   debug_printf("\t\t.wm5.enable_16_pix = 0x%x\n", (*ptr).wm5.enable_16_pix);
+   debug_printf("\t\t.wm5.enable_32_pix = 0x%x\n", (*ptr).wm5.enable_32_pix);
+   debug_printf("\t\t.wm5.enable_con_32_pix = 0x%x\n", (*ptr).wm5.enable_con_32_pix);
+   debug_printf("\t\t.wm5.enable_con_64_pix = 0x%x\n", (*ptr).wm5.enable_con_64_pix);
+   debug_printf("\t\t.wm5.legacy_global_depth_bias = 0x%x\n", (*ptr).wm5.legacy_global_depth_bias);
+   debug_printf("\t\t.wm5.line_stipple = 0x%x\n", (*ptr).wm5.line_stipple);
+   debug_printf("\t\t.wm5.depth_offset = 0x%x\n", (*ptr).wm5.depth_offset);
+   debug_printf("\t\t.wm5.polygon_stipple = 0x%x\n", (*ptr).wm5.polygon_stipple);
+   debug_printf("\t\t.wm5.line_aa_region_width = 0x%x\n", (*ptr).wm5.line_aa_region_width);
+   debug_printf("\t\t.wm5.line_endcap_aa_region_width = 0x%x\n", (*ptr).wm5.line_endcap_aa_region_width);
+   debug_printf("\t\t.wm5.early_depth_test = 0x%x\n", (*ptr).wm5.early_depth_test);
+   debug_printf("\t\t.wm5.thread_dispatch_enable = 0x%x\n", (*ptr).wm5.thread_dispatch_enable);
+   debug_printf("\t\t.wm5.program_uses_depth = 0x%x\n", (*ptr).wm5.program_uses_depth);
+   debug_printf("\t\t.wm5.program_computes_depth = 0x%x\n", (*ptr).wm5.program_computes_depth);
+   debug_printf("\t\t.wm5.program_uses_killpixel = 0x%x\n", (*ptr).wm5.program_uses_killpixel);
+   debug_printf("\t\t.wm5.legacy_line_rast = 0x%x\n", (*ptr).wm5.legacy_line_rast);
+   debug_printf("\t\t.wm5.transposed_urb_read_enable = 0x%x\n", (*ptr).wm5.transposed_urb_read_enable);
+   debug_printf("\t\t.wm5.max_threads = 0x%x\n", (*ptr).wm5.max_threads);
+   debug_printf("\t\t.global_depth_offset_constant = %f\n", (*ptr).global_depth_offset_constant);
+   debug_printf("\t\t.global_depth_offset_scale = %f\n", (*ptr).global_depth_offset_scale);
+   debug_printf("\t\t.wm8.grf_reg_count_1 = 0x%x\n", (*ptr).wm8.grf_reg_count_1);
+   debug_printf("\t\t.wm8.kernel_start_pointer_1 = 0x%x\n", (*ptr).wm8.kernel_start_pointer_1);
+   debug_printf("\t\t.wm9.grf_reg_count_2 = 0x%x\n", (*ptr).wm9.grf_reg_count_2);
+   debug_printf("\t\t.wm9.kernel_start_pointer_2 = 0x%x\n", (*ptr).wm9.kernel_start_pointer_2);
+   debug_printf("\t\t.wm10.grf_reg_count_3 = 0x%x\n", (*ptr).wm10.grf_reg_count_3);
+   debug_printf("\t\t.wm10.kernel_start_pointer_3 = 0x%x\n", (*ptr).wm10.kernel_start_pointer_3);
+}
+
diff --git a/src/gallium/drivers/i965/brw_structs_dump.h b/src/gallium/drivers/i965/brw_structs_dump.h
new file mode 100644 (file)
index 0000000..7c02dbf
--- /dev/null
@@ -0,0 +1,276 @@
+/**************************************************************************
+ *
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ **************************************************************************/
+
+/**
+ * @file
+ * Dump i965 data structures.
+ *
+ * Generated automatically from brw_structs.h by brw_structs_dump.py.
+ */
+
+#ifndef BRW_STRUCTS_DUMP_H
+#define BRW_STRUCTS_DUMP_H
+
+struct brw_3d_control;
+struct brw_3d_primitive;
+struct brw_aa_line_parameters;
+struct brw_binding_table_pointers;
+struct brw_blend_constant_color;
+struct brw_cc0;
+struct brw_cc1;
+struct brw_cc2;
+struct brw_cc3;
+struct brw_cc4;
+struct brw_cc5;
+struct brw_cc6;
+struct brw_cc7;
+struct brw_cc_unit_state;
+struct brw_cc_viewport;
+struct brw_clip_unit_state;
+struct brw_clipper_viewport;
+struct brw_constant_buffer;
+struct brw_cs_urb_state;
+struct brw_depthbuffer;
+struct brw_depthbuffer_g4x;
+struct brw_drawrect;
+struct brw_global_depth_offset_clamp;
+struct brw_gs_unit_state;
+struct brw_indexbuffer;
+struct brw_line_stipple;
+struct brw_mi_flush;
+struct brw_pipe_control;
+struct brw_pipeline_select;
+struct brw_pipelined_state_pointers;
+struct brw_polygon_stipple;
+struct brw_polygon_stipple_offset;
+struct brw_sampler_default_color;
+struct brw_sampler_state;
+struct brw_sf_unit_state;
+struct brw_sf_viewport;
+struct brw_ss0;
+struct brw_ss1;
+struct brw_ss2;
+struct brw_ss3;
+struct brw_state_base_address;
+struct brw_state_prefetch;
+struct brw_surf_ss0;
+struct brw_surf_ss1;
+struct brw_surf_ss2;
+struct brw_surf_ss3;
+struct brw_surf_ss4;
+struct brw_surf_ss5;
+struct brw_surface_state;
+struct brw_system_instruction_pointer;
+struct brw_urb_fence;
+struct brw_urb_immediate;
+struct brw_vb_array_state;
+struct brw_vertex_buffer_state;
+struct brw_vertex_element_packet;
+struct brw_vertex_element_state;
+struct brw_vf_statistics;
+struct brw_vs_unit_state;
+struct brw_wm_unit_state;
+
+void
+brw_dump_3d_control(const struct brw_3d_control *ptr);
+
+void
+brw_dump_3d_primitive(const struct brw_3d_primitive *ptr);
+
+void
+brw_dump_aa_line_parameters(const struct brw_aa_line_parameters *ptr);
+
+void
+brw_dump_binding_table_pointers(const struct brw_binding_table_pointers *ptr);
+
+void
+brw_dump_blend_constant_color(const struct brw_blend_constant_color *ptr);
+
+void
+brw_dump_cc0(const struct brw_cc0 *ptr);
+
+void
+brw_dump_cc1(const struct brw_cc1 *ptr);
+
+void
+brw_dump_cc2(const struct brw_cc2 *ptr);
+
+void
+brw_dump_cc3(const struct brw_cc3 *ptr);
+
+void
+brw_dump_cc4(const struct brw_cc4 *ptr);
+
+void
+brw_dump_cc5(const struct brw_cc5 *ptr);
+
+void
+brw_dump_cc6(const struct brw_cc6 *ptr);
+
+void
+brw_dump_cc7(const struct brw_cc7 *ptr);
+
+void
+brw_dump_cc_unit_state(const struct brw_cc_unit_state *ptr);
+
+void
+brw_dump_cc_viewport(const struct brw_cc_viewport *ptr);
+
+void
+brw_dump_clip_unit_state(const struct brw_clip_unit_state *ptr);
+
+void
+brw_dump_clipper_viewport(const struct brw_clipper_viewport *ptr);
+
+void
+brw_dump_constant_buffer(const struct brw_constant_buffer *ptr);
+
+void
+brw_dump_cs_urb_state(const struct brw_cs_urb_state *ptr);
+
+void
+brw_dump_depthbuffer(const struct brw_depthbuffer *ptr);
+
+void
+brw_dump_depthbuffer_g4x(const struct brw_depthbuffer_g4x *ptr);
+
+void
+brw_dump_drawrect(const struct brw_drawrect *ptr);
+
+void
+brw_dump_global_depth_offset_clamp(const struct brw_global_depth_offset_clamp *ptr);
+
+void
+brw_dump_gs_unit_state(const struct brw_gs_unit_state *ptr);
+
+void
+brw_dump_indexbuffer(const struct brw_indexbuffer *ptr);
+
+void
+brw_dump_line_stipple(const struct brw_line_stipple *ptr);
+
+void
+brw_dump_mi_flush(const struct brw_mi_flush *ptr);
+
+void
+brw_dump_pipe_control(const struct brw_pipe_control *ptr);
+
+void
+brw_dump_pipeline_select(const struct brw_pipeline_select *ptr);
+
+void
+brw_dump_pipelined_state_pointers(const struct brw_pipelined_state_pointers *ptr);
+
+void
+brw_dump_polygon_stipple(const struct brw_polygon_stipple *ptr);
+
+void
+brw_dump_polygon_stipple_offset(const struct brw_polygon_stipple_offset *ptr);
+
+void
+brw_dump_sampler_default_color(const struct brw_sampler_default_color *ptr);
+
+void
+brw_dump_sampler_state(const struct brw_sampler_state *ptr);
+
+void
+brw_dump_sf_unit_state(const struct brw_sf_unit_state *ptr);
+
+void
+brw_dump_sf_viewport(const struct brw_sf_viewport *ptr);
+
+void
+brw_dump_ss0(const struct brw_ss0 *ptr);
+
+void
+brw_dump_ss1(const struct brw_ss1 *ptr);
+
+void
+brw_dump_ss2(const struct brw_ss2 *ptr);
+
+void
+brw_dump_ss3(const struct brw_ss3 *ptr);
+
+void
+brw_dump_state_base_address(const struct brw_state_base_address *ptr);
+
+void
+brw_dump_state_prefetch(const struct brw_state_prefetch *ptr);
+
+void
+brw_dump_surf_ss0(const struct brw_surf_ss0 *ptr);
+
+void
+brw_dump_surf_ss1(const struct brw_surf_ss1 *ptr);
+
+void
+brw_dump_surf_ss2(const struct brw_surf_ss2 *ptr);
+
+void
+brw_dump_surf_ss3(const struct brw_surf_ss3 *ptr);
+
+void
+brw_dump_surf_ss4(const struct brw_surf_ss4 *ptr);
+
+void
+brw_dump_surf_ss5(const struct brw_surf_ss5 *ptr);
+
+void
+brw_dump_surface_state(const struct brw_surface_state *ptr);
+
+void
+brw_dump_system_instruction_pointer(const struct brw_system_instruction_pointer *ptr);
+
+void
+brw_dump_urb_fence(const struct brw_urb_fence *ptr);
+
+void
+brw_dump_urb_immediate(const struct brw_urb_immediate *ptr);
+
+void
+brw_dump_vb_array_state(const struct brw_vb_array_state *ptr);
+
+void
+brw_dump_vertex_buffer_state(const struct brw_vertex_buffer_state *ptr);
+
+void
+brw_dump_vertex_element_packet(const struct brw_vertex_element_packet *ptr);
+
+void
+brw_dump_vertex_element_state(const struct brw_vertex_element_state *ptr);
+
+void
+brw_dump_vf_statistics(const struct brw_vf_statistics *ptr);
+
+void
+brw_dump_vs_unit_state(const struct brw_vs_unit_state *ptr);
+
+void
+brw_dump_wm_unit_state(const struct brw_wm_unit_state *ptr);
+
+
+#endif /* BRW_STRUCTS_DUMP_H */
diff --git a/src/gallium/drivers/i965/brw_structs_dump.py b/src/gallium/drivers/i965/brw_structs_dump.py
new file mode 100755 (executable)
index 0000000..6dba49a
--- /dev/null
@@ -0,0 +1,291 @@
+#!/usr/bin/env python
+'''
+Generates dumpers for the i965 state strucutures using pygccxml.
+
+Run as 
+
+  PYTHONPATH=/path/to/pygccxml-1.0.0 python brw_structs_dump.py
+
+Jose Fonseca <jfonseca@vmware.com>
+'''
+
+copyright = '''
+/**************************************************************************
+ *
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ **************************************************************************/
+ '''
+
+import os
+import sys
+import re
+
+from pygccxml import parser
+from pygccxml import declarations
+
+from pygccxml.declarations import algorithm
+from pygccxml.declarations import decl_visitor
+from pygccxml.declarations import type_traits
+from pygccxml.declarations import type_visitor
+
+
+enums = True
+
+
+def vars_filter(variable):
+    name = variable.name
+    return not re.match('^pad\d*', name) and name != 'dword' 
+
+
+class decl_dumper_t(decl_visitor.decl_visitor_t):
+
+    def __init__(self, stream, instance = '', decl = None):
+        decl_visitor.decl_visitor_t.__init__(self)
+        self.stream = stream
+        self._instance = instance
+        self.decl = decl
+
+    def clone(self):
+        return decl_dumper_t(self.stream, self._instance, self.decl)
+
+    def visit_class(self):
+        class_ = self.decl
+        assert self.decl.class_type in ('struct', 'union')
+
+        for variable in class_.variables(recursive = False):
+            if vars_filter(variable):
+                dump_type(self.stream, self._instance + '.' + variable.name, variable.type)
+
+    def visit_enumeration(self):
+        if enums:
+            self.stream.write('   switch(%s) {\n' % ("(*ptr)" + self._instance,))
+            for name, value in self.decl.values:
+                self.stream.write('   case %s:\n' % (name,))
+                self.stream.write('      debug_printf("\\t\\t%s = %s\\n");\n' % (self._instance, name))
+                self.stream.write('      break;\n')
+            self.stream.write('   default:\n')
+            self.stream.write('      debug_printf("\\t\\t%s = %%i\\n", %s);\n' % (self._instance, "(*ptr)" + self._instance))
+            self.stream.write('      break;\n')
+            self.stream.write('   }\n')
+        else:
+            self.stream.write('   debug_printf("\\t\\t%s = %%i\\n", %s);\n' % (self._instance, "(*ptr)" + self._instance))
+
+
+def dump_decl(stream, instance, decl):
+    dumper = decl_dumper_t(stream, instance, decl)
+    algorithm.apply_visitor(dumper, decl)
+
+
+class type_dumper_t(type_visitor.type_visitor_t):
+
+    def __init__(self, stream, instance, type_):
+        type_visitor.type_visitor_t.__init__(self)
+        self.stream = stream
+        self.instance = instance
+        self.type = type_
+
+    def clone(self):
+        return type_dumper_t(self.instance, self.type)
+
+    def visit_bool(self):
+        self.print_instance('%i')
+        
+    def visit_char(self):
+        #self.print_instance('%i')
+        self.print_instance('0x%x')
+        
+    def visit_unsigned_char(self):
+        #self.print_instance('%u')
+        self.print_instance('0x%x')
+
+    def visit_signed_char(self):
+        #self.print_instance('%i')
+        self.print_instance('0x%x')
+    
+    def visit_wchar(self):
+        self.print_instance('0x%x')
+        
+    def visit_short_int(self):
+        #self.print_instance('%i')
+        self.print_instance('0x%x')
+        
+    def visit_short_unsigned_int(self):
+        #self.print_instance('%u')
+        self.print_instance('0x%x')
+        
+    def visit_int(self):
+        #self.print_instance('%i')
+        self.print_instance('0x%x')
+        
+    def visit_unsigned_int(self):
+        #self.print_instance('%u')
+        self.print_instance('0x%x')
+        
+    def visit_long_int(self):
+        #self.print_instance('%li')
+        self.print_instance('0x%lx')
+        
+    def visit_long_unsigned_int(self):
+        #self.print_instance('%lu')
+        self.print_instance('%0xlx')
+        
+    def visit_long_long_int(self):
+        #self.print_instance('%lli')
+        self.print_instance('%0xllx')
+        
+    def visit_long_long_unsigned_int(self):
+        #self.print_instance('%llu')
+        self.print_instance('0x%llx')
+        
+    def visit_float(self):
+        self.print_instance('%f')
+        
+    def visit_double(self):
+        self.print_instance('%f')
+        
+    def visit_array(self):
+        for i in range(type_traits.array_size(self.type)):
+            dump_type(self.stream, self.instance + '[%i]' % i, type_traits.base_type(self.type))
+
+    def visit_pointer(self):
+        self.print_instance('%p')
+
+    def visit_declarated(self):
+        #stream.write('decl = %r\n' % self.type.decl_string)
+        decl = type_traits.remove_declarated(self.type)
+        dump_decl(self.stream, self.instance, decl)
+
+    def print_instance(self, format):
+        self.stream.write('   debug_printf("\\t\\t%s = %s\\n", %s);\n' % (self.instance, format, "(*ptr)" + self.instance))
+
+
+
+def dump_type(stream, instance, type_):
+    type_ = type_traits.remove_alias(type_)
+    visitor = type_dumper_t(stream, instance, type_)
+    algorithm.apply_visitor(visitor, type_)
+
+
+def dump_struct_interface(stream, class_, suffix = ';'):
+    name = class_.name
+    assert name.startswith('brw_');
+    name = name[:4] + 'dump_' + name[4:]
+    stream.write('void\n')
+    stream.write('%s(const struct %s *ptr)%s\n' % (name, class_.name, suffix))
+
+
+def dump_struct_implementation(stream, decls, class_):
+    dump_struct_interface(stream, class_, suffix = '')
+    stream.write('{\n')
+    dump_decl(stream, '', class_)
+    stream.write('}\n')
+    stream.write('\n')
+
+
+def dump_header(stream):
+    stream.write(copyright.strip() + '\n')
+    stream.write('\n')
+    stream.write('/**\n')
+    stream.write(' * @file\n')
+    stream.write(' * Dump i965 data structures.\n')
+    stream.write(' *\n')
+    stream.write(' * Generated automatically from brw_structs.h by brw_structs_dump.py.\n')
+    stream.write(' */\n')
+    stream.write('\n')
+
+
+def dump_interfaces(decls, global_ns, names):
+    stream = open('brw_structs_dump.h', 'wt')
+    
+    dump_header(stream)
+    
+    stream.write('#ifndef BRW_STRUCTS_DUMP_H\n')
+    stream.write('#define BRW_STRUCTS_DUMP_H\n')
+    stream.write('\n')
+    
+    for name in names:
+        stream.write('struct %s;\n' % (name,))
+    stream.write('\n')
+
+    for name in names:
+        (class_,) = global_ns.classes(name = name)
+        dump_struct_interface(stream, class_)
+        stream.write('\n')
+    stream.write('\n')
+
+    stream.write('#endif /* BRW_STRUCTS_DUMP_H */\n')
+
+
+def dump_implementations(decls, global_ns, names):
+    stream = open('brw_structs_dump.c', 'wt')
+    
+    dump_header(stream)
+
+    stream.write('#include "util/u_debug.h"\n')
+    stream.write('\n')
+    stream.write('#include "brw_types.h"\n')
+    stream.write('#include "brw_structs.h"\n')
+    stream.write('#include "brw_structs_dump.h"\n')
+    stream.write('\n')
+
+    for name in names:
+        (class_,) = global_ns.classes(name = name)
+        dump_struct_implementation(stream, decls, class_)
+
+
+def decl_filter(decl):
+    '''Filter the declarations we're interested in'''
+    name = decl.name
+    return name.startswith('brw_') and name not in ('brw_instruction',) 
+
+
+def main():
+
+    config = parser.config_t(
+        include_paths = [
+            '../../include',
+        ],
+        compiler = 'gcc',
+    )
+
+    headers = [
+        'brw_types.h', 
+        'brw_structs.h', 
+    ]
+
+    decls = parser.parse(headers, config, parser.COMPILATION_MODE.ALL_AT_ONCE)
+    global_ns = declarations.get_global_namespace(decls)
+
+    names = []
+    for class_ in global_ns.classes(decl_filter):
+        names.append(class_.name)
+    names.sort()
+
+    dump_interfaces(decls, global_ns, names)
+    dump_implementations(decls, global_ns, names)
+
+
+if __name__ == '__main__':
+    main()
diff --git a/src/gallium/drivers/i965/brw_swtnl.c b/src/gallium/drivers/i965/brw_swtnl.c
new file mode 100644 (file)
index 0000000..464013e
--- /dev/null
@@ -0,0 +1,95 @@
+
+#include "brw_context.h"
+#include "brw_pipe_rast.h"
+
+
+static GLboolean need_swtnl( struct brw_context *brw )
+{
+   const struct pipe_rasterizer_state *rast = &brw->curr.rast->templ;
+
+   /* If we don't require strict OpenGL conformance, never 
+    * use fallbacks.  If we're forcing fallbacks, always
+    * use fallfacks.
+    */
+   if (brw->flags.no_swtnl)
+      return FALSE;
+
+   if (brw->flags.force_swtnl)
+      return TRUE;
+
+   /* Exceeding hw limits on number of VS inputs?
+    */
+   if (brw->curr.num_vertex_elements == 0 ||
+       brw->curr.num_vertex_elements >= BRW_VEP_MAX) {
+      return TRUE;
+   }
+
+   /* Position array with zero stride?
+    *
+    * XXX: position isn't always at zero...
+    * XXX: eliminate zero-stride arrays
+    */
+   {
+      int ve0_vb = brw->curr.vertex_element[0].vertex_buffer_index;
+      
+      if (brw->curr.vertex_buffer[ve0_vb].stride == 0)
+        return TRUE;
+   }
+
+   /* XXX: short-circuit
+    */
+   return FALSE;
+
+   if (brw->reduced_primitive == PIPE_PRIM_TRIANGLES) {
+      if (rast->poly_smooth)
+        return TRUE;
+
+   }
+   
+   if (brw->reduced_primitive == PIPE_PRIM_LINES ||
+       (brw->reduced_primitive == PIPE_PRIM_TRIANGLES &&
+       (rast->fill_cw == PIPE_POLYGON_MODE_LINE ||
+        rast->fill_ccw == PIPE_POLYGON_MODE_LINE)))
+   {
+      /* BRW hardware will do AA lines, but they are non-conformant it
+       * seems.  TBD whether we keep this fallback:
+       */
+      if (rast->line_smooth)
+        return TRUE;
+
+      /* XXX: was a fallback in mesa (gs doesn't get enough
+       * information to know when to reset stipple counter), but there
+       * must be a way around it.
+       */
+      if (rast->line_stipple_enable &&
+         (brw->reduced_primitive == PIPE_PRIM_TRIANGLES ||
+          brw->primitive == PIPE_PRIM_LINE_LOOP || 
+          brw->primitive == PIPE_PRIM_LINE_STRIP))
+        return TRUE;
+   }
+
+   
+   if (brw->reduced_primitive == PIPE_PRIM_POINTS ||
+       (brw->reduced_primitive == PIPE_PRIM_TRIANGLES &&
+       (rast->fill_cw == PIPE_POLYGON_MODE_POINT ||
+        rast->fill_ccw == PIPE_POLYGON_MODE_POINT)))
+   {
+      if (rast->point_smooth)
+        return TRUE;
+   }
+
+   /* BRW hardware doesn't handle CLAMP texturing correctly;
+    * brw_wm_sampler_state:translate_wrap_mode() treats CLAMP
+    * as CLAMP_TO_EDGE instead.  If we're using CLAMP, and
+    * we want strict conformance, force the fallback.
+    *
+    * XXX: need a workaround for this.
+    */
+      
+   /* Nothing stopping us from the fast path now */
+   return FALSE;
+}
+
+
+
+
diff --git a/src/gallium/drivers/i965/brw_types.h b/src/gallium/drivers/i965/brw_types.h
new file mode 100644 (file)
index 0000000..89e08a5
--- /dev/null
@@ -0,0 +1,21 @@
+#ifndef BRW_TYPES_H
+#define BRW_TYPES_H
+
+#include "pipe/p_compiler.h"
+
+typedef uint32_t GLuint;
+typedef uint8_t GLubyte;
+typedef uint16_t GLushort;
+typedef int32_t GLint;
+typedef int8_t GLbyte;
+typedef int16_t GLshort;
+typedef float GLfloat;
+
+/* no GLenum, translate all away */
+
+typedef uint8_t GLboolean;
+
+#define GL_FALSE FALSE
+#define GL_TRUE TRUE
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_urb.c b/src/gallium/drivers/i965/brw_urb.c
new file mode 100644 (file)
index 0000000..907ec56
--- /dev/null
@@ -0,0 +1,263 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+        
+
+
+#include "brw_batchbuffer.h"
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+#include "brw_debug.h"
+
+#define VS 0
+#define GS 1
+#define CLP 2
+#define SF 3
+#define CS 4
+
+/** @file brw_urb.c
+ *
+ * Manages the division of the URB space between the various fixed-function
+ * units.
+ *
+ * See the Thread Initiation Management section of the GEN4 B-Spec, and
+ * the individual *_STATE structures for restrictions on numbers of
+ * entries and threads.
+ */
+
+/*
+ * Generally, a unit requires a min_nr_entries based on how many entries
+ * it produces before the downstream unit gets unblocked and can use and
+ * dereference some of its handles.
+ *
+ * The SF unit preallocates a PUE at the start of thread dispatch, and only
+ * uses that one.  So it requires one entry per thread.
+ *
+ * For CLIP, the SF unit will hold the previous primitive while the
+ * next is getting assembled, meaning that linestrips require 3 CLIP VUEs
+ * (vertices) to ensure continued processing, trifans require 4, and tristrips
+ * require 5.  There can be 1 or 2 threads, and each has the same requirement.
+ *
+ * GS has the same requirement as CLIP, but it never handles tristrips,
+ * so we can lower the minimum to 4 for the POLYGONs (trifans) it produces.
+ * We only run it single-threaded.
+ *
+ * For VS, the number of entries may be 8, 12, 16, or 32 (or 64 on G4X).
+ * Each thread processes 2 preallocated VUEs (vertices) at a time, and they
+ * get streamed down as soon as threads processing earlier vertices get
+ * theirs accepted.
+ *
+ * Each unit will take the number of URB entries we give it (based on the
+ * entry size calculated in brw_vs_emit.c for VUEs, brw_sf_emit.c for PUEs,
+ * and brw_curbe.c for the CURBEs) and decide its maximum number of
+ * threads it can support based on that. in brw_*_state.c.
+ *
+ * XXX: Are the min_entry_size numbers useful?
+ * XXX: Verify min_nr_entries, esp for VS.
+ * XXX: Verify SF min_entry_size.
+ */
+static const struct urb_limits {
+   GLuint min_nr_entries;
+   GLuint preferred_nr_entries;
+   GLuint min_entry_size;
+   GLuint max_entry_size;
+} limits[CS+1] = {
+   { 16, 32, 1, 5 },                   /* vs */
+   { 4, 8,  1, 5 },                    /* gs */
+   { 5, 10,  1, 5 },                   /* clp */
+   { 1, 8,  1, 12 },                   /* sf */
+   { 1, 4,  1, 32 }                    /* cs */
+};
+
+
+static GLboolean check_urb_layout( struct brw_context *brw )
+{
+   brw->urb.vs_start = 0;
+   brw->urb.gs_start = brw->urb.nr_vs_entries * brw->urb.vsize;
+   brw->urb.clip_start = brw->urb.gs_start + brw->urb.nr_gs_entries * brw->urb.vsize;
+   brw->urb.sf_start = brw->urb.clip_start + brw->urb.nr_clip_entries * brw->urb.vsize;
+   brw->urb.cs_start = brw->urb.sf_start + brw->urb.nr_sf_entries * brw->urb.sfsize;
+
+   return brw->urb.cs_start + brw->urb.nr_cs_entries * brw->urb.csize <= URB_SIZES(brw);
+}
+
+/* Most minimal update, forces re-emit of URB fence packet after GS
+ * unit turned on/off.
+ */
+static int recalculate_urb_fence( struct brw_context *brw )
+{
+   GLuint csize = brw->curbe.total_size;
+   GLuint vsize = brw->vs.prog_data->urb_entry_size;
+   GLuint sfsize = brw->sf.prog_data->urb_entry_size;
+
+   if (csize < limits[CS].min_entry_size)
+      csize = limits[CS].min_entry_size;
+
+   if (vsize < limits[VS].min_entry_size)
+      vsize = limits[VS].min_entry_size;
+
+   if (sfsize < limits[SF].min_entry_size)
+      sfsize = limits[SF].min_entry_size;
+
+   if (brw->urb.vsize < vsize ||
+       brw->urb.sfsize < sfsize ||
+       brw->urb.csize < csize ||
+       (brw->urb.constrained && (brw->urb.vsize > vsize ||
+                                brw->urb.sfsize > sfsize ||
+                                brw->urb.csize > csize))) {
+      
+
+      brw->urb.csize = csize;
+      brw->urb.sfsize = sfsize;
+      brw->urb.vsize = vsize;
+
+      brw->urb.nr_vs_entries = limits[VS].preferred_nr_entries;        
+      brw->urb.nr_gs_entries = limits[GS].preferred_nr_entries;        
+      brw->urb.nr_clip_entries = limits[CLP].preferred_nr_entries;
+      brw->urb.nr_sf_entries = limits[SF].preferred_nr_entries;        
+      brw->urb.nr_cs_entries = limits[CS].preferred_nr_entries;        
+
+      brw->urb.constrained = 0;
+
+      if (BRW_IS_IGDNG(brw)) {
+         brw->urb.nr_vs_entries = 128;
+         brw->urb.nr_sf_entries = 48;
+         if (check_urb_layout(brw)) {
+            goto done;
+         } else {
+            brw->urb.constrained = 1;
+            brw->urb.nr_vs_entries = limits[VS].preferred_nr_entries;
+            brw->urb.nr_sf_entries = limits[SF].preferred_nr_entries;
+         }
+      } else if (BRW_IS_G4X(brw)) {
+        brw->urb.nr_vs_entries = 64;
+        if (check_urb_layout(brw)) {
+           goto done;
+        } else {
+           brw->urb.constrained = 1;
+           brw->urb.nr_vs_entries = limits[VS].preferred_nr_entries;
+        }
+      }
+
+      if (BRW_DEBUG & DEBUG_MIN_URB) {
+        brw->urb.nr_vs_entries = limits[VS].min_nr_entries;    
+        brw->urb.nr_gs_entries = limits[GS].min_nr_entries;    
+        brw->urb.nr_clip_entries = limits[CLP].min_nr_entries;
+        brw->urb.nr_sf_entries = limits[SF].min_nr_entries;    
+        brw->urb.nr_cs_entries = limits[CS].min_nr_entries;    
+        brw->urb.constrained = 1;
+      }
+
+      if (!check_urb_layout(brw)) {
+        brw->urb.nr_vs_entries = limits[VS].min_nr_entries;    
+        brw->urb.nr_gs_entries = limits[GS].min_nr_entries;    
+        brw->urb.nr_clip_entries = limits[CLP].min_nr_entries;
+        brw->urb.nr_sf_entries = limits[SF].min_nr_entries;    
+        brw->urb.nr_cs_entries = limits[CS].min_nr_entries;    
+
+        /* Mark us as operating with constrained nr_entries, so that next
+         * time we recalculate we'll resize the fences in the hope of
+         * escaping constrained mode and getting back to normal performance.
+         */
+        brw->urb.constrained = 1;
+        
+        if (!check_urb_layout(brw)) {
+           /* This is impossible, given the maximal sizes of urb
+            * entries and the values for minimum nr of entries
+            * provided above.
+            */
+           debug_printf("couldn't calculate URB layout!\n");
+           exit(1);
+        }
+        
+        if (BRW_DEBUG & (DEBUG_URB|DEBUG_FALLBACKS))
+           debug_printf("URB CONSTRAINED\n");
+      }
+
+done:
+      if (BRW_DEBUG & DEBUG_URB)
+        debug_printf("URB fence: %d ..VS.. %d ..GS.. %d ..CLP.. %d ..SF.. %d ..CS.. %d\n",
+                     brw->urb.vs_start,
+                     brw->urb.gs_start,
+                     brw->urb.clip_start,
+                     brw->urb.sf_start,
+                     brw->urb.cs_start, 
+                     URB_SIZES(brw));
+      
+      brw->state.dirty.brw |= BRW_NEW_URB_FENCE;
+   }
+
+   return 0;
+}
+
+
+const struct brw_tracked_state brw_recalculate_urb_fence = {
+   .dirty = {
+      .mesa = 0,
+      .brw = BRW_NEW_CURBE_OFFSETS,
+      .cache = (CACHE_NEW_VS_PROG |
+               CACHE_NEW_SF_PROG)
+   },
+   .prepare = recalculate_urb_fence
+};
+
+
+
+
+
+int brw_upload_urb_fence(struct brw_context *brw)
+{
+   struct brw_urb_fence uf;
+   memset(&uf, 0, sizeof(uf));
+
+   uf.header.opcode = CMD_URB_FENCE;
+   uf.header.length = sizeof(uf)/4-2;
+   uf.header.vs_realloc = 1;
+   uf.header.gs_realloc = 1;
+   uf.header.clp_realloc = 1;
+   uf.header.sf_realloc = 1;
+   uf.header.vfe_realloc = 1;
+   uf.header.cs_realloc = 1;
+
+   /* The ordering below is correct, not the layout in the
+    * instruction.
+    *
+    * There are 256/384 urb reg pairs in total.
+    */
+   uf.bits0.vs_fence  = brw->urb.gs_start;
+   uf.bits0.gs_fence  = brw->urb.clip_start; 
+   uf.bits0.clp_fence = brw->urb.sf_start; 
+   uf.bits1.sf_fence  = brw->urb.cs_start; 
+   uf.bits1.cs_fence  = URB_SIZES(brw);
+
+   BRW_BATCH_STRUCT(brw, &uf);
+   return 0;
+}
diff --git a/src/gallium/drivers/i965/brw_util.c b/src/gallium/drivers/i965/brw_util.c
new file mode 100644 (file)
index 0000000..458058d
--- /dev/null
@@ -0,0 +1,38 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+         
+
+#include "brw_util.h"
+#include "brw_defines.h"
+
+
+
+
diff --git a/src/gallium/drivers/i965/brw_util.h b/src/gallium/drivers/i965/brw_util.h
new file mode 100644 (file)
index 0000000..b5f9a36
--- /dev/null
@@ -0,0 +1,44 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+          
+
+#ifndef BRW_UTIL_H
+#define BRW_UTIL_H
+
+#include "brw_types.h"
+
+extern GLuint brw_count_bits( GLuint val );
+extern GLuint brw_translate_blend_factor( unsigned factor );
+extern GLuint brw_translate_blend_equation( unsigned mode );
+
+
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_vs.c b/src/gallium/drivers/i965/brw_vs.c
new file mode 100644 (file)
index 0000000..e3ea5a3
--- /dev/null
@@ -0,0 +1,131 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "tgsi/tgsi_dump.h"           
+
+#include "brw_context.h"
+#include "brw_vs.h"
+#include "brw_util.h"
+#include "brw_state.h"
+#include "brw_pipe_rast.h"
+
+
+
+static enum pipe_error do_vs_prog( struct brw_context *brw, 
+                                   struct brw_vertex_shader *vp,
+                                   struct brw_vs_prog_key *key,
+                                   struct brw_winsys_buffer **bo_out)
+{
+   enum pipe_error ret;
+   GLuint program_size;
+   const GLuint *program;
+   struct brw_vs_compile c;
+
+   memset(&c, 0, sizeof(c));
+   memcpy(&c.key, key, sizeof(*key));
+
+   brw_init_compile(brw, &c.func);
+   c.vp = vp;
+
+   c.prog_data.nr_outputs = vp->info.num_outputs;
+   c.prog_data.nr_inputs = vp->info.num_inputs;
+
+   if (1)
+      tgsi_dump(c.vp->tokens, 0);
+
+   /* Emit GEN4 code.
+    */
+   brw_vs_emit(&c);
+
+   /* get the program
+    */
+   ret = brw_get_program(&c.func, &program, &program_size);
+   if (ret)
+      return ret;
+
+   ret = brw_upload_cache( &brw->cache, BRW_VS_PROG,
+                           &c.key, brw_vs_prog_key_size(&c.key),
+                           NULL, 0,
+                           program, program_size,
+                           &c.prog_data,
+                           &brw->vs.prog_data,
+                           bo_out);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+static enum pipe_error brw_upload_vs_prog(struct brw_context *brw)
+{
+   struct brw_vs_prog_key key;
+   struct brw_vertex_shader *vp = brw->curr.vertex_shader;
+   struct brw_fs_signature *sig = &brw->curr.fragment_shader->signature;
+   enum pipe_error ret;
+
+   memset(&key, 0, sizeof(key));
+
+   key.program_string_id = vp->id;
+   key.nr_userclip = brw->curr.ucp.nr;
+
+   memcpy(&key.fs_signature, sig, brw_fs_signature_size(sig));
+
+
+   /* Make an early check for the key.
+    */
+   if (brw_search_cache(&brw->cache, BRW_VS_PROG,
+                        &key, brw_vs_prog_key_size(&key),
+                        NULL, 0,
+                        &brw->vs.prog_data,
+                        &brw->vs.prog_bo))
+      return PIPE_OK;
+
+   ret = do_vs_prog(brw, vp, &key, &brw->vs.prog_bo);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+/* See brw_vs.c:
+ */
+const struct brw_tracked_state brw_vs_prog = {
+   .dirty = {
+      .mesa  = (PIPE_NEW_CLIP | 
+                PIPE_NEW_RAST |
+                PIPE_NEW_FRAGMENT_SIGNATURE),
+      .brw   = BRW_NEW_VERTEX_PROGRAM,
+      .cache = 0
+   },
+   .prepare = brw_upload_vs_prog
+};
diff --git a/src/gallium/drivers/i965/brw_vs.h b/src/gallium/drivers/i965/brw_vs.h
new file mode 100644 (file)
index 0000000..944d88c
--- /dev/null
@@ -0,0 +1,106 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+            
+
+#ifndef BRW_VS_H
+#define BRW_VS_H
+
+
+#include "brw_context.h"
+#include "brw_eu.h"
+
+
+struct brw_vs_prog_key {
+   GLuint program_string_id;
+   GLuint nr_userclip:4;
+   GLuint pad:26;
+   struct brw_fs_signature fs_signature;
+};
+
+#define brw_vs_prog_key_size(s) (offsetof(struct brw_vs_prog_key, fs_signature) + \
+                                 brw_fs_signature_size(&(s)->fs_signature))
+
+
+#define MAX_IF_DEPTH 32
+#define MAX_LOOP_DEPTH 32
+
+struct brw_vs_compile {
+   struct brw_compile func;
+   struct brw_vs_prog_key key;
+   struct brw_vs_prog_data prog_data;
+   struct brw_chipset chipset;
+
+   struct brw_vertex_shader *vp;
+
+   GLuint nr_inputs;
+   GLuint nr_outputs;
+   GLuint nr_immediates;
+   GLfloat immediate[128][4];
+
+   GLuint overflow_grf_start;
+   GLuint overflow_count;
+
+   GLuint first_tmp;
+   GLuint last_tmp;
+
+   struct brw_reg r0;
+   struct brw_reg r1;
+   struct brw_reg regs[TGSI_FILE_COUNT][128];
+   struct brw_reg tmp;
+   struct brw_reg stack;
+
+   struct {    
+       GLboolean used_in_src;
+       struct brw_reg reg;
+   } output_regs[128];
+
+   struct brw_reg userplane[6];
+
+   /** we may need up to 3 constants per instruction (if use_const_buffer) */
+   struct {
+      GLint index;
+      struct brw_reg reg;
+   } current_const[3];
+
+   struct brw_instruction *if_inst[MAX_IF_DEPTH];
+   struct brw_instruction *loop_inst[MAX_LOOP_DEPTH];
+   GLuint insn;
+   GLuint if_depth;
+   GLuint loop_depth;
+   GLuint end_offset;
+
+   struct brw_indirect stack_index;
+};
+
+
+void brw_vs_emit( struct brw_vs_compile *c );
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_vs_emit.c b/src/gallium/drivers/i965/brw_vs_emit.c
new file mode 100644 (file)
index 0000000..8a16205
--- /dev/null
@@ -0,0 +1,1654 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "pipe/p_shader_tokens.h"
+            
+#include "util/u_memory.h"
+#include "util/u_math.h"
+
+#include "tgsi/tgsi_parse.h"
+#include "tgsi/tgsi_dump.h"
+#include "tgsi/tgsi_info.h"
+
+#include "brw_context.h"
+#include "brw_vs.h"
+#include "brw_debug.h"
+#include "brw_disasm.h"
+
+/* Choose one of the 4 vec4's which can be packed into each 16-wide reg.
+ */
+static INLINE struct brw_reg brw_vec4_grf_repeat( GLuint reg, GLuint slot )
+{
+   int nr = reg + slot/2;
+   int subnr = (slot%2) * 4;
+
+   return stride(brw_vec4_grf(nr, subnr), 0, 4, 1);
+}
+
+
+static struct brw_reg get_tmp( struct brw_vs_compile *c )
+{
+   struct brw_reg tmp = brw_vec8_grf(c->last_tmp, 0);
+
+   if (++c->last_tmp > c->prog_data.total_grf)
+      c->prog_data.total_grf = c->last_tmp;
+
+   return tmp;
+}
+
+static void release_tmp( struct brw_vs_compile *c, struct brw_reg tmp )
+{
+   if (tmp.nr == c->last_tmp-1)
+      c->last_tmp--;
+}
+                              
+static void release_tmps( struct brw_vs_compile *c )
+{
+   c->last_tmp = c->first_tmp;
+}
+
+
+static boolean is_position_output( struct brw_vs_compile *c,
+                                   unsigned vs_output )
+{
+   const struct brw_vertex_shader *vs = c->vp;
+   unsigned semantic = vs->info.output_semantic_name[vs_output];
+   unsigned index = vs->info.output_semantic_index[vs_output];
+      
+   return (semantic == TGSI_SEMANTIC_POSITION &&
+           index == 0);
+}
+
+
+static boolean find_output_slot( struct brw_vs_compile *c,
+                                  unsigned vs_output,
+                                  unsigned *fs_input_slot )
+{
+   const struct brw_vertex_shader *vs = c->vp;
+   unsigned semantic = vs->info.output_semantic_name[vs_output];
+   unsigned index = vs->info.output_semantic_index[vs_output];
+   unsigned i;
+
+   for (i = 0; i < c->key.fs_signature.nr_inputs; i++) {
+      if (c->key.fs_signature.input[i].semantic == semantic &&
+          c->key.fs_signature.input[i].semantic_index == index) {
+         *fs_input_slot = i;
+         return TRUE;
+      }
+   }
+
+   return FALSE;
+}
+
+
+/**
+ * Preallocate GRF register before code emit.
+ * Do things as simply as possible.  Allocate and populate all regs
+ * ahead of time.
+ */
+static void brw_vs_alloc_regs( struct brw_vs_compile *c )
+{
+   GLuint i, reg = 0, subreg = 0, mrf;
+   int attributes_in_vue;
+
+   /* Determine whether to use a real constant buffer or use a block
+    * of GRF registers for constants.  The later is faster but only
+    * works if everything fits in the GRF.
+    * XXX this heuristic/check may need some fine tuning...
+    */
+   if (c->vp->info.file_max[TGSI_FILE_CONSTANT] + 1 +
+       c->vp->info.file_max[TGSI_FILE_IMMEDIATE] + 1 +
+       c->vp->info.file_max[TGSI_FILE_TEMPORARY] + 1 + 21 > BRW_MAX_GRF)
+      c->vp->use_const_buffer = GL_TRUE;
+   else {
+      /* XXX: immediates can go elsewhere if necessary:
+       */
+      assert(c->vp->info.file_max[TGSI_FILE_IMMEDIATE] + 1 +
+            c->vp->info.file_max[TGSI_FILE_TEMPORARY] + 1 + 21 <= BRW_MAX_GRF);
+
+      c->vp->use_const_buffer = GL_FALSE;
+   }
+
+   /*printf("use_const_buffer = %d\n", c->vp->use_const_buffer);*/
+
+   /* r0 -- reserved as usual
+    */
+   c->r0 = brw_vec8_grf(reg, 0);
+   reg++;
+
+   /* User clip planes from curbe: 
+    */
+   if (c->key.nr_userclip) {
+      /* Skip over fixed planes:  Or never read them into vs unit?
+       */
+      subreg += 6;
+
+      for (i = 0; i < c->key.nr_userclip; i++, subreg++) {
+        c->userplane[i] = 
+            stride( brw_vec4_grf(reg+subreg/2, (subreg%2) * 4), 0, 4, 1);
+      }     
+
+      /* Deal with curbe alignment:
+       */
+      subreg = align(subreg, 2);
+      /*reg += ((6 + c->key.nr_userclip + 3) / 4) * 2;*/
+   }
+
+
+   /* Immediates: always in the curbe.
+    *
+    * XXX: Can try to encode some immediates as brw immediates
+    * XXX: Make sure ureg sets minimal immediate size and respect it
+    * here.
+    */
+   for (i = 0; i < c->vp->info.immediate_count; i++, subreg++) {
+      c->regs[TGSI_FILE_IMMEDIATE][i] = 
+         stride( brw_vec4_grf(reg+subreg/2, (subreg%2) * 4), 0, 4, 1);
+   }
+   c->prog_data.nr_params = c->vp->info.immediate_count * 4;
+
+
+   /* Vertex constant buffer.
+    *
+    * Constants from the buffer can be either cached in the curbe or
+    * loaded as needed from the actual constant buffer.
+    */
+   if (!c->vp->use_const_buffer) {
+      GLuint nr_params = c->vp->info.file_max[TGSI_FILE_CONSTANT] + 1;
+
+      for (i = 0; i < nr_params; i++, subreg++) {
+         c->regs[TGSI_FILE_CONSTANT][i] =
+            stride( brw_vec4_grf(reg+subreg/2, (subreg%2) * 4), 0, 4, 1);
+      }
+
+      c->prog_data.nr_params += nr_params * 4;
+   }
+
+   /* All regs allocated
+    */
+   reg += (subreg + 1) / 2;
+   c->prog_data.curb_read_length = reg - 1;
+
+
+   /* Allocate input regs:  
+    */
+   c->nr_inputs = c->vp->info.num_inputs;
+   for (i = 0; i < c->nr_inputs; i++) {
+      c->regs[TGSI_FILE_INPUT][i] = brw_vec8_grf(reg, 0);
+      reg++;
+   }
+
+   /* If there are no inputs, we'll still be reading one attribute's worth
+    * because it's required -- see urb_read_length setting.
+    */
+   if (c->nr_inputs == 0)
+      reg++;
+
+
+
+   /* Allocate outputs.  The non-position outputs go straight into message regs.
+    */
+   c->nr_outputs = c->prog_data.nr_outputs;
+
+   if (c->chipset.is_igdng)
+      mrf = 8;
+   else
+      mrf = 4;
+
+   
+   if (c->key.fs_signature.nr_inputs > BRW_MAX_MRF) {
+      c->overflow_grf_start = reg;
+      c->overflow_count = c->key.fs_signature.nr_inputs - BRW_MAX_MRF;
+      reg += c->overflow_count;
+   }
+
+   /* XXX: need to access vertex output semantics here:
+    */
+   for (i = 0; i < c->nr_outputs; i++) {
+      unsigned slot;
+
+      /* XXX: Put output position in slot zero always.  Clipper, etc,
+       * need access to this reg.
+       */
+      if (is_position_output(c, i)) {
+        c->regs[TGSI_FILE_OUTPUT][i] = brw_vec8_grf(reg, 0); /* copy to mrf 0 */
+        reg++;
+      }
+      else if (find_output_slot(c, i, &slot)) {
+         
+         if (0 /* is_psize_output(c, i) */ ) {
+            /* c->psize_out.grf = reg; */
+            /* c->psize_out.mrf = i; */
+         }
+         
+         /* The first (16-4) outputs can go straight into the message regs.
+          */
+         if (slot + mrf < BRW_MAX_MRF) {
+            c->regs[TGSI_FILE_OUTPUT][i] = brw_message_reg(slot + mrf);
+         }
+         else {
+            int grf = c->overflow_grf_start + slot - BRW_MAX_MRF;
+            c->regs[TGSI_FILE_OUTPUT][i] = brw_vec8_grf(grf, 0);
+         }
+      }
+      else {
+         c->regs[TGSI_FILE_OUTPUT][i] = brw_null_reg();
+      }
+   }     
+
+   /* Allocate program temporaries:
+    */
+   
+   for (i = 0; i < c->vp->info.file_max[TGSI_FILE_TEMPORARY]+1; i++) {
+      c->regs[TGSI_FILE_TEMPORARY][i] = brw_vec8_grf(reg, 0);
+      reg++;
+   }
+
+   /* Address reg(s).  Don't try to use the internal address reg until
+    * deref time.
+    */
+   for (i = 0; i < c->vp->info.file_max[TGSI_FILE_ADDRESS]+1; i++) {
+      c->regs[TGSI_FILE_ADDRESS][i] =  brw_reg(BRW_GENERAL_REGISTER_FILE,
+                                            reg,
+                                            0,
+                                            BRW_REGISTER_TYPE_D,
+                                            BRW_VERTICAL_STRIDE_8,
+                                            BRW_WIDTH_8,
+                                            BRW_HORIZONTAL_STRIDE_1,
+                                            BRW_SWIZZLE_XXXX,
+                                            BRW_WRITEMASK_X);
+      reg++;
+   }
+
+   if (c->vp->use_const_buffer) {
+      for (i = 0; i < 3; i++) {
+         c->current_const[i].index = -1;
+         c->current_const[i].reg = brw_vec8_grf(reg, 0);
+         reg++;
+      }
+   }
+
+#if 0
+   for (i = 0; i < 128; i++) {
+      if (c->output_regs[i].used_in_src) {
+         c->output_regs[i].reg = brw_vec8_grf(reg, 0);
+         reg++;
+      }
+   }
+#endif
+
+   if (c->vp->has_flow_control) {
+      c->stack =  brw_uw16_reg(BRW_GENERAL_REGISTER_FILE, reg, 0);
+      reg += 2;
+   }
+
+   /* Some opcodes need an internal temporary:
+    */
+   c->first_tmp = reg;
+   c->last_tmp = reg;          /* for allocation purposes */
+
+   /* Each input reg holds data from two vertices.  The
+    * urb_read_length is the number of registers read from *each*
+    * vertex urb, so is half the amount:
+    */
+   c->prog_data.urb_read_length = (c->nr_inputs + 1) / 2;
+
+   /* Setting this field to 0 leads to undefined behavior according to the
+    * the VS_STATE docs.  Our VUEs will always have at least one attribute
+    * sitting in them, even if it's padding.
+    */
+   if (c->prog_data.urb_read_length == 0)
+      c->prog_data.urb_read_length = 1;
+
+   /* The VS VUEs are shared by VF (outputting our inputs) and VS, so size
+    * them to fit the biggest thing they need to.
+    */
+   attributes_in_vue = MAX2(c->nr_outputs, c->nr_inputs);
+
+   if (c->chipset.is_igdng)
+      c->prog_data.urb_entry_size = (attributes_in_vue + 6 + 3) / 4;
+   else
+      c->prog_data.urb_entry_size = (attributes_in_vue + 2 + 3) / 4;
+
+   c->prog_data.total_grf = reg;
+
+   if (BRW_DEBUG & DEBUG_VS) {
+      debug_printf("%s NumAddrRegs %d\n", __FUNCTION__, 
+                  c->vp->info.file_max[TGSI_FILE_ADDRESS]+1);
+      debug_printf("%s NumTemps %d\n", __FUNCTION__,
+                  c->vp->info.file_max[TGSI_FILE_TEMPORARY]+1);
+      debug_printf("%s reg = %d\n", __FUNCTION__, reg);
+   }
+}
+
+
+/**
+ * If an instruction uses a temp reg both as a src and the dest, we
+ * sometimes need to allocate an intermediate temporary.
+ */
+static void unalias1( struct brw_vs_compile *c,
+                     struct brw_reg dst,
+                     struct brw_reg arg0,
+                     void (*func)( struct brw_vs_compile *,
+                                   struct brw_reg,
+                                   struct brw_reg ))
+{
+   if (dst.file == arg0.file && dst.nr == arg0.nr) {
+      struct brw_compile *p = &c->func;
+      struct brw_reg tmp = brw_writemask(get_tmp(c), dst.dw1.bits.writemask);
+      func(c, tmp, arg0);
+      brw_MOV(p, dst, tmp);
+      release_tmp(c, tmp);
+   }
+   else {
+      func(c, dst, arg0);
+   }
+}
+
+/**
+ * \sa unalias2
+ * Checkes if 2-operand instruction needs an intermediate temporary.
+ */
+static void unalias2( struct brw_vs_compile *c,
+                     struct brw_reg dst,
+                     struct brw_reg arg0,
+                     struct brw_reg arg1,
+                     void (*func)( struct brw_vs_compile *,
+                                   struct brw_reg,
+                                   struct brw_reg,
+                                   struct brw_reg ))
+{
+   if ((dst.file == arg0.file && dst.nr == arg0.nr) ||
+       (dst.file == arg1.file && dst.nr == arg1.nr)) {
+      struct brw_compile *p = &c->func;
+      struct brw_reg tmp = brw_writemask(get_tmp(c), dst.dw1.bits.writemask);
+      func(c, tmp, arg0, arg1);
+      brw_MOV(p, dst, tmp);
+      release_tmp(c, tmp);
+   }
+   else {
+      func(c, dst, arg0, arg1);
+   }
+}
+
+/**
+ * \sa unalias2
+ * Checkes if 3-operand instruction needs an intermediate temporary.
+ */
+static void unalias3( struct brw_vs_compile *c,
+                     struct brw_reg dst,
+                     struct brw_reg arg0,
+                     struct brw_reg arg1,
+                     struct brw_reg arg2,
+                     void (*func)( struct brw_vs_compile *,
+                                   struct brw_reg,
+                                   struct brw_reg,
+                                   struct brw_reg,
+                                   struct brw_reg ))
+{
+   if ((dst.file == arg0.file && dst.nr == arg0.nr) ||
+       (dst.file == arg1.file && dst.nr == arg1.nr) ||
+       (dst.file == arg2.file && dst.nr == arg2.nr)) {
+      struct brw_compile *p = &c->func;
+      struct brw_reg tmp = brw_writemask(get_tmp(c), dst.dw1.bits.writemask);
+      func(c, tmp, arg0, arg1, arg2);
+      brw_MOV(p, dst, tmp);
+      release_tmp(c, tmp);
+   }
+   else {
+      func(c, dst, arg0, arg1, arg2);
+   }
+}
+
+static void emit_sop( struct brw_compile *p,
+                      struct brw_reg dst,
+                      struct brw_reg arg0,
+                      struct brw_reg arg1, 
+                     GLuint cond)
+{
+   brw_MOV(p, dst, brw_imm_f(0.0f));
+   brw_CMP(p, brw_null_reg(), cond, arg0, arg1);
+   brw_MOV(p, dst, brw_imm_f(1.0f));
+   brw_set_predicate_control_flag_value(p, 0xff);
+}
+
+static void emit_seq( struct brw_compile *p,
+                      struct brw_reg dst,
+                      struct brw_reg arg0,
+                      struct brw_reg arg1 )
+{
+   emit_sop(p, dst, arg0, arg1, BRW_CONDITIONAL_EQ);
+}
+
+static void emit_sne( struct brw_compile *p,
+                      struct brw_reg dst,
+                      struct brw_reg arg0,
+                      struct brw_reg arg1 )
+{
+   emit_sop(p, dst, arg0, arg1, BRW_CONDITIONAL_NEQ);
+}
+static void emit_slt( struct brw_compile *p, 
+                     struct brw_reg dst,
+                     struct brw_reg arg0,
+                     struct brw_reg arg1 )
+{
+   emit_sop(p, dst, arg0, arg1, BRW_CONDITIONAL_L);
+}
+
+static void emit_sle( struct brw_compile *p, 
+                     struct brw_reg dst,
+                     struct brw_reg arg0,
+                     struct brw_reg arg1 )
+{
+   emit_sop(p, dst, arg0, arg1, BRW_CONDITIONAL_LE);
+}
+
+static void emit_sgt( struct brw_compile *p, 
+                     struct brw_reg dst,
+                     struct brw_reg arg0,
+                     struct brw_reg arg1 )
+{
+   emit_sop(p, dst, arg0, arg1, BRW_CONDITIONAL_G);
+}
+
+static void emit_sge( struct brw_compile *p, 
+                     struct brw_reg dst,
+                     struct brw_reg arg0,
+                     struct brw_reg arg1 )
+{
+  emit_sop(p, dst, arg0, arg1, BRW_CONDITIONAL_GE);
+}
+
+static void emit_max( struct brw_compile *p, 
+                     struct brw_reg dst,
+                     struct brw_reg arg0,
+                     struct brw_reg arg1 )
+{
+   brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_L, arg0, arg1);
+   brw_SEL(p, dst, arg1, arg0);
+   brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+}
+
+static void emit_min( struct brw_compile *p, 
+                     struct brw_reg dst,
+                     struct brw_reg arg0,
+                     struct brw_reg arg1 )
+{
+   brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_L, arg0, arg1);
+   brw_SEL(p, dst, arg0, arg1);
+   brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+}
+
+
+static void emit_math1( struct brw_vs_compile *c,
+                       GLuint function,
+                       struct brw_reg dst,
+                       struct brw_reg arg0,
+                       GLuint precision)
+{
+   /* There are various odd behaviours with SEND on the simulator.  In
+    * addition there are documented issues with the fact that the GEN4
+    * processor doesn't do dependency control properly on SEND
+    * results.  So, on balance, this kludge to get around failures
+    * with writemasked math results looks like it might be necessary
+    * whether that turns out to be a simulator bug or not:
+    */
+   struct brw_compile *p = &c->func;
+   struct brw_reg tmp = dst;
+   GLboolean need_tmp = (dst.dw1.bits.writemask != 0xf ||
+                        dst.file != BRW_GENERAL_REGISTER_FILE);
+
+   if (need_tmp) 
+      tmp = get_tmp(c);
+
+   brw_math(p, 
+           tmp,
+           function,
+           BRW_MATH_SATURATE_NONE,
+           2,
+           arg0,
+           BRW_MATH_DATA_SCALAR,
+           precision);
+
+   if (need_tmp) {
+      brw_MOV(p, dst, tmp);
+      release_tmp(c, tmp);
+   }
+}
+
+
+static void emit_math2( struct brw_vs_compile *c, 
+                       GLuint function,
+                       struct brw_reg dst,
+                       struct brw_reg arg0,
+                       struct brw_reg arg1,
+                       GLuint precision)
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg tmp = dst;
+   GLboolean need_tmp = (dst.dw1.bits.writemask != 0xf ||
+                        dst.file != BRW_GENERAL_REGISTER_FILE);
+
+   if (need_tmp) 
+      tmp = get_tmp(c);
+
+   brw_MOV(p, brw_message_reg(3), arg1);
+   
+   brw_math(p, 
+           tmp,
+           function,
+           BRW_MATH_SATURATE_NONE,
+           2,
+           arg0,
+           BRW_MATH_DATA_SCALAR,
+           precision);
+
+   if (need_tmp) {
+      brw_MOV(p, dst, tmp);
+      release_tmp(c, tmp);
+   }
+}
+
+
+static void emit_exp_noalias( struct brw_vs_compile *c,
+                             struct brw_reg dst,
+                             struct brw_reg arg0 )
+{
+   struct brw_compile *p = &c->func;
+   
+
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_X) {
+      struct brw_reg tmp = get_tmp(c);
+      struct brw_reg tmp_d = retype(tmp, BRW_REGISTER_TYPE_D);
+
+      /* tmp_d = floor(arg0.x) */
+      brw_RNDD(p, tmp_d, brw_swizzle1(arg0, 0));
+
+      /* result[0] = 2.0 ^ tmp */
+
+      /* Adjust exponent for floating point: 
+       * exp += 127 
+       */
+      brw_ADD(p, brw_writemask(tmp_d, BRW_WRITEMASK_X), tmp_d, brw_imm_d(127));
+
+      /* Install exponent and sign.  
+       * Excess drops off the edge: 
+       */
+      brw_SHL(p, brw_writemask(retype(dst, BRW_REGISTER_TYPE_D), BRW_WRITEMASK_X), 
+             tmp_d, brw_imm_d(23));
+
+      release_tmp(c, tmp);
+   }
+
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_Y) {
+      /* result[1] = arg0.x - floor(arg0.x) */
+      brw_FRC(p, brw_writemask(dst, BRW_WRITEMASK_Y), brw_swizzle1(arg0, 0));
+   }
+   
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_Z) {
+      /* As with the LOG instruction, we might be better off just
+       * doing a taylor expansion here, seeing as we have to do all
+       * the prep work.
+       *
+       * If mathbox partial precision is too low, consider also:
+       * result[3] = result[0] * EXP(result[1])
+       */
+      emit_math1(c, 
+                BRW_MATH_FUNCTION_EXP, 
+                brw_writemask(dst, BRW_WRITEMASK_Z),
+                brw_swizzle1(arg0, 0), 
+                BRW_MATH_PRECISION_FULL);
+   }  
+
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_W) {
+      /* result[3] = 1.0; */
+      brw_MOV(p, brw_writemask(dst, BRW_WRITEMASK_W), brw_imm_f(1));
+   }
+}
+
+
+static void emit_log_noalias( struct brw_vs_compile *c,
+                             struct brw_reg dst,
+                             struct brw_reg arg0 )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg tmp = dst;
+   struct brw_reg tmp_ud = retype(tmp, BRW_REGISTER_TYPE_UD);
+   struct brw_reg arg0_ud = retype(arg0, BRW_REGISTER_TYPE_UD);
+   GLboolean need_tmp = (dst.dw1.bits.writemask != 0xf ||
+                        dst.file != BRW_GENERAL_REGISTER_FILE);
+
+   if (need_tmp) {
+      tmp = get_tmp(c);
+      tmp_ud = retype(tmp, BRW_REGISTER_TYPE_UD);
+   }
+   
+   /* Perform mant = frexpf(fabsf(x), &exp), adjust exp and mnt
+    * according to spec:
+    *
+    * These almost look likey they could be joined up, but not really
+    * practical:
+    *
+    * result[0].f = (x.i & ((1<<31)-1) >> 23) - 127
+    * result[1].i = (x.i & ((1<<23)-1)        + (127<<23)
+    */
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_XZ) {
+      brw_AND(p, 
+             brw_writemask(tmp_ud, BRW_WRITEMASK_X),
+             brw_swizzle1(arg0_ud, 0),
+             brw_imm_ud((1U<<31)-1));
+
+      brw_SHR(p, 
+             brw_writemask(tmp_ud, BRW_WRITEMASK_X), 
+             tmp_ud,
+             brw_imm_ud(23));
+
+      brw_ADD(p, 
+             brw_writemask(tmp, BRW_WRITEMASK_X), 
+             retype(tmp_ud, BRW_REGISTER_TYPE_D),      /* does it matter? */
+             brw_imm_d(-127));
+   }
+
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_YZ) {
+      brw_AND(p, 
+             brw_writemask(tmp_ud, BRW_WRITEMASK_Y),
+             brw_swizzle1(arg0_ud, 0),
+             brw_imm_ud((1<<23)-1));
+
+      brw_OR(p, 
+            brw_writemask(tmp_ud, BRW_WRITEMASK_Y), 
+            tmp_ud,
+            brw_imm_ud(127<<23));
+   }
+   
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_Z) {
+      /* result[2] = result[0] + LOG2(result[1]); */
+
+      /* Why bother?  The above is just a hint how to do this with a
+       * taylor series.  Maybe we *should* use a taylor series as by
+       * the time all the above has been done it's almost certainly
+       * quicker than calling the mathbox, even with low precision.
+       * 
+       * Options are:
+       *    - result[0] + mathbox.LOG2(result[1])
+       *    - mathbox.LOG2(arg0.x)
+       *    - result[0] + inline_taylor_approx(result[1])
+       */
+      emit_math1(c, 
+                BRW_MATH_FUNCTION_LOG, 
+                brw_writemask(tmp, BRW_WRITEMASK_Z), 
+                brw_swizzle1(tmp, 1), 
+                BRW_MATH_PRECISION_FULL);
+      
+      brw_ADD(p, 
+             brw_writemask(tmp, BRW_WRITEMASK_Z), 
+             brw_swizzle1(tmp, 2), 
+             brw_swizzle1(tmp, 0));
+   }  
+
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_W) {
+      /* result[3] = 1.0; */
+      brw_MOV(p, brw_writemask(tmp, BRW_WRITEMASK_W), brw_imm_f(1));
+   }
+
+   if (need_tmp) {
+      brw_MOV(p, dst, tmp);
+      release_tmp(c, tmp);
+   }
+}
+
+
+/* Need to unalias - consider swizzles:   r0 = DST r0.xxxx r1
+ */
+static void emit_dst_noalias( struct brw_vs_compile *c, 
+                             struct brw_reg dst,
+                             struct brw_reg arg0,
+                             struct brw_reg arg1)
+{
+   struct brw_compile *p = &c->func;
+
+   /* There must be a better way to do this: 
+    */
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_X)
+      brw_MOV(p, brw_writemask(dst, BRW_WRITEMASK_X), brw_imm_f(1.0));
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_Y)
+      brw_MUL(p, brw_writemask(dst, BRW_WRITEMASK_Y), arg0, arg1);
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_Z)
+      brw_MOV(p, brw_writemask(dst, BRW_WRITEMASK_Z), arg0);
+   if (dst.dw1.bits.writemask & BRW_WRITEMASK_W)
+      brw_MOV(p, brw_writemask(dst, BRW_WRITEMASK_W), arg1);
+}
+
+
+static void emit_xpd( struct brw_compile *p,
+                     struct brw_reg dst,
+                     struct brw_reg t,
+                     struct brw_reg u)
+{
+   brw_MUL(p, brw_null_reg(), brw_swizzle(t, 1,2,0,3),  brw_swizzle(u,2,0,1,3));
+   brw_MAC(p, dst,     negate(brw_swizzle(t, 2,0,1,3)), brw_swizzle(u,1,2,0,3));
+}
+
+
+static void emit_lit_noalias( struct brw_vs_compile *c, 
+                             struct brw_reg dst,
+                             struct brw_reg arg0 )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_instruction *if_insn;
+   struct brw_reg tmp = dst;
+   GLboolean need_tmp = (dst.file != BRW_GENERAL_REGISTER_FILE);
+
+   if (need_tmp) 
+      tmp = get_tmp(c);
+   
+   brw_MOV(p, brw_writemask(dst, BRW_WRITEMASK_YZ), brw_imm_f(0)); 
+   brw_MOV(p, brw_writemask(dst, BRW_WRITEMASK_XW), brw_imm_f(1)); 
+
+   /* Need to use BRW_EXECUTE_8 and also do an 8-wide compare in order
+    * to get all channels active inside the IF.  In the clipping code
+    * we run with NoMask, so it's not an option and we can use
+    * BRW_EXECUTE_1 for all comparisions.
+    */
+   brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_G, brw_swizzle1(arg0,0), brw_imm_f(0));
+   if_insn = brw_IF(p, BRW_EXECUTE_8);
+   {
+      brw_MOV(p, brw_writemask(dst, BRW_WRITEMASK_Y), brw_swizzle1(arg0,0));
+
+      brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_G, brw_swizzle1(arg0,1), brw_imm_f(0));
+      brw_MOV(p, brw_writemask(tmp, BRW_WRITEMASK_Z),  brw_swizzle1(arg0,1));
+      brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+
+      emit_math2(c, 
+                BRW_MATH_FUNCTION_POW, 
+                brw_writemask(dst, BRW_WRITEMASK_Z),
+                brw_swizzle1(tmp, 2),
+                brw_swizzle1(arg0, 3),
+                BRW_MATH_PRECISION_PARTIAL);      
+   }
+
+   brw_ENDIF(p, if_insn);
+
+   release_tmp(c, tmp);
+}
+
+static void emit_lrp_noalias(struct brw_vs_compile *c,
+                            struct brw_reg dst,
+                            struct brw_reg arg0,
+                            struct brw_reg arg1,
+                            struct brw_reg arg2)
+{
+   struct brw_compile *p = &c->func;
+
+   brw_ADD(p, dst, negate(arg0), brw_imm_f(1.0));
+   brw_MUL(p, brw_null_reg(), dst, arg2);
+   brw_MAC(p, dst, arg0, arg1);
+}
+
+/** 3 or 4-component vector normalization */
+static void emit_nrm( struct brw_vs_compile *c, 
+                      struct brw_reg dst,
+                      struct brw_reg arg0,
+                      int num_comps)
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg tmp = get_tmp(c);
+
+   /* tmp = dot(arg0, arg0) */
+   if (num_comps == 3)
+      brw_DP3(p, tmp, arg0, arg0);
+   else
+      brw_DP4(p, tmp, arg0, arg0);
+
+   /* tmp = 1 / sqrt(tmp) */
+   emit_math1(c, BRW_MATH_FUNCTION_RSQ, tmp, tmp, BRW_MATH_PRECISION_FULL);
+
+   /* dst = arg0 * tmp */
+   brw_MUL(p, dst, arg0, tmp);
+
+   release_tmp(c, tmp);
+}
+
+
+static struct brw_reg
+get_constant(struct brw_vs_compile *c,
+            GLuint argIndex,
+            GLuint index,
+            GLboolean relAddr)
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg const_reg;
+   struct brw_reg const2_reg;
+
+   assert(argIndex < 3);
+
+   if (c->current_const[argIndex].index != index || relAddr) {
+      struct brw_reg addrReg = c->regs[TGSI_FILE_ADDRESS][0];
+
+      c->current_const[argIndex].index = index;
+
+#if 0
+      printf("  fetch const[%d] for arg %d into reg %d\n",
+             src.Index, argIndex, c->current_const[argIndex].reg.nr);
+#endif
+      /* need to fetch the constant now */
+      brw_dp_READ_4_vs(p,
+                       c->current_const[argIndex].reg,/* writeback dest */
+                       0,                             /* oword */
+                       relAddr,                       /* relative indexing? */
+                       addrReg,                       /* address register */
+                       16 * index,               /* byte offset */
+                       SURF_INDEX_VERT_CONST_BUFFER   /* binding table index */
+                       );
+
+      if (relAddr) {
+         /* second read */
+         const2_reg = get_tmp(c);
+
+         /* use upper half of address reg for second read */
+         addrReg = stride(addrReg, 0, 4, 0);
+         addrReg.subnr = 16;
+
+         brw_dp_READ_4_vs(p,
+                          const2_reg,              /* writeback dest */
+                          1,                       /* oword */
+                          relAddr,                 /* relative indexing? */
+                          addrReg,                 /* address register */
+                          16 * index,         /* byte offset */
+                          SURF_INDEX_VERT_CONST_BUFFER
+                          );
+      }
+   }
+
+   const_reg = c->current_const[argIndex].reg;
+
+   if (relAddr) {
+      /* merge the two Owords into the constant register */
+      /* const_reg[7..4] = const2_reg[7..4] */
+      brw_MOV(p,
+              suboffset(stride(const_reg, 0, 4, 1), 4),
+              suboffset(stride(const2_reg, 0, 4, 1), 4));
+      release_tmp(c, const2_reg);
+   }
+   else {
+      /* replicate lower four floats into upper half (to get XYZWXYZW) */
+      const_reg = stride(const_reg, 0, 4, 0);
+      const_reg.subnr = 0;
+   }
+
+   return const_reg;
+}
+
+
+
+/* TODO: relative addressing!
+ */
+static struct brw_reg get_reg( struct brw_vs_compile *c,
+                              enum tgsi_file_type file,
+                              GLuint index )
+{
+   switch (file) {
+   case TGSI_FILE_TEMPORARY:
+   case TGSI_FILE_INPUT:
+   case TGSI_FILE_OUTPUT:
+   case TGSI_FILE_CONSTANT:
+      assert(c->regs[file][index].nr != 0);
+      return c->regs[file][index];
+
+   case TGSI_FILE_ADDRESS:
+      assert(index == 0);
+      return c->regs[file][index];
+
+   case TGSI_FILE_NULL:                        /* undef values */
+      return brw_null_reg();
+
+   default:
+      assert(0);
+      return brw_null_reg();
+   }
+}
+
+
+/**
+ * Indirect addressing:  get reg[[arg] + offset].
+ */
+static struct brw_reg deref( struct brw_vs_compile *c,
+                            struct brw_reg arg,
+                            GLint offset)
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg tmp = vec4(get_tmp(c));
+   struct brw_reg addr_reg = c->regs[TGSI_FILE_ADDRESS][0];
+   struct brw_reg vp_address = retype(vec1(addr_reg), BRW_REGISTER_TYPE_UW);
+   GLuint byte_offset = arg.nr * 32 + arg.subnr + offset * 16;
+   struct brw_reg indirect = brw_vec4_indirect(0,0);
+
+   {
+      brw_push_insn_state(p);
+      brw_set_access_mode(p, BRW_ALIGN_1);
+
+      /* This is pretty clunky - load the address register twice and
+       * fetch each 4-dword value in turn.  There must be a way to do
+       * this in a single pass, but I couldn't get it to work.
+       */
+      brw_ADD(p, brw_address_reg(0), vp_address, brw_imm_d(byte_offset));
+      brw_MOV(p, tmp, indirect);
+
+      brw_ADD(p, brw_address_reg(0), suboffset(vp_address, 8), brw_imm_d(byte_offset));
+      brw_MOV(p, suboffset(tmp, 4), indirect);
+
+      brw_pop_insn_state(p);
+   }
+   
+   /* NOTE: tmp not released */
+   return vec8(tmp);
+}
+
+
+/**
+ * Get brw reg corresponding to the instruction's [argIndex] src reg.
+ * TODO: relative addressing!
+ */
+static struct brw_reg
+get_src_reg( struct brw_vs_compile *c,
+            GLuint argIndex,
+            GLuint file,
+            GLint index,
+            GLboolean relAddr )
+{
+
+   switch (file) {
+   case TGSI_FILE_TEMPORARY:
+   case TGSI_FILE_INPUT:
+   case TGSI_FILE_OUTPUT:
+      if (relAddr) {
+         return deref(c, c->regs[file][0], index);
+      }
+      else {
+         assert(c->regs[file][index].nr != 0);
+         return c->regs[file][index];
+      }
+
+   case TGSI_FILE_IMMEDIATE:
+      return c->regs[file][index];
+
+   case TGSI_FILE_CONSTANT:
+      if (c->vp->use_const_buffer) {
+         return get_constant(c, argIndex, index, relAddr);
+      }
+      else if (relAddr) {
+         return deref(c, c->regs[TGSI_FILE_CONSTANT][0], index);
+      }
+      else {
+         assert(c->regs[TGSI_FILE_CONSTANT][index].nr != 0);
+         return c->regs[TGSI_FILE_CONSTANT][index];
+      }
+   case TGSI_FILE_ADDRESS:
+      assert(index == 0);
+      return c->regs[file][index];
+
+   case TGSI_FILE_NULL:
+      /* this is a normal case since we loop over all three src args */
+      return brw_null_reg();
+
+   default:
+      assert(0);
+      return brw_null_reg();
+   }
+}
+
+
+static void emit_arl( struct brw_vs_compile *c,
+                     struct brw_reg dst,
+                     struct brw_reg arg0 )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg tmp = dst;
+   GLboolean need_tmp = (dst.file != BRW_GENERAL_REGISTER_FILE);
+   
+   if (need_tmp) 
+      tmp = get_tmp(c);
+
+   brw_RNDD(p, tmp, arg0);               /* tmp = round(arg0) */
+   brw_MUL(p, dst, tmp, brw_imm_d(16));  /* dst = tmp * 16 */
+
+   if (need_tmp)
+      release_tmp(c, tmp);
+}
+
+
+/**
+ * Return the brw reg for the given instruction's src argument.
+ */
+static struct brw_reg get_arg( struct brw_vs_compile *c,
+                               const struct tgsi_full_src_register *src,
+                               GLuint argIndex )
+{
+   struct brw_reg reg;
+
+   if (src->Register.File == TGSI_FILE_NULL)
+      return brw_null_reg();
+
+   reg = get_src_reg(c, argIndex,
+                    src->Register.File,
+                    src->Register.Index,
+                    src->Register.Indirect);
+
+   /* Convert 3-bit swizzle to 2-bit.  
+    */
+   reg.dw1.bits.swizzle = BRW_SWIZZLE4(src->Register.SwizzleX,
+                                      src->Register.SwizzleY,
+                                      src->Register.SwizzleZ,
+                                      src->Register.SwizzleW);
+
+   reg.negate = src->Register.Negate ? 1 : 0;   
+
+   /* XXX: abs, absneg
+    */
+
+   return reg;
+}
+
+
+/**
+ * Get brw register for the given program dest register.
+ */
+static struct brw_reg get_dst( struct brw_vs_compile *c,
+                              unsigned file,
+                              unsigned index,
+                              unsigned writemask )
+{
+   struct brw_reg reg;
+
+   switch (file) {
+   case TGSI_FILE_TEMPORARY:
+   case TGSI_FILE_OUTPUT:
+      assert(c->regs[file][index].nr != 0);
+      reg = c->regs[file][index];
+      break;
+   case TGSI_FILE_ADDRESS:
+      assert(index == 0);
+      reg = c->regs[file][index];
+      break;
+   case TGSI_FILE_NULL:
+      /* we may hit this for OPCODE_END, OPCODE_KIL, etc */
+      reg = brw_null_reg();
+      break;
+   default:
+      assert(0);
+      reg = brw_null_reg();
+   }
+
+   reg.dw1.bits.writemask = writemask;
+
+   return reg;
+}
+
+
+
+
+/**
+ * Post-vertex-program processing.  Send the results to the URB.
+ */
+static void emit_vertex_write( struct brw_vs_compile *c)
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg m0 = brw_message_reg(0);
+   struct brw_reg pos = c->regs[TGSI_FILE_OUTPUT][VERT_RESULT_HPOS];
+   struct brw_reg ndc;
+   int eot;
+   int i;
+   GLuint len_vertext_header = 2;
+
+   /* Build ndc coords */
+   ndc = get_tmp(c);
+   /* ndc = 1.0 / pos.w */
+   emit_math1(c, BRW_MATH_FUNCTION_INV, ndc, brw_swizzle1(pos, 3), BRW_MATH_PRECISION_FULL);
+   /* ndc.xyz = pos * ndc */
+   brw_MUL(p, brw_writemask(ndc, BRW_WRITEMASK_XYZ), pos, ndc);
+
+   /* Update the header for point size, user clipping flags, and -ve rhw
+    * workaround.
+    */
+   if (c->prog_data.writes_psiz ||
+       c->key.nr_userclip || 
+       c->chipset.is_965)
+   {
+      struct brw_reg header1 = retype(get_tmp(c), BRW_REGISTER_TYPE_UD);
+      GLuint i;
+
+      brw_MOV(p, header1, brw_imm_ud(0));
+
+      brw_set_access_mode(p, BRW_ALIGN_16);    
+
+      if (c->prog_data.writes_psiz) {
+        struct brw_reg psiz = c->regs[TGSI_FILE_OUTPUT][VERT_RESULT_PSIZ];
+        brw_MUL(p, brw_writemask(header1, BRW_WRITEMASK_W), brw_swizzle1(psiz, 0), brw_imm_f(1<<11));
+        brw_AND(p, brw_writemask(header1, BRW_WRITEMASK_W), header1, brw_imm_ud(0x7ff<<8));
+      }
+
+      for (i = 0; i < c->key.nr_userclip; i++) {
+        brw_set_conditionalmod(p, BRW_CONDITIONAL_L);
+        brw_DP4(p, brw_null_reg(), pos, c->userplane[i]);
+        brw_OR(p, brw_writemask(header1, BRW_WRITEMASK_W), header1, brw_imm_ud(1<<i));
+        brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+      }
+
+      /* i965 clipping workaround: 
+       * 1) Test for -ve rhw
+       * 2) If set, 
+       *      set ndc = (0,0,0,0)
+       *      set ucp[6] = 1
+       *
+       * Later, clipping will detect ucp[6] and ensure the primitive is
+       * clipped against all fixed planes.
+       */
+      if (c->chipset.is_965) {
+        brw_CMP(p,
+                vec8(brw_null_reg()),
+                BRW_CONDITIONAL_L,
+                brw_swizzle1(ndc, 3),
+                brw_imm_f(0));
+   
+        brw_OR(p, brw_writemask(header1, BRW_WRITEMASK_W), header1, brw_imm_ud(1<<6));
+        brw_MOV(p, ndc, brw_imm_f(0));
+        brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+      }
+
+      brw_set_access_mode(p, BRW_ALIGN_1);     /* why? */
+      brw_MOV(p, retype(brw_message_reg(1), BRW_REGISTER_TYPE_UD), header1);
+      brw_set_access_mode(p, BRW_ALIGN_16);
+
+      release_tmp(c, header1);
+   }
+   else {
+      brw_MOV(p, retype(brw_message_reg(1), BRW_REGISTER_TYPE_UD), brw_imm_ud(0));
+   }
+
+   /* Emit the (interleaved) headers for the two vertices - an 8-reg
+    * of zeros followed by two sets of NDC coordinates:
+    */
+   brw_set_access_mode(p, BRW_ALIGN_1);
+   brw_MOV(p, offset(m0, 2), ndc);
+
+   if (c->chipset.is_igdng) {
+       /* There are 20 DWs (D0-D19) in VUE vertex header on IGDNG */
+       brw_MOV(p, offset(m0, 3), pos); /* a portion of vertex header */
+       /* m4, m5 contain the distances from vertex to the user clip planeXXX. 
+        * Seems it is useless for us.
+        * m6 is used for aligning, so that the remainder of vertex element is 
+        * reg-aligned.
+        */
+       brw_MOV(p, offset(m0, 7), pos); /* the remainder of vertex element */
+       len_vertext_header = 6;
+   } else {
+       brw_MOV(p, offset(m0, 3), pos);
+       len_vertext_header = 2;
+   }
+
+   eot = (c->overflow_count == 0);
+
+   brw_urb_WRITE(p, 
+                brw_null_reg(), /* dest */
+                0,             /* starting mrf reg nr */
+                c->r0,         /* src */
+                0,             /* allocate */
+                1,             /* used */
+                MIN2(c->nr_outputs + 1 + len_vertext_header, (BRW_MAX_MRF-1)), /* msg len */
+                0,             /* response len */
+                eot,           /* eot */
+                eot,           /* writes complete */
+                0,             /* urb destination offset */
+                BRW_URB_SWIZZLE_INTERLEAVE);
+
+   /* Not all of the vertex outputs/results fit into the MRF.
+    * Move the overflowed attributes from the GRF to the MRF and
+    * issue another brw_urb_WRITE().
+    */
+   for (i = 0; i < c->overflow_count; i += BRW_MAX_MRF) {
+      unsigned nr = MIN2(c->overflow_count - i, BRW_MAX_MRF);
+      GLuint j;
+
+      eot = (i + nr >= c->overflow_count);
+
+      /* XXX I'm not 100% sure about which MRF regs to use here.  Starting
+       * at mrf[4] atm...
+       */
+      for (j = 0; j < nr; j++) {
+        brw_MOV(p, brw_message_reg(4+j), 
+                 brw_vec8_grf(c->overflow_grf_start + i + j, 0));
+      }
+
+      brw_urb_WRITE(p,
+                    brw_null_reg(), /* dest */
+                    4,              /* starting mrf reg nr */
+                    c->r0,          /* src */
+                    0,              /* allocate */
+                    1,              /* used */
+                    nr+1,          /* msg len */
+                    0,              /* response len */
+                    eot,            /* eot */
+                    eot,            /* writes complete */
+                    i-1,            /* urb destination offset */
+                    BRW_URB_SWIZZLE_INTERLEAVE);
+   }
+}
+
+
+/**
+ * Called after code generation to resolve subroutine calls and the
+ * END instruction.
+ * \param end_inst  points to brw code for END instruction
+ * \param last_inst  points to last instruction emitted before vertex write
+ */
+static void 
+post_vs_emit( struct brw_vs_compile *c,
+              struct brw_instruction *end_inst,
+              struct brw_instruction *last_inst )
+{
+   GLint offset;
+
+   brw_resolve_cals(&c->func);
+
+   /* patch up the END code to jump past subroutines, etc */
+   offset = last_inst - end_inst;
+   if (offset > 1) {
+      brw_set_src1(end_inst, brw_imm_d(offset * 16));
+   } else {
+      end_inst->header.opcode = BRW_OPCODE_NOP;
+   }
+}
+
+static uint32_t
+get_predicate(const struct tgsi_full_instruction *inst)
+{
+   /* XXX: disabling for now
+    */
+#if 0
+   if (inst->dst.CondMask == COND_TR)
+      return BRW_PREDICATE_NONE;
+
+   /* All of GLSL only produces predicates for COND_NE and one channel per
+    * vector.  Fail badly if someone starts doing something else, as it might
+    * mean infinite looping or something.
+    *
+    * We'd like to support all the condition codes, but our hardware doesn't
+    * quite match the Mesa IR, which is modeled after the NV extensions.  For
+    * those, the instruction may update the condition codes or not, then any
+    * later instruction may use one of those condition codes.  For gen4, the
+    * instruction may update the flags register based on one of the condition
+    * codes output by the instruction, and then further instructions may
+    * predicate on that.  We can probably support this, but it won't
+    * necessarily be easy.
+    */
+/*   assert(inst->dst.CondMask == COND_NE); */
+
+   switch (inst->dst.CondSwizzle) {
+   case SWIZZLE_XXXX:
+      return BRW_PREDICATE_ALIGN16_REPLICATE_X;
+   case SWIZZLE_YYYY:
+      return BRW_PREDICATE_ALIGN16_REPLICATE_Y;
+   case SWIZZLE_ZZZZ:
+      return BRW_PREDICATE_ALIGN16_REPLICATE_Z;
+   case SWIZZLE_WWWW:
+      return BRW_PREDICATE_ALIGN16_REPLICATE_W;
+   default:
+      debug_printf("Unexpected predicate: 0x%08x\n",
+                   inst->dst.CondMask);
+      return BRW_PREDICATE_NORMAL;
+   }
+#else
+   return BRW_PREDICATE_NORMAL;
+#endif
+}
+
+static void emit_insn(struct brw_vs_compile *c,
+                     const struct tgsi_full_instruction *inst)
+{
+   unsigned opcode = inst->Instruction.Opcode;
+   unsigned label = inst->Label.Label;
+   struct brw_compile *p = &c->func;
+   struct brw_reg args[3], dst;
+   GLuint i;
+
+#if 0
+   printf("%d: ", insn);
+   _mesa_print_instruction(inst);
+#endif
+
+   /* Get argument regs.
+    */
+   for (i = 0; i < 3; i++) {
+      args[i] = get_arg(c, &inst->Src[i], i);
+   }
+
+   /* Get dest regs.  Note that it is possible for a reg to be both
+    * dst and arg, given the static allocation of registers.  So
+    * care needs to be taken emitting multi-operation instructions.
+    */ 
+   dst = get_dst(c, 
+                inst->Dst[0].Register.File,
+                inst->Dst[0].Register.Index,
+                inst->Dst[0].Register.WriteMask);
+
+   /* XXX: saturate
+    */
+   if (inst->Instruction.Saturate != TGSI_SAT_NONE) {
+      debug_printf("Unsupported saturate in vertex shader");
+   }
+
+   switch (opcode) {
+   case TGSI_OPCODE_ABS:
+      brw_MOV(p, dst, brw_abs(args[0]));
+      break;
+   case TGSI_OPCODE_ADD:
+      brw_ADD(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_COS:
+      emit_math1(c, BRW_MATH_FUNCTION_COS, dst, args[0], BRW_MATH_PRECISION_FULL);
+      break;
+   case TGSI_OPCODE_DP3:
+      brw_DP3(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_DP4:
+      brw_DP4(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_DPH:
+      brw_DPH(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_NRM:
+      emit_nrm(c, dst, args[0], 3);
+      break;
+   case TGSI_OPCODE_NRM4:
+      emit_nrm(c, dst, args[0], 4);
+      break;
+   case TGSI_OPCODE_DST:
+      unalias2(c, dst, args[0], args[1], emit_dst_noalias); 
+      break;
+   case TGSI_OPCODE_EXP:
+      unalias1(c, dst, args[0], emit_exp_noalias);
+      break;
+   case TGSI_OPCODE_EX2:
+      emit_math1(c, BRW_MATH_FUNCTION_EXP, dst, args[0], BRW_MATH_PRECISION_FULL);
+      break;
+   case TGSI_OPCODE_ARL:
+      emit_arl(c, dst, args[0]);
+      break;
+   case TGSI_OPCODE_FLR:
+      brw_RNDD(p, dst, args[0]);
+      break;
+   case TGSI_OPCODE_FRC:
+      brw_FRC(p, dst, args[0]);
+      break;
+   case TGSI_OPCODE_LOG:
+      unalias1(c, dst, args[0], emit_log_noalias);
+      break;
+   case TGSI_OPCODE_LG2:
+      emit_math1(c, BRW_MATH_FUNCTION_LOG, dst, args[0], BRW_MATH_PRECISION_FULL);
+      break;
+   case TGSI_OPCODE_LIT:
+      unalias1(c, dst, args[0], emit_lit_noalias);
+      break;
+   case TGSI_OPCODE_LRP:
+      unalias3(c, dst, args[0], args[1], args[2], emit_lrp_noalias);
+      break;
+   case TGSI_OPCODE_MAD:
+      brw_MOV(p, brw_acc_reg(), args[2]);
+      brw_MAC(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_MAX:
+      emit_max(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_MIN:
+      emit_min(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_MOV:
+      brw_MOV(p, dst, args[0]);
+      break;
+   case TGSI_OPCODE_MUL:
+      brw_MUL(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_POW:
+      emit_math2(c, BRW_MATH_FUNCTION_POW, dst, args[0], args[1], BRW_MATH_PRECISION_FULL); 
+      break;
+   case TGSI_OPCODE_RCP:
+      emit_math1(c, BRW_MATH_FUNCTION_INV, dst, args[0], BRW_MATH_PRECISION_FULL);
+      break;
+   case TGSI_OPCODE_RSQ:
+      emit_math1(c, BRW_MATH_FUNCTION_RSQ, dst, 
+                 brw_swizzle(args[0], 0,0,0,0), BRW_MATH_PRECISION_FULL);
+      break;
+   case TGSI_OPCODE_SEQ:
+      emit_seq(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_SIN:
+      emit_math1(c, BRW_MATH_FUNCTION_SIN, dst, args[0], BRW_MATH_PRECISION_FULL);
+      break;
+   case TGSI_OPCODE_SNE:
+      emit_sne(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_SGE:
+      emit_sge(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_SGT:
+      emit_sgt(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_SLT:
+      emit_slt(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_SLE:
+      emit_sle(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_SUB:
+      brw_ADD(p, dst, args[0], negate(args[1]));
+      break;
+   case TGSI_OPCODE_TRUNC:
+      /* round toward zero */
+      brw_RNDZ(p, dst, args[0]);
+      break;
+   case TGSI_OPCODE_XPD:
+      emit_xpd(p, dst, args[0], args[1]);
+      break;
+   case TGSI_OPCODE_IF:
+      assert(c->if_depth < MAX_IF_DEPTH);
+      c->if_inst[c->if_depth] = brw_IF(p, BRW_EXECUTE_8);
+      /* Note that brw_IF smashes the predicate_control field. */
+      c->if_inst[c->if_depth]->header.predicate_control = get_predicate(inst);
+      c->if_depth++;
+      break;
+   case TGSI_OPCODE_ELSE:
+      c->if_inst[c->if_depth-1] = brw_ELSE(p, c->if_inst[c->if_depth-1]);
+      break;
+   case TGSI_OPCODE_ENDIF:
+      assert(c->if_depth > 0);
+      brw_ENDIF(p, c->if_inst[--c->if_depth]);
+      break;                   
+   case TGSI_OPCODE_BGNLOOP:
+      c->loop_inst[c->loop_depth++] = brw_DO(p, BRW_EXECUTE_8);
+      break;
+   case TGSI_OPCODE_BRK:
+      brw_set_predicate_control(p, get_predicate(inst));
+      brw_BREAK(p);
+      brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+      break;
+   case TGSI_OPCODE_CONT:
+      brw_set_predicate_control(p, get_predicate(inst));
+      brw_CONT(p);
+      brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+      break;
+   case TGSI_OPCODE_ENDLOOP: 
+   {
+      struct brw_instruction *inst0, *inst1;
+      GLuint br = 1;
+
+      c->loop_depth--;
+
+      if (c->chipset.is_igdng)
+        br = 2;
+
+      inst0 = inst1 = brw_WHILE(p, c->loop_inst[c->loop_depth]);
+      /* patch all the BREAK/CONT instructions from last BEGINLOOP */
+      while (inst0 > c->loop_inst[c->loop_depth]) {
+        inst0--;
+        if (inst0->header.opcode == TGSI_OPCODE_BRK) {
+           inst0->bits3.if_else.jump_count = br * (inst1 - inst0 + 1);
+           inst0->bits3.if_else.pop_count = 0;
+        }
+        else if (inst0->header.opcode == TGSI_OPCODE_CONT) {
+           inst0->bits3.if_else.jump_count = br * (inst1 - inst0);
+           inst0->bits3.if_else.pop_count = 0;
+        }
+      }
+   }
+   break;
+   case TGSI_OPCODE_BRA:
+      brw_set_predicate_control(p, get_predicate(inst));
+      brw_ADD(p, brw_ip_reg(), brw_ip_reg(), brw_imm_d(1*16));
+      brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+      break;
+   case TGSI_OPCODE_CAL:
+      brw_set_access_mode(p, BRW_ALIGN_1);
+      brw_ADD(p, deref_1d(c->stack_index, 0), brw_ip_reg(), brw_imm_d(3*16));
+      brw_set_access_mode(p, BRW_ALIGN_16);
+      brw_ADD(p, get_addr_reg(c->stack_index),
+             get_addr_reg(c->stack_index), brw_imm_d(4));
+      brw_save_call(p, label, p->nr_insn);
+      brw_ADD(p, brw_ip_reg(), brw_ip_reg(), brw_imm_d(1*16));
+      break;
+   case TGSI_OPCODE_RET:
+      brw_ADD(p, get_addr_reg(c->stack_index),
+             get_addr_reg(c->stack_index), brw_imm_d(-4));
+      brw_set_access_mode(p, BRW_ALIGN_1);
+      brw_MOV(p, brw_ip_reg(), deref_1d(c->stack_index, 0));
+      brw_set_access_mode(p, BRW_ALIGN_16);
+      break;
+   case TGSI_OPCODE_END:       
+      c->end_offset = p->nr_insn;
+      /* this instruction will get patched later to jump past subroutine
+       * code, etc.
+       */
+      brw_ADD(p, brw_ip_reg(), brw_ip_reg(), brw_imm_d(1*16));
+      break;
+   case TGSI_OPCODE_BGNSUB:
+      brw_save_label(p, p->nr_insn, p->nr_insn);
+      break;
+   case TGSI_OPCODE_ENDSUB:
+      /* no-op */
+      break;
+   default:
+      debug_printf("Unsupported opcode %i (%s) in vertex shader",
+                  opcode, 
+                  tgsi_get_opcode_name(opcode));
+   }
+
+   /* Set the predication update on the last instruction of the native
+    * instruction sequence.
+    *
+    * This would be problematic if it was set on a math instruction,
+    * but that shouldn't be the case with the current GLSL compiler.
+    */
+#if 0
+   /* XXX: disabled
+    */
+   if (inst->CondUpdate) {
+      struct brw_instruction *hw_insn = &p->store[p->nr_insn - 1];
+
+      assert(hw_insn->header.destreg__conditionalmod == 0);
+      hw_insn->header.destreg__conditionalmod = BRW_CONDITIONAL_NZ;
+   }
+#endif
+
+   release_tmps(c);
+}
+
+
+/* Emit the vertex program instructions here.
+ */
+void brw_vs_emit(struct brw_vs_compile *c)
+{
+   struct brw_compile *p = &c->func;
+   const struct tgsi_token *tokens = c->vp->tokens;
+   struct brw_instruction *end_inst, *last_inst;
+   struct tgsi_parse_context parse;
+   struct tgsi_full_instruction *inst;
+
+   if (BRW_DEBUG & DEBUG_VS)
+      tgsi_dump(c->vp->tokens, 0); 
+
+   c->stack_index = brw_indirect(0, 0);
+
+   brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+   brw_set_access_mode(p, BRW_ALIGN_16);
+   
+
+   /* Static register allocation
+    */
+   brw_vs_alloc_regs(c);
+
+   if (c->vp->has_flow_control) {
+      brw_MOV(p, get_addr_reg(c->stack_index), brw_address(c->stack));
+   }
+
+   /* Instructions
+    */
+   tgsi_parse_init( &parse, tokens );
+   while( !tgsi_parse_end_of_tokens( &parse ) ) {
+      tgsi_parse_token( &parse );
+
+      switch( parse.FullToken.Token.Type ) {
+      case TGSI_TOKEN_TYPE_DECLARATION:
+      case TGSI_TOKEN_TYPE_IMMEDIATE:
+        break;
+
+      case TGSI_TOKEN_TYPE_INSTRUCTION:
+         inst = &parse.FullToken.FullInstruction;
+        emit_insn( c, inst );
+         break;
+
+      default:
+         assert( 0 );
+      }
+   }
+   tgsi_parse_free( &parse );
+
+   end_inst = &p->store[c->end_offset];
+   last_inst = &p->store[p->nr_insn];
+
+   /* The END instruction will be patched to jump to this code */
+   emit_vertex_write(c);
+
+   post_vs_emit(c, end_inst, last_inst);
+
+   if (BRW_DEBUG & DEBUG_VS) {
+      debug_printf("vs-native:\n");
+      brw_disasm(stderr, p->store, p->nr_insn);
+   }
+}
diff --git a/src/gallium/drivers/i965/brw_vs_state.c b/src/gallium/drivers/i965/brw_vs_state.c
new file mode 100644 (file)
index 0000000..dadbb62
--- /dev/null
@@ -0,0 +1,201 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+            
+#include "util/u_math.h"
+
+
+#include "brw_debug.h"
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+
+struct brw_vs_unit_key {
+   unsigned int total_grf;
+   unsigned int urb_entry_read_length;
+   unsigned int curb_entry_read_length;
+
+   unsigned int curbe_offset;
+
+   unsigned int nr_urb_entries, urb_size;
+
+   unsigned int nr_surfaces;
+};
+
+static void
+vs_unit_populate_key(struct brw_context *brw, struct brw_vs_unit_key *key)
+{
+   memset(key, 0, sizeof(*key));
+
+   /* CACHE_NEW_VS_PROG */
+   key->total_grf = brw->vs.prog_data->total_grf;
+   key->urb_entry_read_length = brw->vs.prog_data->urb_read_length;
+   key->curb_entry_read_length = brw->vs.prog_data->curb_read_length;
+
+   /* BRW_NEW_URB_FENCE */
+   key->nr_urb_entries = brw->urb.nr_vs_entries;
+   key->urb_size = brw->urb.vsize;
+
+   /* BRW_NEW_NR_VS_SURFACES */
+   key->nr_surfaces = brw->vs.nr_surfaces;
+
+   /* PIPE_NEW_CLIP */
+   if (brw->curr.ucp.nr) {
+      /* Note that we read in the userclip planes as well, hence
+       * clip_start:
+       */
+      key->curbe_offset = brw->curbe.clip_start;
+   }
+   else {
+      key->curbe_offset = brw->curbe.vs_start;
+   }
+}
+
+static enum pipe_error
+vs_unit_create_from_key(struct brw_context *brw, 
+                        struct brw_vs_unit_key *key,
+                        struct brw_winsys_reloc *reloc,
+                        struct brw_winsys_buffer **bo_out)
+{
+   enum pipe_error ret;
+   struct brw_vs_unit_state vs;
+   int chipset_max_threads;
+
+   memset(&vs, 0, sizeof(vs));
+
+   vs.thread0.kernel_start_pointer = 0; /* reloc */
+   vs.thread0.grf_reg_count = align(key->total_grf, 16) / 16 - 1;
+   vs.thread1.floating_point_mode = BRW_FLOATING_POINT_NON_IEEE_754;
+   /* Choosing multiple program flow means that we may get 2-vertex threads,
+    * which will have the channel mask for dwords 4-7 enabled in the thread,
+    * and those dwords will be written to the second URB handle when we
+    * brw_urb_WRITE() results.
+    */
+   vs.thread1.single_program_flow = 0;
+
+   if (BRW_IS_IGDNG(brw))
+      vs.thread1.binding_table_entry_count = 0; /* hardware requirement */
+   else
+      vs.thread1.binding_table_entry_count = key->nr_surfaces;
+
+   vs.thread3.urb_entry_read_length = key->urb_entry_read_length;
+   vs.thread3.const_urb_entry_read_length = key->curb_entry_read_length;
+   vs.thread3.dispatch_grf_start_reg = 1;
+   vs.thread3.urb_entry_read_offset = 0;
+   vs.thread3.const_urb_entry_read_offset = key->curbe_offset * 2;
+
+   if (BRW_IS_IGDNG(brw))
+       vs.thread4.nr_urb_entries = key->nr_urb_entries >> 2;
+   else
+       vs.thread4.nr_urb_entries = key->nr_urb_entries;
+
+   vs.thread4.urb_entry_allocation_size = key->urb_size - 1;
+
+   if (BRW_IS_IGDNG(brw))
+      chipset_max_threads = 72;
+   else if (BRW_IS_G4X(brw))
+      chipset_max_threads = 32;
+   else
+      chipset_max_threads = 16;
+
+   vs.thread4.max_threads = CLAMP(key->nr_urb_entries / 2,
+                                 1, chipset_max_threads) - 1;
+
+   if (BRW_DEBUG & DEBUG_SINGLE_THREAD)
+      vs.thread4.max_threads = 0;
+
+   /* No samplers for ARB_vp programs:
+    */
+   /* It has to be set to 0 for IGDNG
+    */
+   vs.vs5.sampler_count = 0;
+
+   if (BRW_DEBUG & DEBUG_STATS)
+      vs.thread4.stats_enable = 1;
+
+   /* Vertex program always enabled:
+    */
+   vs.vs6.vs_enable = 1;
+
+   ret = brw_upload_cache(&brw->cache, BRW_VS_UNIT,
+                          key, sizeof(*key),
+                          reloc, 1,
+                          &vs, sizeof(vs),
+                          NULL, NULL,
+                          bo_out);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+static int prepare_vs_unit(struct brw_context *brw)
+{
+   struct brw_vs_unit_key key;
+   enum pipe_error ret;
+   struct brw_winsys_reloc reloc[1];
+   unsigned grf_reg_count;
+
+   vs_unit_populate_key(brw, &key);
+
+   grf_reg_count = (align(key.total_grf, 16) / 16 - 1);
+
+   /* Emit VS program relocation */
+   make_reloc(&reloc[0],
+              BRW_USAGE_STATE,
+              grf_reg_count << 1,
+              offsetof(struct brw_vs_unit_state, thread0),
+              brw->vs.prog_bo);
+
+
+   if (brw_search_cache(&brw->cache, BRW_VS_UNIT,
+                        &key, sizeof(key),
+                        reloc, 1,
+                        NULL,
+                        &brw->vs.state_bo))
+      return PIPE_OK;
+
+   ret = vs_unit_create_from_key(brw, &key, reloc, &brw->vs.state_bo);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+const struct brw_tracked_state brw_vs_unit = {
+   .dirty = {
+      .mesa  = (PIPE_NEW_CLIP),
+      .brw   = (BRW_NEW_CURBE_OFFSETS |
+                BRW_NEW_NR_VS_SURFACES |
+               BRW_NEW_URB_FENCE),
+      .cache = CACHE_NEW_VS_PROG
+   },
+   .prepare = prepare_vs_unit,
+};
diff --git a/src/gallium/drivers/i965/brw_vs_surface_state.c b/src/gallium/drivers/i965/brw_vs_surface_state.c
new file mode 100644 (file)
index 0000000..177a517
--- /dev/null
@@ -0,0 +1,232 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+#include "brw_winsys.h"
+
+/* XXX: disabled true constant buffer functionality
+ */
+
+
+/* Creates a new VS constant buffer reflecting the current VS program's
+ * constants, if needed by the VS program.
+ *
+ * Otherwise, constants go through the CURBEs using the brw_constant_buffer
+ * state atom.
+ */
+#if 0
+static struct brw_winsys_buffer *
+brw_vs_update_constant_buffer(struct brw_context *brw)
+{
+   /* XXX: true constant buffers
+    */
+   struct brw_vertex_program *vp =
+      (struct brw_vertex_program *) brw->vertex_program;
+   const struct gl_program_parameter_list *params = vp->program.Base.Parameters;
+   const int size = params->NumParameters * 4 * sizeof(GLfloat);
+   drm_intel_bo *const_buffer;
+
+   /* BRW_NEW_VERTEX_PROGRAM */
+   if (!vp->use_const_buffer)
+      return NULL;
+
+   const_buffer = brw->sws->bo_alloc(brw->sws, 
+                                    BRW_BUFFER_TYPE_SHADER_CONSTANTS,
+                                    size, 64);
+
+   /* _NEW_PROGRAM_CONSTANTS */
+   brw->sws->bo_subdata(const_buffer, 0, size, params->ParameterValues,
+                        NULL, 0);
+
+   return const_buffer;
+}
+#endif
+
+/**
+ * Update the surface state for a VS constant buffer.
+ *
+ * Sets brw->vs.surf_bo[surf] and brw->vp->const_buffer.
+ */
+#if 0
+static void
+brw_update_vs_constant_surface( struct brw_context *brw,
+                                GLuint surf)
+{
+   struct brw_surface_key key;
+   struct pipe_buffer *cb = brw->curr.vs_constants;
+   enum pipe_error ret;
+
+   assert(surf == 0);
+
+   /* If we're in this state update atom, we need to update VS constants, so
+    * free the old buffer and create a new one for the new contents.
+    */
+   ret = brw_vs_update_constant_buffer(brw, &vp->const_buffer);
+   if (ret)
+      return ret;
+
+   /* If there's no constant buffer, then no surface BO is needed to point at
+    * it.
+    */
+   if (vp->const_buffer == NULL) {
+      bo_reference(brw->vs.surf_bo[surf], NULL);
+      return PIPE_OK;
+   }
+
+   memset(&key, 0, sizeof(key));
+
+   key.format = PIPE_FORMAT_R32G32B32A32_FLOAT;
+   key.bo = vp->const_buffer;
+   key.depthmode = GL_NONE;
+   key.pitch = params->NumParameters;
+   key.width = params->NumParameters;
+   key.height = 1;
+   key.depth = 1;
+   key.cpp = 16;
+
+   /*
+   printf("%s:\n", __FUNCTION__);
+   printf("  width %d  height %d  depth %d  cpp %d  pitch %d\n",
+          key.width, key.height, key.depth, key.cpp, key.pitch);
+   */
+
+   if (brw_search_cache(&brw->surface_cache,
+                        BRW_SS_SURFACE,
+                        &key, sizeof(key),
+                        &key.bo, key.bo ? 1 : 0,
+                        NULL,
+                        &brw->vs.surf_bo[surf]))
+      return PIPE_OK;
+
+   ret = brw_create_constant_surface(brw, &key
+                                     &brw->vs.surf_bo[surf]);
+   if (ret)
+      return ret;
+   
+   return PIPE_OK;
+}
+#endif
+
+
+/**
+ * Constructs the binding table for the VS surface state.
+ */
+static enum pipe_error
+brw_vs_get_binding_table(struct brw_context *brw,
+                         struct brw_winsys_buffer **bo_out)
+{
+#if 0
+   static GLuint data[BRW_VS_MAX_SURF]; /* always zero */
+   struct brw_winsys_reloc reloc[BRW_VS_MAX_SURF];
+   int i;
+
+   /* Emit binding table relocations to surface state */
+   for (i = 0; i < BRW_VS_MAX_SURF; i++) {
+      make_reloc(&reloc[i],
+                 BRW_USAGE_STATE,
+                 0,
+                 i * 4,
+                 brw->vs.surf_bo[i]);
+   }
+   
+   ret = brw_cache_data( &brw->surface_cache, 
+                         BRW_SS_SURF_BIND,
+                         NULL, 0,
+                         reloc, nr_reloc,
+                         data, sizeof data,
+                         NULL, NULL,
+                         bo_out);
+   if (ret)
+      return ret;
+
+   FREE(data);
+   return PIPE_OK;
+#else
+   return PIPE_OK;
+#endif
+}
+
+/**
+ * Vertex shader surfaces (constant buffer).
+ *
+ * This consumes the state updates for the constant buffer needing
+ * to be updated, and produces BRW_NEW_NR_VS_SURFACES for the VS unit and
+ * CACHE_NEW_SURF_BIND for the binding table upload.
+ */
+static enum pipe_error prepare_vs_surfaces(struct brw_context *brw )
+{
+   enum pipe_error ret;
+
+#if 0
+   int i;
+   int nr_surfaces = 0;
+
+   brw_update_vs_constant_surface(ctx, SURF_INDEX_VERT_CONST_BUFFER);
+
+   for (i = 0; i < BRW_VS_MAX_SURF; i++) {
+      if (brw->vs.surf_bo[i] != NULL) {
+        nr_surfaces = i + 1;
+      }
+   }
+
+   if (brw->vs.nr_surfaces != nr_surfaces) {
+      brw->state.dirty.brw |= BRW_NEW_NR_VS_SURFACES;
+      brw->vs.nr_surfaces = nr_surfaces;
+   }
+#endif
+
+   /* Note that we don't end up updating the bind_bo if we don't have a
+    * surface to be pointing at.  This should be relatively harmless, as it
+    * just slightly increases our working set size.
+    */
+   if (brw->vs.nr_surfaces != 0) {
+      ret = brw_vs_get_binding_table(brw, &brw->vs.bind_bo);
+      if (ret)
+         return ret;
+   }
+
+   return PIPE_OK;
+}
+
+const struct brw_tracked_state brw_vs_surfaces = {
+   .dirty = {
+      .mesa = (PIPE_NEW_VERTEX_CONSTANTS |
+              PIPE_NEW_VERTEX_SHADER),
+      .brw = 0,
+      .cache = 0
+   },
+   .prepare = prepare_vs_surfaces,
+};
+
+
+
diff --git a/src/gallium/drivers/i965/brw_winsys.h b/src/gallium/drivers/i965/brw_winsys.h
new file mode 100644 (file)
index 0000000..a242e31
--- /dev/null
@@ -0,0 +1,309 @@
+/**************************************************************************
+ *
+ * Copyright Â© 2009 Jakob Bornecrantz
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER
+ * DEALINGS IN THE SOFTWARE.
+ *
+ **************************************************************************/
+
+#ifndef BRW_WINSYS_H
+#define BRW_WINSYS_H
+
+#include "pipe/p_compiler.h"
+#include "pipe/p_defines.h"
+#include "pipe/p_refcnt.h"
+
+struct brw_winsys;
+struct pipe_fence_handle;
+
+/* Not sure why the winsys needs this:
+ */
+#define BRW_BATCH_SIZE (32*1024)
+
+struct brw_winsys_screen;
+
+/* Need a tiny bit of information inside the abstract buffer struct:
+ */
+struct brw_winsys_buffer {
+   struct pipe_reference reference;
+   struct brw_winsys_screen *sws;
+   unsigned size;
+};
+
+
+/* Should be possible to validate usages above against buffer creation
+ * types, below:
+ */
+enum brw_buffer_type
+{
+   BRW_BUFFER_TYPE_TEXTURE,
+   BRW_BUFFER_TYPE_SCANOUT,          /**< a texture used for scanning out from */
+   BRW_BUFFER_TYPE_VERTEX,
+   BRW_BUFFER_TYPE_CURBE,
+   BRW_BUFFER_TYPE_QUERY,
+   BRW_BUFFER_TYPE_SHADER_CONSTANTS,
+   BRW_BUFFER_TYPE_SHADER_SCRATCH,
+   BRW_BUFFER_TYPE_BATCH,
+   BRW_BUFFER_TYPE_GENERAL_STATE,
+   BRW_BUFFER_TYPE_SURFACE_STATE,
+   BRW_BUFFER_TYPE_PIXEL,            /* image uploads, pbo's, etc */
+   BRW_BUFFER_TYPE_GENERIC,          /* unknown */
+   BRW_BUFFER_TYPE_MAX               /* Count of possible values */
+};
+
+
+/* Describe the usage of a particular buffer in a relocation.  The DRM
+ * winsys will translate these back to GEM read/write domain flags.
+ */
+enum brw_buffer_usage {
+   BRW_USAGE_STATE,         /* INSTRUCTION, 0 */
+   BRW_USAGE_QUERY_RESULT,  /* INSTRUCTION, INSTRUCTION */
+   BRW_USAGE_RENDER_TARGET, /* RENDER,      0 */
+   BRW_USAGE_DEPTH_BUFFER,  /* RENDER,      RENDER */
+   BRW_USAGE_BLIT_SOURCE,   /* RENDER,      0 */
+   BRW_USAGE_BLIT_DEST,     /* RENDER,      RENDER */
+   BRW_USAGE_SAMPLER,       /* SAMPLER,     0 */
+   BRW_USAGE_VERTEX,        /* VERTEX,      0 */
+   BRW_USAGE_SCRATCH,       /* 0,           0 */
+   BRW_USAGE_MAX
+};
+
+enum brw_buffer_data_type {
+   BRW_DATA_GS_CC_VP,
+   BRW_DATA_GS_CC_UNIT,
+   BRW_DATA_GS_WM_PROG,
+   BRW_DATA_GS_SAMPLER_DEFAULT_COLOR,
+   BRW_DATA_GS_SAMPLER,
+   BRW_DATA_GS_WM_UNIT,
+   BRW_DATA_GS_SF_PROG,
+   BRW_DATA_GS_SF_VP,
+   BRW_DATA_GS_SF_UNIT,
+   BRW_DATA_GS_VS_UNIT,
+   BRW_DATA_GS_VS_PROG,
+   BRW_DATA_GS_GS_UNIT,
+   BRW_DATA_GS_GS_PROG,
+   BRW_DATA_GS_CLIP_VP,
+   BRW_DATA_GS_CLIP_UNIT,
+   BRW_DATA_GS_CLIP_PROG,
+   BRW_DATA_SS_SURFACE,
+   BRW_DATA_SS_SURF_BIND,
+   BRW_DATA_CONSTANT_BUFFER,
+   BRW_DATA_BATCH_BUFFER,
+   BRW_DATA_OTHER,
+   BRW_DATA_MAX
+};
+
+
+/* Matches the i915_drm definitions:
+ */
+#define BRW_TILING_NONE  0
+#define BRW_TILING_X     1
+#define BRW_TILING_Y     2
+
+
+/* Relocations to be applied with subdata in a call to sws->bo_subdata, below.
+ *
+ * Effectively this encodes:
+ *
+ *    (unsigned *)(subdata + offset) = bo->offset + delta
+ */
+struct brw_winsys_reloc {
+   enum brw_buffer_usage usage; /* debug only */
+   unsigned delta;
+   unsigned offset;
+   struct brw_winsys_buffer *bo;
+};
+
+static INLINE void make_reloc(struct brw_winsys_reloc *reloc,
+                              enum brw_buffer_usage usage,
+                              unsigned delta,
+                              unsigned offset,
+                              struct brw_winsys_buffer *bo)
+{
+   reloc->usage = usage;
+   reloc->delta = delta;
+   reloc->offset = offset;
+   reloc->bo = bo;              /* Note - note taking a reference yet */
+}
+
+
+
+struct brw_winsys_screen {
+
+
+   /**
+    * Buffer functions.
+    */
+
+   /*@{*/
+   /**
+    * Create a buffer.
+    */
+   enum pipe_error (*bo_alloc)(struct brw_winsys_screen *sws,
+                               enum brw_buffer_type type,
+                               unsigned size,
+                               unsigned alignment,
+                               struct brw_winsys_buffer **bo_out);
+
+   /* Destroy a buffer when our refcount goes to zero:
+    */
+   void (*bo_destroy)(struct brw_winsys_buffer *buffer);
+
+   /* delta -- added to b2->offset, and written into buffer
+    * offset -- location above value is written to within buffer
+    */
+   enum pipe_error (*bo_emit_reloc)(struct brw_winsys_buffer *buffer,
+                                    enum brw_buffer_usage usage,
+                                    unsigned delta,
+                                    unsigned offset,
+                                    struct brw_winsys_buffer *b2);
+
+   enum pipe_error (*bo_exec)(struct brw_winsys_buffer *buffer,
+                              unsigned bytes_used);
+
+   enum pipe_error (*bo_subdata)(struct brw_winsys_buffer *buffer,
+                                 enum brw_buffer_data_type data_type,
+                                 size_t offset,
+                                 size_t size,
+                                 const void *data,
+                                 const struct brw_winsys_reloc *reloc,
+                                 unsigned nr_reloc );
+
+   boolean (*bo_is_busy)(struct brw_winsys_buffer *buffer);
+   boolean (*bo_references)(struct brw_winsys_buffer *a,
+                            struct brw_winsys_buffer *b);
+
+   /* XXX: couldn't this be handled by returning true/false on
+    * bo_emit_reloc?
+    */
+   enum pipe_error (*check_aperture_space)(struct brw_winsys_screen *iws,
+                                           struct brw_winsys_buffer **buffers,
+                                           unsigned count);
+
+   /**
+    * Map a buffer.
+    */
+   void *(*bo_map)(struct brw_winsys_buffer *buffer,
+                   enum brw_buffer_data_type data_type,
+                   unsigned offset,
+                   unsigned length,
+                   boolean write,
+                   boolean discard,
+                   boolean flush_explicit);
+
+   void (*bo_flush_range)(struct brw_winsys_buffer *buffer,
+                          unsigned offset,
+                          unsigned length);
+
+   /**
+    * Unmap a buffer.
+    */
+   void (*bo_unmap)(struct brw_winsys_buffer *buffer);
+   /*@}*/
+
+   
+   /* Wait for buffer to go idle.  Similar to map+unmap, but doesn't
+    * mark buffer contents as dirty.
+    */
+   void (*bo_wait_idle)(struct brw_winsys_buffer *buffer);
+   
+   /**
+    * Destroy the winsys.
+    */
+   void (*destroy)(struct brw_winsys_screen *iws);
+};
+
+static INLINE void *
+bo_map_read(struct brw_winsys_screen *sws, struct brw_winsys_buffer *buf)
+{
+   return sws->bo_map( buf,
+                       BRW_DATA_OTHER,
+                       0, buf->size,
+                       FALSE, FALSE, FALSE );
+}
+
+static INLINE void
+bo_reference(struct brw_winsys_buffer **ptr, struct brw_winsys_buffer *buf)
+{
+   struct brw_winsys_buffer *old_buf = *ptr;
+
+   if (pipe_reference(&(*ptr)->reference, &buf->reference))
+      old_buf->sws->bo_destroy(old_buf);
+
+   *ptr = buf;
+}
+
+
+/**
+ * Create brw pipe_screen.
+ */
+struct pipe_screen *brw_create_screen(struct brw_winsys_screen *iws, unsigned pci_id);
+
+/**
+ * Create a brw pipe_context.
+ */
+struct pipe_context *brw_create_context(struct pipe_screen *screen);
+
+/**
+ * Get the brw_winsys buffer backing the texture.
+ *
+ * TODO UGLY
+ */
+struct pipe_texture;
+boolean brw_texture_get_winsys_buffer(struct pipe_texture *texture,
+                                      struct brw_winsys_buffer **buffer,
+                                      unsigned *stride);
+
+/**
+ * Wrap a brw_winsys buffer with a texture blanket.
+ *
+ * TODO UGLY
+ */
+struct pipe_texture * 
+brw_texture_blanket_winsys_buffer(struct pipe_screen *screen,
+                                  const struct pipe_texture *template,
+                                  unsigned pitch,
+                                 unsigned tiling,
+                                  struct brw_winsys_buffer *buffer);
+
+
+/*************************************************************************
+ * Cooperative dumping between winsys and driver.  TODO: make this
+ * driver-only by wrapping calls to winsys->bo_subdata().
+ */
+
+#ifdef DEBUG
+extern int BRW_DUMP;
+#else
+#define BRW_DUMP 0
+#endif 
+
+#define DUMP_ASM               0x1
+#define DUMP_STATE             0x2
+#define DUMP_BATCH             0x4
+
+void brw_dump_data( unsigned pci_id,
+                   enum brw_buffer_data_type data_type,
+                   unsigned offset,
+                   const void *data,
+                   size_t size );
+
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_winsys_debug.c b/src/gallium/drivers/i965/brw_winsys_debug.c
new file mode 100644 (file)
index 0000000..f8f6a53
--- /dev/null
@@ -0,0 +1,87 @@
+#include "brw_winsys.h"
+#include "brw_disasm.h"
+#include "brw_structs_dump.h"
+#include "brw_structs.h"
+#include "intel_decode.h"
+
+
+void brw_dump_data( unsigned pci_id,
+                   enum brw_buffer_data_type data_type,
+                   unsigned offset,
+                   const void *data,
+                   size_t size )
+{
+   if (BRW_DUMP & DUMP_ASM) {
+      switch (data_type) {
+      case BRW_DATA_GS_WM_PROG:
+      case BRW_DATA_GS_SF_PROG:
+      case BRW_DATA_GS_VS_PROG:
+      case BRW_DATA_GS_GS_PROG:
+      case BRW_DATA_GS_CLIP_PROG:
+         brw_disasm( stderr, data, size / sizeof(struct brw_instruction) );
+         break;
+      default:
+         break;
+      }
+   }
+
+   if (BRW_DUMP & DUMP_STATE) {
+      switch (data_type) {
+      case BRW_DATA_GS_CC_VP:
+         brw_dump_cc_viewport( data );
+         break;
+      case BRW_DATA_GS_CC_UNIT:
+         brw_dump_cc_unit_state( data );
+         break;
+      case BRW_DATA_GS_SAMPLER_DEFAULT_COLOR:
+         brw_dump_sampler_default_color( data );
+         break;
+      case BRW_DATA_GS_SAMPLER:
+         brw_dump_sampler_state( data );
+         break;
+      case BRW_DATA_GS_WM_UNIT:
+         brw_dump_wm_unit_state( data );
+         break;
+      case BRW_DATA_GS_SF_VP:
+         brw_dump_sf_viewport( data );
+         break;
+      case BRW_DATA_GS_SF_UNIT:
+         brw_dump_sf_unit_state( data );
+         break;
+      case BRW_DATA_GS_VS_UNIT:
+         brw_dump_vs_unit_state( data );
+         break;
+      case BRW_DATA_GS_GS_UNIT:
+         brw_dump_gs_unit_state( data );
+         break;
+      case BRW_DATA_GS_CLIP_VP:
+         brw_dump_clipper_viewport( data );
+         break;
+      case BRW_DATA_GS_CLIP_UNIT:
+         brw_dump_clip_unit_state( data );
+         break;
+      case BRW_DATA_SS_SURFACE:
+         brw_dump_surface_state( data );
+         break;
+      case BRW_DATA_SS_SURF_BIND:
+         break;
+      case BRW_DATA_OTHER:
+         break;
+      case BRW_DATA_CONSTANT_BUFFER:
+         break;
+      default:
+         break;
+      }
+   }
+
+   if (BRW_DUMP & DUMP_BATCH) {
+      switch (data_type) {
+      case BRW_DATA_BATCH_BUFFER:
+         intel_decode(data, size / 4, offset, pci_id);
+         break;
+      default:
+         break;
+      }
+   }
+}
+
diff --git a/src/gallium/drivers/i965/brw_wm.c b/src/gallium/drivers/i965/brw_wm.c
new file mode 100644 (file)
index 0000000..fdf820a
--- /dev/null
@@ -0,0 +1,319 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+#include "tgsi/tgsi_info.h"
+
+#include "brw_context.h"
+#include "brw_screen.h"
+#include "brw_util.h"
+#include "brw_wm.h"
+#include "brw_state.h"
+#include "brw_debug.h"
+#include "brw_pipe_rast.h"
+
+
+/** Return number of src args for given instruction */
+GLuint brw_wm_nr_args( GLuint opcode )
+{
+   switch (opcode) {
+   case WM_FRONTFACING:
+   case WM_PIXELXY:
+      return 0;
+   case WM_CINTERP:
+   case WM_WPOSXY:
+   case WM_DELTAXY:
+      return 1;
+   case WM_LINTERP:
+   case WM_PIXELW:
+      return 2;
+   case WM_FB_WRITE:
+   case WM_PINTERP:
+      return 3;
+   case TGSI_OPCODE_TEX:
+   case TGSI_OPCODE_TXP:
+   case TGSI_OPCODE_TXB:
+   case TGSI_OPCODE_TXD:
+      /* sampler arg is held as a field in the instruction, not in an
+       * actual register:
+       */
+      return tgsi_get_opcode_info(opcode)->num_src - 1;
+
+   default:
+      assert(opcode < MAX_OPCODE);
+      return tgsi_get_opcode_info(opcode)->num_src;
+   }
+}
+
+
+GLuint brw_wm_is_scalar_result( GLuint opcode )
+{
+   switch (opcode) {
+   case TGSI_OPCODE_COS:
+   case TGSI_OPCODE_EX2:
+   case TGSI_OPCODE_LG2:
+   case TGSI_OPCODE_POW:
+   case TGSI_OPCODE_RCP:
+   case TGSI_OPCODE_RSQ:
+   case TGSI_OPCODE_SIN:
+   case TGSI_OPCODE_DP3:
+   case TGSI_OPCODE_DP4:
+   case TGSI_OPCODE_DPH:
+   case TGSI_OPCODE_DST:
+      return 1;
+      
+   default:
+      return 0;
+   }
+}
+
+
+/**
+ * Do GPU code generation for shaders without flow control.  Shaders
+ * without flow control instructions can more readily be analysed for
+ * SSA-style optimizations.
+ */
+static void
+brw_wm_linear_shader_emit(struct brw_context *brw, struct brw_wm_compile *c)
+{
+   /* Augment fragment program.  Add instructions for pre- and
+    * post-fragment-program tasks such as interpolation and fogging.
+    */
+   brw_wm_pass_fp(c);
+
+   /* Translate to intermediate representation.  Build register usage
+    * chains.
+    */
+   brw_wm_pass0(c);
+
+   /* Dead code removal.
+    */
+   brw_wm_pass1(c);
+
+   /* Register allocation.
+    * Divide by two because we operate on 16 pixels at a time and require
+    * two GRF entries for each logical shader register.
+    */
+   c->grf_limit = BRW_WM_MAX_GRF / 2;
+
+   brw_wm_pass2(c);
+
+   /* how many general-purpose registers are used */
+   c->prog_data.total_grf = c->max_wm_grf;
+
+   /* Scratch space is used for register spilling */
+   if (c->last_scratch) {
+      c->prog_data.total_scratch = c->last_scratch + 0x40;
+   }
+   else {
+      c->prog_data.total_scratch = 0;
+   }
+
+   /* Emit GEN4 code.
+    */
+   brw_wm_emit(c);
+}
+
+
+/**
+ * All Mesa program -> GPU code generation goes through this function.
+ * Depending on the instructions used (i.e. flow control instructions)
+ * we'll use one of two code generators.
+ */
+static enum pipe_error do_wm_prog( struct brw_context *brw,
+                                   struct brw_fragment_shader *fp, 
+                                   struct brw_wm_prog_key *key,
+                                   struct brw_winsys_buffer **bo_out)
+{
+   enum pipe_error ret;
+   struct brw_wm_compile *c;
+   const GLuint *program;
+   GLuint program_size;
+
+   if (brw->wm.compile_data == NULL) {
+      brw->wm.compile_data = MALLOC(sizeof(*brw->wm.compile_data));
+      if (!brw->wm.compile_data) 
+         return PIPE_ERROR_OUT_OF_MEMORY;
+   }
+
+   c = brw->wm.compile_data;
+   memset(c, 0, sizeof *c);
+
+   c->key = *key;
+   c->fp = fp;
+   c->env_param = NULL; /*brw->intel.ctx.FragmentProgram.Parameters;*/
+
+   brw_init_compile(brw, &c->func);
+
+   /*
+    * Shader which use GLSL features such as flow control are handled
+    * differently from "simple" shaders.
+    */
+   if (fp->has_flow_control) {
+      c->dispatch_width = 8;
+      /* XXX: GLSL support
+       */
+      exit(1);
+      /* brw_wm_branching_shader_emit(brw, c); */
+   }
+   else {
+      c->dispatch_width = 16;
+      brw_wm_linear_shader_emit(brw, c);
+   }
+
+   if (BRW_DEBUG & DEBUG_WM)
+      debug_printf("\n");
+
+   /* get the program
+    */
+   ret = brw_get_program(&c->func, &program, &program_size);
+   if (ret)
+      return ret;
+
+   ret = brw_upload_cache( &brw->cache, BRW_WM_PROG,
+                           &c->key, sizeof(c->key),
+                           NULL, 0,
+                           program, program_size,
+                           &c->prog_data,
+                           &brw->wm.prog_data,
+                           bo_out );
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+
+static void brw_wm_populate_key( struct brw_context *brw,
+                                struct brw_wm_prog_key *key )
+{
+   unsigned lookup, line_aa;
+   unsigned i;
+
+   memset(key, 0, sizeof(*key));
+
+   /* PIPE_NEW_FRAGMENT_SHADER
+    * PIPE_NEW_DEPTH_STENCIL_ALPHA
+    */
+   lookup = (brw->curr.zstencil->iz_lookup |
+            brw->curr.fragment_shader->iz_lookup);
+
+
+   /* PIPE_NEW_RAST
+    * BRW_NEW_REDUCED_PRIMITIVE 
+    */
+   switch (brw->reduced_primitive) {
+   case PIPE_PRIM_POINTS:
+      line_aa = AA_NEVER;
+      break;
+   case PIPE_PRIM_LINES:
+      line_aa = (brw->curr.rast->templ.line_smooth ? 
+                 AA_ALWAYS : AA_NEVER);
+      break;
+   default:
+      line_aa = brw->curr.rast->unfilled_aa_line;
+      break;
+   }
+        
+   brw_wm_lookup_iz(line_aa,
+                   lookup,
+                   brw->curr.fragment_shader->uses_depth,
+                   key);
+
+   /* PIPE_NEW_RAST */
+   key->flat_shade = brw->curr.rast->templ.flatshade;
+
+
+   /* PIPE_NEW_BOUND_TEXTURES */
+   for (i = 0; i < brw->curr.num_textures; i++) {
+      const struct brw_texture *tex = brw_texture(brw->curr.texture[i]);
+        
+      if (tex->base.format == PIPE_FORMAT_YCBCR)
+        key->yuvtex_mask |= 1 << i;
+
+      if (tex->base.format == PIPE_FORMAT_YCBCR_REV)
+        key->yuvtex_swap_mask |= 1 << i;
+
+      /* XXX: shadow texture
+       */
+      /* key->shadowtex_mask |= 1<<i; */
+   }
+
+   /* CACHE_NEW_VS_PROG */
+   key->vp_nr_outputs = brw->vs.prog_data->nr_outputs;
+
+   key->nr_cbufs = brw->curr.fb.nr_cbufs;
+
+   key->nr_inputs = brw->curr.fragment_shader->info.num_inputs;
+
+   /* The unique fragment program ID */
+   key->program_string_id = brw->curr.fragment_shader->id;
+}
+
+
+static enum pipe_error brw_prepare_wm_prog(struct brw_context *brw)
+{
+   struct brw_wm_prog_key key;
+   struct brw_fragment_shader *fs = brw->curr.fragment_shader;
+   enum pipe_error ret;
+     
+   brw_wm_populate_key(brw, &key);
+
+   /* Make an early check for the key.
+    */
+   if (brw_search_cache(&brw->cache, BRW_WM_PROG,
+                        &key, sizeof(key),
+                        NULL, 0,
+                        &brw->wm.prog_data,
+                        &brw->wm.prog_bo))
+      return PIPE_OK;
+
+   ret = do_wm_prog(brw, fs, &key, &brw->wm.prog_bo);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+const struct brw_tracked_state brw_wm_prog = {
+   .dirty = {
+      .mesa  = (PIPE_NEW_FRAGMENT_SHADER |
+               PIPE_NEW_DEPTH_STENCIL_ALPHA |
+               PIPE_NEW_RAST |
+               PIPE_NEW_NR_CBUFS |
+               PIPE_NEW_BOUND_TEXTURES),
+      .brw   = (BRW_NEW_WM_INPUT_DIMENSIONS |
+               BRW_NEW_REDUCED_PRIMITIVE),
+      .cache = CACHE_NEW_VS_PROG,
+   },
+   .prepare = brw_prepare_wm_prog
+};
+
diff --git a/src/gallium/drivers/i965/brw_wm.h b/src/gallium/drivers/i965/brw_wm.h
new file mode 100644 (file)
index 0000000..f1ca9f6
--- /dev/null
@@ -0,0 +1,344 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+              
+
+#ifndef BRW_WM_H
+#define BRW_WM_H
+
+#include "brw_context.h"
+#include "brw_eu.h"
+
+#define SATURATE (1<<5)
+
+/* A big lookup table is used to figure out which and how many
+ * additional regs will inserted before the main payload in the WM
+ * program execution.  These mainly relate to depth and stencil
+ * processing and the early-depth-test optimization.
+ */
+#define IZ_PS_KILL_ALPHATEST_BIT    0x1
+#define IZ_PS_COMPUTES_DEPTH_BIT    0x2
+#define IZ_DEPTH_WRITE_ENABLE_BIT   0x4
+#define IZ_DEPTH_TEST_ENABLE_BIT    0x8
+#define IZ_STENCIL_WRITE_ENABLE_BIT 0x10
+#define IZ_STENCIL_TEST_ENABLE_BIT  0x20
+#define IZ_BIT_MAX                  0x40
+
+#define AA_NEVER     0
+#define AA_SOMETIMES 1
+#define AA_ALWAYS    2
+
+struct brw_wm_prog_key {
+   GLuint source_depth_reg:3;
+   GLuint aa_dest_stencil_reg:3;
+   GLuint dest_depth_reg:3;
+   GLuint nr_depth_regs:3;
+   GLuint computes_depth:1;
+   GLuint source_depth_to_render_target:1;
+   GLuint flat_shade:1;
+   GLuint runtime_check_aads_emit:1;
+
+   GLuint shadowtex_mask:16;
+   GLuint yuvtex_mask:16;
+   GLuint yuvtex_swap_mask:16; /* UV swaped */
+
+   GLuint vp_nr_outputs:6;
+   GLuint nr_inputs:6;
+   GLuint nr_cbufs:3;
+   GLuint has_flow_control:1;
+
+   GLuint program_string_id;
+};
+
+
+/* A bit of a glossary:
+ *
+ * brw_wm_value: A computed value or program input.  Values are
+ * constant, they are created once and are never modified.  When a
+ * fragment program register is written or overwritten, new values are
+ * created fresh, preserving the rule that values are constant.
+ *
+ * brw_wm_ref: A reference to a value.  Wherever a value used is by an
+ * instruction or as a program output, that is tracked with an
+ * instance of this struct.  All references to a value occur after it
+ * is created.  After the last reference, a value is dead and can be
+ * discarded.
+ *
+ * brw_wm_grf: Represents a physical hardware register.  May be either
+ * empty or hold a value.  Register allocation is the process of
+ * assigning values to grf registers.  This occurs in pass2 and the
+ * brw_wm_grf struct is not used before that.
+ *
+ * Fragment program registers: These are time-varying constructs that
+ * are hard to reason about and which we translate away in pass0.  A
+ * single fragment program register element (eg. temp[0].x) will be
+ * translated to one or more brw_wm_value structs, one for each time
+ * that temp[0].x is written to during the program. 
+ */
+
+
+
+/* Used in pass2 to track register allocation.
+ */
+struct brw_wm_grf {
+   struct brw_wm_value *value;
+   GLuint nextuse;
+};
+
+struct brw_wm_value {
+   struct brw_reg hw_reg;      /* emitted to this reg, may not always be there */
+   struct brw_wm_ref *lastuse;
+   struct brw_wm_grf *resident; 
+   GLuint contributes_to_output:1;
+   GLuint spill_slot:16;       /* if non-zero, spill immediately after calculation */
+};
+
+struct brw_wm_ref {
+   struct brw_reg hw_reg;      /* nr filled in in pass2, everything else, pass0 */
+   struct brw_wm_value *value;
+   struct brw_wm_ref *prevuse;
+   GLuint unspill_reg:7;       /* unspill to reg */
+   GLuint emitted:1;
+   GLuint insn:24;
+};
+
+struct brw_wm_instruction {
+   struct brw_wm_value *dst[4];
+   struct brw_wm_ref *src[3][4];
+   GLuint opcode:8;
+   GLuint saturate:1;
+   GLuint writemask:4;
+   GLuint sampler:4;
+   GLuint tex_unit:4;   /* texture/sampler unit for texture instructions */
+   GLuint target:4;     /* TGSI_TEXTURE_x for texture instructions,
+                         * target binding table index for FB_WRITE
+                         */
+   GLuint eot:1;       /* End of thread indicator for FB_WRITE*/
+};
+
+
+#define BRW_WM_MAX_INSN  2048
+#define BRW_WM_MAX_GRF   128           /* hardware limit */
+#define BRW_WM_MAX_VREG  (BRW_WM_MAX_INSN * 4)
+#define BRW_WM_MAX_REF   (BRW_WM_MAX_INSN * 12)
+#define BRW_WM_MAX_PARAM 256
+#define BRW_WM_MAX_CONST 256
+#define BRW_WM_MAX_KILLS MAX_NV_FRAGMENT_PROGRAM_INSTRUCTIONS
+#define BRW_WM_MAX_SUBROUTINE 16
+
+
+/* New opcodes to track internal operations required for WM unit.
+ * These are added early so that the registers used can be tracked,
+ * freed and reused like those of other instructions.
+ */
+#define MAX_OPCODE        TGSI_OPCODE_LAST
+#define WM_PIXELXY        (MAX_OPCODE)
+#define WM_DELTAXY        (MAX_OPCODE + 1)
+#define WM_PIXELW         (MAX_OPCODE + 2)
+#define WM_LINTERP        (MAX_OPCODE + 3)
+#define WM_PINTERP        (MAX_OPCODE + 4)
+#define WM_CINTERP        (MAX_OPCODE + 5)
+#define WM_WPOSXY         (MAX_OPCODE + 6)
+#define WM_FB_WRITE       (MAX_OPCODE + 7)
+#define WM_FRONTFACING    (MAX_OPCODE + 8)
+#define MAX_WM_OPCODE     (MAX_OPCODE + 9)
+
+#define BRW_FILE_PAYLOAD   (TGSI_FILE_COUNT)
+#define PAYLOAD_DEPTH      (PIPE_MAX_SHADER_INPUTS) /* ?? */
+
+#define X    0
+#define Y    1
+#define Z    2
+#define W    3
+
+
+struct brw_fp_src {
+   unsigned file:4;
+   unsigned index:16;
+   unsigned swizzle:8;
+   unsigned indirect:1;
+   unsigned negate:1;
+   unsigned abs:1;
+};
+
+struct brw_fp_dst {
+   unsigned file:4;
+   unsigned index:16;
+   unsigned writemask:4;
+   unsigned indirect:1;
+   unsigned saturate:1;
+};
+
+struct brw_fp_instruction {
+   struct brw_fp_dst dst;
+   struct brw_fp_src src[3];
+   unsigned opcode:8;
+   unsigned target:8; /* XXX: special usage for FB_WRITE */
+   unsigned tex_unit:4;
+   unsigned sampler:4;
+   unsigned pad:8;
+};
+
+
+struct brw_wm_compile {
+   struct brw_compile func;
+   struct brw_wm_prog_key key;
+   struct brw_wm_prog_data prog_data;
+
+   struct brw_fragment_shader *fp;
+
+   GLfloat (*env_param)[4];
+
+   enum {
+      START,
+      PASS2_DONE
+   } state;
+
+   /* Initial pass - translate fp instructions to fp instructions,
+    * simplifying and adding instructions for interpolation and
+    * framebuffer writes.
+    */
+   struct {
+      GLfloat v[4];
+      unsigned nr;
+   } immediate[BRW_WM_MAX_CONST+3];
+   GLuint nr_immediates;
+   
+   struct brw_fp_instruction fp_instructions[BRW_WM_MAX_INSN];
+   GLuint nr_fp_insns;
+   GLuint fp_temp;
+   GLuint fp_interp_emitted;
+   GLuint fp_fragcolor_emitted;
+   GLuint fp_first_internal_temp;
+
+   struct brw_fp_src fp_pixel_xy;
+   struct brw_fp_src fp_delta_xy;
+   struct brw_fp_src fp_pixel_w;
+
+
+   /* Subsequent passes using SSA representation:
+    */
+   struct brw_wm_value vreg[BRW_WM_MAX_VREG];
+   GLuint nr_vreg;
+
+   struct brw_wm_value creg[BRW_WM_MAX_PARAM];
+   GLuint nr_creg;
+
+   struct {
+      struct brw_wm_value depth[4]; /* includes r0/r1 */
+      struct brw_wm_value input_interp[PIPE_MAX_SHADER_INPUTS];
+   } payload;
+
+
+   const struct brw_wm_ref *pass0_fp_reg[BRW_FILE_PAYLOAD+1][256][4];
+
+   struct brw_wm_ref undef_ref;
+   struct brw_wm_value undef_value;
+
+   struct brw_wm_ref refs[BRW_WM_MAX_REF];
+   GLuint nr_refs;
+
+   struct brw_wm_instruction instruction[BRW_WM_MAX_INSN];
+   GLuint nr_insns;
+
+   struct brw_wm_grf pass2_grf[BRW_WM_MAX_GRF/2];
+
+   GLuint grf_limit;
+   GLuint max_wm_grf;
+   GLuint last_scratch;
+
+   GLuint cur_inst;  /**< index of current instruction */
+
+   GLboolean out_of_regs;  /**< ran out of GRF registers? */
+
+   /** Mapping from Mesa registers to hardware registers */
+   struct {
+      GLboolean inited;
+      struct brw_reg reg;
+   } wm_regs[BRW_FILE_PAYLOAD+1][256][4];
+
+   GLboolean used_grf[BRW_WM_MAX_GRF];
+   GLuint first_free_grf;
+   struct brw_reg stack;
+   struct brw_reg emit_mask_reg;
+   GLuint tmp_regs[BRW_WM_MAX_GRF];
+   GLuint tmp_index;
+   GLuint tmp_max;
+   GLuint subroutines[BRW_WM_MAX_SUBROUTINE];
+   GLuint dispatch_width;
+
+   /** we may need up to 3 constants per instruction (if use_const_buffer) */
+   struct {
+      GLint index;
+      struct brw_reg reg;
+   } current_const[3];
+
+   GLuint error;
+};
+
+
+GLuint brw_wm_nr_args( GLuint opcode );
+GLuint brw_wm_is_scalar_result( GLuint opcode );
+
+int brw_wm_pass_fp( struct brw_wm_compile *c );
+void brw_wm_pass0( struct brw_wm_compile *c );
+void brw_wm_pass1( struct brw_wm_compile *c );
+void brw_wm_pass2( struct brw_wm_compile *c );
+void brw_wm_emit( struct brw_wm_compile *c );
+
+void brw_wm_print_value( struct brw_wm_compile *c,
+                        struct brw_wm_value *value );
+
+void brw_wm_print_ref( struct brw_wm_compile *c,
+                      struct brw_wm_ref *ref );
+
+void brw_wm_print_insn( struct brw_wm_compile *c,
+                       struct brw_wm_instruction *inst );
+
+void brw_wm_print_program( struct brw_wm_compile *c,
+                          const char *stage );
+
+void brw_wm_print_fp_program( struct brw_wm_compile *c,
+                              const char *stage );
+
+void brw_wm_lookup_iz( GLuint line_aa,
+                      GLuint lookup,
+                      GLboolean ps_uses_depth,
+                      struct brw_wm_prog_key *key );
+
+void brw_wm_branching_shader_emit(struct brw_context *brw, struct brw_wm_compile *c);
+
+void emit_ddxy(struct brw_compile *p,
+              const struct brw_reg *dst,
+              GLuint mask,
+              GLboolean is_ddx,
+              const struct brw_reg *arg0);
+
+#endif
diff --git a/src/gallium/drivers/i965/brw_wm_constant_buffer.c b/src/gallium/drivers/i965/brw_wm_constant_buffer.c
new file mode 100644 (file)
index 0000000..6434c6a
--- /dev/null
@@ -0,0 +1,165 @@
+/* XXX: Constant buffers disabled
+ */
+
+
+/**
+ * Create the constant buffer surface.  Vertex/fragment shader constants will be
+ * read from this buffer with Data Port Read instructions/messages.
+ */
+enum pipe_error
+brw_create_constant_surface( struct brw_context *brw,
+                             struct brw_surface_key *key,
+                             struct brw_winsys_buffer **bo_out )
+{
+   const GLint w = key->width - 1;
+   struct brw_winsys_buffer *bo;
+   struct brw_winsys_reloc reloc[1];
+   enum pipe_error ret;
+
+      /* Emit relocation to surface contents */
+   make_reloc(&reloc[0],
+              BRW_USAGE_SAMPLER,
+              0,
+              offsetof(struct brw_surface_state, ss1),
+              key->bo);
+
+   
+   memset(&surf, 0, sizeof(surf));
+
+   surf.ss0.mipmap_layout_mode = BRW_SURFACE_MIPMAPLAYOUT_BELOW;
+   surf.ss0.surface_type = BRW_SURFACE_BUFFER;
+   surf.ss0.surface_format = BRW_SURFACEFORMAT_R32G32B32A32_FLOAT;
+
+   surf.ss1.base_addr = 0; /* reloc */
+
+   surf.ss2.width = w & 0x7f;            /* bits 6:0 of size or width */
+   surf.ss2.height = (w >> 7) & 0x1fff;  /* bits 19:7 of size or width */
+   surf.ss3.depth = (w >> 20) & 0x7f;    /* bits 26:20 of size or width */
+   surf.ss3.pitch = (key->pitch * key->cpp) - 1; /* ignored?? */
+   brw_set_surface_tiling(&surf, key->tiling); /* tiling now allowed */
+   ret = brw_upload_cache(&brw->surface_cache, BRW_SS_SURFACE,
+                          key, sizeof(*key),
+                          reloc, Elements(reloc),
+                          &surf, sizeof(surf),
+                          NULL, NULL,
+                          &bo_out);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+
+/**
+ * Update the surface state for a WM constant buffer.
+ * The constant buffer will be (re)allocated here if needed.
+ */
+static enum pipe_error
+brw_update_wm_constant_surface( struct brw_context *brw,
+                                GLuint surf)
+{
+   struct brw_surface_key key;
+   struct brw_fragment_shader *fp = brw->curr.fragment_shader;
+   struct pipe_buffer *cbuf = brw->curr.fragment_constants;
+   int pitch = cbuf->size / (4 * sizeof(float));
+   enum pipe_error ret;
+
+   /* If we're in this state update atom, we need to update WM constants, so
+    * free the old buffer and create a new one for the new contents.
+    */
+   ret = brw_wm_update_constant_buffer(brw, &fp->const_buffer);
+   if (ret)
+      return ret;
+
+   /* If there's no constant buffer, then no surface BO is needed to point at
+    * it.
+    */
+   if (cbuf == NULL) {
+      bo_reference(&brw->wm.surf_bo[surf], NULL);
+      return PIPE_OK;
+   }
+
+   memset(&key, 0, sizeof(key));
+
+   key.ss0.mipmap_layout_mode = BRW_SURFACE_MIPMAPLAYOUT_BELOW;
+   key.ss0.surface_type = BRW_SURFACE_BUFFER;
+   key.ss0.surface_format = BRW_SURFACEFORMAT_R32G32B32A32_FLOAT;
+
+   key.bo = brw_buffer(cbuf)->bo;
+
+   key.ss2.width = (pitch-1) & 0x7f;            /* bits 6:0 of size or width */
+   key.ss2.height = ((pitch-1) >> 7) & 0x1fff;  /* bits 19:7 of size or width */
+   key.ss3.depth = ((pitch-1) >> 20) & 0x7f;    /* bits 26:20 of size or width */
+   key.ss3.pitch = (pitch * 4 * sizeof(float)) - 1; /* ignored?? */
+   brw_set_surface_tiling(&surf, key->tiling); /* tiling now allowed */
+
+
+   /*
+   printf("%s:\n", __FUNCTION__);
+   printf("  width %d  height %d  depth %d  cpp %d  pitch %d\n",
+          key.width, key.height, key.depth, key.cpp, key.pitch);
+   */
+
+   if (brw_search_cache(&brw->surface_cache,
+                        BRW_SS_SURFACE,
+                        &key, sizeof(key),
+                        &key.bo, 1,
+                        NULL,
+                        &brw->wm.surf_bo[surf]))
+      return PIPE_OK;
+
+   ret = brw_create_constant_surface(brw, &key, &brw->wm.surf_bo[surf]);
+   if (ret)
+      return ret;
+
+   brw->state.dirty.brw |= BRW_NEW_WM_SURFACES;
+   return PIPE_OK;
+}
+
+/**
+ * Updates surface / buffer for fragment shader constant buffer, if
+ * one is required.
+ *
+ * This consumes the state updates for the constant buffer, and produces
+ * BRW_NEW_WM_SURFACES to get picked up by brw_prepare_wm_surfaces for
+ * inclusion in the binding table.
+ */
+static enum pipe_error prepare_wm_constant_surface(struct brw_context *brw )
+{
+   struct brw_fragment_program *fp =
+      (struct brw_fragment_program *) brw->fragment_program;
+   GLuint surf = SURF_INDEX_FRAG_CONST_BUFFER;
+
+   ret = brw_wm_update_constant_buffer(brw,
+                                       &fp->const_buffer);
+   if (ret)
+      return ret;
+
+   /* If there's no constant buffer, then no surface BO is needed to point at
+    * it.
+    */
+   if (fp->const_buffer == 0) {
+      if (brw->wm.surf_bo[surf] != NULL) {
+        bo_reference(&brw->wm.surf_bo[surf], NULL);
+        brw->state.dirty.brw |= BRW_NEW_WM_SURFACES;
+      }
+      return PIPE_OK;
+   }
+
+   ret = brw_update_wm_constant_surface(ctx, surf);
+   if (ret)
+      return ret;
+
+   return PIPE_OK
+}
+
+const struct brw_tracked_state brw_wm_constant_surface = {
+   .dirty = {
+      .mesa = (_NEW_PROGRAM_CONSTANTS),
+      .brw = (BRW_NEW_FRAGMENT_PROGRAM),
+      .cache = 0
+   },
+   .prepare = prepare_wm_constant_surface,
+};
diff --git a/src/gallium/drivers/i965/brw_wm_debug.c b/src/gallium/drivers/i965/brw_wm_debug.c
new file mode 100644 (file)
index 0000000..3d11fa0
--- /dev/null
@@ -0,0 +1,256 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "tgsi/tgsi_info.h"
+
+#include "brw_context.h"
+#include "brw_wm.h"
+
+static void print_writemask( unsigned writemask )
+{
+   if (writemask != BRW_WRITEMASK_XYZW)
+      debug_printf(".%s%s%s%s", 
+                  (writemask & BRW_WRITEMASK_X) ? "x" : "",
+                  (writemask & BRW_WRITEMASK_Y) ? "y" : "",
+                  (writemask & BRW_WRITEMASK_Z) ? "z" : "",
+                  (writemask & BRW_WRITEMASK_W) ? "w" : "");
+}
+
+static void print_swizzle( unsigned swizzle )
+{
+   char *swz = "xyzw";
+   if (swizzle != BRW_SWIZZLE_XYZW)
+      debug_printf(".%c%c%c%c", 
+                  swz[BRW_GET_SWZ(swizzle, X)],
+                  swz[BRW_GET_SWZ(swizzle, Y)],
+                  swz[BRW_GET_SWZ(swizzle, Z)],
+                  swz[BRW_GET_SWZ(swizzle, W)]);
+}
+
+static void print_opcode( unsigned opcode )
+{
+   switch (opcode) {
+   case WM_PIXELXY:
+      debug_printf("PIXELXY");
+      break;
+   case WM_DELTAXY:
+      debug_printf("DELTAXY");
+      break;
+   case WM_PIXELW:
+      debug_printf("PIXELW");
+      break;
+   case WM_WPOSXY:
+      debug_printf("WPOSXY");
+      break;
+   case WM_PINTERP:
+      debug_printf("PINTERP");
+      break;
+   case WM_LINTERP:
+      debug_printf("LINTERP");
+      break;
+   case WM_CINTERP:
+      debug_printf("CINTERP");
+      break;
+   case WM_FB_WRITE:
+      debug_printf("FB_WRITE");
+      break;
+   case WM_FRONTFACING:
+      debug_printf("FRONTFACING");
+      break;
+   default:
+      debug_printf("%s", tgsi_get_opcode_info(opcode)->mnemonic);
+      break;
+   }
+}
+
+void brw_wm_print_value( struct brw_wm_compile *c,
+                      struct brw_wm_value *value )
+{
+   assert(value);
+   if (c->state >= PASS2_DONE) 
+      brw_print_reg(value->hw_reg);
+   else if( value == &c->undef_value )
+      debug_printf("undef");
+   else if( value - c->vreg >= 0 &&
+           value - c->vreg < BRW_WM_MAX_VREG)
+      debug_printf("r%d", value - c->vreg);
+   else if (value - c->creg >= 0 &&
+           value - c->creg < BRW_WM_MAX_PARAM)
+      debug_printf("c%d", value - c->creg);
+   else if (value - c->payload.input_interp >= 0 &&
+           value - c->payload.input_interp < PIPE_MAX_SHADER_INPUTS)
+      debug_printf("i%d", value - c->payload.input_interp);
+   else if (value - c->payload.depth >= 0 &&
+           value - c->payload.depth < PIPE_MAX_SHADER_INPUTS)
+      debug_printf("d%d", value - c->payload.depth);
+   else 
+      debug_printf("?");
+}
+
+void brw_wm_print_ref( struct brw_wm_compile *c,
+                      struct brw_wm_ref *ref )
+{
+   struct brw_reg hw_reg = ref->hw_reg;
+
+   if (ref->unspill_reg)
+      debug_printf("UNSPILL(%x)/", ref->value->spill_slot);
+
+   if (c->state >= PASS2_DONE)
+      brw_print_reg(ref->hw_reg);
+   else {
+      debug_printf("%s", hw_reg.negate ? "-" : "");
+      debug_printf("%s", hw_reg.abs ? "abs/" : "");
+      brw_wm_print_value(c, ref->value);
+      if ((hw_reg.nr&1) || hw_reg.subnr) {
+        debug_printf("->%d.%d", (hw_reg.nr&1), hw_reg.subnr);
+      }
+   }
+}
+
+void brw_wm_print_insn( struct brw_wm_compile *c,
+                       struct brw_wm_instruction *inst )
+{
+   GLuint i, arg;
+   GLuint nr_args = brw_wm_nr_args(inst->opcode);
+
+   debug_printf("[");
+   for (i = 0; i < 4; i++) {
+      if (inst->dst[i]) {
+        brw_wm_print_value(c, inst->dst[i]);
+        if (inst->dst[i]->spill_slot)
+           debug_printf("/SPILL(%x)",inst->dst[i]->spill_slot);
+      }
+      else
+        debug_printf("#");
+      if (i < 3)      
+        debug_printf(",");
+   }
+   debug_printf("]");
+   print_writemask(inst->writemask);
+   
+   debug_printf(" = ");
+   print_opcode(inst->opcode);
+  
+   if (inst->saturate)
+      debug_printf("_SAT");
+
+   for (arg = 0; arg < nr_args; arg++) {
+
+      debug_printf(" [");
+
+      for (i = 0; i < 4; i++) {
+        if (inst->src[arg][i]) {
+           brw_wm_print_ref(c, inst->src[arg][i]);
+        }
+        else
+           debug_printf("%%");
+
+        if (i < 3) 
+           debug_printf(",");
+        else
+           debug_printf("]");
+      }
+   }
+   debug_printf("\n");
+}
+
+void brw_wm_print_program( struct brw_wm_compile *c,
+                          const char *stage )
+{
+   GLuint insn;
+
+   debug_printf("%s:\n", stage);
+   for (insn = 0; insn < c->nr_insns; insn++)
+      brw_wm_print_insn(c, &c->instruction[insn]);
+   debug_printf("\n");
+}
+
+static const char *file_strings[TGSI_FILE_COUNT+1] = {
+   "NULL",
+   "CONST",
+   "IN",
+   "OUT",
+   "TEMP",
+   "SAMPLER",
+   "ADDR",
+   "IMM",
+   "LOOP",
+   "PAYLOAD"
+};
+
+static void brw_wm_print_fp_insn( struct brw_wm_compile *c,
+                                  struct brw_fp_instruction *inst )
+{
+   GLuint i;
+   GLuint nr_args = brw_wm_nr_args(inst->opcode);
+
+   print_opcode(inst->opcode);
+   if (inst->dst.saturate)
+      debug_printf("_SAT");
+   debug_printf(" ");
+
+   if (inst->dst.indirect)
+      debug_printf("[");
+
+   debug_printf("%s[%d]",
+                file_strings[inst->dst.file],
+                inst->dst.index );
+   print_writemask(inst->dst.writemask);
+
+   if (inst->dst.indirect)
+      debug_printf("]");
+
+   debug_printf(nr_args ? ", " : "\n");
+   
+   for (i = 0; i < nr_args; i++) {
+      debug_printf("%s%s%s[%d]%s",
+                   inst->src[i].negate ? "-" : "",
+                   inst->src[i].abs ? "ABS(" : "",
+                   file_strings[inst->src[i].file],
+                   inst->src[i].index,
+                   inst->src[i].abs ? ")" : "");
+      print_swizzle(inst->src[i].swizzle);
+      debug_printf("%s", i == nr_args - 1 ? "\n" : ", ");
+   }
+}
+
+
+void brw_wm_print_fp_program( struct brw_wm_compile *c,
+                              const char *stage )
+{
+   GLuint insn;
+
+   debug_printf("%s:\n", stage);
+   for (insn = 0; insn < c->nr_fp_insns; insn++)
+      brw_wm_print_fp_insn(c, &c->fp_instructions[insn]);
+   debug_printf("\n");
+}
+
diff --git a/src/gallium/drivers/i965/brw_wm_emit.c b/src/gallium/drivers/i965/brw_wm_emit.c
new file mode 100644 (file)
index 0000000..7e57d03
--- /dev/null
@@ -0,0 +1,1521 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "util/u_math.h"
+#include "tgsi/tgsi_info.h"
+
+#include "brw_context.h"
+#include "brw_wm.h"
+#include "brw_debug.h"
+#include "brw_disasm.h"
+
+/* Not quite sure how correct this is - need to understand horiz
+ * vs. vertical strides a little better.
+ */
+static INLINE struct brw_reg sechalf( struct brw_reg reg )
+{
+   if (reg.vstride)
+      reg.nr++;
+   return reg;
+}
+
+/* Payload R0:
+ *
+ * R0.0 -- pixel mask, one bit for each of 4 pixels in 4 quads,
+ *         corresponding to each of the 16 execution channels.
+ * R0.1..8 -- ?
+ * R1.0 -- triangle vertex 0.X
+ * R1.1 -- triangle vertex 0.Y
+ * R1.2 -- quad 0 x,y coords (2 packed uwords)
+ * R1.3 -- quad 1 x,y coords (2 packed uwords)
+ * R1.4 -- quad 2 x,y coords (2 packed uwords)
+ * R1.5 -- quad 3 x,y coords (2 packed uwords)
+ * R1.6 -- ?
+ * R1.7 -- ?
+ * R1.8 -- ?
+ */
+
+
+static void emit_pixel_xy(struct brw_compile *p,
+                         const struct brw_reg *dst,
+                         GLuint mask)
+{
+   struct brw_reg r1 = brw_vec1_grf(1, 0);
+   struct brw_reg r1_uw = retype(r1, BRW_REGISTER_TYPE_UW);
+
+   brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+
+   /* Calculate pixel centers by adding 1 or 0 to each of the
+    * micro-tile coordinates passed in r1.
+    */
+   if (mask & BRW_WRITEMASK_X) {
+      brw_ADD(p,
+             vec16(retype(dst[0], BRW_REGISTER_TYPE_UW)),
+             stride(suboffset(r1_uw, 4), 2, 4, 0),
+             brw_imm_v(0x10101010));
+   }
+
+   if (mask & BRW_WRITEMASK_Y) {
+      brw_ADD(p,
+             vec16(retype(dst[1], BRW_REGISTER_TYPE_UW)),
+             stride(suboffset(r1_uw,5), 2, 4, 0),
+             brw_imm_v(0x11001100));
+   }
+
+   brw_set_compression_control(p, BRW_COMPRESSION_COMPRESSED);
+}
+
+
+
+static void emit_delta_xy(struct brw_compile *p,
+                         const struct brw_reg *dst,
+                         GLuint mask,
+                         const struct brw_reg *arg0)
+{
+   struct brw_reg r1 = brw_vec1_grf(1, 0);
+
+   /* Calc delta X,Y by subtracting origin in r1 from the pixel
+    * centers.
+    */
+   if (mask & BRW_WRITEMASK_X) {
+      brw_ADD(p,
+             dst[0],
+             retype(arg0[0], BRW_REGISTER_TYPE_UW),
+             negate(r1));
+   }
+
+   if (mask & BRW_WRITEMASK_Y) {
+      brw_ADD(p,
+             dst[1],
+             retype(arg0[1], BRW_REGISTER_TYPE_UW),
+             negate(suboffset(r1,1)));
+
+   }
+}
+
+static void emit_wpos_xy(struct brw_wm_compile *c,
+                        const struct brw_reg *dst,
+                        GLuint mask,
+                        const struct brw_reg *arg0)
+{
+   struct brw_compile *p = &c->func;
+
+   if (mask & BRW_WRITEMASK_X) {
+      /* X' = X */
+      brw_MOV(p,
+             dst[0],
+             retype(arg0[0], BRW_REGISTER_TYPE_W));
+   }
+
+   /* XXX: is this needed any more, or is this a NOOP?
+    */
+   if (mask & BRW_WRITEMASK_Y) {
+#if 0
+      /* Y' = height - 1 - Y */
+      brw_ADD(p,
+             dst[1],
+             negate(retype(arg0[1], BRW_REGISTER_TYPE_W)),
+             brw_imm_d(c->key.drawable_height - 1));
+#else
+      brw_MOV(p,
+             dst[0],
+             retype(arg0[0], BRW_REGISTER_TYPE_W));
+#endif
+   }
+}
+
+
+static void emit_pixel_w( struct brw_compile *p,
+                         const struct brw_reg *dst,
+                         GLuint mask,
+                         const struct brw_reg *arg0,
+                         const struct brw_reg *deltas)
+{
+   /* Don't need this if all you are doing is interpolating color, for
+    * instance.
+    */
+   if (mask & BRW_WRITEMASK_W) {      
+      struct brw_reg interp3 = brw_vec1_grf(arg0[0].nr+1, 4);
+
+      /* Calc 1/w - just linterp wpos[3] optimized by putting the
+       * result straight into a message reg.
+       */
+      brw_LINE(p, brw_null_reg(), interp3, deltas[0]);
+      brw_MAC(p, brw_message_reg(2), suboffset(interp3, 1), deltas[1]);
+
+      /* Calc w */
+      brw_math_16( p, dst[3],
+                  BRW_MATH_FUNCTION_INV,
+                  BRW_MATH_SATURATE_NONE,
+                  2, brw_null_reg(),
+                  BRW_MATH_PRECISION_FULL);
+   }
+}
+
+
+
+static void emit_linterp( struct brw_compile *p, 
+                        const struct brw_reg *dst,
+                        GLuint mask,
+                        const struct brw_reg *arg0,
+                        const struct brw_reg *deltas )
+{
+   struct brw_reg interp[4];
+   GLuint nr = arg0[0].nr;
+   GLuint i;
+
+   interp[0] = brw_vec1_grf(nr, 0);
+   interp[1] = brw_vec1_grf(nr, 4);
+   interp[2] = brw_vec1_grf(nr+1, 0);
+   interp[3] = brw_vec1_grf(nr+1, 4);
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {
+        brw_LINE(p, brw_null_reg(), interp[i], deltas[0]);
+        brw_MAC(p, dst[i], suboffset(interp[i],1), deltas[1]);
+      }
+   }
+}
+
+
+static void emit_pinterp( struct brw_compile *p, 
+                         const struct brw_reg *dst,
+                         GLuint mask,
+                         const struct brw_reg *arg0,
+                         const struct brw_reg *deltas,
+                         const struct brw_reg *w)
+{
+   struct brw_reg interp[4];
+   GLuint nr = arg0[0].nr;
+   GLuint i;
+
+   interp[0] = brw_vec1_grf(nr, 0);
+   interp[1] = brw_vec1_grf(nr, 4);
+   interp[2] = brw_vec1_grf(nr+1, 0);
+   interp[3] = brw_vec1_grf(nr+1, 4);
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {
+        brw_LINE(p, brw_null_reg(), interp[i], deltas[0]);
+        brw_MAC(p, dst[i], suboffset(interp[i],1), deltas[1]);
+      }
+   }
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {
+        brw_MUL(p, dst[i], dst[i], w[3]);
+      }
+   }
+}
+
+
+static void emit_cinterp( struct brw_compile *p, 
+                        const struct brw_reg *dst,
+                        GLuint mask,
+                        const struct brw_reg *arg0 )
+{
+   struct brw_reg interp[4];
+   GLuint nr = arg0[0].nr;
+   GLuint i;
+
+   interp[0] = brw_vec1_grf(nr, 0);
+   interp[1] = brw_vec1_grf(nr, 4);
+   interp[2] = brw_vec1_grf(nr+1, 0);
+   interp[3] = brw_vec1_grf(nr+1, 4);
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {
+         brw_MOV(p, dst[i], suboffset(interp[i],3));   /* TODO: optimize away like other moves */
+      }
+   }
+}
+
+/* Sets the destination channels to 1.0 or 0.0 according to glFrontFacing. */
+static void emit_frontfacing( struct brw_compile *p,
+                             const struct brw_reg *dst,
+                             GLuint mask )
+{
+   struct brw_reg r1_6ud = retype(brw_vec1_grf(1, 6), BRW_REGISTER_TYPE_UD);
+   GLuint i;
+
+   if (!(mask & BRW_WRITEMASK_XYZW))
+      return;
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {
+        brw_MOV(p, dst[i], brw_imm_f(0.0));
+      }
+   }
+
+   /* bit 31 is "primitive is back face", so checking < (1 << 31) gives
+    * us front face
+    */
+   brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_L, r1_6ud, brw_imm_ud(1 << 31));
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {
+        brw_MOV(p, dst[i], brw_imm_f(1.0));
+      }
+   }
+   brw_set_predicate_control_flag_value(p, 0xff);
+}
+
+/* For OPCODE_DDX and OPCODE_DDY, per channel of output we've got input
+ * looking like:
+ *
+ * arg0: q0.tl q0.tr q0.bl q0.br q1.tl q1.tr q1.bl q1.br
+ *
+ * and we're trying to produce:
+ *
+ *           DDX                     DDY
+ * dst: (q0.tr - q0.tl)     (q0.tl - q0.bl)
+ *      (q0.tr - q0.tl)     (q0.tr - q0.br)
+ *      (q0.br - q0.bl)     (q0.tl - q0.bl)
+ *      (q0.br - q0.bl)     (q0.tr - q0.br)
+ *      (q1.tr - q1.tl)     (q1.tl - q1.bl)
+ *      (q1.tr - q1.tl)     (q1.tr - q1.br)
+ *      (q1.br - q1.bl)     (q1.tl - q1.bl)
+ *      (q1.br - q1.bl)     (q1.tr - q1.br)
+ *
+ * and add two more quads if in 16-pixel dispatch mode.
+ *
+ * For DDX, it ends up being easy: width = 2, horiz=0 gets us the same result
+ * for each pair, and vertstride = 2 jumps us 2 elements after processing a
+ * pair. But for DDY, it's harder, as we want to produce the pairs swizzled
+ * between each other.  We could probably do it like ddx and swizzle the right
+ * order later, but bail for now and just produce
+ * ((q0.tl - q0.bl)x4 (q1.tl - q1.bl)x4)
+ */
+void emit_ddxy(struct brw_compile *p,
+              const struct brw_reg *dst,
+              GLuint mask,
+              GLboolean is_ddx,
+              const struct brw_reg *arg0)
+{
+   int i;
+   struct brw_reg src0, src1;
+
+   if (mask & SATURATE)
+      brw_set_saturate(p, 1);
+   for (i = 0; i < 4; i++ ) {
+      if (mask & (1<<i)) {
+        if (is_ddx) {
+           src0 = brw_reg(arg0[i].file, arg0[i].nr, 1,
+                          BRW_REGISTER_TYPE_F,
+                          BRW_VERTICAL_STRIDE_2,
+                          BRW_WIDTH_2,
+                          BRW_HORIZONTAL_STRIDE_0,
+                          BRW_SWIZZLE_XYZW, BRW_WRITEMASK_XYZW);
+           src1 = brw_reg(arg0[i].file, arg0[i].nr, 0,
+                          BRW_REGISTER_TYPE_F,
+                          BRW_VERTICAL_STRIDE_2,
+                          BRW_WIDTH_2,
+                          BRW_HORIZONTAL_STRIDE_0,
+                          BRW_SWIZZLE_XYZW, BRW_WRITEMASK_XYZW);
+        } else {
+           src0 = brw_reg(arg0[i].file, arg0[i].nr, 0,
+                          BRW_REGISTER_TYPE_F,
+                          BRW_VERTICAL_STRIDE_4,
+                          BRW_WIDTH_4,
+                          BRW_HORIZONTAL_STRIDE_0,
+                          BRW_SWIZZLE_XYZW, BRW_WRITEMASK_XYZW);
+           src1 = brw_reg(arg0[i].file, arg0[i].nr, 2,
+                          BRW_REGISTER_TYPE_F,
+                          BRW_VERTICAL_STRIDE_4,
+                          BRW_WIDTH_4,
+                          BRW_HORIZONTAL_STRIDE_0,
+                          BRW_SWIZZLE_XYZW, BRW_WRITEMASK_XYZW);
+        }
+        brw_ADD(p, dst[i], src0, negate(src1));
+      }
+   }
+   if (mask & SATURATE)
+      brw_set_saturate(p, 0);
+}
+
+static void emit_alu1( struct brw_compile *p, 
+                      struct brw_instruction *(*func)(struct brw_compile *, 
+                                                      struct brw_reg, 
+                                                      struct brw_reg),
+                      const struct brw_reg *dst,
+                      GLuint mask,
+                      const struct brw_reg *arg0 )
+{
+   GLuint i;
+
+   if (mask & SATURATE)
+      brw_set_saturate(p, 1);
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {
+        func(p, dst[i], arg0[i]);
+      }
+   }
+
+   if (mask & SATURATE)
+      brw_set_saturate(p, 0);
+}
+
+
+static void emit_alu2( struct brw_compile *p, 
+                      struct brw_instruction *(*func)(struct brw_compile *, 
+                                                      struct brw_reg, 
+                                                      struct brw_reg, 
+                                                      struct brw_reg),
+                      const struct brw_reg *dst,
+                      GLuint mask,
+                      const struct brw_reg *arg0,
+                      const struct brw_reg *arg1 )
+{
+   GLuint i;
+
+   if (mask & SATURATE)
+      brw_set_saturate(p, 1);
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {
+        func(p, dst[i], arg0[i], arg1[i]);
+      }
+   }
+
+   if (mask & SATURATE)
+      brw_set_saturate(p, 0);
+}
+
+
+static void emit_mad( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1,
+                     const struct brw_reg *arg2 )
+{
+   GLuint i;
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {
+        brw_MUL(p, dst[i], arg0[i], arg1[i]);
+
+        brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+        brw_ADD(p, dst[i], dst[i], arg2[i]);
+        brw_set_saturate(p, 0);
+      }
+   }
+}
+
+static void emit_trunc( struct brw_compile *p,
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0)
+{
+   GLuint i;
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {
+        brw_RNDZ(p, dst[i], arg0[i]);
+      }
+   }
+}
+
+static void emit_lrp( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1,
+                     const struct brw_reg *arg2 )
+{
+   GLuint i;
+
+   /* Uses dst as a temporary:
+    */
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {     
+        /* Can I use the LINE instruction for this? 
+         */
+        brw_ADD(p, dst[i], negate(arg0[i]), brw_imm_f(1.0));
+        brw_MUL(p, brw_null_reg(), dst[i], arg2[i]);
+
+        brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+        brw_MAC(p, dst[i], arg0[i], arg1[i]);
+        brw_set_saturate(p, 0);
+      }
+   }
+}
+
+static void emit_sop( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     GLuint cond,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   GLuint i;
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {     
+        brw_MOV(p, dst[i], brw_imm_f(0));
+        brw_CMP(p, brw_null_reg(), cond, arg0[i], arg1[i]);
+        brw_MOV(p, dst[i], brw_imm_f(1.0));
+        brw_set_predicate_control_flag_value(p, 0xff);
+      }
+   }
+}
+
+static void emit_slt( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   emit_sop(p, dst, mask, BRW_CONDITIONAL_L, arg0, arg1);
+}
+
+static void emit_sle( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   emit_sop(p, dst, mask, BRW_CONDITIONAL_LE, arg0, arg1);
+}
+
+static void emit_sgt( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   emit_sop(p, dst, mask, BRW_CONDITIONAL_G, arg0, arg1);
+}
+
+static void emit_sge( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   emit_sop(p, dst, mask, BRW_CONDITIONAL_GE, arg0, arg1);
+}
+
+static void emit_seq( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   emit_sop(p, dst, mask, BRW_CONDITIONAL_EQ, arg0, arg1);
+}
+
+static void emit_sne( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   emit_sop(p, dst, mask, BRW_CONDITIONAL_NEQ, arg0, arg1);
+}
+
+static void emit_cmp( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1,
+                     const struct brw_reg *arg2 )
+{
+   GLuint i;
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {     
+        brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+        brw_MOV(p, dst[i], arg2[i]);
+        brw_set_saturate(p, 0);
+
+        brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_L, arg0[i], brw_imm_f(0));
+
+        brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+        brw_MOV(p, dst[i], arg1[i]);
+        brw_set_saturate(p, 0);
+        brw_set_predicate_control_flag_value(p, 0xff);
+      }
+   }
+}
+
+static void emit_max( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   GLuint i;
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {     
+        brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+        brw_MOV(p, dst[i], arg0[i]);
+        brw_set_saturate(p, 0);
+
+        brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_L, arg0[i], arg1[i]);
+
+        brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+        brw_MOV(p, dst[i], arg1[i]);
+        brw_set_saturate(p, 0);
+        brw_set_predicate_control_flag_value(p, 0xff);
+      }
+   }
+}
+
+static void emit_min( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   GLuint i;
+
+   for (i = 0; i < 4; i++) {
+      if (mask & (1<<i)) {     
+        brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+        brw_MOV(p, dst[i], arg1[i]);
+        brw_set_saturate(p, 0);
+
+        brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_L, arg0[i], arg1[i]);
+
+        brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+        brw_MOV(p, dst[i], arg0[i]);
+        brw_set_saturate(p, 0);
+        brw_set_predicate_control_flag_value(p, 0xff);
+      }
+   }
+}
+
+
+static void emit_dp3( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   int dst_chan = ffs(mask & BRW_WRITEMASK_XYZW) - 1;
+
+   if (!(mask & BRW_WRITEMASK_XYZW))
+      return; /* Do not emit dead code */
+
+   assert(util_is_power_of_two(mask & BRW_WRITEMASK_XYZW));
+
+   brw_MUL(p, brw_null_reg(), arg0[0], arg1[0]);
+   brw_MAC(p, brw_null_reg(), arg0[1], arg1[1]);
+
+   brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+   brw_MAC(p, dst[dst_chan], arg0[2], arg1[2]);
+   brw_set_saturate(p, 0);
+}
+
+
+static void emit_dp4( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   int dst_chan = ffs(mask & BRW_WRITEMASK_XYZW) - 1;
+
+   if (!(mask & BRW_WRITEMASK_XYZW))
+      return; /* Do not emit dead code */
+
+   assert(util_is_power_of_two(mask & BRW_WRITEMASK_XYZW));
+
+   brw_MUL(p, brw_null_reg(), arg0[0], arg1[0]);
+   brw_MAC(p, brw_null_reg(), arg0[1], arg1[1]);
+   brw_MAC(p, brw_null_reg(), arg0[2], arg1[2]);
+
+   brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+   brw_MAC(p, dst[dst_chan], arg0[3], arg1[3]);
+   brw_set_saturate(p, 0);
+}
+
+
+static void emit_dph( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   const int dst_chan = ffs(mask & BRW_WRITEMASK_XYZW) - 1;
+
+   if (!(mask & BRW_WRITEMASK_XYZW))
+      return; /* Do not emit dead code */
+
+   assert(util_is_power_of_two(mask & BRW_WRITEMASK_XYZW));
+
+   brw_MUL(p, brw_null_reg(), arg0[0], arg1[0]);
+   brw_MAC(p, brw_null_reg(), arg0[1], arg1[1]);
+   brw_MAC(p, dst[dst_chan], arg0[2], arg1[2]);
+
+   brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+   brw_ADD(p, dst[dst_chan], dst[dst_chan], arg1[3]);
+   brw_set_saturate(p, 0);
+}
+
+
+static void emit_xpd( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0,
+                     const struct brw_reg *arg1 )
+{
+   GLuint i;
+
+   assert(!(mask & BRW_WRITEMASK_W) == BRW_WRITEMASK_X);
+   
+   for (i = 0 ; i < 3; i++) {
+      if (mask & (1<<i)) {
+        GLuint i2 = (i+2)%3;
+        GLuint i1 = (i+1)%3;
+
+        brw_MUL(p, brw_null_reg(), negate(arg0[i2]), arg1[i1]);
+
+        brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+        brw_MAC(p, dst[i], arg0[i1], arg1[i2]);
+        brw_set_saturate(p, 0);
+      }
+   }
+}
+
+
+static void emit_math1( struct brw_compile *p, 
+                       GLuint function,
+                       const struct brw_reg *dst,
+                       GLuint mask,
+                       const struct brw_reg *arg0 )
+{
+   int dst_chan = ffs(mask & BRW_WRITEMASK_XYZW) - 1;
+
+   if (!(mask & BRW_WRITEMASK_XYZW))
+      return; /* Do not emit dead code */
+
+   assert(util_is_power_of_two(mask & BRW_WRITEMASK_XYZW));
+
+   brw_MOV(p, brw_message_reg(2), arg0[0]);
+
+   /* Send two messages to perform all 16 operations:
+    */
+   brw_math_16(p, 
+              dst[dst_chan],
+              function,
+              (mask & SATURATE) ? BRW_MATH_SATURATE_SATURATE : BRW_MATH_SATURATE_NONE,
+              2,
+              brw_null_reg(),
+              BRW_MATH_PRECISION_FULL);
+}
+
+
+static void emit_math2( struct brw_compile *p, 
+                       GLuint function,
+                       const struct brw_reg *dst,
+                       GLuint mask,
+                       const struct brw_reg *arg0,
+                       const struct brw_reg *arg1)
+{
+   int dst_chan = ffs(mask & BRW_WRITEMASK_XYZW) - 1;
+
+   if (!(mask & BRW_WRITEMASK_XYZW))
+      return; /* Do not emit dead code */
+
+   assert(util_is_power_of_two(mask & BRW_WRITEMASK_XYZW));
+
+   brw_push_insn_state(p);
+
+   brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+   brw_MOV(p, brw_message_reg(2), arg0[0]);
+   brw_set_compression_control(p, BRW_COMPRESSION_2NDHALF);
+   brw_MOV(p, brw_message_reg(4), sechalf(arg0[0]));
+
+   brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+   brw_MOV(p, brw_message_reg(3), arg1[0]);
+   brw_set_compression_control(p, BRW_COMPRESSION_2NDHALF);
+   brw_MOV(p, brw_message_reg(5), sechalf(arg1[0]));
+
+   
+   /* Send two messages to perform all 16 operations:
+    */
+   brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+   brw_math(p, 
+           dst[dst_chan],
+           function,
+           (mask & SATURATE) ? BRW_MATH_SATURATE_SATURATE : BRW_MATH_SATURATE_NONE,
+           2,
+           brw_null_reg(),
+           BRW_MATH_DATA_VECTOR,
+           BRW_MATH_PRECISION_FULL);
+
+   brw_set_compression_control(p, BRW_COMPRESSION_2NDHALF);
+   brw_math(p, 
+           offset(dst[dst_chan],1),
+           function,
+           (mask & SATURATE) ? BRW_MATH_SATURATE_SATURATE : BRW_MATH_SATURATE_NONE,
+           4,
+           brw_null_reg(),
+           BRW_MATH_DATA_VECTOR,
+           BRW_MATH_PRECISION_FULL);
+   
+   brw_pop_insn_state(p);
+}
+                    
+
+
+static void emit_tex( struct brw_wm_compile *c,
+                     const struct brw_wm_instruction *inst,
+                     struct brw_reg *dst,
+                     GLuint dst_flags,
+                     struct brw_reg *coord,
+                     GLuint sampler)
+{
+   struct brw_compile *p = &c->func;
+   GLuint msgLength, responseLength;
+   GLuint i, nr;
+   GLuint emit;
+   GLuint msg_type;
+   GLboolean shadow = FALSE;
+
+   /* How many input regs are there?
+    */
+   switch (inst->target) {
+   case TGSI_TEXTURE_1D:
+      emit = BRW_WRITEMASK_X;
+      nr = 1;
+      break;
+   case TGSI_TEXTURE_SHADOW1D:
+      emit = BRW_WRITEMASK_XW;
+      nr = 4;
+      shadow = TRUE;
+      break;
+   case TGSI_TEXTURE_2D:
+      emit = BRW_WRITEMASK_XY;
+      nr = 2;
+      break;
+   case TGSI_TEXTURE_SHADOW2D:
+   case TGSI_TEXTURE_SHADOWRECT:
+      emit = BRW_WRITEMASK_XYW;
+      nr = 4;
+      shadow = TRUE;
+      break;
+   case TGSI_TEXTURE_3D:
+   case TGSI_TEXTURE_CUBE:
+      emit = BRW_WRITEMASK_XYZ;
+      nr = 3;
+      break;
+   default:
+      /* unexpected target */
+      abort();
+   }
+
+   msgLength = 1;
+
+   for (i = 0; i < nr; i++) {
+      static const GLuint swz[4] = {0,1,2,2};
+      if (emit & (1<<i)) 
+        brw_MOV(p, brw_message_reg(msgLength+1), coord[swz[i]]);
+      else
+        brw_MOV(p, brw_message_reg(msgLength+1), brw_imm_f(0));
+      msgLength += 2;
+   }
+
+   responseLength = 8;         /* always */
+
+   if (BRW_IS_IGDNG(p->brw)) {
+       if (shadow)
+           msg_type = BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_COMPARE_IGDNG;
+       else
+           msg_type = BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_IGDNG;
+   } else {
+       if (shadow)
+           msg_type = BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_COMPARE;
+       else
+           msg_type = BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE;
+   }
+
+   brw_SAMPLE(p, 
+             retype(vec16(dst[0]), BRW_REGISTER_TYPE_UW),
+             1,
+             retype(c->payload.depth[0].hw_reg, BRW_REGISTER_TYPE_UW),
+              BTI_TEXTURE(inst->tex_unit),
+             sampler,          /* sampler index */
+             inst->writemask,
+             msg_type, 
+             responseLength,
+             msgLength,
+             0,        
+             1,
+             BRW_SAMPLER_SIMD_MODE_SIMD16);    
+}
+
+
+static void emit_txb( struct brw_wm_compile *c,
+                     const struct brw_wm_instruction *inst,
+                     struct brw_reg *dst,
+                     GLuint dst_flags,
+                     struct brw_reg *coord,
+                     GLuint sampler )
+{
+   struct brw_compile *p = &c->func;
+   GLuint msgLength;
+   GLuint msg_type;
+   /* Shadow ignored for txb.
+    */
+   switch (inst->target) {
+   case TGSI_TEXTURE_1D:
+   case TGSI_TEXTURE_SHADOW1D:
+      brw_MOV(p, brw_message_reg(2), coord[0]);
+      brw_MOV(p, brw_message_reg(4), brw_imm_f(0));
+      brw_MOV(p, brw_message_reg(6), brw_imm_f(0));
+      break;
+   case TGSI_TEXTURE_2D:
+   case TGSI_TEXTURE_RECT:
+   case TGSI_TEXTURE_SHADOW2D:
+   case TGSI_TEXTURE_SHADOWRECT:
+      brw_MOV(p, brw_message_reg(2), coord[0]);
+      brw_MOV(p, brw_message_reg(4), coord[1]);
+      brw_MOV(p, brw_message_reg(6), brw_imm_f(0));
+      break;
+   case TGSI_TEXTURE_3D:
+   case TGSI_TEXTURE_CUBE:
+      brw_MOV(p, brw_message_reg(2), coord[0]);
+      brw_MOV(p, brw_message_reg(4), coord[1]);
+      brw_MOV(p, brw_message_reg(6), coord[2]);
+      break;
+   default:
+      /* unexpected target */
+      abort();
+   }
+
+   brw_MOV(p, brw_message_reg(8), coord[3]);
+   msgLength = 9;
+
+   if (BRW_IS_IGDNG(p->brw))
+       msg_type = BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_BIAS_IGDNG;
+   else
+       msg_type = BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_BIAS;
+
+   brw_SAMPLE(p, 
+             retype(vec16(dst[0]), BRW_REGISTER_TYPE_UW),
+             1,
+             retype(c->payload.depth[0].hw_reg, BRW_REGISTER_TYPE_UW),
+              BTI_TEXTURE(inst->tex_unit),
+             sampler,          /* sampler index */
+             inst->writemask,
+             msg_type,
+             8,                /* responseLength */
+             msgLength,
+             0,        
+             1,
+             BRW_SAMPLER_SIMD_MODE_SIMD16);    
+}
+
+
+static void emit_lit( struct brw_compile *p, 
+                     const struct brw_reg *dst,
+                     GLuint mask,
+                     const struct brw_reg *arg0 )
+{
+   assert((mask & BRW_WRITEMASK_XW) == 0);
+
+   if (mask & BRW_WRITEMASK_Y) {
+      brw_set_saturate(p, (mask & SATURATE) ? 1 : 0);
+      brw_MOV(p, dst[1], arg0[0]);
+      brw_set_saturate(p, 0);
+   }
+
+   if (mask & BRW_WRITEMASK_Z) {
+      emit_math2(p, BRW_MATH_FUNCTION_POW,
+                &dst[2],
+                BRW_WRITEMASK_X | (mask & SATURATE),
+                &arg0[1],
+                &arg0[3]);
+   }
+
+   /* Ordinarily you'd use an iff statement to skip or shortcircuit
+    * some of the POW calculations above, but 16-wide iff statements
+    * seem to lock c1 hardware, so this is a nasty workaround:
+    */
+   brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_LE, arg0[0], brw_imm_f(0));
+   {
+      if (mask & BRW_WRITEMASK_Y) 
+        brw_MOV(p, dst[1], brw_imm_f(0));
+
+      if (mask & BRW_WRITEMASK_Z) 
+        brw_MOV(p, dst[2], brw_imm_f(0)); 
+   }
+   brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+}
+
+
+/* Kill pixel - set execution mask to zero for those pixels which
+ * fail.
+ */
+static void emit_kil( struct brw_wm_compile *c,
+                     struct brw_reg *arg0)
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg r0uw = retype(brw_vec1_grf(0, 0), BRW_REGISTER_TYPE_UW);
+   GLuint i;
+   
+   /* XXX - usually won't need 4 compares!
+    */
+   for (i = 0; i < 4; i++) {
+      brw_push_insn_state(p);
+      brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_GE, arg0[i], brw_imm_f(0));   
+      brw_set_predicate_control_flag_value(p, 0xff);
+      brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+      brw_AND(p, r0uw, brw_flag_reg(), r0uw);
+      brw_pop_insn_state(p);
+   }
+}
+
+/* KILLP kills the pixels that are currently executing, not based on a test
+ * of the arguments.
+ */
+static void emit_killp( struct brw_wm_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   struct brw_reg r0uw = retype(brw_vec1_grf(0, 0), BRW_REGISTER_TYPE_UW);
+
+   brw_push_insn_state(p);
+   brw_set_mask_control(p, BRW_MASK_DISABLE);
+   brw_NOT(p, c->emit_mask_reg, brw_mask_reg(1)); /* IMASK */
+   brw_AND(p, r0uw, c->emit_mask_reg, r0uw);
+   brw_pop_insn_state(p);
+}
+
+static void fire_fb_write( struct brw_wm_compile *c,
+                          GLuint base_reg,
+                          GLuint nr,
+                          GLuint target,
+                          GLuint eot )
+{
+   struct brw_compile *p = &c->func;
+   
+   /* Pass through control information:
+    */
+/*  mov (8) m1.0<1>:ud   r1.0<8;8,1>:ud   { Align1 NoMask } */
+   {
+      brw_push_insn_state(p);
+      brw_set_mask_control(p, BRW_MASK_DISABLE); /* ? */
+      brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+      brw_MOV(p, 
+              brw_message_reg(base_reg + 1),
+              brw_vec8_grf(1, 0));
+      brw_pop_insn_state(p);
+   }
+
+   /* Send framebuffer write message: */
+/*  send (16) null.0<1>:uw m0               r0.0<8;8,1>:uw   0x85a04000:ud    { Align1 EOT } */
+   brw_fb_WRITE(p,
+               retype(vec16(brw_null_reg()), BRW_REGISTER_TYPE_UW),
+               base_reg,
+               retype(brw_vec8_grf(0, 0), BRW_REGISTER_TYPE_UW),
+               target,         
+               nr,
+               0, 
+               eot);
+}
+
+
+static void emit_aa( struct brw_wm_compile *c,
+                    struct brw_reg *arg1,
+                    GLuint reg )
+{
+   struct brw_compile *p = &c->func;
+   GLuint comp = c->key.aa_dest_stencil_reg / 2;
+   GLuint off = c->key.aa_dest_stencil_reg % 2;
+   struct brw_reg aa = offset(arg1[comp], off);
+
+   brw_push_insn_state(p);
+   brw_set_compression_control(p, BRW_COMPRESSION_NONE); /* ?? */
+   brw_MOV(p, brw_message_reg(reg), aa);
+   brw_pop_insn_state(p);
+}
+
+
+/* Post-fragment-program processing.  Send the results to the
+ * framebuffer.
+ * \param arg0  the fragment color
+ * \param arg1  the pass-through depth value
+ * \param arg2  the shader-computed depth value
+ */
+static void emit_fb_write( struct brw_wm_compile *c,
+                          struct brw_reg *arg0,
+                          struct brw_reg *arg1,
+                          struct brw_reg *arg2,
+                          GLuint target,
+                          GLuint eot)
+{
+   struct brw_compile *p = &c->func;
+   GLuint nr = 2;
+   GLuint channel;
+
+   /* Reserve a space for AA - may not be needed:
+    */
+   if (c->key.aa_dest_stencil_reg)
+      nr += 1;
+
+   /* I don't really understand how this achieves the color interleave
+    * (ie RGBARGBA) in the result:  [Do the saturation here]
+    */
+   {
+      brw_push_insn_state(p);
+      
+      for (channel = 0; channel < 4; channel++) {
+        /*  mov (8) m2.0<1>:ud   r28.0<8;8,1>:ud  { Align1 } */
+        /*  mov (8) m6.0<1>:ud   r29.0<8;8,1>:ud  { Align1 SecHalf } */
+
+        brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+        brw_MOV(p,
+                brw_message_reg(nr + channel),
+                arg0[channel]);
+       
+        brw_set_compression_control(p, BRW_COMPRESSION_2NDHALF);
+        brw_MOV(p,
+                brw_message_reg(nr + channel + 4),
+                sechalf(arg0[channel]));
+      }
+
+      /* skip over the regs populated above:
+       */
+      nr += 8;
+   
+      brw_pop_insn_state(p);
+   }
+
+   if (c->key.source_depth_to_render_target)
+   {
+      if (c->key.computes_depth) 
+        brw_MOV(p, brw_message_reg(nr), arg2[2]);
+      else 
+        brw_MOV(p, brw_message_reg(nr), arg1[1]); /* ? */
+
+      nr += 2;
+   }
+
+   if (c->key.dest_depth_reg)
+   {
+      GLuint comp = c->key.dest_depth_reg / 2;
+      GLuint off = c->key.dest_depth_reg % 2;
+
+      if (off != 0) {
+         brw_push_insn_state(p);
+         brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+
+         brw_MOV(p, brw_message_reg(nr), offset(arg1[comp],1));
+         /* 2nd half? */
+         brw_MOV(p, brw_message_reg(nr+1), arg1[comp+1]);
+         brw_pop_insn_state(p);
+      }
+      else {
+         brw_MOV(p, brw_message_reg(nr), arg1[comp]);
+      }
+      nr += 2;
+   }
+
+   if (!c->key.runtime_check_aads_emit) {
+      if (c->key.aa_dest_stencil_reg)
+        emit_aa(c, arg1, 2);
+
+      fire_fb_write(c, 0, nr, target, eot);
+   }
+   else {
+      struct brw_reg v1_null_ud = vec1(retype(brw_null_reg(), BRW_REGISTER_TYPE_UD));
+      struct brw_reg ip = brw_ip_reg();
+      struct brw_instruction *jmp;
+      
+      brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+      brw_set_conditionalmod(p, BRW_CONDITIONAL_Z);
+      brw_AND(p, 
+             v1_null_ud, 
+             get_element_ud(brw_vec8_grf(1,0), 6), 
+             brw_imm_ud(1<<26)); 
+
+      jmp = brw_JMPI(p, ip, ip, brw_imm_d(0));
+      {
+        emit_aa(c, arg1, 2);
+        fire_fb_write(c, 0, nr, target, eot);
+        /* note - thread killed in subroutine */
+      }
+      brw_land_fwd_jump(p, jmp);
+
+      /* ELSE: Shuffle up one register to fill in the hole left for AA:
+       */
+      fire_fb_write(c, 1, nr-1, target, eot);
+   }
+}
+
+
+/**
+ * Move a GPR to scratch memory. 
+ */
+static void emit_spill( struct brw_wm_compile *c,
+                       struct brw_reg reg,
+                       GLuint slot )
+{
+   struct brw_compile *p = &c->func;
+
+   /*
+     mov (16) m2.0<1>:ud   r2.0<8;8,1>:ud   { Align1 Compr }
+   */
+   brw_MOV(p, brw_message_reg(2), reg);
+
+   /*
+     mov (1) r0.2<1>:d    0x00000080:d     { Align1 NoMask }
+     send (16) null.0<1>:uw m1               r0.0<8;8,1>:uw   0x053003ff:ud    { Align1 }
+   */
+   brw_dp_WRITE_16(p, 
+                  retype(vec16(brw_vec8_grf(0, 0)), BRW_REGISTER_TYPE_UW),
+                  slot);
+}
+
+
+/**
+ * Load a GPR from scratch memory. 
+ */
+static void emit_unspill( struct brw_wm_compile *c,
+                         struct brw_reg reg,
+                         GLuint slot )
+{
+   struct brw_compile *p = &c->func;
+
+   /* Slot 0 is the undef value.
+    */
+   if (slot == 0) {
+      brw_MOV(p, reg, brw_imm_f(0));
+      return;
+   }
+
+   /*
+     mov (1) r0.2<1>:d    0x000000c0:d     { Align1 NoMask }
+     send (16) r110.0<1>:uw m1               r0.0<8;8,1>:uw   0x041243ff:ud    { Align1 }
+   */
+
+   brw_dp_READ_16(p,
+                 retype(vec16(reg), BRW_REGISTER_TYPE_UW),
+                 slot);
+}
+
+
+/**
+ * Retrieve up to 4 GEN4 register pairs for the given wm reg:
+ * Args with unspill_reg != 0 will be loaded from scratch memory.
+ */
+static void get_argument_regs( struct brw_wm_compile *c,
+                              struct brw_wm_ref *arg[],
+                              struct brw_reg *regs )
+{
+   GLuint i;
+
+   for (i = 0; i < 4; i++) {
+      if (arg[i]) {
+        if (arg[i]->unspill_reg)
+           emit_unspill(c,
+                        brw_vec8_grf(arg[i]->unspill_reg, 0),
+                        arg[i]->value->spill_slot);
+
+        regs[i] = arg[i]->hw_reg;
+      }
+      else {
+        regs[i] = brw_null_reg();
+      }
+   }
+}
+
+
+/**
+ * For values that have a spill_slot!=0, write those regs to scratch memory.
+ */
+static void spill_values( struct brw_wm_compile *c,
+                         struct brw_wm_value *values,
+                         GLuint nr )
+{
+   GLuint i;
+
+   for (i = 0; i < nr; i++)
+      if (values[i].spill_slot) 
+        emit_spill(c, values[i].hw_reg, values[i].spill_slot);
+}
+
+
+/* Emit the fragment program instructions here.
+ */
+void brw_wm_emit( struct brw_wm_compile *c )
+{
+   struct brw_compile *p = &c->func;
+   GLuint insn;
+
+   brw_set_compression_control(p, BRW_COMPRESSION_COMPRESSED);
+
+   /* Check if any of the payload regs need to be spilled:
+    */
+   spill_values(c, c->payload.depth, 4);
+   spill_values(c, c->creg, c->nr_creg);
+   spill_values(c, c->payload.input_interp, PIPE_MAX_SHADER_INPUTS);
+   
+
+   for (insn = 0; insn < c->nr_insns; insn++) {
+
+      struct brw_wm_instruction *inst = &c->instruction[insn];
+      struct brw_reg args[3][4], dst[4];
+      GLuint i, dst_flags;
+      
+      /* Get argument regs:
+       */
+      for (i = 0; i < 3; i++) 
+        get_argument_regs(c, inst->src[i], args[i]);
+
+      /* Get dest regs:
+       */
+      for (i = 0; i < 4; i++)
+        if (inst->dst[i])
+           dst[i] = inst->dst[i]->hw_reg;
+        else
+           dst[i] = brw_null_reg();
+      
+      /* Flags
+       */
+      dst_flags = inst->writemask;
+      if (inst->saturate) 
+        dst_flags |= SATURATE;
+
+      switch (inst->opcode) {
+        /* Generated instructions for calculating triangle interpolants:
+         */
+      case WM_PIXELXY:
+        emit_pixel_xy(p, dst, dst_flags);
+        break;
+
+      case WM_DELTAXY:
+        emit_delta_xy(p, dst, dst_flags, args[0]);
+        break;
+
+      case WM_WPOSXY:
+        emit_wpos_xy(c, dst, dst_flags, args[0]);
+        break;
+
+      case WM_PIXELW:
+        emit_pixel_w(p, dst, dst_flags, args[0], args[1]);
+        break;
+
+      case WM_LINTERP:
+        emit_linterp(p, dst, dst_flags, args[0], args[1]);
+        break;
+
+      case WM_PINTERP:
+        emit_pinterp(p, dst, dst_flags, args[0], args[1], args[2]);
+        break;
+
+      case WM_CINTERP:
+        emit_cinterp(p, dst, dst_flags, args[0]);
+        break;
+
+      case WM_FB_WRITE:
+        emit_fb_write(c, args[0], args[1], args[2], inst->target, inst->eot);
+        break;
+
+      case WM_FRONTFACING:
+        emit_frontfacing(p, dst, dst_flags);
+        break;
+
+        /* Straightforward arithmetic:
+         */
+      case TGSI_OPCODE_ADD:
+        emit_alu2(p, brw_ADD, dst, dst_flags, args[0], args[1]);
+        break;
+
+      case TGSI_OPCODE_FRC:
+        emit_alu1(p, brw_FRC, dst, dst_flags, args[0]);
+        break;
+
+      case TGSI_OPCODE_FLR:
+        emit_alu1(p, brw_RNDD, dst, dst_flags, args[0]);
+        break;
+
+      case TGSI_OPCODE_DDX:
+        emit_ddxy(p, dst, dst_flags, GL_TRUE, args[0]);
+        break;
+
+      case TGSI_OPCODE_DDY:
+        emit_ddxy(p, dst, dst_flags, GL_FALSE, args[0]);
+        break;
+
+      case TGSI_OPCODE_DP3:
+        emit_dp3(p, dst, dst_flags, args[0], args[1]);
+        break;
+
+      case TGSI_OPCODE_DP4:
+        emit_dp4(p, dst, dst_flags, args[0], args[1]);
+        break;
+
+      case TGSI_OPCODE_DPH:
+        emit_dph(p, dst, dst_flags, args[0], args[1]);
+        break;
+
+      case TGSI_OPCODE_TRUNC:
+        emit_trunc(p, dst, dst_flags, args[0]);
+        break;
+
+      case TGSI_OPCODE_LRP:
+        emit_lrp(p, dst, dst_flags, args[0], args[1], args[2]);
+        break;
+
+      case TGSI_OPCODE_MAD:    
+        emit_mad(p, dst, dst_flags, args[0], args[1], args[2]);
+        break;
+
+      case TGSI_OPCODE_MOV:
+        emit_alu1(p, brw_MOV, dst, dst_flags, args[0]);
+        break;
+
+      case TGSI_OPCODE_MUL:
+        emit_alu2(p, brw_MUL, dst, dst_flags, args[0], args[1]);
+        break;
+
+      case TGSI_OPCODE_XPD:
+        emit_xpd(p, dst, dst_flags, args[0], args[1]);
+        break;
+
+        /* Higher math functions:
+         */
+      case TGSI_OPCODE_RCP:
+        emit_math1(p, BRW_MATH_FUNCTION_INV, dst, dst_flags, args[0]);
+        break;
+
+      case TGSI_OPCODE_RSQ:
+        emit_math1(p, BRW_MATH_FUNCTION_RSQ, dst, dst_flags, args[0]);
+        break;
+
+      case TGSI_OPCODE_SIN:
+        emit_math1(p, BRW_MATH_FUNCTION_SIN, dst, dst_flags, args[0]);
+        break;
+
+      case TGSI_OPCODE_COS:
+        emit_math1(p, BRW_MATH_FUNCTION_COS, dst, dst_flags, args[0]);
+        break;
+
+      case TGSI_OPCODE_EX2:
+        emit_math1(p, BRW_MATH_FUNCTION_EXP, dst, dst_flags, args[0]);
+        break;
+
+      case TGSI_OPCODE_LG2:
+        emit_math1(p, BRW_MATH_FUNCTION_LOG, dst, dst_flags, args[0]);
+        break;
+
+      case TGSI_OPCODE_SCS:
+        /* There is an scs math function, but it would need some
+         * fixup for 16-element execution.
+         */
+        if (dst_flags & BRW_WRITEMASK_X)
+           emit_math1(p, BRW_MATH_FUNCTION_COS, dst, (dst_flags&SATURATE)|BRW_WRITEMASK_X, args[0]);
+        if (dst_flags & BRW_WRITEMASK_Y)
+           emit_math1(p, BRW_MATH_FUNCTION_SIN, dst+1, (dst_flags&SATURATE)|BRW_WRITEMASK_X, args[0]);
+        break;
+
+      case TGSI_OPCODE_POW:
+        emit_math2(p, BRW_MATH_FUNCTION_POW, dst, dst_flags, args[0], args[1]);
+        break;
+
+        /* Comparisons:
+         */
+      case TGSI_OPCODE_CMP:
+        emit_cmp(p, dst, dst_flags, args[0], args[1], args[2]);
+        break;
+
+      case TGSI_OPCODE_MAX:
+        emit_max(p, dst, dst_flags, args[0], args[1]);
+        break;
+
+      case TGSI_OPCODE_MIN:
+        emit_min(p, dst, dst_flags, args[0], args[1]);
+        break;
+
+      case TGSI_OPCODE_SLT:
+        emit_slt(p, dst, dst_flags, args[0], args[1]);
+        break;
+
+      case TGSI_OPCODE_SLE:
+        emit_sle(p, dst, dst_flags, args[0], args[1]);
+       break;
+      case TGSI_OPCODE_SGT:
+        emit_sgt(p, dst, dst_flags, args[0], args[1]);
+       break;
+      case TGSI_OPCODE_SGE:
+        emit_sge(p, dst, dst_flags, args[0], args[1]);
+        break;
+      case TGSI_OPCODE_SEQ:
+        emit_seq(p, dst, dst_flags, args[0], args[1]);
+       break;
+      case TGSI_OPCODE_SNE:
+        emit_sne(p, dst, dst_flags, args[0], args[1]);
+       break;
+
+      case TGSI_OPCODE_LIT:
+        emit_lit(p, dst, dst_flags, args[0]);
+        break;
+
+        /* Texturing operations:
+         */
+      case TGSI_OPCODE_TEX:
+        emit_tex(c, inst, dst, dst_flags, args[0], inst->sampler);
+        break;
+
+      case TGSI_OPCODE_TXB:
+        emit_txb(c, inst, dst, dst_flags, args[0], inst->sampler);
+        break;
+
+      case TGSI_OPCODE_KIL:
+        emit_kil(c, args[0]);
+        break;
+
+      case TGSI_OPCODE_KILP:
+        emit_killp(c);
+        break;
+
+      default:
+        debug_printf("Unsupported opcode %i (%s) in fragment shader\n",
+                     inst->opcode, 
+                     tgsi_get_opcode_info(inst->opcode)->mnemonic);
+      }
+      
+      for (i = 0; i < 4; i++)
+       if (inst->dst[i] && inst->dst[i]->spill_slot) 
+          emit_spill(c, 
+                     inst->dst[i]->hw_reg, 
+                     inst->dst[i]->spill_slot);
+   }
+
+   if (BRW_DEBUG & DEBUG_WM) {
+      debug_printf("wm-native:\n");
+      brw_disasm(stderr, p->store, p->nr_insn);
+   }
+}
diff --git a/src/gallium/drivers/i965/brw_wm_fp.c b/src/gallium/drivers/i965/brw_wm_fp.c
new file mode 100644 (file)
index 0000000..9c5b527
--- /dev/null
@@ -0,0 +1,1224 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+               
+
+#include "pipe/p_shader_tokens.h"
+
+#include "util/u_math.h"
+#include "util/u_memory.h"
+
+#include "tgsi/tgsi_parse.h"
+#include "tgsi/tgsi_dump.h"
+#include "tgsi/tgsi_info.h"
+#include "tgsi/tgsi_util.h"
+
+#include "brw_wm.h"
+#include "brw_util.h"
+#include "brw_debug.h"
+
+
+/***********************************************************************
+ * Source regs
+ */
+
+static struct brw_fp_src src_reg(GLuint file, GLuint idx)
+{
+   struct brw_fp_src reg;
+   reg.file = file;
+   reg.index = idx;
+   reg.swizzle = BRW_SWIZZLE_XYZW;
+   reg.indirect = 0;
+   reg.negate = 0;
+   reg.abs = 0;
+   return reg;
+}
+
+static struct brw_fp_src src_reg_from_dst(struct brw_fp_dst dst)
+{
+   return src_reg(dst.file, dst.index);
+}
+
+static struct brw_fp_src src_undef( void )
+{
+   return src_reg(TGSI_FILE_NULL, 0);
+}
+
+static GLboolean src_is_undef(struct brw_fp_src src)
+{
+   return src.file == TGSI_FILE_NULL;
+}
+
+static struct brw_fp_src src_swizzle( struct brw_fp_src reg, int x, int y, int z, int w )
+{
+   unsigned swz = reg.swizzle;
+
+   reg.swizzle = ( BRW_GET_SWZ(swz, x) << 0 |
+                  BRW_GET_SWZ(swz, y) << 2 |
+                  BRW_GET_SWZ(swz, z) << 4 |
+                  BRW_GET_SWZ(swz, w) << 6 );
+
+   return reg;
+}
+
+static struct brw_fp_src src_scalar( struct brw_fp_src reg, int x )
+{
+   return src_swizzle(reg, x, x, x, x);
+}
+
+static struct brw_fp_src src_abs( struct brw_fp_src src )
+{
+   src.negate = 0;
+   src.abs = 1;
+   return src;
+}
+
+static struct brw_fp_src src_negate( struct brw_fp_src src )
+{
+   src.negate = 1;
+   src.abs = 0;
+   return src;
+}
+
+
+static int match_or_expand_immediate( const float *v,
+                                      unsigned nr,
+                                      float *v2,
+                                      unsigned *nr2,
+                                      unsigned *swizzle )
+{
+   unsigned i, j;
+   
+   *swizzle = 0;
+
+   for (i = 0; i < nr; i++) {
+      boolean found = FALSE;
+
+      for (j = 0; j < *nr2 && !found; j++) {
+         if (v[i] == v2[j]) {
+            *swizzle |= j << (i * 2);
+            found = TRUE;
+         }
+      }
+
+      if (!found) {
+         if (*nr2 >= 4) 
+            return FALSE;
+
+         v2[*nr2] = v[i];
+         *swizzle |= *nr2 << (i * 2);
+         (*nr2)++;
+      }
+   }
+
+   return TRUE;
+}
+
+
+
+/* Internally generated immediates: overkill...
+ */
+static struct brw_fp_src src_imm( struct brw_wm_compile *c, 
+                                 const GLfloat *v, 
+                                 unsigned nr)
+{
+   unsigned i, j;
+   unsigned swizzle;
+
+   /* Could do a first pass where we examine all existing immediates
+    * without expanding.
+    */
+
+   for (i = 0; i < c->nr_immediates; i++) {
+      if (match_or_expand_immediate( v, 
+                                     nr,
+                                     c->immediate[i].v,
+                                     &c->immediate[i].nr, 
+                                     &swizzle ))
+         goto out;
+   }
+
+   if (c->nr_immediates < Elements(c->immediate)) {
+      i = c->nr_immediates++;
+      if (match_or_expand_immediate( v,
+                                     nr,
+                                     c->immediate[i].v,
+                                     &c->immediate[i].nr, 
+                                     &swizzle ))
+         goto out;
+   }
+
+   c->error = 1;
+   return src_undef();
+
+out:
+   /* Make sure that all referenced elements are from this immediate.
+    * Has the effect of making size-one immediates into scalars.
+    */
+   for (j = nr; j < 4; j++)
+      swizzle |= (swizzle & 0x3) << (j * 2);
+
+   return src_swizzle( src_reg( TGSI_FILE_IMMEDIATE, i ),
+                      BRW_GET_SWZ(swizzle, X),
+                      BRW_GET_SWZ(swizzle, Y),
+                      BRW_GET_SWZ(swizzle, Z),
+                      BRW_GET_SWZ(swizzle, W) );
+}
+
+
+
+static struct brw_fp_src src_imm1f( struct brw_wm_compile *c,
+                                   GLfloat f )
+{
+   return src_imm(c, &f, 1);
+}
+
+static struct brw_fp_src src_imm4f( struct brw_wm_compile *c,
+                                   GLfloat x,
+                                   GLfloat y,
+                                   GLfloat z,
+                                   GLfloat w)
+{
+   GLfloat f[4] = {x,y,z,w};
+   return src_imm(c, f, 4);
+}
+
+
+
+/***********************************************************************
+ * Dest regs
+ */
+
+static struct brw_fp_dst dst_reg(GLuint file, GLuint idx)
+{
+   struct brw_fp_dst reg;
+   reg.file = file;
+   reg.index = idx;
+   reg.writemask = BRW_WRITEMASK_XYZW;
+   reg.indirect = 0;
+   reg.saturate = 0;
+   return reg;
+}
+
+static struct brw_fp_dst dst_mask( struct brw_fp_dst reg, int mask )
+{
+   reg.writemask &= mask;
+   return reg;
+}
+
+static struct brw_fp_dst dst_undef( void )
+{
+   return dst_reg(TGSI_FILE_NULL, 0);
+}
+
+static boolean dst_is_undef( struct brw_fp_dst dst )
+{
+   return dst.file == TGSI_FILE_NULL;
+}
+
+static struct brw_fp_dst dst_saturate( struct brw_fp_dst reg, boolean flag )
+{
+   reg.saturate = flag;
+   return reg;
+}
+
+static struct brw_fp_dst get_temp( struct brw_wm_compile *c )
+{
+   int bit = ffs( ~c->fp_temp );
+
+   if (!bit) {
+      debug_printf("%s: out of temporaries\n", __FILE__);
+   }
+
+   c->fp_temp |= 1<<(bit-1);
+   return dst_reg(TGSI_FILE_TEMPORARY, c->fp_first_internal_temp+(bit-1));
+}
+
+
+static void release_temp( struct brw_wm_compile *c, struct brw_fp_dst temp )
+{
+   c->fp_temp &= ~(1 << (temp.index - c->fp_first_internal_temp));
+}
+
+
+/***********************************************************************
+ * Instructions 
+ */
+
+static struct brw_fp_instruction *get_fp_inst(struct brw_wm_compile *c)
+{
+   return &c->fp_instructions[c->nr_fp_insns++];
+}
+
+static struct brw_fp_instruction * emit_tex_op(struct brw_wm_compile *c,
+                                            GLuint op,
+                                            struct brw_fp_dst dest,
+                                            GLuint tex_unit,
+                                            GLuint target,
+                                            GLuint sampler,
+                                            struct brw_fp_src src0,
+                                            struct brw_fp_src src1,
+                                            struct brw_fp_src src2 )
+{
+   struct brw_fp_instruction *inst = get_fp_inst(c);
+
+   if (tex_unit || target)
+      assert(op == TGSI_OPCODE_TXP ||
+             op == TGSI_OPCODE_TXB ||
+             op == TGSI_OPCODE_TEX ||
+             op == WM_FB_WRITE);
+
+   inst->opcode = op;
+   inst->dst = dest;
+   inst->tex_unit = tex_unit;
+   inst->target = target;
+   inst->sampler = sampler;
+   inst->src[0] = src0;
+   inst->src[1] = src1;
+   inst->src[2] = src2;
+
+   return inst;
+}
+   
+
+static INLINE void emit_op3(struct brw_wm_compile *c,
+                           GLuint op,
+                           struct brw_fp_dst dest,
+                           struct brw_fp_src src0,
+                           struct brw_fp_src src1,
+                           struct brw_fp_src src2 )
+{
+   emit_tex_op(c, op, dest, 0, 0, 0, src0, src1, src2);
+}
+
+
+static INLINE void emit_op2(struct brw_wm_compile *c,
+                           GLuint op,
+                           struct brw_fp_dst dest,
+                           struct brw_fp_src src0,
+                           struct brw_fp_src src1)
+{
+   emit_tex_op(c, op, dest, 0, 0, 0, src0, src1, src_undef());
+}
+
+static INLINE void emit_op1(struct brw_wm_compile *c,
+                           GLuint op,
+                           struct brw_fp_dst dest,
+                           struct brw_fp_src src0)
+{
+   emit_tex_op(c, op, dest, 0, 0, 0, src0, src_undef(), src_undef());
+}
+
+static INLINE void emit_op0(struct brw_wm_compile *c,
+                          GLuint op,
+                          struct brw_fp_dst dest)
+{
+   emit_tex_op(c, op, dest, 0, 0, 0, src_undef(), src_undef(), src_undef());
+}
+
+
+
+/* Many opcodes produce the same value across all the result channels.
+ * We'd rather not have to support that splatting in the opcode implementations,
+ * and brw_wm_pass*.c wants to optimize them out by shuffling references around
+ * anyway.  We can easily get both by emitting the opcode to one channel, and
+ * then MOVing it to the others, which brw_wm_pass*.c already understands.
+ */
+static void emit_scalar_insn(struct brw_wm_compile *c,
+                            unsigned opcode,
+                            struct brw_fp_dst dst,
+                            struct brw_fp_src src0,
+                            struct brw_fp_src src1,
+                            struct brw_fp_src src2 )
+{
+   unsigned first_chan = ffs(dst.writemask) - 1;
+   unsigned first_mask = 1 << first_chan;
+
+   if (dst.writemask == 0)
+      return;
+
+   emit_op3( c, opcode,
+            dst_mask(dst, first_mask),
+            src0, src1, src2 );
+
+   if (dst.writemask != first_mask) {
+      emit_op1(c, TGSI_OPCODE_MOV,
+              dst_mask(dst, ~first_mask),
+              src_scalar(src_reg_from_dst(dst), first_chan));
+   }
+}
+
+
+/***********************************************************************
+ * Special instructions for interpolation and other tasks
+ */
+
+static struct brw_fp_src get_pixel_xy( struct brw_wm_compile *c )
+{
+   if (src_is_undef(c->fp_pixel_xy)) {
+      struct brw_fp_dst pixel_xy = get_temp(c);
+      struct brw_fp_src payload_r0_depth = src_reg(BRW_FILE_PAYLOAD, PAYLOAD_DEPTH);
+      
+      
+      /* Emit the out calculations, and hold onto the results.  Use
+       * two instructions as a temporary is required.
+       */   
+      /* pixel_xy.xy = PIXELXY payload[0];
+       */
+      emit_op1(c,
+              WM_PIXELXY,
+              dst_mask(pixel_xy, BRW_WRITEMASK_XY),
+              payload_r0_depth);
+
+      c->fp_pixel_xy = src_reg_from_dst(pixel_xy);
+   }
+
+   return c->fp_pixel_xy;
+}
+
+static struct brw_fp_src get_delta_xy( struct brw_wm_compile *c )
+{
+   if (src_is_undef(c->fp_delta_xy)) {
+      struct brw_fp_dst delta_xy = get_temp(c);
+      struct brw_fp_src pixel_xy = get_pixel_xy(c);
+      struct brw_fp_src payload_r0_depth = src_reg(BRW_FILE_PAYLOAD, PAYLOAD_DEPTH);
+      
+      /* deltas.xy = DELTAXY pixel_xy, payload[0]
+       */
+      emit_op3(c,
+             WM_DELTAXY,
+             dst_mask(delta_xy, BRW_WRITEMASK_XY),
+             pixel_xy, 
+             payload_r0_depth,
+             src_undef());
+      
+      c->fp_delta_xy = src_reg_from_dst(delta_xy);
+   }
+
+   return c->fp_delta_xy;
+}
+
+static struct brw_fp_src get_pixel_w( struct brw_wm_compile *c )
+{
+   if (src_is_undef(c->fp_pixel_w)) {
+      struct brw_fp_dst pixel_w = get_temp(c);
+      struct brw_fp_src deltas = get_delta_xy(c);
+
+      /* XXX: assuming position is always first -- valid? 
+       */
+      struct brw_fp_src interp_wpos = src_reg(BRW_FILE_PAYLOAD, 0);
+
+      /* deltas.xyw = DELTAS2 deltas.xy, payload.interp_wpos.x
+       */
+      emit_op3(c,
+              WM_PIXELW,
+              dst_mask(pixel_w, BRW_WRITEMASK_W),
+              interp_wpos,
+              deltas, 
+              src_undef());
+      
+
+      c->fp_pixel_w = src_reg_from_dst(pixel_w);
+   }
+
+   return c->fp_pixel_w;
+}
+
+
+/***********************************************************************
+ * Emit INTERP instructions ahead of first use of each attrib.
+ */
+
+static void emit_interp( struct brw_wm_compile *c,
+                        GLuint idx,
+                        GLuint semantic,
+                        GLuint interp_mode )
+{
+   struct brw_fp_dst dst = dst_reg(TGSI_FILE_INPUT, idx);
+   struct brw_fp_src interp = src_reg(BRW_FILE_PAYLOAD, idx);
+   struct brw_fp_src deltas = get_delta_xy(c);
+
+   /* Need to use PINTERP on attributes which have been
+    * multiplied by 1/W in the SF program, and LINTERP on those
+    * which have not:
+    */
+   switch (semantic) {
+   case TGSI_SEMANTIC_POSITION:
+      /* Have to treat wpos.xy specially:
+       */
+      emit_op1(c,
+             WM_WPOSXY,
+             dst_mask(dst, BRW_WRITEMASK_XY),
+             get_pixel_xy(c));
+      
+      /* TGSI_FILE_INPUT.attr.xyzw = INTERP payload.interp[attr].x, deltas.xyw
+       */
+      emit_op2(c,
+              WM_LINTERP,
+              dst_mask(dst, BRW_WRITEMASK_ZW),
+              interp,
+              deltas);
+      break;
+
+   case TGSI_SEMANTIC_COLOR:
+      if (c->key.flat_shade) {
+        emit_op1(c,
+                WM_CINTERP,
+                dst,
+                interp);
+      }
+      else if (interp_mode == TGSI_INTERPOLATE_LINEAR) {
+        emit_op2(c,
+                 WM_LINTERP,
+                 dst,
+                 interp,
+                 deltas);
+      }
+      else {
+        emit_op3(c,
+                 WM_PINTERP,
+                 dst,
+                 interp,
+                 deltas,
+                 get_pixel_w(c));
+      }
+
+      break;
+
+   case TGSI_SEMANTIC_FOG:
+      /* Interpolate the fog coordinate */
+      emit_op3(c,
+             WM_PINTERP,
+             dst_mask(dst, BRW_WRITEMASK_X),
+             interp,
+             deltas,
+             get_pixel_w(c));
+
+      emit_op1(c,
+              TGSI_OPCODE_MOV,
+              dst_mask(dst, BRW_WRITEMASK_YZ),
+              src_imm1f(c, 0.0));
+
+      emit_op1(c,
+              TGSI_OPCODE_MOV,
+              dst_mask(dst, BRW_WRITEMASK_W),
+              src_imm1f(c, 1.0));
+      break;
+
+   case TGSI_SEMANTIC_FACE:
+      /* XXX review/test this case */
+      emit_op0(c,
+              WM_FRONTFACING,
+              dst_mask(dst, BRW_WRITEMASK_X));
+      
+      emit_op1(c,
+             TGSI_OPCODE_MOV,
+             dst_mask(dst, BRW_WRITEMASK_YZ),
+              src_imm1f(c, 0.0));
+
+      emit_op1(c,
+             TGSI_OPCODE_MOV,
+             dst_mask(dst, BRW_WRITEMASK_W),
+              src_imm1f(c, 1.0));
+      break;
+
+   case TGSI_SEMANTIC_PSIZE:
+      /* XXX review/test this case */
+      emit_op3(c,
+              WM_PINTERP,
+              dst_mask(dst, BRW_WRITEMASK_XY),
+              interp,
+              deltas,
+              get_pixel_w(c));
+
+      emit_op1(c,
+             TGSI_OPCODE_MOV,
+             dst_mask(dst, BRW_WRITEMASK_Z),
+             src_imm1f(c, 0.0f));
+
+      emit_op1(c,
+             TGSI_OPCODE_MOV,
+             dst_mask(dst, BRW_WRITEMASK_W),
+             src_imm1f(c, 1.0f));
+      break;
+
+   default: 
+      switch (interp_mode) {
+      case TGSI_INTERPOLATE_CONSTANT:
+        emit_op1(c,
+                 WM_CINTERP,
+                 dst,
+                 interp);
+        break;
+
+      case TGSI_INTERPOLATE_LINEAR:
+        emit_op2(c,
+                 WM_LINTERP,
+                 dst,
+                 interp,
+                 deltas);
+        break;
+
+      case TGSI_INTERPOLATE_PERSPECTIVE:
+        emit_op3(c,
+                 WM_PINTERP,
+                 dst,
+                 interp,
+                 deltas,
+                 get_pixel_w(c));
+        break;
+      }
+      break;
+   }
+}
+
+
+/***********************************************************************
+ * Expand various instructions here to simpler forms.  
+ */
+static void precalc_dst( struct brw_wm_compile *c,
+                        struct brw_fp_dst dst,
+                        struct brw_fp_src src0,
+                        struct brw_fp_src src1 )
+{
+   if (dst.writemask & BRW_WRITEMASK_Y) {      
+      /* dst.y = mul src0.y, src1.y
+       */
+      emit_op2(c,
+              TGSI_OPCODE_MUL,
+              dst_mask(dst, BRW_WRITEMASK_Y),
+              src0,
+              src1);
+   }
+
+   if (dst.writemask & BRW_WRITEMASK_XZ) {
+      /* dst.z = mov src0.zzzz
+       */
+      emit_op1(c,
+             TGSI_OPCODE_MOV,
+             dst_mask(dst, BRW_WRITEMASK_Z),
+             src_scalar(src0, Z));
+
+      /* dst.x = imm1f(1.0)
+       */
+      emit_op1(c,
+             TGSI_OPCODE_MOV,
+             dst_saturate(dst_mask(dst, BRW_WRITEMASK_X), 0),
+             src_imm1f(c, 1.0));
+   }
+   if (dst.writemask & BRW_WRITEMASK_W) {
+      /* dst.w = mov src1.w
+       */
+      emit_op1(c,
+              TGSI_OPCODE_MOV,
+              dst_mask(dst, BRW_WRITEMASK_W),
+              src1);
+   }
+}
+
+
+static void precalc_lit( struct brw_wm_compile *c,
+                        struct brw_fp_dst dst,
+                        struct brw_fp_src src0 )
+{
+   if (dst.writemask & BRW_WRITEMASK_XW) {
+      /* dst.xw = imm(1.0f)
+       */
+      emit_op1(c,
+              TGSI_OPCODE_MOV,
+              dst_saturate(dst_mask(dst, BRW_WRITEMASK_XW), 0),
+              src_imm1f(c, 1.0f));
+   }
+
+   if (dst.writemask & BRW_WRITEMASK_YZ) {
+      emit_op1(c,
+              TGSI_OPCODE_LIT,
+              dst_mask(dst, BRW_WRITEMASK_YZ),
+              src0);
+   }
+}
+
+
+/**
+ * Some TEX instructions require extra code, cube map coordinate
+ * normalization, or coordinate scaling for RECT textures, etc.
+ * This function emits those extra instructions and the TEX
+ * instruction itself.
+ */
+static void precalc_tex( struct brw_wm_compile *c,
+                        struct brw_fp_dst dst,
+                        unsigned target,
+                        unsigned unit,
+                        struct brw_fp_src src0,
+                        struct brw_fp_src sampler )
+{
+   struct brw_fp_src coord = src_undef();
+   struct brw_fp_dst tmp = dst_undef();
+
+   assert(unit < BRW_MAX_TEX_UNIT);
+
+   /* Cubemap: find longest component of coord vector and normalize
+    * it.
+    */
+   if (target == TGSI_TEXTURE_CUBE) {
+      struct brw_fp_src tmpsrc;
+
+      tmp = get_temp(c);
+      tmpsrc = src_reg_from_dst(tmp);
+
+      /* tmp = abs(src0) */
+      emit_op1(c, 
+              TGSI_OPCODE_MOV,
+              tmp,
+              src_abs(src0));
+
+      /* tmp.X = MAX(tmp.X, tmp.Y) */
+      emit_op2(c, TGSI_OPCODE_MAX,
+              dst_mask(tmp, BRW_WRITEMASK_X),
+              src_scalar(tmpsrc, X),
+              src_scalar(tmpsrc, Y));
+
+      /* tmp.X = MAX(tmp.X, tmp.Z) */
+      emit_op2(c, TGSI_OPCODE_MAX,
+              dst_mask(tmp, BRW_WRITEMASK_X),
+              tmpsrc,
+              src_scalar(tmpsrc, Z));
+
+      /* tmp.X = 1 / tmp.X */
+      emit_op1(c, TGSI_OPCODE_RCP,
+             dst_mask(tmp, BRW_WRITEMASK_X),
+             tmpsrc);
+
+      /* tmp = src0 * tmp.xxxx */
+      emit_op2(c, TGSI_OPCODE_MUL,
+              tmp,
+              src0,
+              src_scalar(tmpsrc, X));
+
+      coord = tmpsrc;
+   }
+   else if (target == TGSI_TEXTURE_RECT ||
+           target == TGSI_TEXTURE_SHADOWRECT) {
+      /* XXX: need a mechanism for internally generated constants.
+       */
+      coord = src0;
+   }
+   else {
+      coord = src0;
+   }
+
+   /* Need to emit YUV texture conversions by hand.  Probably need to
+    * do this here - the alternative is in brw_wm_emit.c, but the
+    * conversion requires allocating a temporary variable which we
+    * don't have the facility to do that late in the compilation.
+    */
+   if (c->key.yuvtex_mask & (1 << unit)) {
+      /* convert ycbcr to RGBA */
+      GLboolean  swap_uv = c->key.yuvtex_swap_mask & (1<<unit);
+      struct brw_fp_dst tmp = get_temp(c);
+      struct brw_fp_src tmpsrc = src_reg_from_dst(tmp);
+      struct brw_fp_src C0 = src_imm4f( c,  -.5, -.0625, -.5, 1.164 );
+      struct brw_fp_src C1 = src_imm4f( c, 1.596, -0.813, 2.018, -.391 );
+     
+      /* tmp     = TEX ...
+       */
+      emit_tex_op(c, 
+                  TGSI_OPCODE_TEX,
+                  dst_saturate(tmp, dst.saturate),
+                  unit,
+                  target,
+                  sampler.index,
+                  coord,
+                  src_undef(),
+                  src_undef());
+
+      /* tmp.xyz =  ADD TMP, C0
+       */
+      emit_op2(c, TGSI_OPCODE_ADD,
+              dst_mask(tmp, BRW_WRITEMASK_XYZ),
+              tmpsrc,
+              C0);
+
+      /* YUV.y   = MUL YUV.y, C0.w
+       */
+      emit_op2(c, TGSI_OPCODE_MUL,
+              dst_mask(tmp, BRW_WRITEMASK_Y),
+              tmpsrc,
+              src_scalar(C0, W));
+
+      /* 
+       * if (UV swaped)
+       *     RGB.xyz = MAD YUV.zzx, C1, YUV.y
+       * else
+       *     RGB.xyz = MAD YUV.xxz, C1, YUV.y
+       */
+
+      emit_op3(c, TGSI_OPCODE_MAD,
+              dst_mask(dst, BRW_WRITEMASK_XYZ),
+              ( swap_uv ? 
+                src_swizzle(tmpsrc, Z,Z,X,X) : 
+                src_swizzle(tmpsrc, X,X,Z,Z)),
+              C1,
+              src_scalar(tmpsrc, Y));
+
+      /*  RGB.y   = MAD YUV.z, C1.w, RGB.y
+       */
+      emit_op3(c,
+              TGSI_OPCODE_MAD,
+              dst_mask(dst, BRW_WRITEMASK_Y),
+              src_scalar(tmpsrc, Z),
+              src_scalar(C1, W),
+              src_scalar(src_reg_from_dst(dst), Y));
+
+      release_temp(c, tmp);
+   }
+   else {
+      /* ordinary RGBA tex instruction */
+      emit_tex_op(c, 
+                  TGSI_OPCODE_TEX,
+                  dst,
+                  unit,
+                  target,
+                  sampler.index,
+                  coord,
+                  src_undef(),
+                  src_undef());
+   }
+
+   /* XXX: add GL_EXT_texture_swizzle support to gallium -- by
+    * generating shader varients in mesa state tracker.
+    */
+
+   /* Release this temp if we ended up allocating it:
+    */
+   if (!dst_is_undef(tmp))
+      release_temp(c, tmp);
+}
+
+
+/**
+ * Check if the given TXP instruction really needs the divide-by-W step.
+ */
+static GLboolean projtex( struct brw_wm_compile *c,
+                         unsigned target, 
+                         struct brw_fp_src src )
+{
+   /* Only try to detect the simplest cases.  Could detect (later)
+    * cases where we are trying to emit code like RCP {1.0}, MUL x,
+    * {1.0}, and so on.
+    *
+    * More complex cases than this typically only arise from
+    * user-provided fragment programs anyway:
+    */
+   if (target == TGSI_TEXTURE_CUBE)
+      return GL_FALSE;  /* ut2004 gun rendering !?! */
+   
+   if (src.file == TGSI_FILE_INPUT && 
+       BRW_GET_SWZ(src.swizzle, W) == W &&
+       c->fp->info.input_interpolate[src.index] != TGSI_INTERPOLATE_PERSPECTIVE)
+      return GL_FALSE;
+
+   return GL_TRUE;
+}
+
+
+/**
+ * Emit code for TXP.
+ */
+static void precalc_txp( struct brw_wm_compile *c,
+                        struct brw_fp_dst dst,
+                        unsigned target,
+                        unsigned unit,
+                        struct brw_fp_src src0,
+                         struct brw_fp_src sampler )
+{
+   if (projtex(c, target, src0)) {
+      struct brw_fp_dst tmp = get_temp(c);
+
+      /* tmp0.w = RCP inst.arg[0][3]
+       */
+      emit_op1(c,
+             TGSI_OPCODE_RCP,
+             dst_mask(tmp, BRW_WRITEMASK_W),
+             src_scalar(src0, W));
+
+      /* tmp0.xyz =  MUL inst.arg[0], tmp0.wwww
+       */
+      emit_op2(c,
+              TGSI_OPCODE_MUL,
+              dst_mask(tmp, BRW_WRITEMASK_XYZ),
+              src0,
+              src_scalar(src_reg_from_dst(tmp), W));
+
+      /* dst = TEX tmp0
+       */
+      precalc_tex(c, 
+                 dst,
+                 target,
+                 unit,
+                 src_reg_from_dst(tmp),
+                  sampler );
+
+      release_temp(c, tmp);
+   }
+   else
+   {
+      /* dst = TEX src0
+       */
+      precalc_tex(c, dst, target, unit, src0, sampler);
+   }
+}
+
+
+/* XXX: note this returns a src_reg.
+ */
+static struct brw_fp_src
+find_output_by_semantic( struct brw_wm_compile *c,
+                        unsigned semantic,
+                        unsigned index )
+{
+   const struct tgsi_shader_info *info = &c->fp->info;
+   unsigned i;
+
+   for (i = 0; i < info->num_outputs; i++)
+      if (info->output_semantic_name[i] == semantic &&
+         info->output_semantic_index[i] == index)
+        return src_reg( TGSI_FILE_OUTPUT, i );
+
+   /* If not found, return some arbitrary immediate value:
+    *
+    * XXX: this is a good idea but immediates are up generating extra
+    * curbe entries atm, as they would have in the original driver.
+    */
+   return src_reg( TGSI_FILE_OUTPUT, 0 ); /* src_imm1f(c, 1.0); */
+}
+
+
+static void emit_fb_write( struct brw_wm_compile *c )
+{
+   struct brw_fp_src payload_r0_depth = src_reg(BRW_FILE_PAYLOAD, PAYLOAD_DEPTH);
+   struct brw_fp_src outdepth = find_output_by_semantic(c, TGSI_SEMANTIC_POSITION, 0);
+   GLuint i;
+
+
+   outdepth = src_scalar(outdepth, Z);
+
+   for (i = 0 ; i < c->key.nr_cbufs; i++) {
+      struct brw_fp_src outcolor;
+      
+      outcolor = find_output_by_semantic(c, TGSI_SEMANTIC_COLOR, i);
+
+      /* Use emit_tex_op so that we can specify the inst->target
+       * field, which is abused to contain the FB write target and the
+       * EOT marker
+       */
+      emit_tex_op(c, WM_FB_WRITE,
+                 dst_undef(),
+                 (i == c->key.nr_cbufs - 1), /* EOT */
+                 i,
+                  0,            /* no sampler */
+                 outcolor,
+                 payload_r0_depth,
+                 outdepth);
+   }
+}
+
+
+static struct brw_fp_dst translate_dst( struct brw_wm_compile *c,
+                                       const struct tgsi_full_dst_register *dst,
+                                       unsigned saturate )
+{
+   struct brw_fp_dst out;
+
+   out.file = dst->Register.File;
+   out.index = dst->Register.Index;
+   out.writemask = dst->Register.WriteMask;
+   out.indirect = dst->Register.Indirect;
+   out.saturate = (saturate == TGSI_SAT_ZERO_ONE);
+   
+   if (out.indirect) {
+      assert(dst->Indirect.File == TGSI_FILE_ADDRESS);
+      assert(dst->Indirect.Index == 0);
+   }
+   
+   return out;
+}
+
+
+static struct brw_fp_src translate_src( struct brw_wm_compile *c,
+                                       const struct tgsi_full_src_register *src )
+{
+   struct brw_fp_src out;
+
+   out.file = src->Register.File;
+   out.index = src->Register.Index;
+   out.indirect = src->Register.Indirect;
+
+   out.swizzle = ((src->Register.SwizzleX << 0) |
+                 (src->Register.SwizzleY << 2) |
+                 (src->Register.SwizzleZ << 4) |
+                 (src->Register.SwizzleW << 6));
+   
+   switch (tgsi_util_get_full_src_register_sign_mode( src, 0 )) {
+   case TGSI_UTIL_SIGN_CLEAR:
+      out.abs = 1;
+      out.negate = 0;
+      break;
+
+   case TGSI_UTIL_SIGN_SET:
+      out.abs = 1;
+      out.negate = 1;
+      break;
+
+   case TGSI_UTIL_SIGN_TOGGLE:
+      out.abs = 0;
+      out.negate = 1;
+      break;
+
+   case TGSI_UTIL_SIGN_KEEP:
+   default:
+      out.abs = 0;
+      out.negate = 0;
+      break;
+   }
+
+   if (out.indirect) {
+      assert(src->Indirect.File == TGSI_FILE_ADDRESS);
+      assert(src->Indirect.Index == 0);
+   }
+   
+   return out;
+}
+
+
+
+static void emit_insn( struct brw_wm_compile *c,
+                      const struct tgsi_full_instruction *inst )
+{
+   unsigned opcode = inst->Instruction.Opcode;
+   struct brw_fp_dst dst;
+   struct brw_fp_src src[3];
+   int i;
+
+   dst = translate_dst( c, &inst->Dst[0],
+                       inst->Instruction.Saturate );
+
+   for (i = 0; i < inst->Instruction.NumSrcRegs; i++)
+      src[i] = translate_src( c, &inst->Src[i] );
+   
+   switch (opcode) {
+   case TGSI_OPCODE_ABS:
+      emit_op1(c, TGSI_OPCODE_MOV,
+              dst, 
+              src_abs(src[0]));
+      break;
+
+   case TGSI_OPCODE_SUB: 
+      emit_op2(c, TGSI_OPCODE_ADD,
+              dst,
+              src[0],
+              src_negate(src[1]));
+      break;
+
+   case TGSI_OPCODE_SCS: 
+      emit_op1(c, TGSI_OPCODE_SCS,
+              dst_mask(dst, BRW_WRITEMASK_XY),
+              src[0]);
+      break;
+        
+   case TGSI_OPCODE_DST:
+      precalc_dst(c, dst, src[0], src[1]);
+      break;
+
+   case TGSI_OPCODE_LIT:
+      precalc_lit(c, dst, src[0]);
+      break;
+
+   case TGSI_OPCODE_TEX:
+      precalc_tex(c, dst,
+                 inst->Texture.Texture,
+                 src[1].index, /* use sampler unit for tex idx */
+                 src[0],       /* coord */
+                  src[1]);      /* sampler */
+      break;
+
+   case TGSI_OPCODE_TXP:
+      precalc_txp(c, dst,
+                 inst->Texture.Texture,
+                 src[1].index, /* use sampler unit for tex idx */
+                 src[0],       /* coord */
+                  src[1]);      /* sampler */
+      break;
+
+   case TGSI_OPCODE_TXB:
+      /* XXX: TXB not done
+       */
+      precalc_tex(c, dst,
+                 inst->Texture.Texture,
+                 src[1].index, /* use sampler unit for tex idx*/
+                 src[0],
+                  src[1]);
+      break;
+
+   case TGSI_OPCODE_XPD: 
+      emit_op2(c, TGSI_OPCODE_XPD,
+              dst_mask(dst, BRW_WRITEMASK_XYZ),
+              src[0], 
+              src[1]);
+      break;
+
+   case TGSI_OPCODE_KIL: 
+      emit_op1(c, TGSI_OPCODE_KIL,
+              dst_mask(dst_undef(), 0),
+              src[0]);
+      break;
+
+   case TGSI_OPCODE_END:
+      emit_fb_write(c);
+      break;
+   default:
+      if (!c->key.has_flow_control &&
+         brw_wm_is_scalar_result(opcode))
+        emit_scalar_insn(c, opcode, dst, src[0], src[1], src[2]);
+      else
+        emit_op3(c, opcode, dst, src[0], src[1], src[2]);
+      break;
+   }
+}
+
+/**
+ * Initial pass for fragment program code generation.
+ * This function is used by both the GLSL and non-GLSL paths.
+ */
+int brw_wm_pass_fp( struct brw_wm_compile *c )
+{
+   struct brw_fragment_shader *fs = c->fp;
+   struct tgsi_parse_context parse;
+   struct tgsi_full_instruction *inst;
+   struct tgsi_full_declaration *decl;
+   const float *imm;
+   GLuint size;
+   GLuint i;
+
+   if (BRW_DEBUG & DEBUG_WM) {
+      debug_printf("pre-fp:\n");
+      tgsi_dump(fs->tokens, 0); 
+   }
+
+   c->fp_pixel_xy = src_undef();
+   c->fp_delta_xy = src_undef();
+   c->fp_pixel_w = src_undef();
+   c->nr_fp_insns = 0;
+   c->nr_immediates = 0;
+
+
+   /* Loop over all instructions doing assorted simplifications and
+    * transformations.
+    */
+   tgsi_parse_init( &parse, fs->tokens );
+   while( !tgsi_parse_end_of_tokens( &parse ) ) {
+      tgsi_parse_token( &parse );
+
+      switch( parse.FullToken.Token.Type ) {
+      case TGSI_TOKEN_TYPE_DECLARATION:
+        /* Turn intput declarations into special WM_* instructions.
+         *
+         * XXX: For non-branching shaders, consider deferring variable
+         * initialization as late as possible to minimize register
+         * usage.  This is how the original BRW driver worked.
+         *
+         * In a branching shader, must preamble instructions at decl
+         * time, as instruction order in the shader does not
+         * correspond to the order instructions are executed in the
+         * wild.
+         *
+         * This is where special instructions such as WM_CINTERP,
+         * WM_LINTERP, WM_PINTERP and WM_WPOSXY are emitted to
+         * compute shader inputs from the payload registers and pixel
+         * position.
+         */
+         decl = &parse.FullToken.FullDeclaration;
+         if( decl->Declaration.File == TGSI_FILE_INPUT ) {
+            unsigned first, last, mask;
+            unsigned attrib;
+
+            first = decl->Range.First;
+            last = decl->Range.Last;
+            mask = decl->Declaration.UsageMask;
+
+            for (attrib = first; attrib <= last; attrib++) {
+              emit_interp(c, 
+                          attrib, 
+                          decl->Semantic.Name,
+                          decl->Declaration.Interpolate );
+            }
+         }
+        
+         break;
+
+      case TGSI_TOKEN_TYPE_IMMEDIATE:
+        /* Unlike VS programs we can probably manage fine encoding
+         * immediate values directly into the emitted EU
+         * instructions, as we probably only need to reference one
+         * float value per instruction.  Just save the data for now
+         * and use directly later.
+         */
+        i = c->nr_immediates++;
+        imm = &parse.FullToken.FullImmediate.u[i].Float;
+        size = parse.FullToken.FullImmediate.Immediate.NrTokens - 1;
+
+        if (c->nr_immediates >= BRW_WM_MAX_CONST)
+           return PIPE_ERROR_OUT_OF_MEMORY;
+
+        for (i = 0; i < size; i++)
+           c->immediate[c->nr_immediates].v[i] = imm[i];
+
+        for (; i < 4; i++)
+           c->immediate[c->nr_immediates].v[i] = 0.0;
+
+        c->immediate[c->nr_immediates].nr = size;
+        c->nr_immediates++;
+        break;
+
+      case TGSI_TOKEN_TYPE_INSTRUCTION:
+         inst = &parse.FullToken.FullInstruction;
+        emit_insn(c, inst);
+        break;
+      }
+   }
+
+   if (BRW_DEBUG & DEBUG_WM) {
+      brw_wm_print_fp_program( c, "pass_fp" );
+      debug_printf("\n");
+   }
+
+   return c->error;
+}
+
diff --git a/src/gallium/drivers/i965/brw_wm_glsl.c b/src/gallium/drivers/i965/brw_wm_glsl.c
new file mode 100644 (file)
index 0000000..3b3afc3
--- /dev/null
@@ -0,0 +1,2032 @@
+#include "util/u_math.h"
+
+
+#include "brw_context.h"
+#include "brw_eu.h"
+#include "brw_wm.h"
+
+
+static struct brw_reg get_dst_reg(struct brw_wm_compile *c,
+                                  const struct brw_fp_instruction *inst,
+                                  GLuint component);
+
+
+static void
+reclaim_temps(struct brw_wm_compile *c);
+
+
+/** Mark GRF register as used. */
+static void
+prealloc_grf(struct brw_wm_compile *c, int r)
+{
+   c->used_grf[r] = GL_TRUE;
+}
+
+
+/** Mark given GRF register as not in use. */
+static void
+release_grf(struct brw_wm_compile *c, int r)
+{
+   /*assert(c->used_grf[r]);*/
+   c->used_grf[r] = GL_FALSE;
+   c->first_free_grf = MIN2(c->first_free_grf, r);
+}
+
+
+/** Return index of a free GRF, mark it as used. */
+static int
+alloc_grf(struct brw_wm_compile *c)
+{
+   GLuint r;
+   for (r = c->first_free_grf; r < BRW_WM_MAX_GRF; r++) {
+      if (!c->used_grf[r]) {
+         c->used_grf[r] = GL_TRUE;
+         c->first_free_grf = r + 1;  /* a guess */
+         return r;
+      }
+   }
+
+   /* no free temps, try to reclaim some */
+   reclaim_temps(c);
+   c->first_free_grf = 0;
+
+   /* try alloc again */
+   for (r = c->first_free_grf; r < BRW_WM_MAX_GRF; r++) {
+      if (!c->used_grf[r]) {
+         c->used_grf[r] = GL_TRUE;
+         c->first_free_grf = r + 1;  /* a guess */
+         return r;
+      }
+   }
+
+   for (r = 0; r < BRW_WM_MAX_GRF; r++) {
+      assert(c->used_grf[r]);
+   }
+
+   /* really, no free GRF regs found */
+   if (!c->out_of_regs) {
+      /* print warning once per compilation */
+      debug_printf("%s: ran out of registers for fragment program", __FUNCTION__);
+      c->out_of_regs = GL_TRUE;
+   }
+
+   return -1;
+}
+
+
+/** Return number of GRF registers used */
+static int
+num_grf_used(const struct brw_wm_compile *c)
+{
+   int r;
+   for (r = BRW_WM_MAX_GRF - 1; r >= 0; r--)
+      if (c->used_grf[r])
+         return r + 1;
+   return 0;
+}
+
+
+
+/**
+ * Record the mapping of a Mesa register to a hardware register.
+ */
+static void set_reg(struct brw_wm_compile *c, int file, int index, 
+       int component, struct brw_reg reg)
+{
+    c->wm_regs[file][index][component].reg = reg;
+    c->wm_regs[file][index][component].inited = GL_TRUE;
+}
+
+static struct brw_reg alloc_tmp(struct brw_wm_compile *c)
+{
+    struct brw_reg reg;
+
+    /* if we need to allocate another temp, grow the tmp_regs[] array */
+    if (c->tmp_index == c->tmp_max) {
+       int r = alloc_grf(c);
+       if (r < 0) {
+          /*printf("Out of temps in %s\n", __FUNCTION__);*/
+          r = 50; /* XXX random register! */
+       }
+       c->tmp_regs[ c->tmp_max++ ] = r;
+    }
+
+    /* form the GRF register */
+    reg = brw_vec8_grf(c->tmp_regs[ c->tmp_index++ ], 0);
+    /*printf("alloc_temp %d\n", reg.nr);*/
+    assert(reg.nr < BRW_WM_MAX_GRF);
+    return reg;
+
+}
+
+/**
+ * Save current temp register info.
+ * There must be a matching call to release_tmps().
+ */
+static int mark_tmps(struct brw_wm_compile *c)
+{
+    return c->tmp_index;
+}
+
+static struct brw_reg lookup_tmp( struct brw_wm_compile *c, int index )
+{
+    return brw_vec8_grf( c->tmp_regs[ index ], 0 );
+}
+
+static void release_tmps(struct brw_wm_compile *c, int mark)
+{
+    c->tmp_index = mark;
+}
+
+/**
+ * Convert Mesa src register to brw register.
+ *
+ * Since we're running in SOA mode each Mesa register corresponds to four
+ * hardware registers.  We allocate the hardware registers as needed here.
+ *
+ * \param file  register file, one of PROGRAM_x
+ * \param index  register number
+ * \param component  src component (X=0, Y=1, Z=2, W=3)
+ * \param nr  not used?!?
+ * \param neg  negate value?
+ * \param abs  take absolute value?
+ */
+static struct brw_reg 
+get_reg(struct brw_wm_compile *c, int file, int index, int component,
+        int nr, GLuint neg, GLuint abs)
+{
+    struct brw_reg reg;
+    switch (file) {
+       case TGSI_FILE_NULL:
+           return brw_null_reg();      
+
+       case TGSI_FILE_CONSTANT:
+       case TGSI_FILE_TEMPORARY:
+       case TGSI_FILE_INPUT:
+       case TGSI_FILE_OUTPUT:
+       case BRW_FILE_PAYLOAD:
+           break;
+
+       default:
+          debug_printf("%s: Unexpected file type\n", __FUNCTION__);
+          return brw_null_reg();
+    }
+
+    assert(index < 256);
+    assert(component < 4);
+
+    /* see if we've already allocated a HW register for this Mesa register */
+    if (c->wm_regs[file][index][component].inited) {
+       /* yes, re-use */
+       reg = c->wm_regs[file][index][component].reg;
+    }
+    else {
+       /* no, allocate new register */
+       int grf = alloc_grf(c);
+       /*printf("alloc grf %d for reg %d:%d.%d\n", grf, file, index, component);*/
+       if (grf < 0) {
+          /* totally out of temps */
+          grf = 51; /* XXX random register! */
+       }
+
+       reg = brw_vec8_grf(grf, 0);
+       /*printf("Alloc new grf %d for %d.%d\n", reg.nr, index, component);*/
+
+       set_reg(c, file, index, component, reg);
+    }
+
+    if (neg & (1 << component)) {
+       reg = negate(reg);
+    }
+    if (abs)
+       reg = brw_abs(reg);
+    return reg;
+}
+
+
+
+
+/**
+ * Find first/last instruction that references each temporary register.
+ */
+GLboolean
+_mesa_find_temp_intervals(const struct prog_instruction *instructions,
+                          GLuint numInstructions,
+                          GLint intBegin[MAX_PROGRAM_TEMPS],
+                          GLint intEnd[MAX_PROGRAM_TEMPS])
+{
+   struct loop_info
+   {
+      GLuint Start, End;  /**< Start, end instructions of loop */
+   };
+   struct loop_info loopStack[MAX_LOOP_NESTING];
+   GLuint loopStackDepth = 0;
+   GLuint i;
+
+   for (i = 0; i < MAX_PROGRAM_TEMPS; i++){
+      intBegin[i] = intEnd[i] = -1;
+   }
+
+   /* Scan instructions looking for temporary registers */
+   for (i = 0; i < numInstructions; i++) {
+      const struct prog_instruction *inst = instructions + i;
+      if (inst->Opcode == OPCODE_BGNLOOP) {
+         loopStack[loopStackDepth].Start = i;
+         loopStack[loopStackDepth].End = inst->BranchTarget;
+         loopStackDepth++;
+      }
+      else if (inst->Opcode == OPCODE_ENDLOOP) {
+         loopStackDepth--;
+      }
+      else if (inst->Opcode == OPCODE_CAL) {
+         return GL_FALSE;
+      }
+      else {
+         const GLuint numSrc = 3;
+         GLuint j;
+         for (j = 0; j < numSrc; j++) {
+            if (inst->SrcReg[j].File == PROGRAM_TEMPORARY) {
+               const GLuint index = inst->SrcReg[j].Index;
+               if (inst->SrcReg[j].RelAddr)
+                  return GL_FALSE;
+               update_interval(intBegin, intEnd, index, i);
+               if (loopStackDepth > 0) {
+                  /* extend temp register's interval to end of loop */
+                  GLuint loopEnd = loopStack[loopStackDepth - 1].End;
+                  update_interval(intBegin, intEnd, index, loopEnd);
+               }
+            }
+         }
+         if (inst->DstReg.File == PROGRAM_TEMPORARY) {
+            const GLuint index = inst->DstReg.Index;
+            if (inst->DstReg.RelAddr)
+               return GL_FALSE;
+            update_interval(intBegin, intEnd, index, i);
+            if (loopStackDepth > 0) {
+               /* extend temp register's interval to end of loop */
+               GLuint loopEnd = loopStack[loopStackDepth - 1].End;
+               update_interval(intBegin, intEnd, index, loopEnd);
+            }
+         }
+      }
+   }
+
+   return GL_TRUE;
+}
+
+
+/**
+ * This is called if we run out of GRF registers.  Examine the live intervals
+ * of temp regs in the program and free those which won't be used again.
+ */
+static void
+reclaim_temps(struct brw_wm_compile *c)
+{
+   GLint intBegin[BRW_WM_MAX_TEMPS];
+   GLint intEnd[BRW_WM_MAX_TEMPS];
+   int index;
+
+   /*printf("Reclaim temps:\n");*/
+
+   _mesa_find_temp_intervals(c->fp_instructions, c->nr_fp_insns,
+                             intBegin, intEnd);
+
+   for (index = 0; index < BRW_WM_MAX_TEMPS; index++) {
+      if (intEnd[index] != -1 && intEnd[index] < c->cur_inst) {
+         /* program temp[i] can be freed */
+         int component;
+         /*printf("  temp[%d] is dead\n", index);*/
+         for (component = 0; component < 4; component++) {
+            if (c->wm_regs[TGSI_FILE_TEMPORARY][index][component].inited) {
+               int r = c->wm_regs[TGSI_FILE_TEMPORARY][index][component].reg.nr;
+               release_grf(c, r);
+               /*
+               printf("  Reclaim temp %d, reg %d at inst %d\n",
+                      index, r, c->cur_inst);
+               */
+               c->wm_regs[TGSI_FILE_TEMPORARY][index][component].inited = GL_FALSE;
+            }
+         }
+      }
+   }
+}
+
+
+
+
+/**
+ * Preallocate registers.  This sets up the Mesa to hardware register
+ * mapping for certain registers, such as constants (uniforms/state vars)
+ * and shader inputs.
+ */
+static void prealloc_reg(struct brw_wm_compile *c)
+{
+    int i, j;
+    struct brw_reg reg;
+    int urb_read_length = 0;
+    GLuint inputs = FRAG_BIT_WPOS | c->fp_interp_emitted;
+    GLuint reg_index = 0;
+
+    memset(c->used_grf, GL_FALSE, sizeof(c->used_grf));
+    c->first_free_grf = 0;
+
+    for (i = 0; i < 4; i++) {
+        if (i < c->key.nr_depth_regs) 
+            reg = brw_vec8_grf(i * 2, 0);
+        else
+            reg = brw_vec8_grf(0, 0);
+       set_reg(c, TGSI_FILE_PAYLOAD, PAYLOAD_DEPTH, i, reg);
+    }
+    reg_index += 2 * c->key.nr_depth_regs;
+
+    /* constants */
+    {
+        const GLuint nr_params = c->fp->program.Base.Parameters->NumParameters;
+        const GLuint nr_temps = c->fp->program.Base.NumTemporaries;
+
+        /* use a real constant buffer, or just use a section of the GRF? */
+        /* XXX this heuristic may need adjustment... */
+        if ((nr_params + nr_temps) * 4 + reg_index > 80)
+           c->fp->use_const_buffer = GL_TRUE;
+        else
+           c->fp->use_const_buffer = GL_FALSE;
+        /*printf("WM use_const_buffer = %d\n", c->fp->use_const_buffer);*/
+
+        if (c->fp->use_const_buffer) {
+           /* We'll use a real constant buffer and fetch constants from
+            * it with a dataport read message.
+            */
+
+           /* number of float constants in CURBE */
+           c->prog_data.nr_params = 0;
+        }
+        else {
+           const struct gl_program_parameter_list *plist = 
+              c->fp->program.Base.Parameters;
+           int index = 0;
+
+           /* number of float constants in CURBE */
+           c->prog_data.nr_params = 4 * nr_params;
+
+           /* loop over program constants (float[4]) */
+           for (i = 0; i < nr_params; i++) {
+              /* loop over XYZW channels */
+              for (j = 0; j < 4; j++, index++) {
+                 reg = brw_vec1_grf(reg_index + index / 8, index % 8);
+                 /* Save pointer to parameter/constant value.
+                  * Constants will be copied in prepare_constant_buffer()
+                  */
+                 c->prog_data.param[index] = &plist->ParameterValues[i][j];
+                 set_reg(c, TGSI_FILE_STATE_VAR, i, j, reg);
+              }
+           }
+           /* number of constant regs used (each reg is float[8]) */
+           c->nr_creg = 2 * ((4 * nr_params + 15) / 16);
+           reg_index += c->nr_creg;
+        }
+    }
+
+    /* fragment shader inputs */
+    for (i = 0; i < VERT_RESULT_MAX; i++) {
+       int fp_input;
+
+       if (i >= VERT_RESULT_VAR0)
+         fp_input = i - VERT_RESULT_VAR0 + FRAG_ATTRIB_VAR0;
+       else if (i <= VERT_RESULT_TEX7)
+         fp_input = i;
+       else
+         fp_input = -1;
+
+       if (fp_input >= 0 && inputs & (1 << fp_input)) {
+         urb_read_length = reg_index;
+         reg = brw_vec8_grf(reg_index, 0);
+         for (j = 0; j < 4; j++)
+            set_reg(c, TGSI_FILE_PAYLOAD, fp_input, j, reg);
+       }
+       if (c->key.nr_vp_outputs > i) {
+         reg_index += 2;
+       }
+    }
+
+    c->prog_data.first_curbe_grf = c->key.nr_depth_regs * 2;
+    c->prog_data.urb_read_length = urb_read_length;
+    c->prog_data.curb_read_length = c->nr_creg;
+    c->emit_mask_reg = brw_uw1_reg(BRW_GENERAL_REGISTER_FILE, reg_index, 0);
+    reg_index++;
+    c->stack =  brw_uw16_reg(BRW_GENERAL_REGISTER_FILE, reg_index, 0);
+    reg_index += 2;
+
+    /* mark GRF regs [0..reg_index-1] as in-use */
+    for (i = 0; i < reg_index; i++)
+       prealloc_grf(c, i);
+
+    /* Don't use GRF 126, 127.  Using them seems to lead to GPU lock-ups */
+    prealloc_grf(c, 126);
+    prealloc_grf(c, 127);
+
+    for (i = 0; i < c->nr_fp_insns; i++) {
+       const struct brw_fp_instruction *inst = &c->fp_instructions[i];
+       struct brw_reg dst[4];
+
+       switch (inst->Opcode) {
+       case OPCODE_TEX:
+       case OPCODE_TXB:
+           /* Allocate the channels of texture results contiguously,
+            * since they are written out that way by the sampler unit.
+            */
+           for (j = 0; j < 4; j++) {
+               dst[j] = get_dst_reg(c, inst, j);
+               if (j != 0)
+                   assert(dst[j].nr == dst[j - 1].nr + 1);
+           }
+           break;
+       default:
+           break;
+       }
+    }
+
+    /* An instruction may reference up to three constants.
+     * They'll be found in these registers.
+     * XXX alloc these on demand!
+     */
+    if (c->fp->use_const_buffer) {
+       for (i = 0; i < 3; i++) {
+          c->current_const[i].index = -1;
+          c->current_const[i].reg = brw_vec8_grf(alloc_grf(c), 0);
+       }
+    }
+#if 0
+    printf("USE CONST BUFFER? %d\n", c->fp->use_const_buffer);
+    printf("AFTER PRE_ALLOC, reg_index = %d\n", reg_index);
+#endif
+}
+
+
+/**
+ * Check if any of the instruction's src registers are constants, uniforms,
+ * or statevars.  If so, fetch any constants that we don't already have in
+ * the three GRF slots.
+ */
+static void fetch_constants(struct brw_wm_compile *c,
+                            const struct brw_fp_instruction *inst)
+{
+   struct brw_compile *p = &c->func;
+   GLuint i;
+
+   /* loop over instruction src regs */
+   for (i = 0; i < 3; i++) {
+      const struct prog_src_register *src = &inst->SrcReg[i];
+      if (src->File == TGSI_FILE_IMMEDIATE ||
+          src->File == TGSI_FILE_CONSTANT) {
+        c->current_const[i].index = src->Index;
+
+#if 0
+        printf("  fetch const[%d] for arg %d into reg %d\n",
+               src->Index, i, c->current_const[i].reg.nr);
+#endif
+
+        /* need to fetch the constant now */
+        brw_dp_READ_4(p,
+                      c->current_const[i].reg,  /* writeback dest */
+                      src->RelAddr,             /* relative indexing? */
+                      16 * src->Index,          /* byte offset */
+                      SURF_INDEX_FRAG_CONST_BUFFER/* binding table index */
+                      );
+      }
+   }
+}
+
+
+/**
+ * Convert Mesa dst register to brw register.
+ */
+static struct brw_reg get_dst_reg(struct brw_wm_compile *c, 
+                                  const struct brw_fp_instruction *inst,
+                                  GLuint component)
+{
+    const int nr = 1;
+    return get_reg(c, inst->DstReg.File, inst->DstReg.Index, component, nr,
+           0, 0);
+}
+
+
+static struct brw_reg
+get_src_reg_const(struct brw_wm_compile *c,
+                  const struct brw_fp_instruction *inst,
+                  GLuint srcRegIndex, GLuint component)
+{
+   /* We should have already fetched the constant from the constant
+    * buffer in fetch_constants().  Now we just have to return a
+    * register description that extracts the needed component and
+    * smears it across all eight vector components.
+    */
+   const struct prog_src_register *src = &inst->SrcReg[srcRegIndex];
+   struct brw_reg const_reg;
+
+   assert(component < 4);
+   assert(srcRegIndex < 3);
+   assert(c->current_const[srcRegIndex].index != -1);
+   const_reg = c->current_const[srcRegIndex].reg;
+
+   /* extract desired float from the const_reg, and smear */
+   const_reg = stride(const_reg, 0, 1, 0);
+   const_reg.subnr = component * 4;
+
+   if (src->Negate)
+      const_reg = negate(const_reg);
+   if (src->Abs)
+      const_reg = brw_abs(const_reg);
+
+#if 0
+   printf("  form const[%d].%d for arg %d, reg %d\n",
+          c->current_const[srcRegIndex].index,
+          component,
+          srcRegIndex,
+          const_reg.nr);
+#endif
+
+   return const_reg;
+}
+
+
+/**
+ * Convert Mesa src register to brw register.
+ */
+static struct brw_reg get_src_reg(struct brw_wm_compile *c, 
+                                  const struct brw_fp_instruction *inst,
+                                  GLuint srcRegIndex, GLuint channel)
+{
+    const struct prog_src_register *src = &inst->SrcReg[srcRegIndex];
+    const GLuint nr = 1;
+    const GLuint component = BRW_GET_SWZ(src->Swizzle, channel);
+
+    /* Extended swizzle terms */
+    if (component == SWIZZLE_ZERO) {
+       return brw_imm_f(0.0F);
+    }
+    else if (component == SWIZZLE_ONE) {
+       return brw_imm_f(1.0F);
+    }
+
+    if (c->fp->use_const_buffer &&
+        (src->File == TGSI_FILE_STATE_VAR ||
+         src->File == TGSI_FILE_CONSTANT ||
+         src->File == TGSI_FILE_UNIFORM)) {
+       return get_src_reg_const(c, inst, srcRegIndex, component);
+    }
+    else {
+       /* other type of source register */
+       return get_reg(c, src->File, src->Index, component, nr, 
+                      src->Negate, src->Abs);
+    }
+}
+
+
+/**
+ * Same as \sa get_src_reg() but if the register is a immediate, emit
+ * a brw_reg encoding the immediate.
+ * Note that a brw instruction only allows one src operand to be a immediate.
+ * For instructions with more than one operand, only the second can be a
+ * immediate.  This means that we treat some immediates as constants
+ * (which why TGSI_FILE_IMMEDIATE is checked in fetch_constants()).
+ * 
+ */
+static struct brw_reg get_src_reg_imm(struct brw_wm_compile *c, 
+                                      const struct brw_fp_instruction *inst,
+                                      GLuint srcRegIndex, GLuint channel)
+{
+    const struct prog_src_register *src = &inst->SrcReg[srcRegIndex];
+    if (src->File == TGSI_FILE_IMMEDIATE) {
+       /* an immediate */
+       const int component = BRW_GET_SWZ(src->Swizzle, channel);
+       const GLfloat *param =
+          c->fp->program.Base.Parameters->ParameterValues[src->Index];
+       GLfloat value = param[component];
+       if (src->Negate)
+          value = -value;
+       if (src->Abs)
+          value = FABSF(value);
+#if 0
+       printf("  form immed value %f for chan %d\n", value, channel);
+#endif
+       return brw_imm_f(value);
+    }
+    else {
+       return get_src_reg(c, inst, srcRegIndex, channel);
+    }
+}
+
+
+/**
+ * Subroutines are minimal support for resusable instruction sequences.
+ * They are implemented as simply as possible to minimise overhead: there
+ * is no explicit support for communication between the caller and callee
+ * other than saving the return address in a temporary register, nor is
+ * there any automatic local storage.  This implies that great care is
+ * required before attempting reentrancy or any kind of nested
+ * subroutine invocations.
+ */
+static void invoke_subroutine( struct brw_wm_compile *c,
+                              enum _subroutine subroutine,
+                              void (*emit)( struct brw_wm_compile * ) )
+{
+    struct brw_compile *p = &c->func;
+
+    assert( subroutine < BRW_WM_MAX_SUBROUTINE );
+    
+    if( c->subroutines[ subroutine ] ) {
+       /* subroutine previously emitted: reuse existing instructions */
+
+       int mark = mark_tmps( c );
+       struct brw_reg return_address = retype( alloc_tmp( c ),
+                                               BRW_REGISTER_TYPE_UD );
+       int here = p->nr_insn;
+       
+       brw_push_insn_state(p);
+       brw_set_mask_control(p, BRW_MASK_DISABLE);
+       brw_ADD( p, return_address, brw_ip_reg(), brw_imm_ud( 2 << 4 ) );
+
+       brw_ADD( p, brw_ip_reg(), brw_ip_reg(),
+                brw_imm_d( ( c->subroutines[ subroutine ] -
+                             here - 1 ) << 4 ) );
+       brw_pop_insn_state(p);
+
+       release_tmps( c, mark );
+    } else {
+       /* previously unused subroutine: emit, and mark for later reuse */
+       
+       int mark = mark_tmps( c );
+       struct brw_reg return_address = retype( alloc_tmp( c ),
+                                               BRW_REGISTER_TYPE_UD );
+       struct brw_instruction *calc;
+       int base = p->nr_insn;
+       
+       brw_push_insn_state(p);
+       brw_set_mask_control(p, BRW_MASK_DISABLE);
+       calc = brw_ADD( p, return_address, brw_ip_reg(), brw_imm_ud( 0 ) );
+       brw_pop_insn_state(p);
+       
+       c->subroutines[ subroutine ] = p->nr_insn;
+
+       emit( c );
+       
+       brw_push_insn_state(p);
+       brw_set_mask_control(p, BRW_MASK_DISABLE);
+       brw_MOV( p, brw_ip_reg(), return_address );
+       brw_pop_insn_state(p);
+
+       brw_set_src1( calc, brw_imm_ud( ( p->nr_insn - base ) << 4 ) );
+       
+       release_tmps( c, mark );
+    }
+}
+
+static void emit_trunc( struct brw_wm_compile *c,
+                        const struct brw_fp_instruction *inst)
+{
+    int i;
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    brw_set_saturate(p, inst->SaturateMode != SATURATE_OFF);
+    for (i = 0; i < 4; i++) {
+       if (mask & (1<<i)) {
+           struct brw_reg src, dst;
+           dst = get_dst_reg(c, inst, i);
+           src = get_src_reg(c, inst, 0, i);
+           brw_RNDZ(p, dst, src);
+       }
+    }
+    brw_set_saturate(p, 0);
+}
+
+static void emit_mov( struct brw_wm_compile *c,
+                      const struct brw_fp_instruction *inst)
+{
+    int i;
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    brw_set_saturate(p, inst->SaturateMode != SATURATE_OFF);
+    for (i = 0; i < 4; i++) {
+       if (mask & (1<<i)) {
+           struct brw_reg src, dst;
+           dst = get_dst_reg(c, inst, i);
+            /* XXX some moves from immediate value don't work reliably!!! */
+            /*src = get_src_reg_imm(c, inst, 0, i);*/
+            src = get_src_reg(c, inst, 0, i);
+           brw_MOV(p, dst, src);
+       }
+    }
+    brw_set_saturate(p, 0);
+}
+
+static void emit_pixel_xy(struct brw_wm_compile *c,
+                          const struct brw_fp_instruction *inst)
+{
+    struct brw_reg r1 = brw_vec1_grf(1, 0);
+    struct brw_reg r1_uw = retype(r1, BRW_REGISTER_TYPE_UW);
+
+    struct brw_reg dst0, dst1;
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+
+    dst0 = get_dst_reg(c, inst, 0);
+    dst1 = get_dst_reg(c, inst, 1);
+    /* Calculate pixel centers by adding 1 or 0 to each of the
+     * micro-tile coordinates passed in r1.
+     */
+    if (mask & WRITEMASK_X) {
+       brw_ADD(p,
+               vec8(retype(dst0, BRW_REGISTER_TYPE_UW)),
+               stride(suboffset(r1_uw, 4), 2, 4, 0),
+               brw_imm_v(0x10101010));
+    }
+
+    if (mask & WRITEMASK_Y) {
+       brw_ADD(p,
+               vec8(retype(dst1, BRW_REGISTER_TYPE_UW)),
+               stride(suboffset(r1_uw, 5), 2, 4, 0),
+               brw_imm_v(0x11001100));
+    }
+}
+
+static void emit_delta_xy(struct brw_wm_compile *c,
+                          const struct brw_fp_instruction *inst)
+{
+    struct brw_reg r1 = brw_vec1_grf(1, 0);
+    struct brw_reg dst0, dst1, src0, src1;
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+
+    dst0 = get_dst_reg(c, inst, 0);
+    dst1 = get_dst_reg(c, inst, 1);
+    src0 = get_src_reg(c, inst, 0, 0);
+    src1 = get_src_reg(c, inst, 0, 1);
+    /* Calc delta X,Y by subtracting origin in r1 from the pixel
+     * centers.
+     */
+    if (mask & WRITEMASK_X) {
+       brw_ADD(p,
+               dst0,
+               retype(src0, BRW_REGISTER_TYPE_UW),
+               negate(r1));
+    }
+
+    if (mask & WRITEMASK_Y) {
+       brw_ADD(p,
+               dst1,
+               retype(src1, BRW_REGISTER_TYPE_UW),
+               negate(suboffset(r1,1)));
+
+    }
+}
+
+static void fire_fb_write( struct brw_wm_compile *c,
+                           GLuint base_reg,
+                           GLuint nr,
+                           GLuint target,
+                           GLuint eot)
+{
+    struct brw_compile *p = &c->func;
+    /* Pass through control information:
+     */
+    /*  mov (8) m1.0<1>:ud   r1.0<8;8,1>:ud   { Align1 NoMask } */
+    {
+       brw_push_insn_state(p);
+       brw_set_mask_control(p, BRW_MASK_DISABLE); /* ? */
+       brw_MOV(p,
+               brw_message_reg(base_reg + 1),
+               brw_vec8_grf(1, 0));
+       brw_pop_insn_state(p);
+    }
+    /* Send framebuffer write message: */
+    brw_fb_WRITE(p,
+           retype(vec8(brw_null_reg()), BRW_REGISTER_TYPE_UW),
+           base_reg,
+           retype(brw_vec8_grf(0, 0), BRW_REGISTER_TYPE_UW),
+           target,              
+           nr,
+           0,
+           eot);
+}
+
+static void emit_fb_write(struct brw_wm_compile *c,
+                          const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    int nr = 2;
+    int channel;
+    GLuint target, eot;
+    struct brw_reg src0;
+
+    /* Reserve a space for AA - may not be needed:
+     */
+    if (c->key.aa_dest_stencil_reg)
+       nr += 1;
+
+    brw_push_insn_state(p);
+    for (channel = 0; channel < 4; channel++) {
+        src0 = get_src_reg(c,  inst, 0, channel);
+        /*  mov (8) m2.0<1>:ud   r28.0<8;8,1>:ud  { Align1 } */
+        /*  mov (8) m6.0<1>:ud   r29.0<8;8,1>:ud  { Align1 SecHalf } */
+        brw_MOV(p, brw_message_reg(nr + channel), src0);
+    }
+    /* skip over the regs populated above: */
+    nr += 8;
+    brw_pop_insn_state(p);
+
+    if (c->key.source_depth_to_render_target) {
+       if (c->key.computes_depth) {
+          src0 = get_src_reg(c, inst, 2, 2);
+          brw_MOV(p, brw_message_reg(nr), src0);
+       }
+       else {
+          src0 = get_src_reg(c, inst, 1, 1);
+          brw_MOV(p, brw_message_reg(nr), src0);
+       }
+
+       nr += 2;
+    }
+
+    if (c->key.dest_depth_reg) {
+        const GLuint comp = c->key.dest_depth_reg / 2;
+        const GLuint off = c->key.dest_depth_reg % 2;
+
+        if (off != 0) {
+            /* XXX this code needs review/testing */
+            struct brw_reg arg1_0 = get_src_reg(c, inst, 1, comp);
+            struct brw_reg arg1_1 = get_src_reg(c, inst, 1, comp+1);
+
+            brw_push_insn_state(p);
+            brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+
+            brw_MOV(p, brw_message_reg(nr), offset(arg1_0, 1));
+            /* 2nd half? */
+            brw_MOV(p, brw_message_reg(nr+1), arg1_1);
+            brw_pop_insn_state(p);
+        }
+        else
+        {
+            struct brw_reg src =  get_src_reg(c, inst, 1, 1);
+            brw_MOV(p, brw_message_reg(nr), src);
+        }
+        nr += 2;
+   }
+
+    target = inst->Aux >> 1;
+    eot = inst->Aux & 1;
+    fire_fb_write(c, 0, nr, target, eot);
+}
+
+static void emit_pixel_w( struct brw_wm_compile *c,
+                          const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    if (mask & WRITEMASK_W) {
+       struct brw_reg dst, src0, delta0, delta1;
+       struct brw_reg interp3;
+
+       dst = get_dst_reg(c, inst, 3);
+       src0 = get_src_reg(c, inst, 0, 0);
+       delta0 = get_src_reg(c, inst, 1, 0);
+       delta1 = get_src_reg(c, inst, 1, 1);
+
+       interp3 = brw_vec1_grf(src0.nr+1, 4);
+       /* Calc 1/w - just linterp wpos[3] optimized by putting the
+        * result straight into a message reg.
+        */
+       brw_LINE(p, brw_null_reg(), interp3, delta0);
+       brw_MAC(p, brw_message_reg(2), suboffset(interp3, 1), delta1);
+
+       /* Calc w */
+       brw_math_16( p, dst,
+               BRW_MATH_FUNCTION_INV,
+               BRW_MATH_SATURATE_NONE,
+               2, brw_null_reg(),
+               BRW_MATH_PRECISION_FULL);
+    }
+}
+
+static void emit_linterp(struct brw_wm_compile *c,
+                         const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    struct brw_reg interp[4];
+    struct brw_reg dst, delta0, delta1;
+    struct brw_reg src0;
+    GLuint nr, i;
+
+    src0 = get_src_reg(c, inst, 0, 0);
+    delta0 = get_src_reg(c, inst, 1, 0);
+    delta1 = get_src_reg(c, inst, 1, 1);
+    nr = src0.nr;
+
+    interp[0] = brw_vec1_grf(nr, 0);
+    interp[1] = brw_vec1_grf(nr, 4);
+    interp[2] = brw_vec1_grf(nr+1, 0);
+    interp[3] = brw_vec1_grf(nr+1, 4);
+
+    for(i = 0; i < 4; i++ ) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           brw_LINE(p, brw_null_reg(), interp[i], delta0);
+           brw_MAC(p, dst, suboffset(interp[i],1), delta1);
+       }
+    }
+}
+
+static void emit_cinterp(struct brw_wm_compile *c,
+                         const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+
+    struct brw_reg interp[4];
+    struct brw_reg dst, src0;
+    GLuint nr, i;
+
+    src0 = get_src_reg(c, inst, 0, 0);
+    nr = src0.nr;
+
+    interp[0] = brw_vec1_grf(nr, 0);
+    interp[1] = brw_vec1_grf(nr, 4);
+    interp[2] = brw_vec1_grf(nr+1, 0);
+    interp[3] = brw_vec1_grf(nr+1, 4);
+
+    for(i = 0; i < 4; i++ ) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           brw_MOV(p, dst, suboffset(interp[i],3));
+       }
+    }
+}
+
+static void emit_pinterp(struct brw_wm_compile *c,
+                         const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+
+    struct brw_reg interp[4];
+    struct brw_reg dst, delta0, delta1;
+    struct brw_reg src0, w;
+    GLuint nr, i;
+
+    src0 = get_src_reg(c, inst, 0, 0);
+    delta0 = get_src_reg(c, inst, 1, 0);
+    delta1 = get_src_reg(c, inst, 1, 1);
+    w = get_src_reg(c, inst, 2, 3);
+    nr = src0.nr;
+
+    interp[0] = brw_vec1_grf(nr, 0);
+    interp[1] = brw_vec1_grf(nr, 4);
+    interp[2] = brw_vec1_grf(nr+1, 0);
+    interp[3] = brw_vec1_grf(nr+1, 4);
+
+    for(i = 0; i < 4; i++ ) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           brw_LINE(p, brw_null_reg(), interp[i], delta0);
+           brw_MAC(p, dst, suboffset(interp[i],1), 
+                   delta1);
+           brw_MUL(p, dst, dst, w);
+       }
+    }
+}
+
+/* Sets the destination channels to 1.0 or 0.0 according to glFrontFacing. */
+static void emit_frontfacing(struct brw_wm_compile *c,
+                            const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg r1_6ud = retype(brw_vec1_grf(1, 6), BRW_REGISTER_TYPE_UD);
+    struct brw_reg dst;
+    GLuint mask = inst->DstReg.WriteMask;
+    int i;
+
+    for (i = 0; i < 4; i++) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           brw_MOV(p, dst, brw_imm_f(0.0));
+       }
+    }
+
+    /* bit 31 is "primitive is back face", so checking < (1 << 31) gives
+     * us front face
+     */
+    brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_L, r1_6ud, brw_imm_ud(1 << 31));
+    for (i = 0; i < 4; i++) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           brw_MOV(p, dst, brw_imm_f(1.0));
+       }
+    }
+    brw_set_predicate_control_flag_value(p, 0xff);
+}
+
+static void emit_xpd(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    int i;
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    for (i = 0; i < 4; i++) {
+       GLuint i2 = (i+2)%3;
+       GLuint i1 = (i+1)%3;
+       if (mask & (1<<i)) {
+           struct brw_reg src0, src1, dst;
+           dst = get_dst_reg(c, inst, i);
+           src0 = negate(get_src_reg(c, inst, 0, i2));
+           src1 = get_src_reg_imm(c, inst, 1, i1);
+           brw_MUL(p, brw_null_reg(), src0, src1);
+           src0 = get_src_reg(c, inst, 0, i1);
+           src1 = get_src_reg_imm(c, inst, 1, i2);
+           brw_set_saturate(p, inst->SaturateMode != SATURATE_OFF);
+           brw_MAC(p, dst, src0, src1);
+           brw_set_saturate(p, 0);
+       }
+    }
+    brw_set_saturate(p, 0);
+}
+
+static void emit_dp3(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_reg src0[3], src1[3], dst;
+    int i;
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    int dst_chan = ffs(mask & WRITEMASK_XYZW) - 1;
+
+    if (!(mask & WRITEMASK_XYZW))
+       return;
+
+    assert(is_power_of_two(mask & WRITEMASK_XYZW));
+
+    for (i = 0; i < 3; i++) {
+       src0[i] = get_src_reg(c, inst, 0, i);
+       src1[i] = get_src_reg_imm(c, inst, 1, i);
+    }
+
+    dst = get_dst_reg(c, inst, dst_chan);
+    brw_MUL(p, brw_null_reg(), src0[0], src1[0]);
+    brw_MAC(p, brw_null_reg(), src0[1], src1[1]);
+    brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+    brw_MAC(p, dst, src0[2], src1[2]);
+    brw_set_saturate(p, 0);
+}
+
+static void emit_dp4(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_reg src0[4], src1[4], dst;
+    int i;
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    int dst_chan = ffs(mask & WRITEMASK_XYZW) - 1;
+
+    if (!(mask & WRITEMASK_XYZW))
+       return;
+
+    assert(is_power_of_two(mask & WRITEMASK_XYZW));
+
+    for (i = 0; i < 4; i++) {
+       src0[i] = get_src_reg(c, inst, 0, i);
+       src1[i] = get_src_reg_imm(c, inst, 1, i);
+    }
+    dst = get_dst_reg(c, inst, dst_chan);
+    brw_MUL(p, brw_null_reg(), src0[0], src1[0]);
+    brw_MAC(p, brw_null_reg(), src0[1], src1[1]);
+    brw_MAC(p, brw_null_reg(), src0[2], src1[2]);
+    brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+    brw_MAC(p, dst, src0[3], src1[3]);
+    brw_set_saturate(p, 0);
+}
+
+static void emit_dph(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_reg src0[4], src1[4], dst;
+    int i;
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    int dst_chan = ffs(mask & WRITEMASK_XYZW) - 1;
+
+    if (!(mask & WRITEMASK_XYZW))
+       return;
+
+    assert(is_power_of_two(mask & WRITEMASK_XYZW));
+
+    for (i = 0; i < 4; i++) {
+       src0[i] = get_src_reg(c, inst, 0, i);
+       src1[i] = get_src_reg_imm(c, inst, 1, i);
+    }
+    dst = get_dst_reg(c, inst, dst_chan);
+    brw_MUL(p, brw_null_reg(), src0[0], src1[0]);
+    brw_MAC(p, brw_null_reg(), src0[1], src1[1]);
+    brw_MAC(p, dst, src0[2], src1[2]);
+    brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+    brw_ADD(p, dst, dst, src1[3]);
+    brw_set_saturate(p, 0);
+}
+
+/**
+ * Emit a scalar instruction, like RCP, RSQ, LOG, EXP.
+ * Note that the result of the function is smeared across the dest
+ * register's X, Y, Z and W channels (subject to writemasking of course).
+ */
+static void emit_math1(struct brw_wm_compile *c,
+                       const struct brw_fp_instruction *inst, GLuint func)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg src0, dst;
+    GLuint mask = inst->DstReg.WriteMask;
+    int dst_chan = ffs(mask & WRITEMASK_XYZW) - 1;
+
+    if (!(mask & WRITEMASK_XYZW))
+       return;
+
+    assert(is_power_of_two(mask & WRITEMASK_XYZW));
+
+    /* Get first component of source register */
+    dst = get_dst_reg(c, inst, dst_chan);
+    src0 = get_src_reg(c, inst, 0, 0);
+
+    brw_MOV(p, brw_message_reg(2), src0);
+    brw_math(p,
+             dst,
+             func,
+             (inst->SaturateMode != SATURATE_OFF) ? BRW_MATH_SATURATE_SATURATE : BRW_MATH_SATURATE_NONE,
+             2,
+             brw_null_reg(),
+             BRW_MATH_DATA_VECTOR,
+             BRW_MATH_PRECISION_FULL);
+}
+
+static void emit_rcp(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_math1(c, inst, BRW_MATH_FUNCTION_INV);
+}
+
+static void emit_rsq(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_math1(c, inst, BRW_MATH_FUNCTION_RSQ);
+}
+
+static void emit_sin(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_math1(c, inst, BRW_MATH_FUNCTION_SIN);
+}
+
+static void emit_cos(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_math1(c, inst, BRW_MATH_FUNCTION_COS);
+}
+
+static void emit_ex2(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_math1(c, inst, BRW_MATH_FUNCTION_EXP);
+}
+
+static void emit_lg2(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_math1(c, inst, BRW_MATH_FUNCTION_LOG);
+}
+
+static void emit_add(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg src0, src1, dst;
+    GLuint mask = inst->DstReg.WriteMask;
+    int i;
+    brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+    for (i = 0 ; i < 4; i++) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           src0 = get_src_reg(c, inst, 0, i);
+           src1 = get_src_reg_imm(c, inst, 1, i);
+           brw_ADD(p, dst, src0, src1);
+       }
+    }
+    brw_set_saturate(p, 0);
+}
+
+static void emit_arl(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg src0, addr_reg;
+    brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+    addr_reg = brw_uw8_reg(BRW_ARCHITECTURE_REGISTER_FILE, 
+                           BRW_ARF_ADDRESS, 0);
+    src0 = get_src_reg(c, inst, 0, 0); /* channel 0 */
+    brw_MOV(p, addr_reg, src0);
+    brw_set_saturate(p, 0);
+}
+
+
+static void emit_mul(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg src0, src1, dst;
+    GLuint mask = inst->DstReg.WriteMask;
+    int i;
+    brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+    for (i = 0 ; i < 4; i++) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           src0 = get_src_reg(c, inst, 0, i);
+           src1 = get_src_reg_imm(c, inst, 1, i);
+           brw_MUL(p, dst, src0, src1);
+       }
+    }
+    brw_set_saturate(p, 0);
+}
+
+static void emit_frc(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg src0, dst;
+    GLuint mask = inst->DstReg.WriteMask;
+    int i;
+    brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+    for (i = 0 ; i < 4; i++) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           src0 = get_src_reg_imm(c, inst, 0, i);
+           brw_FRC(p, dst, src0);
+       }
+    }
+    if (inst->SaturateMode != SATURATE_OFF)
+       brw_set_saturate(p, 0);
+}
+
+static void emit_flr(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg src0, dst;
+    GLuint mask = inst->DstReg.WriteMask;
+    int i;
+    brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+    for (i = 0 ; i < 4; i++) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           src0 = get_src_reg_imm(c, inst, 0, i);
+           brw_RNDD(p, dst, src0);
+       }
+    }
+    brw_set_saturate(p, 0);
+}
+
+
+static void emit_min_max(struct brw_wm_compile *c,
+                         const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    const GLuint mask = inst->DstReg.WriteMask;
+    const int mark = mark_tmps(c);
+    int i;
+    brw_push_insn_state(p);
+    for (i = 0; i < 4; i++) {
+       if (mask & (1<<i)) {
+            struct brw_reg real_dst = get_dst_reg(c, inst, i);
+           struct brw_reg src0 = get_src_reg(c, inst, 0, i);
+           struct brw_reg src1 = get_src_reg(c, inst, 1, i);
+            struct brw_reg dst;
+            /* if dst==src0 or dst==src1 we need to use a temp reg */
+            GLboolean use_temp = brw_same_reg(dst, src0) ||
+                                 brw_same_reg(dst, src1);
+            if (use_temp)
+               dst = alloc_tmp(c);
+            else
+               dst = real_dst;
+
+            /*
+            printf("  Min/max: dst %d  src0 %d  src1 %d\n",
+                   dst.nr, src0.nr, src1.nr);
+            */
+           brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+           brw_MOV(p, dst, src0);
+           brw_set_saturate(p, 0);
+
+            if (inst->Opcode == OPCODE_MIN)
+               brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_L, src1, src0);
+            else
+               brw_CMP(p, brw_null_reg(), BRW_CONDITIONAL_G, src1, src0);
+
+           brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+           brw_set_predicate_control(p, BRW_PREDICATE_NORMAL);
+           brw_MOV(p, dst, src1);
+           brw_set_saturate(p, 0);
+           brw_set_predicate_control_flag_value(p, 0xff);
+            if (use_temp)
+               brw_MOV(p, real_dst, dst);
+       }
+    }
+    brw_pop_insn_state(p);
+    release_tmps(c, mark);
+}
+
+static void emit_pow(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg dst, src0, src1;
+    GLuint mask = inst->DstReg.WriteMask;
+    int dst_chan = ffs(mask & WRITEMASK_XYZW) - 1;
+
+    if (!(mask & WRITEMASK_XYZW))
+       return;
+
+    assert(is_power_of_two(mask & WRITEMASK_XYZW));
+
+    dst = get_dst_reg(c, inst, dst_chan);
+    src0 = get_src_reg_imm(c, inst, 0, 0);
+    src1 = get_src_reg_imm(c, inst, 1, 0);
+
+    brw_MOV(p, brw_message_reg(2), src0);
+    brw_MOV(p, brw_message_reg(3), src1);
+
+    brw_math(p,
+           dst,
+           BRW_MATH_FUNCTION_POW,
+           (inst->SaturateMode != SATURATE_OFF) ? BRW_MATH_SATURATE_SATURATE : BRW_MATH_SATURATE_NONE,
+           2,
+           brw_null_reg(),
+           BRW_MATH_DATA_VECTOR,
+           BRW_MATH_PRECISION_FULL);
+}
+
+static void emit_lrp(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    struct brw_reg dst, tmp1, tmp2, src0, src1, src2;
+    int i;
+    int mark = mark_tmps(c);
+    for (i = 0; i < 4; i++) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           src0 = get_src_reg(c, inst, 0, i);
+
+           src1 = get_src_reg_imm(c, inst, 1, i);
+
+           if (src1.nr == dst.nr) {
+               tmp1 = alloc_tmp(c);
+               brw_MOV(p, tmp1, src1);
+           } else
+               tmp1 = src1;
+
+           src2 = get_src_reg(c, inst, 2, i);
+           if (src2.nr == dst.nr) {
+               tmp2 = alloc_tmp(c);
+               brw_MOV(p, tmp2, src2);
+           } else
+               tmp2 = src2;
+
+           brw_ADD(p, dst, negate(src0), brw_imm_f(1.0));
+           brw_MUL(p, brw_null_reg(), dst, tmp2);
+           brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+           brw_MAC(p, dst, src0, tmp1);
+           brw_set_saturate(p, 0);
+       }
+       release_tmps(c, mark);
+    }
+}
+
+/**
+ * For GLSL shaders, this KIL will be unconditional.
+ * It may be contained inside an IF/ENDIF structure of course.
+ */
+static void emit_kil(struct brw_wm_compile *c)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg depth = retype(brw_vec1_grf(0, 0), BRW_REGISTER_TYPE_UW);
+    brw_push_insn_state(p);
+    brw_set_mask_control(p, BRW_MASK_DISABLE);
+    brw_NOT(p, c->emit_mask_reg, brw_mask_reg(1)); //IMASK
+    brw_AND(p, depth, c->emit_mask_reg, depth);
+    brw_pop_insn_state(p);
+}
+
+static void emit_mad(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    struct brw_reg dst, src0, src1, src2;
+    int i;
+
+    for (i = 0; i < 4; i++) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           src0 = get_src_reg(c, inst, 0, i);
+           src1 = get_src_reg_imm(c, inst, 1, i);
+           src2 = get_src_reg_imm(c, inst, 2, i);
+           brw_MUL(p, dst, src0, src1);
+
+           brw_set_saturate(p, (inst->SaturateMode != SATURATE_OFF) ? 1 : 0);
+           brw_ADD(p, dst, dst, src2);
+           brw_set_saturate(p, 0);
+       }
+    }
+}
+
+static void emit_sop(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst, GLuint cond)
+{
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    struct brw_reg dst, src0, src1;
+    int i;
+
+    for (i = 0; i < 4; i++) {
+       if (mask & (1<<i)) {
+           dst = get_dst_reg(c, inst, i);
+           src0 = get_src_reg(c, inst, 0, i);
+           src1 = get_src_reg_imm(c, inst, 1, i);
+           brw_push_insn_state(p);
+           brw_CMP(p, brw_null_reg(), cond, src0, src1);
+           brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+           brw_MOV(p, dst, brw_imm_f(0.0));
+           brw_set_predicate_control(p, BRW_PREDICATE_NORMAL);
+           brw_MOV(p, dst, brw_imm_f(1.0));
+           brw_pop_insn_state(p);
+       }
+    }
+}
+
+static void emit_slt(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_sop(c, inst, BRW_CONDITIONAL_L);
+}
+
+static void emit_sle(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_sop(c, inst, BRW_CONDITIONAL_LE);
+}
+
+static void emit_sgt(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_sop(c, inst, BRW_CONDITIONAL_G);
+}
+
+static void emit_sge(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_sop(c, inst, BRW_CONDITIONAL_GE);
+}
+
+static void emit_seq(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_sop(c, inst, BRW_CONDITIONAL_EQ);
+}
+
+static void emit_sne(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    emit_sop(c, inst, BRW_CONDITIONAL_NEQ);
+}
+
+static INLINE struct brw_reg high_words( struct brw_reg reg )
+{
+    return stride( suboffset( retype( reg, BRW_REGISTER_TYPE_W ), 1 ),
+                  0, 8, 2 );
+}
+
+static INLINE struct brw_reg low_words( struct brw_reg reg )
+{
+    return stride( retype( reg, BRW_REGISTER_TYPE_W ), 0, 8, 2 );
+}
+
+static INLINE struct brw_reg even_bytes( struct brw_reg reg )
+{
+    return stride( retype( reg, BRW_REGISTER_TYPE_B ), 0, 16, 2 );
+}
+
+static INLINE struct brw_reg odd_bytes( struct brw_reg reg )
+{
+    return stride( suboffset( retype( reg, BRW_REGISTER_TYPE_B ), 1 ),
+                  0, 16, 2 );
+}
+
+
+    
+static void emit_wpos_xy(struct brw_wm_compile *c,
+                         const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    GLuint mask = inst->DstReg.WriteMask;
+    struct brw_reg src0[2], dst[2];
+
+    dst[0] = get_dst_reg(c, inst, 0);
+    dst[1] = get_dst_reg(c, inst, 1);
+
+    src0[0] = get_src_reg(c, inst, 0, 0);
+    src0[1] = get_src_reg(c, inst, 0, 1);
+
+    /* Calculate the pixel offset from window bottom left into destination
+     * X and Y channels.
+     */
+    if (mask & WRITEMASK_X) {
+       /* X' = X */
+       brw_MOV(p,
+               dst[0],
+               retype(src0[0], BRW_REGISTER_TYPE_W));
+    }
+
+    if (mask & WRITEMASK_Y) {
+       /* Y' = height - 1 - Y */
+       brw_ADD(p,
+               dst[1],
+               negate(retype(src0[1], BRW_REGISTER_TYPE_W)),
+               brw_imm_d(c->key.drawable_height - 1));
+    }
+}
+
+/* TODO
+   BIAS on SIMD8 not working yet...
+ */    
+static void emit_txb(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg dst[4], src[4], payload_reg;
+    /* Note: tex_unit was already looked up through SamplerTextures[] */
+    const GLuint unit = inst->tex_unit;
+    GLuint i;
+    GLuint msg_type;
+
+    assert(unit < BRW_MAX_TEX_UNIT);
+
+    payload_reg = get_reg(c, TGSI_FILE_PAYLOAD, PAYLOAD_DEPTH, 0, 1, 0, 0);
+
+    for (i = 0; i < 4; i++) 
+       dst[i] = get_dst_reg(c, inst, i);
+    for (i = 0; i < 4; i++)
+       src[i] = get_src_reg(c, inst, 0, i);
+
+    switch (inst->tex_target) {
+       case TEXTURE_1D_INDEX:
+           brw_MOV(p, brw_message_reg(2), src[0]);         /* s coord */
+           brw_MOV(p, brw_message_reg(3), brw_imm_f(0));   /* t coord */
+           brw_MOV(p, brw_message_reg(4), brw_imm_f(0));   /* r coord */
+           break;
+       case TEXTURE_2D_INDEX:
+       case TEXTURE_RECT_INDEX:
+           brw_MOV(p, brw_message_reg(2), src[0]);
+           brw_MOV(p, brw_message_reg(3), src[1]);
+           brw_MOV(p, brw_message_reg(4), brw_imm_f(0));
+           break;
+       case TEXTURE_3D_INDEX:
+       case TEXTURE_CUBE_INDEX:
+           brw_MOV(p, brw_message_reg(2), src[0]);
+           brw_MOV(p, brw_message_reg(3), src[1]);
+           brw_MOV(p, brw_message_reg(4), src[2]);
+           break;
+       default:
+            /* invalid target */
+            abort();
+    }
+    brw_MOV(p, brw_message_reg(5), src[3]);          /* bias */
+    brw_MOV(p, brw_message_reg(6), brw_imm_f(0));    /* ref (unused?) */
+
+    if (BRW_IS_IGDNG(p->brw)) {
+        msg_type = BRW_SAMPLER_MESSAGE_SIMD8_SAMPLE_BIAS_IGDNG;
+    } else {
+        /* Does it work well on SIMD8? */
+        msg_type = BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_BIAS;
+    }
+
+    brw_SAMPLE(p,
+               retype(vec8(dst[0]), BRW_REGISTER_TYPE_UW),  /* dest */
+               1,                                           /* msg_reg_nr */
+               retype(payload_reg, BRW_REGISTER_TYPE_UW),   /* src0 */
+               SURF_INDEX_TEXTURE(unit),
+               unit,                                        /* sampler */
+               inst->DstReg.WriteMask,                      /* writemask */
+               msg_type,                                    /* msg_type */
+               4,                                           /* response_length */
+               4,                                           /* msg_length */
+               0,                                           /* eot */
+               1,
+               BRW_SAMPLER_SIMD_MODE_SIMD8);   
+}
+
+
+static void emit_tex(struct brw_wm_compile *c,
+                     const struct brw_fp_instruction *inst)
+{
+    struct brw_compile *p = &c->func;
+    struct brw_reg dst[4], src[4], payload_reg;
+    /* Note: tex_unit was already looked up through SamplerTextures[] */
+    const GLuint unit = inst->tex_unit;
+    GLuint msg_len;
+    GLuint i, nr;
+    GLuint emit;
+    GLboolean shadow = (c->key.shadowtex_mask & (1<<unit)) ? 1 : 0;
+    GLuint msg_type;
+
+    assert(unit < BRW_MAX_TEX_UNIT);
+
+    payload_reg = get_reg(c, TGSI_FILE_PAYLOAD, PAYLOAD_DEPTH, 0, 1, 0, 0);
+
+    for (i = 0; i < 4; i++) 
+       dst[i] = get_dst_reg(c, inst, i);
+    for (i = 0; i < 4; i++)
+       src[i] = get_src_reg(c, inst, 0, i);
+
+    switch (inst->tex_target) {
+       case TEXTURE_1D_INDEX:
+           emit = WRITEMASK_X;
+           nr = 1;
+           break;
+       case TEXTURE_2D_INDEX:
+       case TEXTURE_RECT_INDEX:
+           emit = WRITEMASK_XY;
+           nr = 2;
+           break;
+       case TEXTURE_3D_INDEX:
+       case TEXTURE_CUBE_INDEX:
+           emit = WRITEMASK_XYZ;
+           nr = 3;
+           break;
+       default:
+           /* invalid target */
+           abort();
+    }
+    msg_len = 1;
+
+    /* move/load S, T, R coords */
+    for (i = 0; i < nr; i++) {
+       static const GLuint swz[4] = {0,1,2,2};
+       if (emit & (1<<i))
+           brw_MOV(p, brw_message_reg(msg_len+1), src[swz[i]]);
+       else
+           brw_MOV(p, brw_message_reg(msg_len+1), brw_imm_f(0));
+       msg_len += 1;
+    }
+
+    if (shadow) {
+       brw_MOV(p, brw_message_reg(5), brw_imm_f(0));  /* lod / bias */
+       brw_MOV(p, brw_message_reg(6), src[2]);        /* ref value / R coord */
+    }
+
+    if (BRW_IS_IGDNG(p->brw)) {
+        if (shadow)
+            msg_type = BRW_SAMPLER_MESSAGE_SIMD8_SAMPLE_COMPARE_IGDNG;
+        else
+            msg_type = BRW_SAMPLER_MESSAGE_SIMD8_SAMPLE_IGDNG;
+    } else {
+        /* Does it work for shadow on SIMD8 ? */
+        msg_type = BRW_SAMPLER_MESSAGE_SIMD8_SAMPLE;
+    }
+    
+    brw_SAMPLE(p,
+               retype(vec8(dst[0]), BRW_REGISTER_TYPE_UW), /* dest */
+               1,                                          /* msg_reg_nr */
+               retype(payload_reg, BRW_REGISTER_TYPE_UW),  /* src0 */
+               SURF_INDEX_TEXTURE(unit),
+               unit,                                       /* sampler */
+               inst->DstReg.WriteMask,                     /* writemask */
+               msg_type,                                   /* msg_type */
+               4,                                          /* response_length */
+               shadow ? 6 : 4,                             /* msg_length */
+               0,                                          /* eot */
+               1,
+               BRW_SAMPLER_SIMD_MODE_SIMD8);   
+
+    if (shadow)
+       brw_MOV(p, dst[3], brw_imm_f(1.0));
+}
+
+
+/**
+ * Resolve subroutine calls after code emit is done.
+ */
+static void post_wm_emit( struct brw_wm_compile *c )
+{
+    brw_resolve_cals(&c->func);
+}
+
+static void
+get_argument_regs(struct brw_wm_compile *c,
+                 const struct brw_fp_instruction *inst,
+                 int index,
+                 struct brw_reg *regs,
+                 int mask)
+{
+    int i;
+
+    for (i = 0; i < 4; i++) {
+       if (mask & (1 << i))
+           regs[i] = get_src_reg(c, inst, index, i);
+    }
+}
+
+static void brw_wm_emit_branching_shader(struct brw_context *brw, struct brw_wm_compile *c)
+{
+#define MAX_IF_DEPTH 32
+#define MAX_LOOP_DEPTH 32
+    struct brw_instruction *if_inst[MAX_IF_DEPTH], *loop_inst[MAX_LOOP_DEPTH];
+    GLuint i, if_depth = 0, loop_depth = 0;
+    struct brw_compile *p = &c->func;
+    struct brw_indirect stack_index = brw_indirect(0, 0);
+
+    c->out_of_regs = GL_FALSE;
+
+    prealloc_reg(c);
+    brw_set_compression_control(p, BRW_COMPRESSION_NONE);
+    brw_MOV(p, get_addr_reg(stack_index), brw_address(c->stack));
+
+    for (i = 0; i < c->nr_fp_insns; i++) {
+        const struct brw_fp_instruction *inst = &c->fp_instructions[i];
+       int dst_flags;
+       struct brw_reg args[3][4], dst[4];
+       int j;
+
+        c->cur_inst = i;
+
+#if 0
+        debug_printf("Inst %d: ", i);
+        _mesa_print_instruction(inst);
+#endif
+
+        /* fetch any constants that this instruction needs */
+        if (c->fp->use_const_buffer)
+           fetch_constants(c, inst);
+
+       if (inst->CondUpdate)
+           brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
+       else
+           brw_set_conditionalmod(p, BRW_CONDITIONAL_NONE);
+
+       dst_flags = inst->DstReg.WriteMask;
+       if (inst->SaturateMode == SATURATE_ZERO_ONE)
+           dst_flags |= SATURATE;
+
+       switch (inst->Opcode) {
+           case WM_PIXELXY:
+               emit_pixel_xy(c, inst);
+               break;
+           case WM_DELTAXY: 
+               emit_delta_xy(c, inst);
+               break;
+           case WM_PIXELW:
+               emit_pixel_w(c, inst);
+               break;  
+           case WM_LINTERP:
+               emit_linterp(c, inst);
+               break;
+           case WM_PINTERP:
+               emit_pinterp(c, inst);
+               break;
+           case WM_CINTERP:
+               emit_cinterp(c, inst);
+               break;
+           case WM_WPOSXY:
+               emit_wpos_xy(c, inst);
+               break;
+           case WM_FB_WRITE:
+               emit_fb_write(c, inst);
+               break;
+           case WM_FRONTFACING:
+               emit_frontfacing(c, inst);
+               break;
+           case OPCODE_ADD:
+               emit_add(c, inst);
+               break;
+           case OPCODE_ARL:
+               emit_arl(c, inst);
+               break;
+           case OPCODE_FRC:
+               emit_frc(c, inst);
+               break;
+           case OPCODE_FLR:
+               emit_flr(c, inst);
+               break;
+           case OPCODE_LRP:
+               emit_lrp(c, inst);
+               break;
+           case OPCODE_TRUNC:
+               emit_trunc(c, inst);
+               break;
+           case OPCODE_MOV:
+               emit_mov(c, inst);
+               break;
+           case OPCODE_DP3:
+               emit_dp3(c, inst);
+               break;
+           case OPCODE_DP4:
+               emit_dp4(c, inst);
+               break;
+           case OPCODE_XPD:
+               emit_xpd(c, inst);
+               break;
+           case OPCODE_DPH:
+               emit_dph(c, inst);
+               break;
+           case OPCODE_RCP:
+               emit_rcp(c, inst);
+               break;
+           case OPCODE_RSQ:
+               emit_rsq(c, inst);
+               break;
+           case OPCODE_SIN:
+               emit_sin(c, inst);
+               break;
+           case OPCODE_COS:
+               emit_cos(c, inst);
+               break;
+           case OPCODE_EX2:
+               emit_ex2(c, inst);
+               break;
+           case OPCODE_LG2:
+               emit_lg2(c, inst);
+               break;
+           case OPCODE_MIN:    
+           case OPCODE_MAX:    
+               emit_min_max(c, inst);
+               break;
+           case OPCODE_DDX:
+           case OPCODE_DDY:
+               for (j = 0; j < 4; j++) {
+                   if (inst->DstReg.WriteMask & (1 << j))
+                       dst[j] = get_dst_reg(c, inst, j);
+                   else
+                       dst[j] = brw_null_reg();
+               }
+               get_argument_regs(c, inst, 0, args[0], WRITEMASK_XYZW);
+               emit_ddxy(p, dst, dst_flags, (inst->Opcode == OPCODE_DDX),
+                         args[0]);
+                break;
+           case OPCODE_SLT:
+               emit_slt(c, inst);
+               break;
+           case OPCODE_SLE:
+               emit_sle(c, inst);
+               break;
+           case OPCODE_SGT:
+               emit_sgt(c, inst);
+               break;
+           case OPCODE_SGE:
+               emit_sge(c, inst);
+               break;
+           case OPCODE_SEQ:
+               emit_seq(c, inst);
+               break;
+           case OPCODE_SNE:
+               emit_sne(c, inst);
+               break;
+           case OPCODE_MUL:
+               emit_mul(c, inst);
+               break;
+           case OPCODE_POW:
+               emit_pow(c, inst);
+               break;
+           case OPCODE_MAD:
+               emit_mad(c, inst);
+               break;
+           case OPCODE_TEX:
+               emit_tex(c, inst);
+               break;
+           case OPCODE_TXB:
+               emit_txb(c, inst);
+               break;
+           case OPCODE_KIL_NV:
+               emit_kil(c);
+               break;
+           case OPCODE_IF:
+               assert(if_depth < MAX_IF_DEPTH);
+               if_inst[if_depth++] = brw_IF(p, BRW_EXECUTE_8);
+               break;
+           case OPCODE_ELSE:
+               if_inst[if_depth-1]  = brw_ELSE(p, if_inst[if_depth-1]);
+               break;
+           case OPCODE_ENDIF:
+               assert(if_depth > 0);
+               brw_ENDIF(p, if_inst[--if_depth]);
+               break;
+           case OPCODE_BGNSUB:
+               brw_save_label(p, inst->Comment, p->nr_insn);
+               break;
+           case OPCODE_ENDSUB:
+               /* no-op */
+               break;
+           case OPCODE_CAL: 
+               brw_push_insn_state(p);
+               brw_set_mask_control(p, BRW_MASK_DISABLE);
+                brw_set_access_mode(p, BRW_ALIGN_1);
+                brw_ADD(p, deref_1ud(stack_index, 0), brw_ip_reg(), brw_imm_d(3*16));
+                brw_set_access_mode(p, BRW_ALIGN_16);
+                brw_ADD(p, get_addr_reg(stack_index),
+                         get_addr_reg(stack_index), brw_imm_d(4));
+               brw_save_call(&c->func, inst->label, p->nr_insn);
+                brw_ADD(p, brw_ip_reg(), brw_ip_reg(), brw_imm_d(1*16));
+                brw_pop_insn_state(p);
+               break;
+
+           case OPCODE_RET:
+               brw_push_insn_state(p);
+               brw_set_mask_control(p, BRW_MASK_DISABLE);
+                brw_ADD(p, get_addr_reg(stack_index),
+                        get_addr_reg(stack_index), brw_imm_d(-4));
+                brw_set_access_mode(p, BRW_ALIGN_1);
+                brw_MOV(p, brw_ip_reg(), deref_1ud(stack_index, 0));
+                brw_set_access_mode(p, BRW_ALIGN_16);
+               brw_pop_insn_state(p);
+
+               break;
+           case OPCODE_BGNLOOP:
+                /* XXX may need to invalidate the current_constant regs */
+               loop_inst[loop_depth++] = brw_DO(p, BRW_EXECUTE_8);
+               break;
+           case OPCODE_BRK:
+               brw_BREAK(p);
+               brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+               break;
+           case OPCODE_CONT:
+               brw_CONT(p);
+               brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+               break;
+           case OPCODE_ENDLOOP: 
+               {
+                  struct brw_instruction *inst0, *inst1;
+                  GLuint br = 1;
+
+                  if (BRW_IS_IGDNG(brw))
+                     br = 2;
+                  loop_depth--;
+                  inst0 = inst1 = brw_WHILE(p, loop_inst[loop_depth]);
+                  /* patch all the BREAK/CONT instructions from last BGNLOOP */
+                  while (inst0 > loop_inst[loop_depth]) {
+                     inst0--;
+                     if (inst0->header.opcode == BRW_OPCODE_BREAK) {
+                       inst0->bits3.if_else.jump_count = br * (inst1 - inst0 + 1);
+                       inst0->bits3.if_else.pop_count = 0;
+                     }
+                     else if (inst0->header.opcode == BRW_OPCODE_CONTINUE) {
+                        inst0->bits3.if_else.jump_count = br * (inst1 - inst0);
+                        inst0->bits3.if_else.pop_count = 0;
+                     }
+                  }
+               }
+               break;
+           default:
+               debug_printf("unsupported IR in fragment shader %d\n",
+                       inst->Opcode);
+       }
+
+       if (inst->CondUpdate)
+           brw_set_predicate_control(p, BRW_PREDICATE_NORMAL);
+       else
+           brw_set_predicate_control(p, BRW_PREDICATE_NONE);
+    }
+    post_wm_emit(c);
+
+    if (BRW_DEBUG & DEBUG_WM) {
+      debug_printf("wm-native:\n");
+      brw_disasm(stderr, p->store, p->nr_insn);
+    }
+}
+
+/**
+ * Do GPU code generation for shaders that use GLSL features such as
+ * flow control.  Other shaders will be compiled with the 
+ */
+void brw_wm_branching_shader_emit(struct brw_context *brw, struct brw_wm_compile *c)
+{
+    if (BRW_DEBUG & DEBUG_WM) {
+       debug_printf("%s:\n", __FUNCTION__);
+    }
+
+    /* initial instruction translation/simplification */
+    brw_wm_pass_fp(c);
+
+    /* actual code generation */
+    brw_wm_emit_branching_shader(brw, c);
+
+    if (BRW_DEBUG & DEBUG_WM) {
+        brw_wm_print_program(c, "brw_wm_branching_shader_emit done");
+    }
+
+    c->prog_data.total_grf = num_grf_used(c);
+    c->prog_data.total_scratch = 0;
+}
diff --git a/src/gallium/drivers/i965/brw_wm_iz.c b/src/gallium/drivers/i965/brw_wm_iz.c
new file mode 100644 (file)
index 0000000..6f1e9fc
--- /dev/null
@@ -0,0 +1,156 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+                
+
+#include "brw_wm.h"
+
+
+#undef P                       /* prompted depth */
+#undef C                       /* computed */
+#undef N                       /* non-promoted? */
+
+#define P 0
+#define C 1
+#define N 2
+
+const struct {
+   GLuint mode:2;
+   GLuint sd_present:1;
+   GLuint sd_to_rt:1;
+   GLuint dd_present:1;
+   GLuint ds_present:1;
+} wm_iz_table[IZ_BIT_MAX] =
+{
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { N, 1, 1, 0, 0 }, 
+ { N, 0, 1, 0, 0 }, 
+ { N, 0, 1, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { C, 0, 1, 1, 0 }, 
+ { C, 0, 1, 1, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { N, 1, 1, 0, 0 }, 
+ { C, 0, 1, 1, 0 }, 
+ { C, 0, 1, 1, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { N, 1, 1, 0, 0 }, 
+ { N, 0, 1, 0, 0 }, 
+ { N, 0, 1, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { C, 0, 1, 1, 0 }, 
+ { C, 0, 1, 1, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { N, 1, 1, 0, 0 }, 
+ { C, 0, 1, 1, 0 }, 
+ { C, 0, 1, 1, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { N, 1, 1, 0, 1 }, 
+ { N, 0, 1, 0, 1 }, 
+ { N, 0, 1, 0, 1 }, 
+ { P, 0, 0, 0, 0 }, 
+ { P, 0, 0, 0, 0 }, 
+ { C, 0, 1, 1, 1 }, 
+ { C, 0, 1, 1, 1 }, 
+ { P, 0, 0, 0, 0 }, 
+ { N, 1, 1, 0, 1 }, 
+ { C, 0, 1, 1, 1 }, 
+ { C, 0, 1, 1, 1 }, 
+ { P, 0, 0, 0, 0 }, 
+ { C, 0, 0, 0, 1 }, 
+ { P, 0, 0, 0, 0 }, 
+ { C, 0, 1, 0, 1 }, 
+ { P, 0, 0, 0, 0 }, 
+ { C, 1, 1, 0, 1 }, 
+ { C, 0, 1, 0, 1 }, 
+ { C, 0, 1, 0, 1 }, 
+ { P, 0, 0, 0, 0 }, 
+ { C, 1, 1, 1, 1 }, 
+ { C, 0, 1, 1, 1 }, 
+ { C, 0, 1, 1, 1 }, 
+ { P, 0, 0, 0, 0 }, 
+ { C, 1, 1, 1, 1 }, 
+ { C, 0, 1, 1, 1 }, 
+ { C, 0, 1, 1, 1 } 
+};
+
+/**
+ * \param line_aa  AA_NEVER, AA_ALWAYS or AA_SOMETIMES
+ * \param lookup  bitmask of IZ_* flags
+ */
+void brw_wm_lookup_iz( GLuint line_aa,
+                      GLuint lookup,
+                      GLboolean ps_uses_depth,
+                      struct brw_wm_prog_key *key )
+{
+   GLuint reg = 2;
+
+   assert (lookup < IZ_BIT_MAX);
+      
+   if (lookup & IZ_PS_COMPUTES_DEPTH_BIT)
+      key->computes_depth = 1;
+
+   if (wm_iz_table[lookup].sd_present || ps_uses_depth) {
+      key->source_depth_reg = reg;
+      reg += 2;
+   }
+
+   if (wm_iz_table[lookup].sd_to_rt)
+      key->source_depth_to_render_target = 1;
+
+   if (wm_iz_table[lookup].ds_present || line_aa != AA_NEVER) {
+      key->aa_dest_stencil_reg = reg;
+      key->runtime_check_aads_emit = (!wm_iz_table[lookup].ds_present &&
+                                     line_aa == AA_SOMETIMES);
+      reg++;
+   }
+
+   if (wm_iz_table[lookup].dd_present) {
+      key->dest_depth_reg = reg;
+      reg+=2;
+   }
+
+   key->nr_depth_regs = (reg+1)/2;
+}
+
diff --git a/src/gallium/drivers/i965/brw_wm_pass0.c b/src/gallium/drivers/i965/brw_wm_pass0.c
new file mode 100644 (file)
index 0000000..0bacad2
--- /dev/null
@@ -0,0 +1,366 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+
+#include "util/u_memory.h"
+#include "util/u_math.h"
+
+#include "brw_debug.h"
+#include "brw_wm.h"
+
+
+
+/***********************************************************************
+ */
+
+static struct brw_wm_ref *get_ref( struct brw_wm_compile *c )
+{
+   assert(c->nr_refs < BRW_WM_MAX_REF);
+   return &c->refs[c->nr_refs++];
+}
+
+static struct brw_wm_value *get_value( struct brw_wm_compile *c)
+{
+   assert(c->nr_refs < BRW_WM_MAX_VREG);
+   return &c->vreg[c->nr_vreg++];
+}
+
+/** return pointer to a newly allocated instruction */
+static struct brw_wm_instruction *get_instruction( struct brw_wm_compile *c )
+{
+   assert(c->nr_insns < BRW_WM_MAX_INSN);
+   return &c->instruction[c->nr_insns++];
+}
+
+/***********************************************************************
+ */
+
+/** Init the "undef" register */
+static void pass0_init_undef( struct brw_wm_compile *c)
+{
+   struct brw_wm_ref *ref = &c->undef_ref;
+   ref->value = &c->undef_value;
+   ref->hw_reg = brw_vec8_grf(0, 0);
+   ref->insn = 0;
+   ref->prevuse = NULL;
+}
+
+/** Set a FP register to a value */
+static void pass0_set_fpreg_value( struct brw_wm_compile *c,
+                                  GLuint file,
+                                  GLuint idx,
+                                  GLuint component,
+                                  struct brw_wm_value *value )
+{
+   struct brw_wm_ref *ref = get_ref(c);
+   ref->value = value;
+   ref->hw_reg = brw_vec8_grf(0, 0);
+   ref->insn = 0;
+   ref->prevuse = NULL;
+   c->pass0_fp_reg[file][idx][component] = ref;
+}
+
+/** Set a FP register to a ref */
+static void pass0_set_fpreg_ref( struct brw_wm_compile *c,
+                                GLuint file,
+                                GLuint idx,
+                                GLuint component,
+                                const struct brw_wm_ref *src_ref )
+{
+   c->pass0_fp_reg[file][idx][component] = src_ref;
+}
+
+static const struct brw_wm_ref *get_param_ref( struct brw_wm_compile *c, 
+                                              unsigned idx,
+                                               unsigned component)
+{
+   GLuint i = idx * 4 + component;
+   
+   if (i >= BRW_WM_MAX_PARAM) {
+      debug_printf("%s: out of params\n", __FUNCTION__);
+      c->prog_data.error = 1;
+      return NULL;
+   }
+   else {
+      struct brw_wm_ref *ref = get_ref(c);
+
+      c->nr_creg = MAX2(c->nr_creg, (i+16)/16);
+
+      /* Push the offsets into hw_reg.  These will be added to the
+       * real register numbers once one is allocated in pass2.
+       */
+      ref->hw_reg = brw_vec1_grf((i&8)?1:0, i%8);
+      ref->value = &c->creg[i/16];
+      ref->insn = 0;
+      ref->prevuse = NULL;
+
+      return ref;
+   }
+}
+
+
+
+
+/* Lookup our internal registers
+ */
+static const struct brw_wm_ref *pass0_get_reg( struct brw_wm_compile *c,
+                                              GLuint file,
+                                              GLuint idx,
+                                              GLuint component )
+{
+   const struct brw_wm_ref *ref = c->pass0_fp_reg[file][idx][component];
+
+   if (!ref) {
+      switch (file) {
+      case TGSI_FILE_INPUT:
+      case TGSI_FILE_TEMPORARY:
+      case TGSI_FILE_OUTPUT:
+      case BRW_FILE_PAYLOAD:
+        /* should already be done?? */
+        break;
+
+      case TGSI_FILE_CONSTANT:
+        ref = get_param_ref(c, 
+                             c->fp->info.immediate_count + idx,
+                             component);
+        break;
+
+      case TGSI_FILE_IMMEDIATE:
+        ref = get_param_ref(c, 
+                             idx,
+                             component);
+        break;
+
+      default:
+        assert(0);
+        break;
+      }
+
+      c->pass0_fp_reg[file][idx][component] = ref;
+   }
+
+   if (!ref)
+      ref = &c->undef_ref;
+
+   return ref;
+}
+
+
+
+/***********************************************************************
+ * Straight translation to internal instruction format
+ */
+
+static void pass0_set_dst( struct brw_wm_compile *c,
+                          struct brw_wm_instruction *out,
+                          const struct brw_fp_instruction *inst,
+                          GLuint writemask )
+{
+   const struct brw_fp_dst dst = inst->dst;
+   GLuint i;
+
+   for (i = 0; i < 4; i++) {
+      if (writemask & (1<<i)) {
+        out->dst[i] = get_value(c);
+        pass0_set_fpreg_value(c, dst.file, dst.index, i, out->dst[i]);
+      }
+   }
+
+   out->writemask = writemask;
+}
+
+
+static const struct brw_wm_ref *get_fp_src_reg_ref( struct brw_wm_compile *c,
+                                                   struct brw_fp_src src,
+                                                   GLuint i )
+{
+   return pass0_get_reg(c, src.file, src.index, BRW_GET_SWZ(src.swizzle,i));
+}
+
+
+static struct brw_wm_ref *get_new_ref( struct brw_wm_compile *c,
+                                      struct brw_fp_src src,
+                                      GLuint i,
+                                      struct brw_wm_instruction *insn)
+{
+   const struct brw_wm_ref *ref = get_fp_src_reg_ref(c, src, i);
+   struct brw_wm_ref *newref = get_ref(c);
+
+   newref->value = ref->value;
+   newref->hw_reg = ref->hw_reg;
+
+   if (insn) {
+      newref->insn = insn - c->instruction;
+      newref->prevuse = newref->value->lastuse;
+      newref->value->lastuse = newref;
+   }
+
+   if (src.negate)
+      newref->hw_reg.negate ^= 1;
+
+   if (src.abs) {
+      newref->hw_reg.negate = 0;
+      newref->hw_reg.abs = 1;
+   }
+
+   return newref;
+}
+
+
+static void
+translate_insn(struct brw_wm_compile *c,
+               const struct brw_fp_instruction *inst)
+{
+   struct brw_wm_instruction *out = get_instruction(c);
+   GLuint writemask = inst->dst.writemask;
+   GLuint nr_args = brw_wm_nr_args(inst->opcode);
+   GLuint i, j;
+
+   /* Copy some data out of the instruction
+    */
+   out->opcode = inst->opcode;
+   out->saturate = inst->dst.saturate;
+   out->tex_unit = inst->tex_unit;
+   out->target = inst->target;
+
+   /* Nasty hack:
+    */
+   out->eot = (inst->opcode == WM_FB_WRITE &&
+               inst->tex_unit != 0);
+
+
+   /* Args:
+    */
+   for (i = 0; i < nr_args; i++) {
+      for (j = 0; j < 4; j++) {
+        out->src[i][j] = get_new_ref(c, inst->src[i], j, out);
+      }
+   }
+
+   /* Dst:
+    */
+   pass0_set_dst(c, out, inst, writemask);
+}
+
+
+
+/***********************************************************************
+ * Optimize moves and swizzles away:
+ */ 
+static void pass0_precalc_mov( struct brw_wm_compile *c,
+                              const struct brw_fp_instruction *inst )
+{
+   const struct brw_fp_dst dst = inst->dst;
+   GLuint writemask = dst.writemask;
+   struct brw_wm_ref *refs[4];
+   GLuint i;
+
+   /* Get the effect of a MOV by manipulating our register table:
+    * First get all refs, then assign refs.  This ensures that "in-place"
+    * swizzles such as:
+    *   MOV t, t.xxyx
+    * are handled correctly.  Previously, these two steps were done in
+    * one loop and the above case was incorrectly handled.
+    */
+   for (i = 0; i < 4; i++) {
+      refs[i] = get_new_ref(c, inst->src[0], i, NULL);
+   }
+   for (i = 0; i < 4; i++) {
+      if (writemask & (1 << i)) {          
+         pass0_set_fpreg_ref( c, dst.file, dst.index, i, refs[i]);
+      }
+   }
+}
+
+
+/* Initialize payload "registers".
+ */
+static void pass0_init_payload( struct brw_wm_compile *c )
+{
+   GLuint i;
+
+   for (i = 0; i < 4; i++) {
+      GLuint j = i >= c->key.nr_depth_regs ? 0 : i;
+      pass0_set_fpreg_value( c, BRW_FILE_PAYLOAD, PAYLOAD_DEPTH, i, 
+                            &c->payload.depth[j] );
+   }
+
+   for (i = 0; i < c->key.nr_inputs; i++)
+      pass0_set_fpreg_value( c, BRW_FILE_PAYLOAD, i, 0, 
+                            &c->payload.input_interp[i] );      
+}
+
+
+/***********************************************************************
+ * PASS 0
+ *
+ * Work forwards to give each calculated value a unique number.  Where
+ * an instruction produces duplicate values (eg DP3), all are given
+ * the same number.
+ *
+ * Translate away swizzling and eliminate non-saturating moves.
+ *
+ * Translate instructions from our fp_instruction structs to our
+ * internal brw_wm_instruction representation.
+ */
+void brw_wm_pass0( struct brw_wm_compile *c )
+{
+   GLuint insn;
+
+   c->nr_vreg = 0;
+   c->nr_insns = 0;
+
+   pass0_init_undef(c);
+   pass0_init_payload(c);
+
+   for (insn = 0; insn < c->nr_fp_insns; insn++) {
+      const struct brw_fp_instruction *inst = &c->fp_instructions[insn];
+
+      /* Optimize away moves, otherwise emit translated instruction:
+       */      
+      switch (inst->opcode) {
+      case TGSI_OPCODE_MOV: 
+        if (!inst->dst.saturate) {
+           pass0_precalc_mov(c, inst);
+        }
+        else {
+           translate_insn(c, inst);
+        }
+        break;
+      default:
+        translate_insn(c, inst);
+        break;
+      }
+   }
+   if (BRW_DEBUG & DEBUG_WM) {
+      brw_wm_print_program(c, "pass0");
+   }
+}
diff --git a/src/gallium/drivers/i965/brw_wm_pass1.c b/src/gallium/drivers/i965/brw_wm_pass1.c
new file mode 100644 (file)
index 0000000..005747f
--- /dev/null
@@ -0,0 +1,292 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+                  
+
+#include "brw_wm.h"
+#include "brw_debug.h"
+
+
+static GLuint get_tracked_mask(struct brw_wm_compile *c,
+                              struct brw_wm_instruction *inst)
+{
+   GLuint i;
+   for (i = 0; i < 4; i++) {
+      if (inst->writemask & (1<<i)) {
+        if (!inst->dst[i]->contributes_to_output) {
+           inst->writemask &= ~(1<<i);
+           inst->dst[i] = 0;
+        }
+      }
+   }
+
+   return inst->writemask;
+}
+
+/* Remove a reference from a value's usage chain.
+ */
+static void unlink_ref(struct brw_wm_ref *ref)
+{
+   struct brw_wm_value *value = ref->value;
+
+   if (ref == value->lastuse) {
+      value->lastuse = ref->prevuse;
+   }
+   else {
+      struct brw_wm_ref *i = value->lastuse;
+      while (i->prevuse != ref) i = i->prevuse;
+      i->prevuse = ref->prevuse;
+   }
+}
+
+static void track_arg(struct brw_wm_compile *c,
+                     struct brw_wm_instruction *inst,
+                     GLuint arg,
+                     GLuint readmask)
+{
+   GLuint i;
+
+   for (i = 0; i < 4; i++) {
+      struct brw_wm_ref *ref = inst->src[arg][i];
+      if (ref) {
+        if (readmask & (1<<i)) {
+           ref->value->contributes_to_output = 1;
+         }
+        else {
+           unlink_ref(ref);
+           inst->src[arg][i] = NULL;
+        }
+      }
+   }
+}
+
+static GLuint get_texcoord_mask( GLuint tex_idx )
+{
+   switch (tex_idx) {
+   case TGSI_TEXTURE_1D:
+      return BRW_WRITEMASK_X;
+   case TGSI_TEXTURE_2D:
+   case TGSI_TEXTURE_RECT:
+      return BRW_WRITEMASK_XY;
+   case TGSI_TEXTURE_3D:
+      return BRW_WRITEMASK_XYZ;
+   case TGSI_TEXTURE_CUBE:
+      return BRW_WRITEMASK_XYZ;
+
+   case TGSI_TEXTURE_SHADOW1D:
+      return BRW_WRITEMASK_XZ;
+   case TGSI_TEXTURE_SHADOW2D:
+   case TGSI_TEXTURE_SHADOWRECT:
+      return BRW_WRITEMASK_XYZ;
+   default: 
+      assert(0);
+      return 0;
+   }
+}
+
+
+/* Step two: Basically this is dead code elimination.  
+ *
+ * Iterate backwards over instructions, noting which values
+ * contribute to the final result.  Adjust writemasks to only
+ * calculate these values.
+ */
+void brw_wm_pass1( struct brw_wm_compile *c )
+{
+   GLint insn;
+
+   for (insn = c->nr_insns-1; insn >= 0; insn--) {
+      struct brw_wm_instruction *inst = &c->instruction[insn];
+      GLuint writemask;
+      GLuint read0, read1, read2;
+
+      if (inst->opcode == TGSI_OPCODE_KIL) {
+        track_arg(c, inst, 0, BRW_WRITEMASK_XYZW); /* All args contribute to final */
+        continue;
+      }
+
+      if (inst->opcode == WM_FB_WRITE) {
+        track_arg(c, inst, 0, BRW_WRITEMASK_XYZW); 
+        track_arg(c, inst, 1, BRW_WRITEMASK_XYZW); 
+        if (c->key.source_depth_to_render_target &&
+            c->key.computes_depth)
+           track_arg(c, inst, 2, BRW_WRITEMASK_Z); 
+        else
+           track_arg(c, inst, 2, 0); 
+        continue;
+      }
+
+      /* Lookup all the registers which were written by this
+       * instruction and get a mask of those that contribute to the output:
+       */
+      writemask = get_tracked_mask(c, inst);
+      if (!writemask) {
+        GLuint arg;
+        for (arg = 0; arg < 3; arg++)
+           track_arg(c, inst, arg, 0);
+        continue;
+      }
+
+      read0 = 0;
+      read1 = 0;
+      read2 = 0;
+
+      /* Mark all inputs which contribute to the marked outputs:
+       */
+      switch (inst->opcode) {
+      case TGSI_OPCODE_ABS:
+      case TGSI_OPCODE_FLR:
+      case TGSI_OPCODE_FRC:
+      case TGSI_OPCODE_MOV:
+      case TGSI_OPCODE_TRUNC:
+        read0 = writemask;
+        break;
+
+      case TGSI_OPCODE_SUB:
+      case TGSI_OPCODE_SLT:
+      case TGSI_OPCODE_SLE:
+      case TGSI_OPCODE_SGE:
+      case TGSI_OPCODE_SGT:
+      case TGSI_OPCODE_SEQ:
+      case TGSI_OPCODE_SNE:
+      case TGSI_OPCODE_ADD:
+      case TGSI_OPCODE_MAX:
+      case TGSI_OPCODE_MIN:
+      case TGSI_OPCODE_MUL:
+        read0 = writemask;
+        read1 = writemask;
+        break;
+
+      case TGSI_OPCODE_DDX:
+      case TGSI_OPCODE_DDY:
+        read0 = writemask;
+        break;
+
+      case TGSI_OPCODE_MAD:    
+      case TGSI_OPCODE_CMP:
+      case TGSI_OPCODE_LRP:
+        read0 = writemask;
+        read1 = writemask;     
+        read2 = writemask;     
+        break;
+
+      case TGSI_OPCODE_XPD: 
+        if (writemask & BRW_WRITEMASK_X) read0 |= BRW_WRITEMASK_YZ;     
+        if (writemask & BRW_WRITEMASK_Y) read0 |= BRW_WRITEMASK_XZ;     
+        if (writemask & BRW_WRITEMASK_Z) read0 |= BRW_WRITEMASK_XY;
+        read1 = read0;
+        break;
+
+      case TGSI_OPCODE_COS:
+      case TGSI_OPCODE_EX2:
+      case TGSI_OPCODE_LG2:
+      case TGSI_OPCODE_RCP:
+      case TGSI_OPCODE_RSQ:
+      case TGSI_OPCODE_SIN:
+      case TGSI_OPCODE_SCS:
+      case WM_CINTERP:
+      case WM_PIXELXY:
+        read0 = BRW_WRITEMASK_X;
+        break;
+
+      case TGSI_OPCODE_POW:
+        read0 = BRW_WRITEMASK_X;
+        read1 = BRW_WRITEMASK_X;
+        break;
+
+      case TGSI_OPCODE_TEX:
+      case TGSI_OPCODE_TXP:
+        read0 = get_texcoord_mask(inst->target);
+        break;
+
+      case TGSI_OPCODE_TXB:
+        read0 = get_texcoord_mask(inst->target) | BRW_WRITEMASK_W;
+        break;
+
+      case WM_WPOSXY:
+        read0 = writemask & BRW_WRITEMASK_XY;
+        break;
+
+      case WM_DELTAXY:
+        read0 = writemask & BRW_WRITEMASK_XY;
+        read1 = BRW_WRITEMASK_X;
+        break;
+
+      case WM_PIXELW:
+        read0 = BRW_WRITEMASK_X;
+        read1 = BRW_WRITEMASK_XY;
+        break;
+
+      case WM_LINTERP:
+        read0 = BRW_WRITEMASK_X;
+        read1 = BRW_WRITEMASK_XY;
+        break;
+
+      case WM_PINTERP:
+        read0 = BRW_WRITEMASK_X; /* interpolant */
+        read1 = BRW_WRITEMASK_XY; /* deltas */
+        read2 = BRW_WRITEMASK_W; /* pixel w */
+        break;
+
+      case TGSI_OPCODE_DP3:    
+        read0 = BRW_WRITEMASK_XYZ;
+        read1 = BRW_WRITEMASK_XYZ;
+        break;
+
+      case TGSI_OPCODE_DPH:
+        read0 = BRW_WRITEMASK_XYZ;
+        read1 = BRW_WRITEMASK_XYZW;
+        break;
+
+      case TGSI_OPCODE_DP4:
+        read0 = BRW_WRITEMASK_XYZW;
+        read1 = BRW_WRITEMASK_XYZW;
+        break;
+
+      case TGSI_OPCODE_LIT: 
+        read0 = BRW_WRITEMASK_XYW;
+        break;
+
+      case TGSI_OPCODE_DST:
+      case WM_FRONTFACING:
+      case TGSI_OPCODE_KILP:
+      default:
+        break;
+      }
+
+      track_arg(c, inst, 0, read0);
+      track_arg(c, inst, 1, read1);
+      track_arg(c, inst, 2, read2);
+   }
+
+   if (BRW_DEBUG & DEBUG_WM) {
+      brw_wm_print_program(c, "pass1");
+   }
+}
diff --git a/src/gallium/drivers/i965/brw_wm_pass2.c b/src/gallium/drivers/i965/brw_wm_pass2.c
new file mode 100644 (file)
index 0000000..19248b4
--- /dev/null
@@ -0,0 +1,334 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+                   
+
+#include "brw_debug.h"
+#include "brw_wm.h"
+
+
+/* Use these to force spilling so that that functionality can be
+ * tested with known-good examples rather than having to construct new
+ * tests.
+ */
+#define TEST_PAYLOAD_SPILLS 0
+#define TEST_DST_SPILLS 0
+
+static void spill_value(struct brw_wm_compile *c,
+                       struct brw_wm_value *value);
+
+static void prealloc_reg(struct brw_wm_compile *c,
+                        struct brw_wm_value *value,
+                        GLuint reg)
+{
+   if (value->lastuse) {
+      /* Set nextuse to zero, it will be corrected by
+       * update_register_usage().
+       */
+      c->pass2_grf[reg].value = value;
+      c->pass2_grf[reg].nextuse = 0;
+
+      value->resident = &c->pass2_grf[reg];
+      value->hw_reg = brw_vec8_grf(reg*2, 0);
+
+      if (TEST_PAYLOAD_SPILLS)
+        spill_value(c, value);
+   }
+}
+
+
+/* Initialize all the register values.  Do the initial setup
+ * calculations for interpolants.
+ */
+static void init_registers( struct brw_wm_compile *c )
+{
+   GLuint reg = 0;
+   GLuint j;
+
+   for (j = 0; j < c->grf_limit; j++) 
+      c->pass2_grf[j].nextuse = BRW_WM_MAX_INSN;
+
+   /* Pre-allocate incoming payload regs:
+    */
+   for (j = 0; j < c->key.nr_depth_regs; j++) 
+      prealloc_reg(c, &c->payload.depth[j], reg++);
+
+   for (j = 0; j < c->nr_creg; j++) 
+      prealloc_reg(c, &c->creg[j], reg++);
+
+   reg++;                       /* XXX: skip over position output */
+
+   /* XXX: currently just hope the VS outputs line up with FS inputs:
+    */
+   for (j = 0; j < c->key.nr_inputs; j++)
+      prealloc_reg(c, &c->payload.input_interp[j], reg++);
+
+   c->prog_data.first_curbe_grf = c->key.nr_depth_regs * 2;
+   c->prog_data.urb_read_length = (c->key.nr_inputs + 1) * 2;
+   c->prog_data.curb_read_length = c->nr_creg * 2;
+
+   /* Note this allocation:
+    */
+   c->max_wm_grf = reg * 2;
+}
+
+
+/* Update the nextuse value for each register in our file.
+ */
+static void update_register_usage(struct brw_wm_compile *c,
+                                 GLuint thisinsn)
+{
+   GLuint i;
+
+   for (i = 1; i < c->grf_limit; i++) {
+      struct brw_wm_grf *grf = &c->pass2_grf[i];
+
+      /* Only search those which can change:
+       */
+      if (grf->nextuse < thisinsn) {
+        const struct brw_wm_ref *ref = grf->value->lastuse;
+
+        /* Has last use of value been passed?
+         */
+        if (ref->insn < thisinsn) {
+           grf->value->resident = 0;
+           grf->value = 0;
+           grf->nextuse = BRW_WM_MAX_INSN;
+        }
+        else {
+           /* Else loop through chain to update:
+            */
+           while (ref->prevuse && ref->prevuse->insn >= thisinsn)
+              ref = ref->prevuse;
+
+           grf->nextuse = ref->insn;
+        }
+      }
+   }
+}
+
+
+static void spill_value(struct brw_wm_compile *c,
+                       struct brw_wm_value *value)
+{      
+   /* Allocate a spill slot.  Note that allocations start from 0x40 -
+    * the first slot is reserved to mean "undef" in brw_wm_emit.c
+    */
+   if (!value->spill_slot) {
+      c->last_scratch += 0x40; 
+      value->spill_slot = c->last_scratch;
+   }
+
+   /* The spill will be done in brw_wm_emit.c immediately after the
+    * value is calculated, so we can just take this reg without any
+    * further work.
+    */
+   value->resident->value = NULL;
+   value->resident->nextuse = BRW_WM_MAX_INSN;
+   value->resident = NULL;
+}
+
+
+
+/* Search for contiguous region with the most distant nearest
+ * member.  Free regs count as very distant.
+ *
+ * TODO: implement spill-to-reg so that we can rearrange discontigous
+ * free regs and then spill the oldest non-free regs in sequence.
+ * This would mean inserting instructions in this pass.
+ */
+static GLuint search_contiguous_regs(struct brw_wm_compile *c,
+                                    GLuint nr,
+                                    GLuint thisinsn)
+{
+   struct brw_wm_grf *grf = c->pass2_grf;
+   GLuint furthest = 0;
+   GLuint reg = 0;
+   GLuint i, j;
+
+   /* Start search at 1: r0 is special and can't be used or spilled.
+    */
+   for (i = 1; i < c->grf_limit && furthest < BRW_WM_MAX_INSN; i++) {
+      GLuint group_nextuse = BRW_WM_MAX_INSN;
+
+      for (j = 0; j < nr; j++) {
+        if (grf[i+j].nextuse < group_nextuse)
+           group_nextuse = grf[i+j].nextuse;
+      }
+
+      if (group_nextuse > furthest) {
+        furthest = group_nextuse;
+        reg = i;
+      }
+   }
+
+   assert(furthest != thisinsn);
+
+   /* Any non-empty regs will need to be spilled:
+    */
+   for (j = 0; j < nr; j++) 
+      if (grf[reg+j].value)
+        spill_value(c, grf[reg+j].value);
+
+   return reg;
+}
+
+
+static void alloc_contiguous_dest(struct brw_wm_compile *c, 
+                                 struct brw_wm_value *dst[],
+                                 GLuint nr,
+                                 GLuint thisinsn)
+{
+   GLuint reg = search_contiguous_regs(c, nr, thisinsn);
+   GLuint i;
+
+   for (i = 0; i < nr; i++) {
+      if (!dst[i]) {
+        /* Need to grab a dummy value in TEX case.  Don't introduce
+         * it into the tracking scheme.
+         */
+        dst[i] = &c->vreg[c->nr_vreg++];
+      }
+      else {
+        assert(!dst[i]->resident);
+        assert(c->pass2_grf[reg+i].nextuse != thisinsn);
+
+        c->pass2_grf[reg+i].value = dst[i];
+        c->pass2_grf[reg+i].nextuse = thisinsn;
+
+        dst[i]->resident = &c->pass2_grf[reg+i];
+      }
+
+      dst[i]->hw_reg = brw_vec8_grf((reg+i)*2, 0);
+   }
+
+   if ((reg+nr)*2 > c->max_wm_grf)
+      c->max_wm_grf = (reg+nr) * 2;
+}
+
+
+static void load_args(struct brw_wm_compile *c, 
+                     struct brw_wm_instruction *inst)
+{
+   GLuint thisinsn = inst - c->instruction;
+   GLuint i,j;
+
+   for (i = 0; i < 3; i++) {
+      for (j = 0; j < 4; j++) {
+        struct brw_wm_ref *ref = inst->src[i][j];
+
+        if (ref) {
+           if (!ref->value->resident) {
+              /* Need to bring the value in from scratch space.  The code for
+               * this will be done in brw_wm_emit.c, here we just do the
+               * register allocation and mark the ref as requiring a fill.
+               */
+              GLuint reg = search_contiguous_regs(c, 1, thisinsn);
+
+              c->pass2_grf[reg].value = ref->value;
+              c->pass2_grf[reg].nextuse = thisinsn;
+
+              ref->value->resident = &c->pass2_grf[reg];
+
+              /* Note that a fill is required:
+               */
+              ref->unspill_reg = reg*2;
+           }
+
+           /* Adjust the hw_reg to point at the value's current location:
+            */
+           assert(ref->value == ref->value->resident->value);
+           ref->hw_reg.nr += (ref->value->resident - c->pass2_grf) * 2;
+        }
+      }
+   }
+}
+
+
+
+/* Step 3: Work forwards once again.  Perform register allocations,
+ * taking into account instructions like TEX which require contiguous
+ * result registers.  Where necessary spill registers to scratch space
+ * and reload later.
+ */
+void brw_wm_pass2( struct brw_wm_compile *c )
+{
+   GLuint insn;
+   GLuint i;
+
+   init_registers(c);
+
+   for (insn = 0; insn < c->nr_insns; insn++) {
+      struct brw_wm_instruction *inst = &c->instruction[insn];
+
+      /* Update registers' nextuse values:
+       */
+      update_register_usage(c, insn);
+
+      /* May need to unspill some args.
+       */
+      load_args(c, inst);
+
+      /* Allocate registers to hold results:
+       */
+      switch (inst->opcode) {
+      case TGSI_OPCODE_TEX:
+      case TGSI_OPCODE_TXB:
+      case TGSI_OPCODE_TXP:
+        alloc_contiguous_dest(c, inst->dst, 4, insn);
+        break;
+
+      default:
+        for (i = 0; i < 4; i++) {
+           if (inst->writemask & (1<<i)) {
+              assert(inst->dst[i]);
+              alloc_contiguous_dest(c, &inst->dst[i], 1, insn);
+           }
+        }
+        break;
+      }
+
+      if (TEST_DST_SPILLS && inst->opcode != WM_PIXELXY) {
+        for (i = 0; i < 4; i++)        
+           if (inst->dst[i])
+              spill_value(c, inst->dst[i]);
+      }
+   }
+
+   if (BRW_DEBUG & DEBUG_WM) {
+      brw_wm_print_program(c, "pass2");
+   }
+
+   c->state = PASS2_DONE;
+
+   if (BRW_DEBUG & DEBUG_WM) {
+       brw_wm_print_program(c, "pass2/done");
+   }
+}
diff --git a/src/gallium/drivers/i965/brw_wm_sampler_state.c b/src/gallium/drivers/i965/brw_wm_sampler_state.c
new file mode 100644 (file)
index 0000000..a8bc31c
--- /dev/null
@@ -0,0 +1,229 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+                   
+#include "util/u_math.h"
+#include "util/u_format.h"
+
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+#include "brw_screen.h"
+
+
+/* Samplers aren't strictly wm state from the hardware's perspective,
+ * but that is the only situation in which we use them in this driver.
+ */
+
+
+
+static enum pipe_error
+upload_default_color( struct brw_context *brw,
+                     const GLfloat *color,
+                      struct brw_winsys_buffer **bo_out )
+{
+   struct brw_sampler_default_color sdc;
+   enum pipe_error ret;
+
+   COPY_4V(sdc.color, color); 
+   
+   ret = brw_cache_data( &brw->cache, BRW_SAMPLER_DEFAULT_COLOR, &sdc,
+                         NULL, 0, bo_out );
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+struct wm_sampler_key {
+   int sampler_count;
+   struct brw_sampler_state sampler[BRW_MAX_TEX_UNIT];
+};
+
+
+/** Sets up the cache key for sampler state for all texture units */
+static void
+brw_wm_sampler_populate_key(struct brw_context *brw,
+                           struct wm_sampler_key *key)
+{
+   int i;
+
+   memset(key, 0, sizeof(*key));
+
+   key->sampler_count = MIN2(brw->curr.num_textures,
+                           brw->curr.num_samplers);
+
+   for (i = 0; i < key->sampler_count; i++) {
+      const struct brw_texture *tex = brw_texture(brw->curr.texture[i]);
+      const struct brw_sampler *sampler = brw->curr.sampler[i];
+      struct brw_sampler_state *entry = &key->sampler[i];
+
+      entry->ss0 = sampler->ss0;
+      entry->ss1 = sampler->ss1;
+      entry->ss2.default_color_pointer = 0; /* reloc */
+      entry->ss3 = sampler->ss3;
+
+      /* Cube-maps on 965 and later must use the same wrap mode for all 3
+       * coordinate dimensions.  Futher, only CUBE and CLAMP are valid.
+       */
+      if (tex->base.target == PIPE_TEXTURE_CUBE) {
+        if (FALSE &&
+            (sampler->ss0.min_filter != BRW_MAPFILTER_NEAREST || 
+             sampler->ss0.mag_filter != BRW_MAPFILTER_NEAREST)) {
+           entry->ss1.r_wrap_mode = BRW_TEXCOORDMODE_CUBE;
+           entry->ss1.s_wrap_mode = BRW_TEXCOORDMODE_CUBE;
+           entry->ss1.t_wrap_mode = BRW_TEXCOORDMODE_CUBE;
+        } else {
+           entry->ss1.r_wrap_mode = BRW_TEXCOORDMODE_CLAMP;
+           entry->ss1.s_wrap_mode = BRW_TEXCOORDMODE_CLAMP;
+           entry->ss1.t_wrap_mode = BRW_TEXCOORDMODE_CLAMP;
+        }
+      } else if (tex->base.target == PIPE_TEXTURE_1D) {
+        /* There's a bug in 1D texture sampling - it actually pays
+         * attention to the wrap_t value, though it should not.
+         * Override the wrap_t value here to GL_REPEAT to keep
+         * any nonexistent border pixels from floating in.
+         */
+        entry->ss1.t_wrap_mode = BRW_TEXCOORDMODE_WRAP;
+      }
+   }
+}
+
+
+static enum pipe_error
+brw_wm_sampler_update_default_colors(struct brw_context *brw)
+{
+   enum pipe_error ret;
+   int nr = MIN2(brw->curr.num_textures,
+                brw->curr.num_samplers);
+   int i;
+
+   for (i = 0; i < nr; i++) {
+      const struct brw_texture *tex = brw_texture(brw->curr.texture[i]);
+      const struct brw_sampler *sampler = brw->curr.sampler[i];
+      const float *bc;
+
+      if (util_format_is_depth_or_stencil(tex->base.format)) {
+        float bordercolor[4] = {
+           sampler->border_color[0],
+           sampler->border_color[0],
+           sampler->border_color[0],
+           sampler->border_color[0]
+        };
+         
+         bc = bordercolor;
+      }
+      else {
+         bc = sampler->border_color;
+      }
+
+      /* GL specs that border color for depth textures is taken from the
+       * R channel, while the hardware uses A.  Spam R into all the
+       * channels for safety.
+       */
+      ret = upload_default_color(brw, 
+                                 bc,
+                                 &brw->wm.sdc_bo[i]);
+      if (ret) 
+         return ret;
+   }
+
+   return PIPE_OK;
+}
+
+
+
+/* All samplers must be uploaded in a single contiguous array.  
+ */
+static int upload_wm_samplers( struct brw_context *brw )
+{
+   struct wm_sampler_key key;
+   struct brw_winsys_reloc reloc[BRW_MAX_TEX_UNIT];
+   enum pipe_error ret;
+   int i;
+
+   brw_wm_sampler_update_default_colors(brw);
+   brw_wm_sampler_populate_key(brw, &key);
+
+   if (brw->wm.sampler_count != key.sampler_count) {
+      brw->wm.sampler_count = key.sampler_count;
+      brw->state.dirty.cache |= CACHE_NEW_SAMPLER;
+   }
+
+   if (brw->wm.sampler_count == 0) {
+      bo_reference(&brw->wm.sampler_bo, NULL);
+      return PIPE_OK;
+   }
+
+   /* Emit SDC relocations */
+   for (i = 0; i < key.sampler_count; i++) {
+      make_reloc( &reloc[i],
+                  BRW_USAGE_SAMPLER,
+                  0,
+                  i * sizeof(struct brw_sampler_state) +
+                  offsetof(struct brw_sampler_state, ss2),
+                  brw->wm.sdc_bo[i]);
+   }
+
+
+   if (brw_search_cache(&brw->cache, BRW_SAMPLER,
+                        &key, sizeof(key),
+                        reloc, key.sampler_count,
+                        NULL,
+                        &brw->wm.sampler_bo))
+      return PIPE_OK;
+
+   /* If we didnt find it in the cache, compute the state and put it in the
+    * cache.
+    */
+   ret = brw_upload_cache(&brw->cache, BRW_SAMPLER,
+                          &key, sizeof(key),
+                          reloc, key.sampler_count,
+                          &key.sampler, sizeof(key.sampler),
+                          NULL, NULL,
+                          &brw->wm.sampler_bo);
+   if (ret)
+      return ret;
+
+
+   return 0;
+}
+
+const struct brw_tracked_state brw_wm_samplers = {
+   .dirty = {
+      .mesa = PIPE_NEW_BOUND_TEXTURES | PIPE_NEW_SAMPLERS,
+      .brw = 0,
+      .cache = 0
+   },
+   .prepare = upload_wm_samplers,
+};
+
+
diff --git a/src/gallium/drivers/i965/brw_wm_state.c b/src/gallium/drivers/i965/brw_wm_state.c
new file mode 100644 (file)
index 0000000..ee970ac
--- /dev/null
@@ -0,0 +1,339 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+                   
+#include "util/u_math.h"
+
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+#include "brw_wm.h"
+#include "brw_debug.h"
+#include "brw_pipe_rast.h"
+
+/***********************************************************************
+ * WM unit - fragment programs and rasterization
+ */
+
+struct brw_wm_unit_key {
+   unsigned int total_grf, total_scratch;
+   unsigned int urb_entry_read_length;
+   unsigned int curb_entry_read_length;
+   unsigned int dispatch_grf_start_reg;
+
+   unsigned int curbe_offset;
+   unsigned int urb_size;
+
+   unsigned int max_threads;
+
+   unsigned int nr_surfaces, sampler_count;
+   GLboolean uses_depth, computes_depth, uses_kill, has_flow_control;
+   GLboolean polygon_stipple, stats_wm, line_stipple, offset_enable;
+   GLfloat offset_units, offset_factor;
+};
+
+static void
+wm_unit_populate_key(struct brw_context *brw, struct brw_wm_unit_key *key)
+{
+   const struct brw_fragment_shader *fp = brw->curr.fragment_shader;
+
+   memset(key, 0, sizeof(*key));
+
+   if (BRW_DEBUG & DEBUG_SINGLE_THREAD)
+      key->max_threads = 1;
+   else {
+      /* WM maximum threads is number of EUs times number of threads per EU. */
+      if (BRW_IS_IGDNG(brw))
+         key->max_threads = 12 * 6;
+      else if (BRW_IS_G4X(brw))
+        key->max_threads = 10 * 5;
+      else
+        key->max_threads = 8 * 4;
+   }
+
+   /* CACHE_NEW_WM_PROG */
+   key->total_grf = brw->wm.prog_data->total_grf;
+   key->urb_entry_read_length = brw->wm.prog_data->urb_read_length;
+   key->curb_entry_read_length = brw->wm.prog_data->curb_read_length;
+   key->dispatch_grf_start_reg = brw->wm.prog_data->first_curbe_grf;
+   key->total_scratch = align(brw->wm.prog_data->total_scratch, 1024);
+
+   /* BRW_NEW_URB_FENCE */
+   key->urb_size = brw->urb.vsize;
+
+   /* BRW_NEW_CURBE_OFFSETS */
+   key->curbe_offset = brw->curbe.wm_start;
+
+   /* BRW_NEW_NR_SURFACEs */
+   key->nr_surfaces = brw->wm.nr_surfaces;
+
+   /* CACHE_NEW_SAMPLER */
+   key->sampler_count = brw->wm.sampler_count;
+
+   /* PIPE_NEW_RAST */
+   key->polygon_stipple = brw->curr.rast->templ.poly_stipple_enable;
+
+   /* PIPE_NEW_FRAGMENT_PROGRAM */
+   key->uses_depth = fp->uses_depth;
+   key->computes_depth = fp->info.writes_z;
+
+   /* PIPE_NEW_DEPTH_BUFFER
+    *
+    * Override for NULL depthbuffer case, required by the Pixel Shader Computed
+    * Depth field.
+    */
+   if (brw->curr.fb.zsbuf == NULL)
+      key->computes_depth = 0;
+
+   /* PIPE_NEW_DEPTH_STENCIL_ALPHA */
+   key->uses_kill = (fp->info.uses_kill || 
+                    brw->curr.zstencil->cc3.alpha_test);
+
+   key->has_flow_control = fp->has_flow_control;
+
+   /* temporary sanity check assertion */
+   assert(fp->has_flow_control == 0);
+
+   /* PIPE_NEW_QUERY */
+   key->stats_wm = (brw->query.stats_wm != 0);
+
+   /* PIPE_NEW_RAST */
+   key->line_stipple = brw->curr.rast->templ.line_stipple_enable;
+
+
+   key->offset_enable = (brw->curr.rast->templ.offset_cw ||
+                        brw->curr.rast->templ.offset_ccw);
+
+   key->offset_units = brw->curr.rast->templ.offset_units;
+   key->offset_factor = brw->curr.rast->templ.offset_scale;
+}
+
+/**
+ * Setup wm hardware state.  See page 225 of Volume 2
+ */
+static enum pipe_error
+wm_unit_create_from_key(struct brw_context *brw, struct brw_wm_unit_key *key,
+                       struct brw_winsys_reloc *reloc,
+                        unsigned nr_reloc,
+                        struct brw_winsys_buffer **bo_out)
+{
+   struct brw_wm_unit_state wm;
+   enum pipe_error ret;
+
+   memset(&wm, 0, sizeof(wm));
+
+   wm.thread0.grf_reg_count = align(key->total_grf, 16) / 16 - 1;
+   wm.thread0.kernel_start_pointer = 0; /* reloc */
+   wm.thread1.depth_coef_urb_read_offset = 1;
+   wm.thread1.floating_point_mode = BRW_FLOATING_POINT_NON_IEEE_754;
+
+   if (BRW_IS_IGDNG(brw))
+      wm.thread1.binding_table_entry_count = 0; /* hardware requirement */
+   else
+      wm.thread1.binding_table_entry_count = key->nr_surfaces;
+
+   if (key->total_scratch != 0) {
+      wm.thread2.scratch_space_base_pointer = 0; /* reloc */
+      wm.thread2.per_thread_scratch_space = key->total_scratch / 1024 - 1;
+   } else {
+      wm.thread2.scratch_space_base_pointer = 0;
+      wm.thread2.per_thread_scratch_space = 0;
+   }
+
+   wm.thread3.dispatch_grf_start_reg = key->dispatch_grf_start_reg;
+   wm.thread3.urb_entry_read_length = key->urb_entry_read_length;
+   wm.thread3.urb_entry_read_offset = 0;
+   wm.thread3.const_urb_entry_read_length = key->curb_entry_read_length;
+   wm.thread3.const_urb_entry_read_offset = key->curbe_offset * 2;
+
+   if (BRW_IS_IGDNG(brw)) 
+      wm.wm4.sampler_count = 0; /* hardware requirement */
+   else
+      wm.wm4.sampler_count = (key->sampler_count + 1) / 4;
+
+   /* reloc */
+   wm.wm4.sampler_state_pointer = 0;
+
+   wm.wm5.program_uses_depth = key->uses_depth;
+   wm.wm5.program_computes_depth = key->computes_depth;
+   wm.wm5.program_uses_killpixel = key->uses_kill;
+
+   if (key->has_flow_control)
+      wm.wm5.enable_8_pix = 1;
+   else
+      wm.wm5.enable_16_pix = 1;
+
+   wm.wm5.max_threads = key->max_threads - 1;
+   wm.wm5.thread_dispatch_enable = 1;  /* AKA: color_write */
+   wm.wm5.legacy_line_rast = 0;
+   wm.wm5.legacy_global_depth_bias = 0;
+   wm.wm5.early_depth_test = 1;                /* never need to disable */
+   wm.wm5.line_aa_region_width = 0;
+   wm.wm5.line_endcap_aa_region_width = 1;
+
+   wm.wm5.polygon_stipple = key->polygon_stipple;
+
+   if (key->offset_enable) {
+      wm.wm5.depth_offset = 1;
+      /* Something wierd going on with legacy_global_depth_bias,
+       * offset_constant, scaling and MRD.  This value passes glean
+       * but gives some odd results elsewere (eg. the
+       * quad-offset-units test).
+       */
+      wm.global_depth_offset_constant = key->offset_units * 2;
+
+      /* This is the only value that passes glean:
+       */
+      wm.global_depth_offset_scale = key->offset_factor;
+   }
+
+   wm.wm5.line_stipple = key->line_stipple;
+
+   if ((BRW_DEBUG & DEBUG_STATS) || key->stats_wm)
+      wm.wm4.stats_enable = 1;
+
+   ret = brw_upload_cache(&brw->cache, BRW_WM_UNIT,
+                          key, sizeof(*key),
+                          reloc, nr_reloc,
+                          &wm, sizeof(wm),
+                          NULL, NULL,
+                          bo_out);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+static enum pipe_error upload_wm_unit( struct brw_context *brw )
+{
+   struct brw_wm_unit_key key;
+   struct brw_winsys_reloc reloc[3];
+   unsigned nr_reloc = 0;
+   enum pipe_error ret;
+   unsigned grf_reg_count;
+   unsigned per_thread_scratch_space;
+   unsigned stats_enable;
+   unsigned sampler_count;
+
+   wm_unit_populate_key(brw, &key);
+
+
+   /* Allocate the necessary scratch space if we haven't already.  Don't
+    * bother reducing the allocation later, since we use scratch so
+    * rarely.
+    */
+   assert(key.total_scratch <= 12 * 1024);
+   if (key.total_scratch) {
+      GLuint total = key.total_scratch * key.max_threads;
+
+      /* Do we need a new buffer:
+       */
+      if (brw->wm.scratch_bo && total > brw->wm.scratch_bo->size) 
+        bo_reference(&brw->wm.scratch_bo, NULL);
+
+      if (brw->wm.scratch_bo == NULL) {
+        ret = brw->sws->bo_alloc(brw->sws,
+                                  BRW_BUFFER_TYPE_SHADER_SCRATCH,
+                                  total,
+                                  4096,
+                                  &brw->wm.scratch_bo);
+         if (ret)
+            return ret;
+      }
+   }
+
+
+   /* XXX: temporary:
+    */
+   grf_reg_count = (align(key.total_grf, 16) / 16 - 1);
+   per_thread_scratch_space = key.total_scratch / 1024 - 1;
+   stats_enable = (BRW_DEBUG & DEBUG_STATS) || key.stats_wm;
+   sampler_count = BRW_IS_IGDNG(brw) ? 0 :(key.sampler_count + 1) / 4;
+
+   /* Emit WM program relocation */
+   make_reloc(&reloc[nr_reloc++],
+              BRW_USAGE_STATE,
+              grf_reg_count << 1,
+              offsetof(struct brw_wm_unit_state, thread0),
+              brw->wm.prog_bo);
+
+   /* Emit scratch space relocation */
+   if (key.total_scratch != 0) {
+      make_reloc(&reloc[nr_reloc++],
+                 BRW_USAGE_SCRATCH,
+                 per_thread_scratch_space,
+                 offsetof(struct brw_wm_unit_state, thread2),
+                 brw->wm.scratch_bo);
+   }
+
+   /* Emit sampler state relocation */
+   if (key.sampler_count != 0) {
+      make_reloc(&reloc[nr_reloc++],
+                 BRW_USAGE_STATE,
+                 stats_enable | (sampler_count << 2),
+                 offsetof(struct brw_wm_unit_state, wm4),
+                 brw->wm.sampler_bo);
+   }
+
+
+   if (brw_search_cache(&brw->cache, BRW_WM_UNIT,
+                        &key, sizeof(key),
+                        reloc, nr_reloc,
+                        NULL,
+                        &brw->wm.state_bo))
+      return PIPE_OK;
+
+   ret = wm_unit_create_from_key(brw, &key, 
+                                 reloc, nr_reloc,
+                                 &brw->wm.state_bo);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+const struct brw_tracked_state brw_wm_unit = {
+   .dirty = {
+      .mesa = (PIPE_NEW_FRAGMENT_SHADER |
+              PIPE_NEW_DEPTH_BUFFER |
+              PIPE_NEW_RAST | 
+              PIPE_NEW_DEPTH_STENCIL_ALPHA |
+              PIPE_NEW_QUERY),
+
+      .brw = (BRW_NEW_CURBE_OFFSETS |
+             BRW_NEW_NR_WM_SURFACES),
+
+      .cache = (CACHE_NEW_WM_PROG |
+               CACHE_NEW_SAMPLER)
+   },
+   .prepare = upload_wm_unit,
+};
+
diff --git a/src/gallium/drivers/i965/brw_wm_surface_state.c b/src/gallium/drivers/i965/brw_wm_surface_state.c
new file mode 100644 (file)
index 0000000..f92b819
--- /dev/null
@@ -0,0 +1,294 @@
+/*
+ Copyright (C) Intel Corp.  2006.  All Rights Reserved.
+ Intel funded Tungsten Graphics (http://www.tungstengraphics.com) to
+ develop this 3D driver.
+ Permission is hereby granted, free of charge, to any person obtaining
+ a copy of this software and associated documentation files (the
+ "Software"), to deal in the Software without restriction, including
+ without limitation the rights to use, copy, modify, merge, publish,
+ distribute, sublicense, and/or sell copies of the Software, and to
+ permit persons to whom the Software is furnished to do so, subject to
+ the following conditions:
+ The above copyright notice and this permission notice (including the
+ next paragraph) shall be included in all copies or substantial
+ portions of the Software.
+ THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ **********************************************************************/
+ /*
+  * Authors:
+  *   Keith Whitwell <keith@tungstengraphics.com>
+  */
+                   
+#include "pipe/p_format.h"
+
+#include "brw_batchbuffer.h"
+#include "brw_context.h"
+#include "brw_state.h"
+#include "brw_defines.h"
+#include "brw_screen.h"
+
+
+
+
+static enum pipe_error
+brw_update_texture_surface( struct brw_context *brw,
+                           struct brw_texture *tex,
+                            struct brw_winsys_buffer **bo_out)
+{
+   struct brw_winsys_reloc reloc[1];
+   enum pipe_error ret;
+
+   /* Emit relocation to surface contents */
+   make_reloc(&reloc[0],
+              BRW_USAGE_SAMPLER,
+              0,
+              offsetof(struct brw_surface_state, ss1),
+              tex->bo);
+
+   if (brw_search_cache(&brw->surface_cache,
+                        BRW_SS_SURFACE,
+                        &tex->ss, sizeof tex->ss,
+                        reloc, Elements(reloc),
+                        NULL,
+                        bo_out))
+      return PIPE_OK;
+
+   ret = brw_upload_cache(&brw->surface_cache, BRW_SS_SURFACE,
+                          &tex->ss, sizeof tex->ss,
+                          reloc, Elements(reloc),
+                          &tex->ss, sizeof tex->ss,
+                          NULL, NULL,
+                          bo_out);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+
+
+
+
+
+
+/**
+ * Sets up a surface state structure to point at the given region.
+ * While it is only used for the front/back buffer currently, it should be
+ * usable for further buffers when doing ARB_draw_buffer support.
+ */
+static enum pipe_error
+brw_update_render_surface(struct brw_context *brw,
+                          struct brw_surface *surface,
+                          struct brw_winsys_buffer **bo_out)
+{
+   struct brw_surf_ss0 blend_ss0 = brw->curr.blend->ss0;
+   struct brw_surface_state ss;
+   struct brw_winsys_reloc reloc[1];
+   enum pipe_error ret;
+
+   /* XXX: we will only be rendering to this surface:
+    */
+   make_reloc(&reloc[0],
+              BRW_USAGE_RENDER_TARGET,
+              0,
+              offsetof(struct brw_surface_state, ss1),
+              surface->bo);
+
+   /* Surfaces are potentially shared between contexts, so can't
+    * scribble the in-place ss0 value in the surface.
+    */
+   memcpy(&ss, &surface->ss, sizeof ss);
+
+   ss.ss0.color_blend        = blend_ss0.color_blend;
+   ss.ss0.writedisable_blue  = blend_ss0.writedisable_blue;
+   ss.ss0.writedisable_green = blend_ss0.writedisable_green;
+   ss.ss0.writedisable_red   = blend_ss0.writedisable_red;
+   ss.ss0.writedisable_alpha = blend_ss0.writedisable_alpha;
+
+   if (brw_search_cache(&brw->surface_cache,
+                        BRW_SS_SURFACE,
+                        &ss, sizeof(ss),
+                        reloc, Elements(reloc),
+                        NULL,
+                        bo_out))
+      return PIPE_OK;
+       
+   ret = brw_upload_cache(&brw->surface_cache,
+                          BRW_SS_SURFACE,
+                          &ss, sizeof ss,
+                          reloc, Elements(reloc),
+                          &ss, sizeof ss,
+                          NULL, NULL,
+                          bo_out);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+
+/**
+ * Constructs the binding table for the WM surface state, which maps unit
+ * numbers to surface state objects.
+ */
+static enum pipe_error
+brw_wm_get_binding_table(struct brw_context *brw,
+                         struct brw_winsys_buffer **bo_out )
+{
+   enum pipe_error ret;
+   struct brw_winsys_reloc reloc[BRW_WM_MAX_SURF];
+   uint32_t data[BRW_WM_MAX_SURF];
+   GLuint nr_relocs = 0;
+   GLuint data_size = brw->wm.nr_surfaces * sizeof data[0];
+   int i;
+
+   assert(brw->wm.nr_surfaces <= BRW_WM_MAX_SURF);
+   assert(brw->wm.nr_surfaces > 0);
+
+   /* Emit binding table relocations to surface state 
+    */
+   for (i = 0; i < brw->wm.nr_surfaces; i++) {
+      if (brw->wm.surf_bo[i]) {
+         make_reloc(&reloc[nr_relocs++],
+                    BRW_USAGE_STATE,
+                    0,
+                    i * sizeof(GLuint),
+                    brw->wm.surf_bo[i]);
+      }
+   }
+
+   /* Note there is no key for this search beyond the values in the
+    * relocation array:
+    */
+   if (brw_search_cache(&brw->surface_cache, BRW_SS_SURF_BIND,
+                        NULL, 0,
+                        reloc, nr_relocs,
+                        NULL,
+                        bo_out))
+      return PIPE_OK;
+
+   /* Upload zero data, will all be overwitten with relocation
+    * offsets:
+    */
+   for (i = 0; i < brw->wm.nr_surfaces; i++)
+      data[i] = 0;
+
+   ret = brw_upload_cache( &brw->surface_cache, BRW_SS_SURF_BIND,
+                           NULL, 0,
+                           reloc, nr_relocs,
+                           data, data_size,
+                           NULL, NULL,
+                           bo_out);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+static enum pipe_error prepare_wm_surfaces(struct brw_context *brw )
+{
+   enum pipe_error ret;
+   int nr_surfaces = 0;
+   GLuint i;
+
+   /* PIPE_NEW_COLOR_BUFFERS | PIPE_NEW_BLEND
+    *
+    * Update surfaces for drawing buffers.  Mixes in colormask and
+    * blend state.
+    *
+    * XXX: no color buffer case
+    */
+   for (i = 0; i < brw->curr.fb.nr_cbufs; i++) {
+      ret = brw_update_render_surface(brw, 
+                                      brw_surface(brw->curr.fb.cbufs[i]), 
+                                      &brw->wm.surf_bo[BTI_COLOR_BUF(i)]);
+      if (ret)
+         return ret;
+      
+      nr_surfaces = BTI_COLOR_BUF(i) + 1;
+   }
+
+
+
+   /* PIPE_NEW_FRAGMENT_CONSTANTS
+    */
+#if 0
+   if (brw->curr.fragment_constants) {
+      ret = brw_update_fragment_constant_surface(
+         brw, 
+         brw->curr.fragment_constants, 
+         &brw->wm.surf_bo[BTI_FRAGMENT_CONSTANTS]);
+
+      if (ret)
+         return ret;
+
+      nr_surfaces = BTI_FRAGMENT_CONSTANTS + 1;
+   }
+   else {
+      bo_reference(&brw->wm.surf_bo[SURF_FRAG_CONSTANTS], NULL);      
+   }
+#endif
+
+
+   /* PIPE_NEW_TEXTURE 
+    */
+   for (i = 0; i < brw->curr.num_textures; i++) {
+      ret = brw_update_texture_surface(brw, 
+                                       brw_texture(brw->curr.texture[i]),
+                                       &brw->wm.surf_bo[BTI_TEXTURE(i)]);
+      if (ret)
+         return ret;
+
+      nr_surfaces = BTI_TEXTURE(i) + 1;
+   }
+
+   /* Clear any inactive entries:
+    */
+   for (i = brw->curr.fb.nr_cbufs; i < BRW_MAX_DRAW_BUFFERS; i++) 
+      bo_reference(&brw->wm.surf_bo[BTI_COLOR_BUF(i)], NULL);
+
+   if (!brw->curr.fragment_constants)
+      bo_reference(&brw->wm.surf_bo[BTI_FRAGMENT_CONSTANTS], NULL);      
+
+   /* XXX: no pipe_max_textures define?? */
+   for (i = brw->curr.num_textures; i < PIPE_MAX_SAMPLERS; i++)
+      bo_reference(&brw->wm.surf_bo[BTI_TEXTURE(i)], NULL);
+
+   if (brw->wm.nr_surfaces != nr_surfaces) {
+      brw->wm.nr_surfaces = nr_surfaces;
+      brw->state.dirty.brw |= BRW_NEW_NR_WM_SURFACES;
+   }
+
+   ret = brw_wm_get_binding_table(brw, &brw->wm.bind_bo);
+   if (ret)
+      return ret;
+
+   return PIPE_OK;
+}
+
+const struct brw_tracked_state brw_wm_surfaces = {
+   .dirty = {
+      .mesa = (PIPE_NEW_COLOR_BUFFERS |
+               PIPE_NEW_BOUND_TEXTURES |
+               PIPE_NEW_FRAGMENT_CONSTANTS |
+              PIPE_NEW_BLEND),
+      .brw = (BRW_NEW_CONTEXT |
+             BRW_NEW_WM_SURFACES),
+      .cache = 0
+   },
+   .prepare = prepare_wm_surfaces,
+};
+
+
+
diff --git a/src/gallium/drivers/i965/intel_decode.c b/src/gallium/drivers/i965/intel_decode.c
new file mode 100644 (file)
index 0000000..3166958
--- /dev/null
@@ -0,0 +1,1790 @@
+/* -*- c-basic-offset: 4 -*- */
+/*
+ * Copyright Â© 2007 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *
+ */
+
+/** @file intel_decode.c
+ * This file contains code to print out batchbuffer contents in a
+ * human-readable format.
+ *
+ * The current version only supports i915 packets, and only pretty-prints a
+ * subset of them.  The intention is for it to make just a best attempt to
+ * decode, but never crash in the process.
+ */
+
+#include <stdio.h>
+#include <stdarg.h>
+#include <string.h>
+#include <inttypes.h>
+
+#include "intel_decode.h"
+
+/*#include "intel_chipset.h"*/
+#define IS_965(x) 1             /* XXX */
+#define IS_9XX(x) 1             /* XXX */
+
+#define BUFFER_FAIL(_count, _len, _name) do {                  \
+    fprintf(out, "Buffer size too small in %s (%d < %d)\n",    \
+           (_name), (_count), (_len));                         \
+    (*failures)++;                                             \
+    return count;                                              \
+} while (0)
+
+static FILE *out;
+static uint32_t saved_s2 = 0, saved_s4 = 0;
+static char saved_s2_set = 0, saved_s4_set = 0;
+
+static float
+int_as_float(uint32_t intval)
+{
+    union intfloat {
+       uint32_t i;
+       float f;
+    } uval;
+
+    uval.i = intval;
+    return uval.f;
+}
+
+static void
+instr_out(const uint32_t *data, uint32_t hw_offset, unsigned int index,
+         char *fmt, ...)
+{
+    va_list va;
+
+    fprintf(out, "0x%08x: 0x%08x:%s ", hw_offset + index * 4, data[index],
+           index == 0 ? "" : "  ");
+    va_start(va, fmt);
+    vfprintf(out, fmt, va);
+    va_end(va);
+}
+
+
+static int
+decode_mi(const uint32_t *data, int count, uint32_t hw_offset, int *failures)
+{
+    unsigned int opcode;
+
+    struct {
+       uint32_t opcode;
+       int len_mask;
+       int min_len;
+       int max_len;
+       char *name;
+    } opcodes_mi[] = {
+       { 0x08, 0, 1, 1, "MI_ARB_ON_OFF" },
+       { 0x0a, 0, 1, 1, "MI_BATCH_BUFFER_END" },
+       { 0x31, 0x3f, 2, 2, "MI_BATCH_BUFFER_START" },
+       { 0x14, 0x3f, 3, 3, "MI_DISPLAY_BUFFER_INFO" },
+       { 0x04, 0, 1, 1, "MI_FLUSH" },
+       { 0x22, 0, 3, 3, "MI_LOAD_REGISTER_IMM" },
+       { 0x13, 0x3f, 2, 2, "MI_LOAD_SCAN_LINES_EXCL" },
+       { 0x12, 0x3f, 2, 2, "MI_LOAD_SCAN_LINES_INCL" },
+       { 0x00, 0, 1, 1, "MI_NOOP" },
+       { 0x11, 0x3f, 2, 2, "MI_OVERLAY_FLIP" },
+       { 0x07, 0, 1, 1, "MI_REPORT_HEAD" },
+       { 0x18, 0x3f, 2, 2, "MI_SET_CONTEXT" },
+       { 0x20, 0x3f, 3, 4, "MI_STORE_DATA_IMM" },
+       { 0x21, 0x3f, 3, 4, "MI_STORE_DATA_INDEX" },
+       { 0x24, 0x3f, 3, 3, "MI_STORE_REGISTER_MEM" },
+       { 0x02, 0, 1, 1, "MI_USER_INTERRUPT" },
+       { 0x03, 0, 1, 1, "MI_WAIT_FOR_EVENT" },
+    };
+
+
+    for (opcode = 0; opcode < sizeof(opcodes_mi) / sizeof(opcodes_mi[0]);
+        opcode++) {
+       if ((data[0] & 0x1f800000) >> 23 == opcodes_mi[opcode].opcode) {
+           unsigned int len = 1, i;
+
+           instr_out(data, hw_offset, 0, "%s\n", opcodes_mi[opcode].name);
+           if (opcodes_mi[opcode].max_len > 1) {
+               len = (data[0] & opcodes_mi[opcode].len_mask) + 2;
+               if (len < opcodes_mi[opcode].min_len ||
+                   len > opcodes_mi[opcode].max_len)
+               {
+                   fprintf(out, "Bad length (%d) in %s, [%d, %d]\n",
+                           len, opcodes_mi[opcode].name,
+                           opcodes_mi[opcode].min_len,
+                           opcodes_mi[opcode].max_len);
+               }
+           }
+
+           for (i = 1; i < len; i++) {
+               if (i >= count)
+                   BUFFER_FAIL(count, len, opcodes_mi[opcode].name);
+               instr_out(data, hw_offset, i, "dword %d\n", i);
+           }
+
+           return len;
+       }
+    }
+
+    instr_out(data, hw_offset, 0, "MI UNKNOWN\n");
+    (*failures)++;
+    return 1;
+}
+
+static int
+decode_2d(const uint32_t *data, int count, uint32_t hw_offset, int *failures)
+{
+    unsigned int opcode, len;
+    char *format = NULL;
+
+    struct {
+       uint32_t opcode;
+       int min_len;
+       int max_len;
+       char *name;
+    } opcodes_2d[] = {
+       { 0x40, 5, 5, "COLOR_BLT" },
+       { 0x43, 6, 6, "SRC_COPY_BLT" },
+       { 0x01, 8, 8, "XY_SETUP_BLT" },
+       { 0x11, 9, 9, "XY_SETUP_MONO_PATTERN_SL_BLT" },
+       { 0x03, 3, 3, "XY_SETUP_CLIP_BLT" },
+       { 0x24, 2, 2, "XY_PIXEL_BLT" },
+       { 0x25, 3, 3, "XY_SCANLINES_BLT" },
+       { 0x26, 4, 4, "Y_TEXT_BLT" },
+       { 0x31, 5, 134, "XY_TEXT_IMMEDIATE_BLT" },
+       { 0x50, 6, 6, "XY_COLOR_BLT" },
+       { 0x51, 6, 6, "XY_PAT_BLT" },
+       { 0x76, 8, 8, "XY_PAT_CHROMA_BLT" },
+       { 0x72, 7, 135, "XY_PAT_BLT_IMMEDIATE" },
+       { 0x77, 9, 137, "XY_PAT_CHROMA_BLT_IMMEDIATE" },
+       { 0x52, 9, 9, "XY_MONO_PAT_BLT" },
+       { 0x59, 7, 7, "XY_MONO_PAT_FIXED_BLT" },
+       { 0x53, 8, 8, "XY_SRC_COPY_BLT" },
+       { 0x54, 8, 8, "XY_MONO_SRC_COPY_BLT" },
+       { 0x71, 9, 137, "XY_MONO_SRC_COPY_IMMEDIATE_BLT" },
+       { 0x55, 9, 9, "XY_FULL_BLT" },
+       { 0x55, 9, 137, "XY_FULL_IMMEDIATE_PATTERN_BLT" },
+       { 0x56, 9, 9, "XY_FULL_MONO_SRC_BLT" },
+       { 0x75, 10, 138, "XY_FULL_MONO_SRC_IMMEDIATE_PATTERN_BLT" },
+       { 0x57, 12, 12, "XY_FULL_MONO_PATTERN_BLT" },
+       { 0x58, 12, 12, "XY_FULL_MONO_PATTERN_MONO_SRC_BLT" },
+    };
+
+    switch ((data[0] & 0x1fc00000) >> 22) {
+    case 0x50:
+       instr_out(data, hw_offset, 0,
+                 "XY_COLOR_BLT (rgb %sabled, alpha %sabled, dst tile %d)\n",
+                 (data[0] & (1 << 20)) ? "en" : "dis",
+                 (data[0] & (1 << 21)) ? "en" : "dis",
+                 (data[0] >> 11) & 1);
+
+       len = (data[0] & 0x000000ff) + 2;
+       if (len != 6)
+           fprintf(out, "Bad count in XY_COLOR_BLT\n");
+       if (count < 6)
+           BUFFER_FAIL(count, len, "XY_COLOR_BLT");
+
+       switch ((data[1] >> 24) & 0x3) {
+       case 0:
+           format="8";
+           break;
+       case 1:
+           format="565";
+           break;
+       case 2:
+           format="1555";
+           break;
+       case 3:
+           format="8888";
+           break;
+       }
+
+       instr_out(data, hw_offset, 1, "format %s, pitch %d, "
+                 "clipping %sabled\n", format,
+                 (short)(data[1] & 0xffff),
+                 data[1] & (1 << 30) ? "en" : "dis");
+       instr_out(data, hw_offset, 2, "(%d,%d)\n",
+                 data[2] & 0xffff, data[2] >> 16);
+       instr_out(data, hw_offset, 3, "(%d,%d)\n",
+                 data[3] & 0xffff, data[3] >> 16);
+       instr_out(data, hw_offset, 4, "offset 0x%08x\n", data[4]);
+       instr_out(data, hw_offset, 5, "color\n");
+       return len;
+    case 0x53:
+       instr_out(data, hw_offset, 0,
+                 "XY_SRC_COPY_BLT (rgb %sabled, alpha %sabled, "
+                 "src tile %d, dst tile %d)\n",
+                 (data[0] & (1 << 20)) ? "en" : "dis",
+                 (data[0] & (1 << 21)) ? "en" : "dis",
+                 (data[0] >> 15) & 1,
+                 (data[0] >> 11) & 1);
+
+       len = (data[0] & 0x000000ff) + 2;
+       if (len != 8)
+           fprintf(out, "Bad count in XY_SRC_COPY_BLT\n");
+       if (count < 8)
+           BUFFER_FAIL(count, len, "XY_SRC_COPY_BLT");
+
+       switch ((data[1] >> 24) & 0x3) {
+       case 0:
+           format="8";
+           break;
+       case 1:
+           format="565";
+           break;
+       case 2:
+           format="1555";
+           break;
+       case 3:
+           format="8888";
+           break;
+       }
+
+       instr_out(data, hw_offset, 1, "format %s, dst pitch %d, "
+                 "clipping %sabled\n", format,
+                 (short)(data[1] & 0xffff),
+                 data[1] & (1 << 30) ? "en" : "dis");
+       instr_out(data, hw_offset, 2, "dst (%d,%d)\n",
+                 data[2] & 0xffff, data[2] >> 16);
+       instr_out(data, hw_offset, 3, "dst (%d,%d)\n",
+                 data[3] & 0xffff, data[3] >> 16);
+       instr_out(data, hw_offset, 4, "dst offset 0x%08x\n", data[4]);
+       instr_out(data, hw_offset, 5, "src (%d,%d)\n",
+                 data[5] & 0xffff, data[5] >> 16);
+       instr_out(data, hw_offset, 6, "src pitch %d\n",
+                 (short)(data[6] & 0xffff));
+       instr_out(data, hw_offset, 7, "src offset 0x%08x\n", data[7]);
+       return len;
+    }
+
+    for (opcode = 0; opcode < sizeof(opcodes_2d) / sizeof(opcodes_2d[0]);
+        opcode++) {
+       if ((data[0] & 0x1fc00000) >> 22 == opcodes_2d[opcode].opcode) {
+           unsigned int i;
+
+           len = 1;
+           instr_out(data, hw_offset, 0, "%s\n", opcodes_2d[opcode].name);
+           if (opcodes_2d[opcode].max_len > 1) {
+               len = (data[0] & 0x000000ff) + 2;
+               if (len < opcodes_2d[opcode].min_len ||
+                   len > opcodes_2d[opcode].max_len)
+               {
+                   fprintf(out, "Bad count in %s\n", opcodes_2d[opcode].name);
+               }
+           }
+
+           for (i = 1; i < len; i++) {
+               if (i >= count)
+                   BUFFER_FAIL(count, len, opcodes_2d[opcode].name);
+               instr_out(data, hw_offset, i, "dword %d\n", i);
+           }
+
+           return len;
+       }
+    }
+
+    instr_out(data, hw_offset, 0, "2D UNKNOWN\n");
+    (*failures)++;
+    return 1;
+}
+
+static int
+decode_3d_1c(const uint32_t *data, int count, uint32_t hw_offset, int *failures)
+{
+    switch ((data[0] & 0x00f80000) >> 19) {
+    case 0x11:
+       instr_out(data, hw_offset, 0, "3DSTATE_DEPTH_SUBRECTANGLE_DISALBE\n");
+       return 1;
+    case 0x10:
+       instr_out(data, hw_offset, 0, "3DSTATE_SCISSOR_ENABLE\n");
+       return 1;
+    case 0x01:
+       instr_out(data, hw_offset, 0, "3DSTATE_MAP_COORD_SET_I830\n");
+       return 1;
+    case 0x0a:
+       instr_out(data, hw_offset, 0, "3DSTATE_MAP_CUBE_I830\n");
+       return 1;
+    case 0x05:
+       instr_out(data, hw_offset, 0, "3DSTATE_MAP_TEX_STREAM_I830\n");
+       return 1;
+    }
+
+    instr_out(data, hw_offset, 0, "3D UNKNOWN\n");
+    (*failures)++;
+    return 1;
+}
+
+/** Sets the string dstname to describe the destination of the PS instruction */
+static void
+i915_get_instruction_dst(const uint32_t *data, int i, char *dstname, int do_mask)
+{
+    uint32_t a0 = data[i];
+    int dst_nr = (a0 >> 14) & 0xf;
+    char dstmask[8];
+    char *sat;
+
+    if (do_mask) {
+       if (((a0 >> 10) & 0xf) == 0xf) {
+           dstmask[0] = 0;
+       } else {
+           int dstmask_index = 0;
+
+           dstmask[dstmask_index++] = '.';
+           if (a0 & (1 << 10))
+               dstmask[dstmask_index++] = 'x';
+           if (a0 & (1 << 11))
+               dstmask[dstmask_index++] = 'y';
+           if (a0 & (1 << 12))
+               dstmask[dstmask_index++] = 'z';
+           if (a0 & (1 << 13))
+               dstmask[dstmask_index++] = 'w';
+           dstmask[dstmask_index++] = 0;
+       }
+
+       if (a0 & (1 << 22))
+           sat = ".sat";
+       else
+           sat = "";
+    } else {
+       dstmask[0] = 0;
+       sat = "";
+    }
+
+    switch ((a0 >> 19) & 0x7) {
+    case 0:
+       if (dst_nr > 15)
+           fprintf(out, "bad destination reg R%d\n", dst_nr);
+       sprintf(dstname, "R%d%s%s", dst_nr, dstmask, sat);
+       break;
+    case 4:
+       if (dst_nr > 0)
+           fprintf(out, "bad destination reg oC%d\n", dst_nr);
+       sprintf(dstname, "oC%s%s", dstmask, sat);
+       break;
+    case 5:
+       if (dst_nr > 0)
+           fprintf(out, "bad destination reg oD%d\n", dst_nr);
+       sprintf(dstname, "oD%s%s",  dstmask, sat);
+       break;
+    case 6:
+       if (dst_nr > 2)
+           fprintf(out, "bad destination reg U%d\n", dst_nr);
+       sprintf(dstname, "U%d%s%s", dst_nr, dstmask, sat);
+       break;
+    default:
+       sprintf(dstname, "RESERVED");
+       break;
+    }
+}
+
+static char *
+i915_get_channel_swizzle(uint32_t select)
+{
+    switch (select & 0x7) {
+    case 0:
+       return (select & 8) ? "-x" : "x";
+    case 1:
+       return (select & 8) ? "-y" : "y";
+    case 2:
+       return (select & 8) ? "-z" : "z";
+    case 3:
+       return (select & 8) ? "-w" : "w";
+    case 4:
+       return (select & 8) ? "-0" : "0";
+    case 5:
+       return (select & 8) ? "-1" : "1";
+    default:
+       return (select & 8) ? "-bad" : "bad";
+    }
+}
+
+static void
+i915_get_instruction_src_name(uint32_t src_type, uint32_t src_nr, char *name)
+{
+    switch (src_type) {
+    case 0:
+       sprintf(name, "R%d", src_nr);
+       if (src_nr > 15)
+           fprintf(out, "bad src reg %s\n", name);
+       break;
+    case 1:
+       if (src_nr < 8)
+           sprintf(name, "T%d", src_nr);
+       else if (src_nr == 8)
+           sprintf(name, "DIFFUSE");
+       else if (src_nr == 9)
+           sprintf(name, "SPECULAR");
+       else if (src_nr == 10)
+           sprintf(name, "FOG");
+       else {
+           fprintf(out, "bad src reg T%d\n", src_nr);
+           sprintf(name, "RESERVED");
+       }
+       break;
+    case 2:
+       sprintf(name, "C%d", src_nr);
+       if (src_nr > 31)
+           fprintf(out, "bad src reg %s\n", name);
+       break;
+    case 4:
+       sprintf(name, "oC");
+       if (src_nr > 0)
+           fprintf(out, "bad src reg oC%d\n", src_nr);
+       break;
+    case 5:
+       sprintf(name, "oD");
+       if (src_nr > 0)
+           fprintf(out, "bad src reg oD%d\n", src_nr);
+       break;
+    case 6:
+       sprintf(name, "U%d", src_nr);
+       if (src_nr > 2)
+           fprintf(out, "bad src reg %s\n", name);
+       break;
+    default:
+       fprintf(out, "bad src reg type %d\n", src_type);
+       sprintf(name, "RESERVED");
+       break;
+    }
+}
+
+static void
+i915_get_instruction_src0(const uint32_t *data, int i, char *srcname)
+{
+    uint32_t a0 = data[i];
+    uint32_t a1 = data[i + 1];
+    int src_nr = (a0 >> 2) & 0x1f;
+    char *swizzle_x = i915_get_channel_swizzle((a1 >> 28) & 0xf);
+    char *swizzle_y = i915_get_channel_swizzle((a1 >> 24) & 0xf);
+    char *swizzle_z = i915_get_channel_swizzle((a1 >> 20) & 0xf);
+    char *swizzle_w = i915_get_channel_swizzle((a1 >> 16) & 0xf);
+    char swizzle[100];
+
+    i915_get_instruction_src_name((a0 >> 7) & 0x7, src_nr, srcname);
+    sprintf(swizzle, ".%s%s%s%s", swizzle_x, swizzle_y, swizzle_z, swizzle_w);
+    if (strcmp(swizzle, ".xyzw") != 0)
+       strcat(srcname, swizzle);
+}
+
+static void
+i915_get_instruction_src1(const uint32_t *data, int i, char *srcname)
+{
+    uint32_t a1 = data[i + 1];
+    uint32_t a2 = data[i + 2];
+    int src_nr = (a1 >> 8) & 0x1f;
+    char *swizzle_x = i915_get_channel_swizzle((a1 >> 4) & 0xf);
+    char *swizzle_y = i915_get_channel_swizzle((a1 >> 0) & 0xf);
+    char *swizzle_z = i915_get_channel_swizzle((a2 >> 28) & 0xf);
+    char *swizzle_w = i915_get_channel_swizzle((a2 >> 24) & 0xf);
+    char swizzle[100];
+
+    i915_get_instruction_src_name((a1 >> 13) & 0x7, src_nr, srcname);
+    sprintf(swizzle, ".%s%s%s%s", swizzle_x, swizzle_y, swizzle_z, swizzle_w);
+    if (strcmp(swizzle, ".xyzw") != 0)
+       strcat(srcname, swizzle);
+}
+
+static void
+i915_get_instruction_src2(const uint32_t *data, int i, char *srcname)
+{
+    uint32_t a2 = data[i + 2];
+    int src_nr = (a2 >> 16) & 0x1f;
+    char *swizzle_x = i915_get_channel_swizzle((a2 >> 12) & 0xf);
+    char *swizzle_y = i915_get_channel_swizzle((a2 >> 8) & 0xf);
+    char *swizzle_z = i915_get_channel_swizzle((a2 >> 4) & 0xf);
+    char *swizzle_w = i915_get_channel_swizzle((a2 >> 0) & 0xf);
+    char swizzle[100];
+
+    i915_get_instruction_src_name((a2 >> 21) & 0x7, src_nr, srcname);
+    sprintf(swizzle, ".%s%s%s%s", swizzle_x, swizzle_y, swizzle_z, swizzle_w);
+    if (strcmp(swizzle, ".xyzw") != 0)
+       strcat(srcname, swizzle);
+}
+
+static void
+i915_get_instruction_addr(uint32_t src_type, uint32_t src_nr, char *name)
+{
+    switch (src_type) {
+    case 0:
+       sprintf(name, "R%d", src_nr);
+       if (src_nr > 15)
+           fprintf(out, "bad src reg %s\n", name);
+       break;
+    case 1:
+       if (src_nr < 8)
+           sprintf(name, "T%d", src_nr);
+       else if (src_nr == 8)
+           sprintf(name, "DIFFUSE");
+       else if (src_nr == 9)
+           sprintf(name, "SPECULAR");
+       else if (src_nr == 10)
+           sprintf(name, "FOG");
+       else {
+           fprintf(out, "bad src reg T%d\n", src_nr);
+           sprintf(name, "RESERVED");
+       }
+       break;
+    case 4:
+       sprintf(name, "oC");
+       if (src_nr > 0)
+           fprintf(out, "bad src reg oC%d\n", src_nr);
+       break;
+    case 5:
+       sprintf(name, "oD");
+       if (src_nr > 0)
+           fprintf(out, "bad src reg oD%d\n", src_nr);
+       break;
+    default:
+       fprintf(out, "bad src reg type %d\n", src_type);
+       sprintf(name, "RESERVED");
+       break;
+    }
+}
+
+static void
+i915_decode_alu1(const uint32_t *data, uint32_t hw_offset,
+                int i, char *instr_prefix, char *op_name)
+{
+    char dst[100], src0[100];
+
+    i915_get_instruction_dst(data, i, dst, 1);
+    i915_get_instruction_src0(data, i, src0);
+
+    instr_out(data, hw_offset, i++, "%s: %s %s, %s\n", instr_prefix,
+             op_name, dst, src0);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+}
+
+static void
+i915_decode_alu2(const uint32_t *data, uint32_t hw_offset,
+                int i, char *instr_prefix, char *op_name)
+{
+    char dst[100], src0[100], src1[100];
+
+    i915_get_instruction_dst(data, i, dst, 1);
+    i915_get_instruction_src0(data, i, src0);
+    i915_get_instruction_src1(data, i, src1);
+
+    instr_out(data, hw_offset, i++, "%s: %s %s, %s, %s\n", instr_prefix,
+             op_name, dst, src0, src1);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+}
+
+static void
+i915_decode_alu3(const uint32_t *data, uint32_t hw_offset,
+                int i, char *instr_prefix, char *op_name)
+{
+    char dst[100], src0[100], src1[100], src2[100];
+
+    i915_get_instruction_dst(data, i, dst, 1);
+    i915_get_instruction_src0(data, i, src0);
+    i915_get_instruction_src1(data, i, src1);
+    i915_get_instruction_src2(data, i, src2);
+
+    instr_out(data, hw_offset, i++, "%s: %s %s, %s, %s, %s\n", instr_prefix,
+             op_name, dst, src0, src1, src2);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+}
+
+static void
+i915_decode_tex(const uint32_t *data, uint32_t hw_offset, int i, char *instr_prefix,
+               char *tex_name)
+{
+    uint32_t t0 = data[i];
+    uint32_t t1 = data[i + 1];
+    char dst_name[100];
+    char addr_name[100];
+    int sampler_nr;
+
+    i915_get_instruction_dst(data, i, dst_name, 0);
+    i915_get_instruction_addr((t1 >> 24) & 0x7,
+                             (t1 >> 17) & 0xf,
+                             addr_name);
+    sampler_nr = t0 & 0xf;
+
+    instr_out(data, hw_offset, i++, "%s: %s %s, S%d, %s\n", instr_prefix,
+             tex_name, dst_name, sampler_nr, addr_name);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+    instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+}
+
+static void
+i915_decode_dcl(const uint32_t *data, uint32_t hw_offset, int i, char *instr_prefix)
+{
+    uint32_t d0 = data[i];
+    char *sampletype;
+    int dcl_nr = (d0 >> 14) & 0xf;
+    char *dcl_x = d0 & (1 << 10) ? "x" : "";
+    char *dcl_y = d0 & (1 << 11) ? "y" : "";
+    char *dcl_z = d0 & (1 << 12) ? "z" : "";
+    char *dcl_w = d0 & (1 << 13) ? "w" : "";
+    char dcl_mask[10];
+
+    switch ((d0 >> 19) & 0x3) {
+    case 1:
+       sprintf(dcl_mask, ".%s%s%s%s", dcl_x, dcl_y, dcl_z, dcl_w);
+       if (strcmp(dcl_mask, ".") == 0)
+           fprintf(out, "bad (empty) dcl mask\n");
+
+       if (dcl_nr > 10)
+           fprintf(out, "bad T%d dcl register number\n", dcl_nr);
+       if (dcl_nr < 8) {
+           if (strcmp(dcl_mask, ".x") != 0 &&
+               strcmp(dcl_mask, ".xy") != 0 &&
+               strcmp(dcl_mask, ".xz") != 0 &&
+               strcmp(dcl_mask, ".w") != 0 &&
+               strcmp(dcl_mask, ".xyzw") != 0) {
+               fprintf(out, "bad T%d.%s dcl mask\n", dcl_nr, dcl_mask);
+           }
+           instr_out(data, hw_offset, i++, "%s: DCL T%d%s\n", instr_prefix,
+                     dcl_nr, dcl_mask);
+       } else {
+           if (strcmp(dcl_mask, ".xz") == 0)
+               fprintf(out, "errataed bad dcl mask %s\n", dcl_mask);
+           else if (strcmp(dcl_mask, ".xw") == 0)
+               fprintf(out, "errataed bad dcl mask %s\n", dcl_mask);
+           else if (strcmp(dcl_mask, ".xzw") == 0)
+               fprintf(out, "errataed bad dcl mask %s\n", dcl_mask);
+
+           if (dcl_nr == 8) {
+               instr_out(data, hw_offset, i++, "%s: DCL DIFFUSE%s\n", instr_prefix,
+                         dcl_mask);
+           } else if (dcl_nr == 9) {
+               instr_out(data, hw_offset, i++, "%s: DCL SPECULAR%s\n", instr_prefix,
+                         dcl_mask);
+           } else if (dcl_nr == 10) {
+               instr_out(data, hw_offset, i++, "%s: DCL FOG%s\n", instr_prefix,
+                         dcl_mask);
+           }
+       }
+       instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+       instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+       break;
+    case 3:
+       switch ((d0 >> 22) & 0x3) {
+       case 0:
+           sampletype = "2D";
+           break;
+       case 1:
+           sampletype = "CUBE";
+           break;
+       case 2:
+           sampletype = "3D";
+           break;
+       default:
+           sampletype = "RESERVED";
+           break;
+       }
+       if (dcl_nr > 15)
+           fprintf(out, "bad S%d dcl register number\n", dcl_nr);
+       instr_out(data, hw_offset, i++, "%s: DCL S%d %s\n", instr_prefix,
+                 dcl_nr, sampletype);
+       instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+       instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+       break;
+    default:
+       instr_out(data, hw_offset, i++, "%s: DCL RESERVED%d\n", instr_prefix, dcl_nr);
+       instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+       instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+    }
+}
+
+static void
+i915_decode_instruction(const uint32_t *data, uint32_t hw_offset,
+                       int i, char *instr_prefix)
+{
+    switch ((data[i] >> 24) & 0x1f) {
+    case 0x0:
+       instr_out(data, hw_offset, i++, "%s: NOP\n", instr_prefix);
+       instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+       instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+       break;
+    case 0x01:
+       i915_decode_alu2(data, hw_offset, i, instr_prefix, "ADD");
+       break;
+    case 0x02:
+       i915_decode_alu1(data, hw_offset, i, instr_prefix, "MOV");
+       break;
+    case 0x03:
+       i915_decode_alu2(data, hw_offset, i, instr_prefix, "MUL");
+       break;
+    case 0x04:
+       i915_decode_alu3(data, hw_offset, i, instr_prefix, "MAD");
+       break;
+    case 0x05:
+       i915_decode_alu3(data, hw_offset, i, instr_prefix, "DP2ADD");
+       break;
+    case 0x06:
+       i915_decode_alu2(data, hw_offset, i, instr_prefix, "DP3");
+       break;
+    case 0x07:
+       i915_decode_alu2(data, hw_offset, i, instr_prefix, "DP4");
+       break;
+    case 0x08:
+       i915_decode_alu1(data, hw_offset, i, instr_prefix, "FRC");
+       break;
+    case 0x09:
+       i915_decode_alu1(data, hw_offset, i, instr_prefix, "RCP");
+       break;
+    case 0x0a:
+       i915_decode_alu1(data, hw_offset, i, instr_prefix, "RSQ");
+       break;
+    case 0x0b:
+       i915_decode_alu1(data, hw_offset, i, instr_prefix, "EXP");
+       break;
+    case 0x0c:
+       i915_decode_alu1(data, hw_offset, i, instr_prefix, "LOG");
+       break;
+    case 0x0d:
+       i915_decode_alu2(data, hw_offset, i, instr_prefix, "CMP");
+       break;
+    case 0x0e:
+       i915_decode_alu2(data, hw_offset, i, instr_prefix, "MIN");
+       break;
+    case 0x0f:
+       i915_decode_alu2(data, hw_offset, i, instr_prefix, "MAX");
+       break;
+    case 0x10:
+       i915_decode_alu1(data, hw_offset, i, instr_prefix, "FLR");
+       break;
+    case 0x11:
+       i915_decode_alu1(data, hw_offset, i, instr_prefix, "MOD");
+       break;
+    case 0x12:
+       i915_decode_alu1(data, hw_offset, i, instr_prefix, "TRC");
+       break;
+    case 0x13:
+       i915_decode_alu2(data, hw_offset, i, instr_prefix, "SGE");
+       break;
+    case 0x14:
+       i915_decode_alu2(data, hw_offset, i, instr_prefix, "SLT");
+       break;
+    case 0x15:
+       i915_decode_tex(data, hw_offset, i, instr_prefix, "TEXLD");
+       break;
+    case 0x16:
+       i915_decode_tex(data, hw_offset, i, instr_prefix, "TEXLDP");
+       break;
+    case 0x17:
+       i915_decode_tex(data, hw_offset, i, instr_prefix, "TEXLDB");
+       break;
+    case 0x19:
+       i915_decode_dcl(data, hw_offset, i, instr_prefix);
+       break;
+    default:
+       instr_out(data, hw_offset, i++, "%s: unknown\n", instr_prefix);
+       instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+       instr_out(data, hw_offset, i++, "%s\n", instr_prefix);
+       break;
+    }
+}
+
+static int
+decode_3d_1d(const uint32_t *data, int count, uint32_t hw_offset, int *failures, int i830)
+{
+    unsigned int len, i, c, opcode, word, map, sampler, instr;
+    char *format;
+
+    struct {
+       uint32_t opcode;
+       int i830_only;
+       int min_len;
+       int max_len;
+       char *name;
+    } opcodes_3d_1d[] = {
+       { 0x8e, 0, 3, 3, "3DSTATE_BUFFER_INFO" },
+       { 0x86, 0, 4, 4, "3DSTATE_CHROMA_KEY" },
+       { 0x9c, 0, 1, 1, "3DSTATE_CLEAR_PARAMETERS" },
+       { 0x88, 0, 2, 2, "3DSTATE_CONSTANT_BLEND_COLOR" },
+       { 0x99, 0, 2, 2, "3DSTATE_DEFAULT_DIFFUSE" },
+       { 0x9a, 0, 2, 2, "3DSTATE_DEFAULT_SPECULAR" },
+       { 0x98, 0, 2, 2, "3DSTATE_DEFAULT_Z" },
+       { 0x97, 0, 2, 2, "3DSTATE_DEPTH_OFFSET_SCALE" },
+       { 0x85, 0, 2, 2, "3DSTATE_DEST_BUFFER_VARIABLES" },
+       { 0x80, 0, 5, 5, "3DSTATE_DRAWING_RECTANGLE" },
+       { 0x8e, 0, 3, 3, "3DSTATE_BUFFER_INFO" },
+       { 0x9d, 0, 65, 65, "3DSTATE_FILTER_COEFFICIENTS_4X4" },
+       { 0x9e, 0, 4, 4, "3DSTATE_MONO_FILTER" },
+       { 0x89, 0, 4, 4, "3DSTATE_FOG_MODE" },
+       { 0x8f, 0, 2, 16, "3DSTATE_MAP_PALLETE_LOAD_32" },
+       { 0x81, 0, 3, 3, "3DSTATE_SCISSOR_RECTANGLE" },
+       { 0x83, 0, 2, 2, "3DSTATE_SPAN_STIPPLE" },
+       { 0x8c, 1, 2, 2, "3DSTATE_MAP_COORD_TRANSFORM_I830" },
+       { 0x8b, 1, 2, 2, "3DSTATE_MAP_VERTEX_TRANSFORM_I830" },
+       { 0x8d, 1, 3, 3, "3DSTATE_W_STATE_I830" },
+       { 0x01, 1, 2, 2, "3DSTATE_COLOR_FACTOR_I830" },
+       { 0x02, 1, 2, 2, "3DSTATE_MAP_COORD_SETBIND_I830" },
+    };
+
+    switch ((data[0] & 0x00ff0000) >> 16) {
+    case 0x07:
+       /* This instruction is unusual.  A 0 length means just 1 DWORD instead of
+        * 2.  The 0 length is specified in one place to be unsupported, but
+        * stated to be required in another, and 0 length LOAD_INDIRECTs appear
+        * to cause no harm at least.
+        */
+       instr_out(data, hw_offset, 0, "3DSTATE_LOAD_INDIRECT\n");
+       len = (data[0] & 0x000000ff) + 1;
+       i = 1;
+       if (data[0] & (0x01 << 8)) {
+           if (i + 2 >= count)
+               BUFFER_FAIL(count, len, "3DSTATE_LOAD_INDIRECT");
+           instr_out(data, hw_offset, i++, "SIS.0\n");
+           instr_out(data, hw_offset, i++, "SIS.1\n");
+       }
+       if (data[0] & (0x02 << 8)) {
+           if (i + 1 >= count)
+               BUFFER_FAIL(count, len, "3DSTATE_LOAD_INDIRECT");
+           instr_out(data, hw_offset, i++, "DIS.0\n");
+       }
+       if (data[0] & (0x04 << 8)) {
+           if (i + 2 >= count)
+               BUFFER_FAIL(count, len, "3DSTATE_LOAD_INDIRECT");
+           instr_out(data, hw_offset, i++, "SSB.0\n");
+           instr_out(data, hw_offset, i++, "SSB.1\n");
+       }
+       if (data[0] & (0x08 << 8)) {
+           if (i + 2 >= count)
+               BUFFER_FAIL(count, len, "3DSTATE_LOAD_INDIRECT");
+           instr_out(data, hw_offset, i++, "MSB.0\n");
+           instr_out(data, hw_offset, i++, "MSB.1\n");
+       }
+       if (data[0] & (0x10 << 8)) {
+           if (i + 2 >= count)
+               BUFFER_FAIL(count, len, "3DSTATE_LOAD_INDIRECT");
+           instr_out(data, hw_offset, i++, "PSP.0\n");
+           instr_out(data, hw_offset, i++, "PSP.1\n");
+       }
+       if (data[0] & (0x20 << 8)) {
+           if (i + 2 >= count)
+               BUFFER_FAIL(count, len, "3DSTATE_LOAD_INDIRECT");
+           instr_out(data, hw_offset, i++, "PSC.0\n");
+           instr_out(data, hw_offset, i++, "PSC.1\n");
+       }
+       if (len != i) {
+           fprintf(out, "Bad count in 3DSTATE_LOAD_INDIRECT\n");
+           (*failures)++;
+           return len;
+       }
+       return len;
+    case 0x04:
+       instr_out(data, hw_offset, 0, "3DSTATE_LOAD_STATE_IMMEDIATE_1\n");
+       len = (data[0] & 0x0000000f) + 2;
+       i = 1;
+       for (word = 0; word <= 7; word++) {
+           if (data[0] & (1 << (4 + word))) {
+               if (i >= count)
+                   BUFFER_FAIL(count, len, "3DSTATE_LOAD_STATE_IMMEDIATE_1");
+
+               /* save vertex state for decode */
+               if (word == 2) {
+                   saved_s2_set = 1;
+                   saved_s2 = data[i];
+               }
+               if (word == 4) {
+                   saved_s4_set = 1;
+                   saved_s4 = data[i];
+               }
+
+               instr_out(data, hw_offset, i++, "S%d\n", word);
+           }
+       }
+       if (len != i) {
+           fprintf(out, "Bad count in 3DSTATE_LOAD_INDIRECT\n");
+           (*failures)++;
+       }
+       return len;
+    case 0x00:
+       instr_out(data, hw_offset, 0, "3DSTATE_MAP_STATE\n");
+       len = (data[0] & 0x0000003f) + 2;
+       instr_out(data, hw_offset, 1, "mask\n");
+
+       i = 2;
+       for (map = 0; map <= 15; map++) {
+           if (data[1] & (1 << map)) {
+               if (i + 3 >= count)
+                   BUFFER_FAIL(count, len, "3DSTATE_MAP_STATE");
+               instr_out(data, hw_offset, i++, "map %d MS2\n", map);
+               instr_out(data, hw_offset, i++, "map %d MS3\n", map);
+               instr_out(data, hw_offset, i++, "map %d MS4\n", map);
+           }
+       }
+       if (len != i) {
+           fprintf(out, "Bad count in 3DSTATE_MAP_STATE\n");
+           (*failures)++;
+           return len;
+       }
+       return len;
+    case 0x06:
+       instr_out(data, hw_offset, 0, "3DSTATE_PIXEL_SHADER_CONSTANTS\n");
+       len = (data[0] & 0x000000ff) + 2;
+
+       i = 2;
+       for (c = 0; c <= 31; c++) {
+           if (data[1] & (1 << c)) {
+               if (i + 4 >= count)
+                   BUFFER_FAIL(count, len, "3DSTATE_PIXEL_SHADER_CONSTANTS");
+               instr_out(data, hw_offset, i, "C%d.X = %f\n",
+                         c, int_as_float(data[i]));
+               i++;
+               instr_out(data, hw_offset, i, "C%d.Y = %f\n",
+                         c, int_as_float(data[i]));
+               i++;
+               instr_out(data, hw_offset, i, "C%d.Z = %f\n",
+                         c, int_as_float(data[i]));
+               i++;
+               instr_out(data, hw_offset, i, "C%d.W = %f\n",
+                         c, int_as_float(data[i]));
+               i++;
+           }
+       }
+       if (len != i) {
+           fprintf(out, "Bad count in 3DSTATE_PIXEL_SHADER_CONSTANTS\n");
+           (*failures)++;
+       }
+       return len;
+    case 0x05:
+       instr_out(data, hw_offset, 0, "3DSTATE_PIXEL_SHADER_PROGRAM\n");
+       len = (data[0] & 0x000000ff) + 2;
+       if ((len - 1) % 3 != 0 || len > 370) {
+           fprintf(out, "Bad count in 3DSTATE_PIXEL_SHADER_PROGRAM\n");
+           (*failures)++;
+       }
+       i = 1;
+       for (instr = 0; instr < (len - 1) / 3; instr++) {
+           char instr_prefix[10];
+
+           if (i + 3 >= count)
+               BUFFER_FAIL(count, len, "3DSTATE_PIXEL_SHADER_PROGRAM");
+           sprintf(instr_prefix, "PS%03d", instr);
+           i915_decode_instruction(data, hw_offset, i, instr_prefix);
+           i += 3;
+       }
+       return len;
+    case 0x01:
+       if (i830)
+           break;
+       instr_out(data, hw_offset, 0, "3DSTATE_SAMPLER_STATE\n");
+       instr_out(data, hw_offset, 1, "mask\n");
+       len = (data[0] & 0x0000003f) + 2;
+       i = 2;
+       for (sampler = 0; sampler <= 15; sampler++) {
+           if (data[1] & (1 << sampler)) {
+               if (i + 3 >= count)
+                   BUFFER_FAIL(count, len, "3DSTATE_SAMPLER_STATE");
+               instr_out(data, hw_offset, i++, "sampler %d SS2\n",
+                         sampler);
+               instr_out(data, hw_offset, i++, "sampler %d SS3\n",
+                         sampler);
+               instr_out(data, hw_offset, i++, "sampler %d SS4\n",
+                         sampler);
+           }
+       }
+       if (len != i) {
+           fprintf(out, "Bad count in 3DSTATE_SAMPLER_STATE\n");
+           (*failures)++;
+       }
+       return len;
+    case 0x85:
+       len = (data[0] & 0x0000000f) + 2;
+
+       if (len != 2)
+           fprintf(out, "Bad count in 3DSTATE_DEST_BUFFER_VARIABLES\n");
+       if (count < 2)
+           BUFFER_FAIL(count, len, "3DSTATE_DEST_BUFFER_VARIABLES");
+
+       instr_out(data, hw_offset, 0,
+                 "3DSTATE_DEST_BUFFER_VARIABLES\n");
+
+       switch ((data[1] >> 8) & 0xf) {
+       case 0x0: format = "g8"; break;
+       case 0x1: format = "x1r5g5b5"; break;
+       case 0x2: format = "r5g6b5"; break;
+       case 0x3: format = "a8r8g8b8"; break;
+       case 0x4: format = "ycrcb_swapy"; break;
+       case 0x5: format = "ycrcb_normal"; break;
+       case 0x6: format = "ycrcb_swapuv"; break;
+       case 0x7: format = "ycrcb_swapuvy"; break;
+       case 0x8: format = "a4r4g4b4"; break;
+       case 0x9: format = "a1r5g5b5"; break;
+       case 0xa: format = "a2r10g10b10"; break;
+       default: format = "BAD"; break;
+       }
+       instr_out(data, hw_offset, 1, "%s format, early Z %sabled\n",
+                 format,
+                 (data[1] & (1 << 31)) ? "en" : "dis");
+       return len;
+    }
+
+    for (opcode = 0; opcode < sizeof(opcodes_3d_1d) / sizeof(opcodes_3d_1d[0]);
+        opcode++)
+    {
+       if (opcodes_3d_1d[opcode].i830_only && !i830)
+           continue;
+
+       if (((data[0] & 0x00ff0000) >> 16) == opcodes_3d_1d[opcode].opcode) {
+           len = 1;
+
+           instr_out(data, hw_offset, 0, "%s\n", opcodes_3d_1d[opcode].name);
+           if (opcodes_3d_1d[opcode].max_len > 1) {
+               len = (data[0] & 0x0000ffff) + 2;
+               if (len < opcodes_3d_1d[opcode].min_len ||
+                   len > opcodes_3d_1d[opcode].max_len)
+               {
+                   fprintf(out, "Bad count in %s\n",
+                           opcodes_3d_1d[opcode].name);
+                   (*failures)++;
+               }
+           }
+
+           for (i = 1; i < len; i++) {
+               if (i >= count)
+                   BUFFER_FAIL(count, len,  opcodes_3d_1d[opcode].name);
+               instr_out(data, hw_offset, i, "dword %d\n", i);
+           }
+
+           return len;
+       }
+    }
+
+    instr_out(data, hw_offset, 0, "3D UNKNOWN\n");
+    (*failures)++;
+    return 1;
+}
+
+static int
+decode_3d_primitive(const uint32_t *data, int count, uint32_t hw_offset,
+                   int *failures)
+{
+    char immediate = (data[0] & (1 << 23)) == 0;
+    unsigned int len, i;
+    char *primtype;
+
+    switch ((data[0] >> 18) & 0xf) {
+    case 0x0: primtype = "TRILIST"; break;
+    case 0x1: primtype = "TRISTRIP"; break;
+    case 0x2: primtype = "TRISTRIP_REVERSE"; break;
+    case 0x3: primtype = "TRIFAN"; break;
+    case 0x4: primtype = "POLYGON"; break;
+    case 0x5: primtype = "LINELIST"; break;
+    case 0x6: primtype = "LINESTRIP"; break;
+    case 0x7: primtype = "RECTLIST"; break;
+    case 0x8: primtype = "POINTLIST"; break;
+    case 0x9: primtype = "DIB"; break;
+    case 0xa: primtype = "CLEAR_RECT"; break;
+    default: primtype = "unknown"; break;
+    }
+
+    /* XXX: 3DPRIM_DIB not supported */
+    if (immediate) {
+       len = (data[0] & 0x0003ffff) + 2;
+       instr_out(data, hw_offset, 0, "3DPRIMITIVE inline %s\n", primtype);
+       if (count < len)
+           BUFFER_FAIL(count, len,  "3DPRIMITIVE inline");
+       if (!saved_s2_set || !saved_s4_set) {
+           fprintf(out, "unknown vertex format\n");
+           for (i = 1; i < len; i++) {
+               instr_out(data, hw_offset, i,
+                         "           vertex data (%f float)\n",
+                         int_as_float(data[i]));
+           }
+       } else {
+           unsigned int vertex = 0;
+           for (i = 1; i < len;) {
+               unsigned int tc;
+
+#define VERTEX_OUT(fmt, ...) do {                                      \
+    if (i < len)                                                       \
+       instr_out(data, hw_offset, i, " V%d."fmt"\n", vertex, __VA_ARGS__); \
+    else                                                               \
+       fprintf(out, " missing data in V%d\n", vertex);                 \
+    i++;                                                               \
+} while (0)
+
+               VERTEX_OUT("X = %f", int_as_float(data[i]));
+               VERTEX_OUT("Y = %f", int_as_float(data[i]));
+               switch (saved_s4 >> 6 & 0x7) {
+               case 0x1:
+                   VERTEX_OUT("Z = %f", int_as_float(data[i]));
+                   break;
+               case 0x2:
+                   VERTEX_OUT("Z = %f", int_as_float(data[i]));
+                   VERTEX_OUT("W = %f", int_as_float(data[i]));
+                   break;
+               case 0x3:
+                   break;
+               case 0x4:
+                   VERTEX_OUT("W = %f", int_as_float(data[i]));
+                   break;
+               default:
+                   fprintf(out, "bad S4 position mask\n");
+               }
+
+               if (saved_s4 & (1 << 10)) {
+                   VERTEX_OUT("color = (A=0x%02x, R=0x%02x, G=0x%02x, "
+                              "B=0x%02x)",
+                              data[i] >> 24,
+                              (data[i] >> 16) & 0xff,
+                              (data[i] >> 8) & 0xff,
+                              data[i] & 0xff);
+               }
+               if (saved_s4 & (1 << 11)) {
+                   VERTEX_OUT("spec = (A=0x%02x, R=0x%02x, G=0x%02x, "
+                              "B=0x%02x)",
+                              data[i] >> 24,
+                              (data[i] >> 16) & 0xff,
+                              (data[i] >> 8) & 0xff,
+                              data[i] & 0xff);
+               }
+               if (saved_s4 & (1 << 12))
+                   VERTEX_OUT("width = 0x%08x)", data[i]);
+
+               for (tc = 0; tc <= 7; tc++) {
+                   switch ((saved_s2 >> (tc * 4)) & 0xf) {
+                   case 0x0:
+                       VERTEX_OUT("T%d.X = %f", tc, int_as_float(data[i]));
+                       VERTEX_OUT("T%d.Y = %f", tc, int_as_float(data[i]));
+                       break;
+                   case 0x1:
+                       VERTEX_OUT("T%d.X = %f", tc, int_as_float(data[i]));
+                       VERTEX_OUT("T%d.Y = %f", tc, int_as_float(data[i]));
+                       VERTEX_OUT("T%d.Z = %f", tc, int_as_float(data[i]));
+                       break;
+                   case 0x2:
+                       VERTEX_OUT("T%d.X = %f", tc, int_as_float(data[i]));
+                       VERTEX_OUT("T%d.Y = %f", tc, int_as_float(data[i]));
+                       VERTEX_OUT("T%d.Z = %f", tc, int_as_float(data[i]));
+                       VERTEX_OUT("T%d.W = %f", tc, int_as_float(data[i]));
+                       break;
+                   case 0x3:
+                       VERTEX_OUT("T%d.X = %f", tc, int_as_float(data[i]));
+                       break;
+                   case 0x4:
+                       VERTEX_OUT("T%d.XY = 0x%08x half-float", tc, data[i]);
+                       break;
+                   case 0x5:
+                       VERTEX_OUT("T%d.XY = 0x%08x half-float", tc, data[i]);
+                       VERTEX_OUT("T%d.ZW = 0x%08x half-float", tc, data[i]);
+                       break;
+                   case 0xf:
+                       break;
+                   default:
+                       fprintf(out, "bad S2.T%d format\n", tc);
+                   }
+               }
+               vertex++;
+           }
+       }
+    } else {
+       /* indirect vertices */
+       len = data[0] & 0x0000ffff; /* index count */
+       if (data[0] & (1 << 17)) {
+           /* random vertex access */
+           if (count < (len + 1) / 2 + 1) {
+               BUFFER_FAIL(count, (len + 1) / 2 + 1,
+                           "3DPRIMITIVE random indirect");
+           }
+           instr_out(data, hw_offset, 0,
+                     "3DPRIMITIVE random indirect %s (%d)\n", primtype, len);
+           if (len == 0) {
+               /* vertex indices continue until 0xffff is found */
+               for (i = 1; i < count; i++) {
+                   if ((data[i] & 0xffff) == 0xffff) {
+                       instr_out(data, hw_offset, i,
+                                 "            indices: (terminator)\n");
+                       return i;
+                   } else if ((data[i] >> 16) == 0xffff) {
+                       instr_out(data, hw_offset, i,
+                                 "            indices: 0x%04x, "
+                                 "(terminator)\n",
+                                 data[i] & 0xffff);
+                       return i;
+                   } else {
+                       instr_out(data, hw_offset, i,
+                                 "            indices: 0x%04x, 0x%04x\n",
+                                 data[i] & 0xffff, data[i] >> 16);
+                   }
+               }
+               fprintf(out,
+                       "3DPRIMITIVE: no terminator found in index buffer\n");
+               (*failures)++;
+               return count;
+           } else {
+               /* fixed size vertex index buffer */
+               for (i = 0; i < len; i += 2) {
+                   if (i * 2 == len - 1) {
+                       instr_out(data, hw_offset, i,
+                                 "            indices: 0x%04x\n",
+                                 data[i] & 0xffff);
+                   } else {
+                       instr_out(data, hw_offset, i,
+                                 "            indices: 0x%04x, 0x%04x\n",
+                                 data[i] & 0xffff, data[i] >> 16);
+                   }
+               }
+           }
+           return (len + 1) / 2 + 1;
+       } else {
+           /* sequential vertex access */
+           if (count < 2)
+               BUFFER_FAIL(count, 2, "3DPRIMITIVE seq indirect");
+           instr_out(data, hw_offset, 0,
+                     "3DPRIMITIVE sequential indirect %s, %d starting from "
+                     "%d\n", primtype, len, data[1] & 0xffff);
+           instr_out(data, hw_offset, 1, "           start\n");
+           return 2;
+       }
+    }
+
+    return len;
+}
+
+static int
+decode_3d(const uint32_t *data, int count, uint32_t hw_offset, int *failures)
+{
+    unsigned int opcode;
+
+    struct {
+       uint32_t opcode;
+       int min_len;
+       int max_len;
+       char *name;
+    } opcodes_3d[] = {
+       { 0x06, 1, 1, "3DSTATE_ANTI_ALIASING" },
+       { 0x08, 1, 1, "3DSTATE_BACKFACE_STENCIL_OPS" },
+       { 0x09, 1, 1, "3DSTATE_BACKFACE_STENCIL_MASKS" },
+       { 0x16, 1, 1, "3DSTATE_COORD_SET_BINDINGS" },
+       { 0x15, 1, 1, "3DSTATE_FOG_COLOR" },
+       { 0x0b, 1, 1, "3DSTATE_INDEPENDENT_ALPHA_BLEND" },
+       { 0x0d, 1, 1, "3DSTATE_MODES_4" },
+       { 0x0c, 1, 1, "3DSTATE_MODES_5" },
+       { 0x07, 1, 1, "3DSTATE_RASTERIZATION_RULES" },
+    };
+
+    switch ((data[0] & 0x1f000000) >> 24) {
+    case 0x1f:
+       return decode_3d_primitive(data, count, hw_offset, failures);
+    case 0x1d:
+       return decode_3d_1d(data, count, hw_offset, failures, 0);
+    case 0x1c:
+       return decode_3d_1c(data, count, hw_offset, failures);
+    }
+
+    for (opcode = 0; opcode < sizeof(opcodes_3d) / sizeof(opcodes_3d[0]);
+        opcode++) {
+       if ((data[0] & 0x1f000000) >> 24 == opcodes_3d[opcode].opcode) {
+           unsigned int len = 1, i;
+
+           instr_out(data, hw_offset, 0, "%s\n", opcodes_3d[opcode].name);
+           if (opcodes_3d[opcode].max_len > 1) {
+               len = (data[0] & 0xff) + 2;
+               if (len < opcodes_3d[opcode].min_len ||
+                   len > opcodes_3d[opcode].max_len)
+               {
+                   fprintf(out, "Bad count in %s\n", opcodes_3d[opcode].name);
+               }
+           }
+
+           for (i = 1; i < len; i++) {
+               if (i >= count)
+                   BUFFER_FAIL(count, len, opcodes_3d[opcode].name);
+               instr_out(data, hw_offset, i, "dword %d\n", i);
+           }
+           return len;
+       }
+    }
+
+    instr_out(data, hw_offset, 0, "3D UNKNOWN\n");
+    (*failures)++;
+    return 1;
+}
+
+static const char *
+get_965_surfacetype(unsigned int surfacetype)
+{
+    switch (surfacetype) {
+    case 0: return "1D";
+    case 1: return "2D";
+    case 2: return "3D";
+    case 3: return "CUBE";
+    case 4: return "BUFFER";
+    case 7: return "NULL";
+    default: return "unknown";
+    }
+}
+
+static const char *
+get_965_depthformat(unsigned int depthformat)
+{
+    switch (depthformat) {
+    case 0: return "s8_z24float";
+    case 1: return "z32float";
+    case 2: return "z24s8";
+    case 5: return "z16";
+    default: return "unknown";
+    }
+}
+
+static const char *
+get_965_element_component(uint32_t data, int component)
+{
+    uint32_t component_control = (data >> (16 + (3 - component) * 4)) & 0x7;
+
+    switch (component_control) {
+    case 0:
+       return "nostore";
+    case 1:
+       switch (component) {
+       case 0: return "X";
+       case 1: return "Y";
+       case 2: return "Z";
+       case 3: return "W";
+       default: return "fail";
+       }
+    case 2:
+       return "0.0";
+    case 3:
+       return "1.0";
+    case 4:
+       return "0x1";
+    case 5:
+       return "VID";
+    default:
+       return "fail";
+    }
+}
+
+static const char *
+get_965_prim_type(uint32_t data)
+{
+    uint32_t primtype = (data >> 10) & 0x1f;
+
+    switch (primtype) {
+    case 0x01: return "point list";
+    case 0x02: return "line list";
+    case 0x03: return "line strip";
+    case 0x04: return "tri list";
+    case 0x05: return "tri strip";
+    case 0x06: return "tri fan";
+    case 0x07: return "quad list";
+    case 0x08: return "quad strip";
+    case 0x09: return "line list adj";
+    case 0x0a: return "line strip adj";
+    case 0x0b: return "tri list adj";
+    case 0x0c: return "tri strip adj";
+    case 0x0d: return "tri strip reverse";
+    case 0x0e: return "polygon";
+    case 0x0f: return "rect list";
+    case 0x10: return "line loop";
+    case 0x11: return "point list bf";
+    case 0x12: return "line strip cont";
+    case 0x13: return "line strip bf";
+    case 0x14: return "line strip cont bf";
+    case 0x15: return "tri fan no stipple";
+    default: return "fail";
+    }
+}
+
+static int
+decode_3d_965(const uint32_t *data, int count, uint32_t hw_offset, int *failures)
+{
+    unsigned int opcode, len;
+    int i;
+
+    struct {
+       uint32_t opcode;
+       int min_len;
+       int max_len;
+       char *name;
+    } opcodes_3d[] = {
+       { 0x6000, 3, 3, "URB_FENCE" },
+       { 0x6001, 2, 2, "CS_URB_STATE" },
+       { 0x6002, 2, 2, "CONSTANT_BUFFER" },
+       { 0x6101, 6, 6, "STATE_BASE_ADDRESS" },
+       { 0x6102, 2, 2 , "STATE_SIP" },
+       { 0x6104, 1, 1, "3DSTATE_PIPELINE_SELECT" },
+       { 0x680b, 1, 1, "3DSTATE_VF_STATISTICS" },
+       { 0x6904, 1, 1, "3DSTATE_PIPELINE_SELECT" },
+       { 0x7800, 7, 7, "3DSTATE_PIPELINED_POINTERS" },
+       { 0x7801, 6, 6, "3DSTATE_BINDING_TABLE_POINTERS" },
+       { 0x780b, 1, 1, "3DSTATE_VF_STATISTICS" },
+       { 0x7808, 5, 257, "3DSTATE_VERTEX_BUFFERS" },
+       { 0x7809, 3, 256, "3DSTATE_VERTEX_ELEMENTS" },
+       { 0x780a, 3, 3, "3DSTATE_INDEX_BUFFER" },
+       { 0x7900, 4, 4, "3DSTATE_DRAWING_RECTANGLE" },
+       { 0x7901, 5, 5, "3DSTATE_CONSTANT_COLOR" },
+       { 0x7905, 5, 7, "3DSTATE_DEPTH_BUFFER" },
+       { 0x7906, 2, 2, "3DSTATE_POLY_STIPPLE_OFFSET" },
+       { 0x7907, 33, 33, "3DSTATE_POLY_STIPPLE_PATTERN" },
+       { 0x7908, 3, 3, "3DSTATE_LINE_STIPPLE" },
+       { 0x7909, 2, 2, "3DSTATE_GLOBAL_DEPTH_OFFSET_CLAMP" },
+       { 0x790a, 3, 3, "3DSTATE_AA_LINE_PARAMETERS" },
+       { 0x7b00, 6, 6, "3DPRIMITIVE" },
+    };
+
+    len = (data[0] & 0x0000ffff) + 2;
+
+    switch ((data[0] & 0xffff0000) >> 16) {
+    case 0x6101:
+       if (len != 6)
+           fprintf(out, "Bad count in STATE_BASE_ADDRESS\n");
+       if (count < 6)
+           BUFFER_FAIL(count, len, "STATE_BASE_ADDRESS");
+
+       instr_out(data, hw_offset, 0,
+                 "STATE_BASE_ADDRESS\n");
+
+       if (data[1] & 1) {
+           instr_out(data, hw_offset, 1, "General state at 0x%08x\n",
+                     data[1] & ~1);
+       } else
+           instr_out(data, hw_offset, 1, "General state not updated\n");
+
+       if (data[2] & 1) {
+           instr_out(data, hw_offset, 2, "Surface state at 0x%08x\n",
+                     data[2] & ~1);
+       } else
+           instr_out(data, hw_offset, 2, "Surface state not updated\n");
+
+       if (data[3] & 1) {
+           instr_out(data, hw_offset, 3, "Indirect state at 0x%08x\n",
+                     data[3] & ~1);
+       } else
+           instr_out(data, hw_offset, 3, "Indirect state not updated\n");
+
+       if (data[4] & 1) {
+           instr_out(data, hw_offset, 4, "General state upper bound 0x%08x\n",
+                     data[4] & ~1);
+       } else
+           instr_out(data, hw_offset, 4, "General state not updated\n");
+
+       if (data[5] & 1) {
+           instr_out(data, hw_offset, 5, "Indirect state upper bound 0x%08x\n",
+                     data[5] & ~1);
+       } else
+           instr_out(data, hw_offset, 5, "Indirect state not updated\n");
+
+       return len;
+    case 0x7800:
+       if (len != 7)
+           fprintf(out, "Bad count in 3DSTATE_PIPELINED_POINTERS\n");
+       if (count < 7)
+           BUFFER_FAIL(count, len, "3DSTATE_PIPELINED_POINTERS");
+
+       instr_out(data, hw_offset, 0,
+                 "3DSTATE_PIPELINED_POINTERS\n");
+       instr_out(data, hw_offset, 1, "VS state\n");
+       instr_out(data, hw_offset, 2, "GS state\n");
+       instr_out(data, hw_offset, 3, "Clip state\n");
+       instr_out(data, hw_offset, 4, "SF state\n");
+       instr_out(data, hw_offset, 5, "WM state\n");
+       instr_out(data, hw_offset, 6, "CC state\n");
+       return len;
+    case 0x7801:
+       if (len != 6)
+           fprintf(out, "Bad count in 3DSTATE_BINDING_TABLE_POINTERS\n");
+       if (count < 6)
+           BUFFER_FAIL(count, len, "3DSTATE_BINDING_TABLE_POINTERS");
+
+       instr_out(data, hw_offset, 0,
+                 "3DSTATE_BINDING_TABLE_POINTERS\n");
+       instr_out(data, hw_offset, 1, "VS binding table\n");
+       instr_out(data, hw_offset, 2, "GS binding table\n");
+       instr_out(data, hw_offset, 3, "Clip binding table\n");
+       instr_out(data, hw_offset, 4, "SF binding table\n");
+       instr_out(data, hw_offset, 5, "WM binding table\n");
+
+       return len;
+
+    case 0x7808:
+       len = (data[0] & 0xff) + 2;
+       if ((len - 1) % 4 != 0)
+           fprintf(out, "Bad count in 3DSTATE_VERTEX_BUFFERS\n");
+       if (count < len)
+           BUFFER_FAIL(count, len, "3DSTATE_VERTEX_BUFFERS");
+       instr_out(data, hw_offset, 0, "3DSTATE_VERTEX_BUFFERS\n");
+
+       for (i = 1; i < len;) {
+           instr_out(data, hw_offset, i, "buffer %d: %s, pitch %db\n",
+                     data[i] >> 27,
+                     data[i] & (1 << 26) ? "random" : "sequential",
+                     data[i] & 0x07ff);
+           i++;
+           instr_out(data, hw_offset, i++, "buffer address\n");
+           instr_out(data, hw_offset, i++, "max index\n");
+           instr_out(data, hw_offset, i++, "mbz\n");
+       }
+       return len;
+
+    case 0x7809:
+       len = (data[0] & 0xff) + 2;
+       if ((len + 1) % 2 != 0)
+           fprintf(out, "Bad count in 3DSTATE_VERTEX_ELEMENTS\n");
+       if (count < len)
+           BUFFER_FAIL(count, len, "3DSTATE_VERTEX_ELEMENTS");
+       instr_out(data, hw_offset, 0, "3DSTATE_VERTEX_ELEMENTS\n");
+
+       for (i = 1; i < len;) {
+           instr_out(data, hw_offset, i, "buffer %d: %svalid, type 0x%04x, "
+                     "src offset 0x%04x bytes\n",
+                     data[i] >> 27,
+                     data[i] & (1 << 26) ? "" : "in",
+                     (data[i] >> 16) & 0x1ff,
+                     data[i] & 0x07ff);
+           i++;
+           instr_out(data, hw_offset, i, "(%s, %s, %s, %s), "
+                     "dst offset 0x%02x bytes\n",
+                     get_965_element_component(data[i], 0),
+                     get_965_element_component(data[i], 1),
+                     get_965_element_component(data[i], 2),
+                     get_965_element_component(data[i], 3),
+                     (data[i] & 0xff) * 4);
+           i++;
+       }
+       return len;
+
+    case 0x780a:
+       len = (data[0] & 0xff) + 2;
+       if (len != 3)
+           fprintf(out, "Bad count in 3DSTATE_INDEX_BUFFER\n");
+       if (count < len)
+           BUFFER_FAIL(count, len, "3DSTATE_INDEX_BUFFER");
+       instr_out(data, hw_offset, 0, "3DSTATE_INDEX_BUFFER\n");
+       instr_out(data, hw_offset, 1, "beginning buffer address\n");
+       instr_out(data, hw_offset, 2, "ending buffer address\n");
+       return len;
+
+    case 0x7900:
+       if (len != 4)
+           fprintf(out, "Bad count in 3DSTATE_DRAWING_RECTANGLE\n");
+       if (count < 4)
+           BUFFER_FAIL(count, len, "3DSTATE_DRAWING_RECTANGLE");
+
+       instr_out(data, hw_offset, 0,
+                 "3DSTATE_DRAWING_RECTANGLE\n");
+       instr_out(data, hw_offset, 1, "top left: %d,%d\n",
+                 data[1] & 0xffff,
+                 (data[1] >> 16) & 0xffff);
+       instr_out(data, hw_offset, 2, "bottom right: %d,%d\n",
+                 data[2] & 0xffff,
+                 (data[2] >> 16) & 0xffff);
+       instr_out(data, hw_offset, 3, "origin: %d,%d\n",
+                 (int)data[3] & 0xffff,
+                 ((int)data[3] >> 16) & 0xffff);
+
+       return len;
+
+    case 0x7905:
+       if (len != 5 && len != 6)
+           fprintf(out, "Bad count in 3DSTATE_DEPTH_BUFFER\n");
+       if (count < len)
+           BUFFER_FAIL(count, len, "3DSTATE_DEPTH_BUFFER");
+
+       instr_out(data, hw_offset, 0,
+                 "3DSTATE_DEPTH_BUFFER\n");
+       instr_out(data, hw_offset, 1, "%s, %s, pitch = %d bytes, %stiled\n",
+                 get_965_surfacetype(data[1] >> 29),
+                 get_965_depthformat((data[1] >> 18) & 0x7),
+                 (data[1] & 0x0001ffff) + 1,
+                 data[1] & (1 << 27) ? "" : "not ");
+       instr_out(data, hw_offset, 2, "depth offset\n");
+       instr_out(data, hw_offset, 3, "%dx%d\n",
+                 ((data[3] & 0x0007ffc0) >> 6) + 1,
+                 ((data[3] & 0xfff80000) >> 19) + 1);
+       instr_out(data, hw_offset, 4, "volume depth\n");
+       if (len == 6)
+           instr_out(data, hw_offset, 5, "\n");
+
+       return len;
+
+    case 0x7b00:
+       len = (data[0] & 0xff) + 2;
+       if (len != 6)
+           fprintf(out, "Bad count in 3DPRIMITIVE\n");
+       if (count < len)
+           BUFFER_FAIL(count, len, "3DPRIMITIVE");
+
+       instr_out(data, hw_offset, 0,
+                 "3DPRIMITIVE: %s %s\n",
+                 get_965_prim_type(data[0]),
+                 (data[0] & (1 << 15)) ? "random" : "sequential");
+       instr_out(data, hw_offset, 1, "vertex count\n");
+       instr_out(data, hw_offset, 2, "start vertex\n");
+       instr_out(data, hw_offset, 3, "instance count\n");
+       instr_out(data, hw_offset, 4, "start instance\n");
+       instr_out(data, hw_offset, 5, "index bias\n");
+       return len;
+    }
+
+    for (opcode = 0; opcode < sizeof(opcodes_3d) / sizeof(opcodes_3d[0]);
+        opcode++) {
+       if ((data[0] & 0xffff0000) >> 16 == opcodes_3d[opcode].opcode) {
+           unsigned int i;
+           len = 1;
+
+           instr_out(data, hw_offset, 0, "%s\n", opcodes_3d[opcode].name);
+           if (opcodes_3d[opcode].max_len > 1) {
+               len = (data[0] & 0xff) + 2;
+               if (len < opcodes_3d[opcode].min_len ||
+                   len > opcodes_3d[opcode].max_len)
+               {
+                   fprintf(out, "Bad count in %s\n", opcodes_3d[opcode].name);
+               }
+           }
+
+           for (i = 1; i < len; i++) {
+               if (i >= count)
+                   BUFFER_FAIL(count, len, opcodes_3d[opcode].name);
+               instr_out(data, hw_offset, i, "dword %d\n", i);
+           }
+           return len;
+       }
+    }
+
+    instr_out(data, hw_offset, 0, "3D UNKNOWN\n");
+    (*failures)++;
+    return 1;
+}
+
+static int
+decode_3d_i830(const uint32_t *data, int count, uint32_t hw_offset, int *failures)
+{
+    unsigned int opcode;
+
+    struct {
+       uint32_t opcode;
+       int min_len;
+       int max_len;
+       char *name;
+    } opcodes_3d[] = {
+       { 0x02, 1, 1, "3DSTATE_MODES_3" },
+       { 0x03, 1, 1, "3DSTATE_ENABLES_1"},
+       { 0x04, 1, 1, "3DSTATE_ENABLES_2"},
+       { 0x05, 1, 1, "3DSTATE_VFT0"},
+       { 0x06, 1, 1, "3DSTATE_AA"},
+       { 0x07, 1, 1, "3DSTATE_RASTERIZATION_RULES" },
+       { 0x08, 1, 1, "3DSTATE_MODES_1" },
+       { 0x09, 1, 1, "3DSTATE_STENCIL_TEST" },
+       { 0x0a, 1, 1, "3DSTATE_VFT1"},
+       { 0x0b, 1, 1, "3DSTATE_INDPT_ALPHA_BLEND" },
+       { 0x0c, 1, 1, "3DSTATE_MODES_5" },
+       { 0x0d, 1, 1, "3DSTATE_MAP_BLEND_OP" },
+       { 0x0e, 1, 1, "3DSTATE_MAP_BLEND_ARG" },
+       { 0x0f, 1, 1, "3DSTATE_MODES_2" },
+       { 0x15, 1, 1, "3DSTATE_FOG_COLOR" },
+       { 0x16, 1, 1, "3DSTATE_MODES_4" },
+    };
+
+    switch ((data[0] & 0x1f000000) >> 24) {
+    case 0x1f:
+       return decode_3d_primitive(data, count, hw_offset, failures);
+    case 0x1d:
+       return decode_3d_1d(data, count, hw_offset, failures, 1);
+    case 0x1c:
+       return decode_3d_1c(data, count, hw_offset, failures);
+    }
+
+    for (opcode = 0; opcode < sizeof(opcodes_3d) / sizeof(opcodes_3d[0]);
+        opcode++) {
+       if ((data[0] & 0x1f000000) >> 24 == opcodes_3d[opcode].opcode) {
+           unsigned int len = 1, i;
+
+           instr_out(data, hw_offset, 0, "%s\n", opcodes_3d[opcode].name);
+           if (opcodes_3d[opcode].max_len > 1) {
+               len = (data[0] & 0xff) + 2;
+               if (len < opcodes_3d[opcode].min_len ||
+                   len > opcodes_3d[opcode].max_len)
+               {
+                   fprintf(out, "Bad count in %s\n", opcodes_3d[opcode].name);
+               }
+           }
+
+           for (i = 1; i < len; i++) {
+               if (i >= count)
+                   BUFFER_FAIL(count, len, opcodes_3d[opcode].name);
+               instr_out(data, hw_offset, i, "dword %d\n", i);
+           }
+           return len;
+       }
+    }
+
+    instr_out(data, hw_offset, 0, "3D UNKNOWN\n");
+    (*failures)++;
+    return 1;
+}
+
+/**
+ * Decodes an i830-i915 batch buffer, writing the output to stdout.
+ *
+ * \param data batch buffer contents
+ * \param count number of DWORDs to decode in the batch buffer
+ * \param hw_offset hardware address for the buffer
+ */
+int
+intel_decode(const uint32_t *data, int count, uint32_t hw_offset, uint32_t devid)
+{
+    int index = 0;
+    int failures = 0;
+
+    out = stderr;
+
+    while (index < count) {
+       switch ((data[index] & 0xe0000000) >> 29) {
+       case 0x0:
+           index += decode_mi(data + index, count - index,
+                              hw_offset + index * 4, &failures);
+           break;
+       case 0x2:
+           index += decode_2d(data + index, count - index,
+                              hw_offset + index * 4, &failures);
+           break;
+       case 0x3:
+           if (IS_965(devid)) {
+               index += decode_3d_965(data + index, count - index,
+                                      hw_offset + index * 4, &failures);
+           } else if (IS_9XX(devid)) {
+               index += decode_3d(data + index, count - index,
+                                  hw_offset + index * 4, &failures);
+           } else {
+               index += decode_3d_i830(data + index, count - index,
+                                       hw_offset + index * 4, &failures);
+           }
+           break;
+       default:
+           instr_out(data, hw_offset, index, "UNKNOWN\n");
+           failures++;
+           index++;
+           break;
+       }
+       fflush(out);
+    }
+
+    return failures;
+}
+
+void intel_decode_context_reset(void)
+{
+    saved_s2_set = 0;
+    saved_s4_set = 1;
+}
+
diff --git a/src/gallium/drivers/i965/intel_decode.h b/src/gallium/drivers/i965/intel_decode.h
new file mode 100644 (file)
index 0000000..7683097
--- /dev/null
@@ -0,0 +1,29 @@
+/*
+ * Copyright Â© 2007 Intel Corporation
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * the rights to use, copy, modify, merge, publish, distribute, sublicense,
+ * and/or sell copies of the Software, and to permit persons to whom the
+ * Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
+ * THE AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING
+ * FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS
+ * IN THE SOFTWARE.
+ *
+ * Authors:
+ *    Eric Anholt <eric@anholt.net>
+ *
+ */
+
+int intel_decode(const uint32_t *data, int count, uint32_t hw_offset, uint32_t devid);
+void intel_decode_context_reset(void);
diff --git a/src/gallium/drivers/i965/intel_structs.h b/src/gallium/drivers/i965/intel_structs.h
new file mode 100644 (file)
index 0000000..522e3bd
--- /dev/null
@@ -0,0 +1,132 @@
+#ifndef INTEL_STRUCTS_H
+#define INTEL_STRUCTS_H
+
+struct br0 {
+   GLuint length:8;
+   GLuint pad0:3;
+   GLuint dst_tiled:1;
+   GLuint pad1:8;
+   GLuint write_rgb:1;
+   GLuint write_alpha:1;
+   GLuint opcode:7;
+   GLuint client:3;
+};
+
+   
+struct br13 {
+   GLint dest_pitch:16;
+   GLuint rop:8;
+   GLuint color_depth:2;
+   GLuint pad1:3;
+   GLuint mono_source_transparency:1;
+   GLuint clipping_enable:1;
+   GLuint pad0:1;
+};
+
+
+
+/* This is an attempt to move some of the 2D interaction in this
+ * driver to using structs for packets rather than a bunch of #defines
+ * and dwords.
+ */
+struct xy_color_blit {
+   struct br0 br0;
+   struct br13 br13;
+
+   struct {
+      GLuint dest_x1:16;
+      GLuint dest_y1:16;
+   } dw2;
+
+   struct {
+      GLuint dest_x2:16;
+      GLuint dest_y2:16;
+   } dw3;
+   
+   GLuint dest_base_addr;
+   GLuint color;
+};
+
+struct xy_src_copy_blit {
+   struct br0 br0;
+   struct br13 br13;
+
+   struct {
+      GLuint dest_x1:16;
+      GLuint dest_y1:16;
+   } dw2;
+
+   struct {
+      GLuint dest_x2:16;
+      GLuint dest_y2:16;
+   } dw3;
+   
+   GLuint dest_base_addr;
+
+   struct {
+      GLuint src_x1:16;
+      GLuint src_y1:16;
+   } dw5;
+
+   struct {
+      GLint src_pitch:16;
+      GLuint pad:16;
+   } dw6;
+   
+   GLuint src_base_addr;
+};
+
+struct xy_setup_blit {
+   struct br0 br0;
+   struct br13 br13;
+
+   struct {
+      GLuint clip_x1:16;
+      GLuint clip_y1:16;
+   } dw2;
+
+   struct {
+      GLuint clip_x2:16;
+      GLuint clip_y2:16;
+   } dw3;
+      
+   GLuint dest_base_addr;
+   GLuint background_color;
+   GLuint foreground_color;
+   GLuint pattern_base_addr;
+};
+
+
+struct xy_text_immediate_blit {
+   struct {
+      GLuint length:8;
+      GLuint pad2:3;
+      GLuint dst_tiled:1;
+      GLuint pad1:4;
+      GLuint byte_packed:1;
+      GLuint pad0:5;
+      GLuint opcode:7;
+      GLuint client:3;
+   } dw0;
+
+   struct {
+      GLuint dest_x1:16;
+      GLuint dest_y1:16;
+   } dw1;
+
+   struct {
+      GLuint dest_x2:16;
+      GLuint dest_y2:16;
+   } dw2;   
+
+   /* Src bitmap data follows as inline dwords.
+    */
+};
+
+
+#define CLIENT_2D 0x2
+#define OPCODE_XY_SETUP_BLT 0x1
+#define OPCODE_XY_COLOR_BLT 0x50
+#define OPCODE_XY_TEXT_IMMEDIATE_BLT 0x31
+
+#endif
index 4e700089e331010ae0d717858a08a73e360f9387..bdbaae59875b3e94d4ca6ded65433c541fd47ce0 100644 (file)
@@ -45,17 +45,6 @@ identity_destroy(struct pipe_context *_pipe)
    free(id_pipe);
 }
 
-static void
-identity_set_edgeflags(struct pipe_context *_pipe,
-                       const unsigned *bitfield)
-{
-   struct identity_context *id_pipe = identity_context(_pipe);
-   struct pipe_context *pipe = id_pipe->pipe;
-
-   pipe->set_edgeflags(pipe,
-                       bitfield);
-}
-
 static boolean
 identity_draw_arrays(struct pipe_context *_pipe,
                      unsigned prim,
@@ -221,16 +210,29 @@ identity_create_sampler_state(struct pipe_context *_pipe,
 }
 
 static void
-identity_bind_sampler_states(struct pipe_context *_pipe,
-                             unsigned num,
-                             void **samplers)
+identity_bind_fragment_sampler_states(struct pipe_context *_pipe,
+                                      unsigned num_samplers,
+                                      void **samplers)
+{
+   struct identity_context *id_pipe = identity_context(_pipe);
+   struct pipe_context *pipe = id_pipe->pipe;
+
+   pipe->bind_fragment_sampler_states(pipe,
+                                      num_samplers,
+                                      samplers);
+}
+
+static void
+identity_bind_vertex_sampler_states(struct pipe_context *_pipe,
+                                    unsigned num_samplers,
+                                    void **samplers)
 {
    struct identity_context *id_pipe = identity_context(_pipe);
    struct pipe_context *pipe = id_pipe->pipe;
 
-   pipe->bind_sampler_states(pipe,
-                             num,
-                             samplers);
+   pipe->bind_vertex_sampler_states(pipe,
+                                    num_samplers,
+                                    samplers);
 }
 
 static void
@@ -480,9 +482,9 @@ identity_set_viewport_state(struct pipe_context *_pipe,
 }
 
 static void
-identity_set_sampler_textures(struct pipe_context *_pipe,
-                              unsigned num_textures,
-                              struct pipe_texture **_textures)
+identity_set_fragment_sampler_textures(struct pipe_context *_pipe,
+                                       unsigned num_textures,
+                                       struct pipe_texture **_textures)
 {
    struct identity_context *id_pipe = identity_context(_pipe);
    struct pipe_context *pipe = id_pipe->pipe;
@@ -499,9 +501,34 @@ identity_set_sampler_textures(struct pipe_context *_pipe,
       textures = unwrapped_textures;
    }
 
-   pipe->set_sampler_textures(pipe,
-                              num_textures,
-                              textures);
+   pipe->set_fragment_sampler_textures(pipe,
+                                       num_textures,
+                                       textures);
+}
+
+static void
+identity_set_vertex_sampler_textures(struct pipe_context *_pipe,
+                                     unsigned num_textures,
+                                     struct pipe_texture **_textures)
+{
+   struct identity_context *id_pipe = identity_context(_pipe);
+   struct pipe_context *pipe = id_pipe->pipe;
+   struct pipe_texture *unwrapped_textures[PIPE_MAX_VERTEX_SAMPLERS];
+   struct pipe_texture **textures = NULL;
+   unsigned i;
+
+   if (_textures) {
+      for (i = 0; i < num_textures; i++)
+         unwrapped_textures[i] = identity_texture_unwrap(_textures[i]);
+      for (; i < PIPE_MAX_VERTEX_SAMPLERS; i++)
+         unwrapped_textures[i] = NULL;
+
+      textures = unwrapped_textures;
+   }
+
+   pipe->set_vertex_sampler_textures(pipe,
+                                     num_textures,
+                                     textures);
 }
 
 static void
@@ -669,7 +696,6 @@ identity_context_create(struct pipe_screen *_screen, struct pipe_context *pipe)
    id_pipe->base.draw = NULL;
 
    id_pipe->base.destroy = identity_destroy;
-   id_pipe->base.set_edgeflags = identity_set_edgeflags;
    id_pipe->base.draw_arrays = identity_draw_arrays;
    id_pipe->base.draw_elements = identity_draw_elements;
    id_pipe->base.draw_range_elements = identity_draw_range_elements;
@@ -682,7 +708,8 @@ identity_context_create(struct pipe_screen *_screen, struct pipe_context *pipe)
    id_pipe->base.bind_blend_state = identity_bind_blend_state;
    id_pipe->base.delete_blend_state = identity_delete_blend_state;
    id_pipe->base.create_sampler_state = identity_create_sampler_state;
-   id_pipe->base.bind_sampler_states = identity_bind_sampler_states;
+   id_pipe->base.bind_fragment_sampler_states = identity_bind_fragment_sampler_states;
+   id_pipe->base.bind_vertex_sampler_states = identity_bind_vertex_sampler_states;
    id_pipe->base.delete_sampler_state = identity_delete_sampler_state;
    id_pipe->base.create_rasterizer_state = identity_create_rasterizer_state;
    id_pipe->base.bind_rasterizer_state = identity_bind_rasterizer_state;
@@ -703,7 +730,8 @@ identity_context_create(struct pipe_screen *_screen, struct pipe_context *pipe)
    id_pipe->base.set_polygon_stipple = identity_set_polygon_stipple;
    id_pipe->base.set_scissor_state = identity_set_scissor_state;
    id_pipe->base.set_viewport_state = identity_set_viewport_state;
-   id_pipe->base.set_sampler_textures = identity_set_sampler_textures;
+   id_pipe->base.set_fragment_sampler_textures = identity_set_fragment_sampler_textures;
+   id_pipe->base.set_vertex_sampler_textures = identity_set_vertex_sampler_textures;
    id_pipe->base.set_vertex_buffers = identity_set_vertex_buffers;
    id_pipe->base.set_vertex_elements = identity_set_vertex_elements;
    id_pipe->base.surface_copy = identity_surface_copy;
index e893e59940816913b900de77329c730e59a803a8..bc9bc7121d58caf5ac2b23267d64c9c5e9b05f3b 100644 (file)
@@ -180,3 +180,42 @@ identity_transfer_destroy(struct identity_transfer *id_transfer)
    screen->tex_transfer_destroy(id_transfer->transfer);
    FREE(id_transfer);
 }
+
+struct pipe_video_surface *
+identity_video_surface_create(struct identity_screen *id_screen,
+                              struct pipe_video_surface *video_surface)
+{
+   struct identity_video_surface *id_video_surface;
+
+   if (!video_surface) {
+      goto error;
+   }
+
+   assert(video_surface->screen == id_screen->screen);
+
+   id_video_surface = CALLOC_STRUCT(identity_video_surface);
+   if (!id_video_surface) {
+      goto error;
+   }
+
+   memcpy(&id_video_surface->base,
+          video_surface,
+          sizeof(struct pipe_video_surface));
+
+   pipe_reference_init(&id_video_surface->base.reference, 1);
+   id_video_surface->base.screen = &id_screen->base;
+   id_video_surface->video_surface = video_surface;
+
+   return &id_video_surface->base;
+
+error:
+   pipe_video_surface_reference(&video_surface, NULL);
+   return NULL;
+}
+
+void
+identity_video_surface_destroy(struct identity_video_surface *id_video_surface)
+{
+   pipe_video_surface_reference(&id_video_surface->video_surface, NULL);
+   FREE(id_video_surface);
+}
index ce58faa3c7c813c1287ad04a9f5004f1d3f63bba..77cc7190798c75810cecfa076b8987ec6c95c7b4 100644 (file)
@@ -31,6 +31,7 @@
 
 #include "pipe/p_compiler.h"
 #include "pipe/p_state.h"
+#include "pipe/p_video_state.h"
 
 #include "id_screen.h"
 
@@ -67,6 +68,14 @@ struct identity_transfer
 };
 
 
+struct identity_video_surface
+{
+   struct pipe_video_surface base;
+
+   struct pipe_video_surface *video_surface;
+};
+
+
 static INLINE struct identity_buffer *
 identity_buffer(struct pipe_buffer *_buffer)
 {
@@ -103,6 +112,15 @@ identity_transfer(struct pipe_transfer *_transfer)
    return (struct identity_transfer *)_transfer;
 }
 
+static INLINE struct identity_video_surface *
+identity_video_surface(struct pipe_video_surface *_video_surface)
+{
+   if (!_video_surface) {
+      return NULL;
+   }
+   (void)identity_screen(_video_surface->screen);
+   return (struct identity_video_surface *)_video_surface;
+}
 
 static INLINE struct pipe_buffer *
 identity_buffer_unwrap(struct pipe_buffer *_buffer)
@@ -165,5 +183,12 @@ identity_transfer_create(struct identity_texture *id_texture,
 void
 identity_transfer_destroy(struct identity_transfer *id_transfer);
 
+struct pipe_video_surface *
+identity_video_surface_create(struct identity_screen *id_screen,
+                              struct pipe_video_surface *video_surface);
+
+void
+identity_video_surface_destroy(struct identity_video_surface *id_video_surface);
+
 
 #endif /* ID_OBJECTS_H */
index cac14cfd60466db6b6827841017e236563e6ce42..3d2862eaa01e506ce5fe292bcb49feb93b767dc1 100644 (file)
@@ -37,4 +37,4 @@ identity_screen_create(struct pipe_screen *screen);
 struct pipe_context *
 identity_context_create(struct pipe_screen *screen, struct pipe_context *pipe);
 
-#endif /* PT_PUBLIC_H */
+#endif /* ID_PUBLIC_H */
index 26439637d08ffffeb98acb4d9a7c59bab458b592..53eae3ef5440c821bb78e1d9471befadaaa75b1d 100644 (file)
@@ -379,6 +379,33 @@ identity_screen_buffer_destroy(struct pipe_buffer *_buffer)
    identity_buffer_destroy(identity_buffer(_buffer));
 }
 
+static struct pipe_video_surface *
+identity_screen_video_surface_create(struct pipe_screen *_screen,
+                                     enum pipe_video_chroma_format chroma_format,
+                                     unsigned width,
+                                     unsigned height)
+{
+   struct identity_screen *id_screen = identity_screen(_screen);
+   struct pipe_screen *screen = id_screen->screen;
+   struct pipe_video_surface *result;
+
+   result = screen->video_surface_create(screen,
+                                         chroma_format,
+                                         width,
+                                         height);
+
+   if (result) {
+      return identity_video_surface_create(id_screen, result);
+   }
+   return NULL;
+}
+
+static void
+identity_screen_video_surface_destroy(struct pipe_video_surface *_vsfc)
+{
+   identity_video_surface_destroy(identity_video_surface(_vsfc));
+}
+
 static void
 identity_screen_flush_frontbuffer(struct pipe_screen *_screen,
                                   struct pipe_surface *_surface,
@@ -472,6 +499,12 @@ identity_screen_create(struct pipe_screen *screen)
    if (screen->buffer_unmap)
       id_screen->base.buffer_unmap = identity_screen_buffer_unmap;
    id_screen->base.buffer_destroy = identity_screen_buffer_destroy;
+   if (screen->video_surface_create) {
+      id_screen->base.video_surface_create = identity_screen_video_surface_create;
+   }
+   if (screen->video_surface_destroy) {
+      id_screen->base.video_surface_destroy = identity_screen_video_surface_destroy;
+   }
    id_screen->base.flush_frontbuffer = identity_screen_flush_frontbuffer;
    id_screen->base.fence_reference = identity_screen_fence_reference;
    id_screen->base.fence_signalled = identity_screen_fence_signalled;
index 89d08834a3c408a4c78276024a6b1b5f6f1608ef..0c3f00fd58f89b69b57821c803a37d00f3b7c3aa 100644 (file)
@@ -51,21 +51,22 @@ Requirements
 
  - Linux
  
- - udis86, http://udis86.sourceforge.net/ . Use my repository, which decodes
-   opcodes not yet supported by upstream.
+ - A x86 or amd64 processor.  64bit mode is preferred.
  
-     git clone git://people.freedesktop.org/~jrfonseca/udis86
-     cd udis86
-     ./configure --with-pic
-     make
-     sudo make install
+   Support for sse2 is strongly encouraged.  Support for ssse3, and sse4.1 will
+   yield the most efficient code.  The less features the CPU has the more
+   likely is that you ran into underperforming, buggy, or incomplete code.  
+   
+   See /proc/cpuinfo to know what your CPU supports.
+ - LLVM 2.5 or greater. LLVM 2.6 is preferred.
  
- LLVM 2.5. On Debian based distributions do:
  On Debian based distributions do:
  
      aptitude install llvm-dev
 
-   There is a typo in one of the llvm-dev 2.5 headers, that causes compilation
-   errors in the debug build:
+   There is a typo in one of the llvm 2.5 headers, that may cause compilation
+   errors. To fix it apply the change:
 
      --- /usr/include/llvm-c/Core.h.orig       2009-08-10 15:38:54.000000000 +0100
      +++ /usr/include/llvm-c/Core.h    2009-08-10 15:38:25.000000000 +0100
@@ -79,12 +80,17 @@ Requirements
           #endif
           return reinterpret_cast<T**>(Vals);
  
- - A x86 or amd64 processor with support for sse2, sse3, and sse4.1 SIMD
-   instructions. This is necessary because we emit several SSE intrinsics for
-   convenience. See /proc/cpuinfo to know what your CPU supports.
- - scons
+ - scons (optional)
 
+ - udis86, http://udis86.sourceforge.net/ (optional):
+     git clone git://udis86.git.sourceforge.net/gitroot/udis86/udis86
+     cd udis86
+     ./autogen.sh
+     ./configure --with-pic
+     make
+     sudo make install
 
 Building
 ========
index 3bd2e7001386e933f6b6bd6d16e2d7e37c1d77cd..de6156795d391dd3eab651e2d2a4d2e2556599a2 100644 (file)
@@ -76,19 +76,17 @@ env = env.Clone()
 
 env.Prepend(LIBS = [llvmpipe] + auxiliaries)
 
-env.Program(
-    target = 'lp_test_format',
-    source = ['lp_test_format.c', 'lp_test_main.c'],
-)
-
-env.Program(
-    target = 'lp_test_blend',
-    source = ['lp_test_blend.c', 'lp_test_main.c'],
-)
+tests = [
+    'format',
+    'blend',
+    'conv',
+]
 
-env.Program(
-    target = 'lp_test_conv',
-    source = ['lp_test_conv.c', 'lp_test_main.c'],
-)
+for test in tests:
+    target = env.Program(
+        target = 'lp_test_' + test,
+        source = ['lp_test_' + test + '.c', 'lp_test_main.c'],
+    )
+    env.InstallProgram(target)
 
 Export('llvmpipe')
index 9c59677a741adc361320751f2af0333fb274e5aa..eea6b5d6a5c9d2d1002fa9ff5b6edac8dba0c07c 100644 (file)
@@ -629,7 +629,8 @@ lp_build_abs(struct lp_build_context *bld,
    if(type.floating) {
       /* Mask out the sign bit */
       LLVMTypeRef int_vec_type = lp_build_int_vec_type(type);
-      LLVMValueRef mask = lp_build_int_const_scalar(type, ((unsigned long long)1 << type.width) - 1);
+      unsigned long absMask = ~(1 << (type.width - 1));
+      LLVMValueRef mask = lp_build_int_const_scalar(type, ((unsigned long long) absMask));
       a = LLVMBuildBitCast(bld->builder, a, int_vec_type, "");
       a = LLVMBuildAnd(bld->builder, a, mask, "");
       a = LLVMBuildBitCast(bld->builder, a, vec_type, "");
@@ -1083,7 +1084,7 @@ lp_build_log(struct lp_build_context *bld,
              LLVMValueRef x)
 {
    /* log(2) */
-   LLVMValueRef log2 = lp_build_const_scalar(bld->type, 1.4426950408889634);
+   LLVMValueRef log2 = lp_build_const_scalar(bld->type, 0.69314718055994529);
 
    return lp_build_mul(bld, log2, lp_build_exp2(bld, x));
 }
@@ -1095,7 +1096,7 @@ lp_build_log(struct lp_build_context *bld,
 
 /**
  * Generate polynomial.
- * Ex:  x^2 * coeffs[0] + x * coeffs[1] + coeffs[2].
+ * Ex:  coeffs[0] + x * coeffs[1] + x^2 * coeffs[2].
  */
 static LLVMValueRef
 lp_build_polynomial(struct lp_build_context *bld,
@@ -1285,13 +1286,13 @@ lp_build_log2_approx(struct lp_build_context *bld,
       /* mant = (float) mantissa(x) */
       mant = LLVMBuildAnd(bld->builder, i, mantmask, "");
       mant = LLVMBuildOr(bld->builder, mant, one, "");
-      mant = LLVMBuildSIToFP(bld->builder, mant, vec_type, "");
+      mant = LLVMBuildBitCast(bld->builder, mant, vec_type, "");
 
       logmant = lp_build_polynomial(bld, mant, lp_build_log2_polynomial,
                                     Elements(lp_build_log2_polynomial));
 
       /* This effectively increases the polynomial degree by one, but ensures that log2(1) == 0*/
-      logmant = LLVMBuildMul(bld->builder, logmant, LLVMBuildMul(bld->builder, mant, bld->one, ""), "");
+      logmant = LLVMBuildMul(bld->builder, logmant, LLVMBuildSub(bld->builder, mant, bld->one, ""), "");
 
       res = LLVMBuildAdd(bld->builder, logmant, logexp, "");
    }
index 59d8f492e60af0af4585c1fcaa5ac95f8154e5f6..39dfc51e50331b6cabeaa2f076a2fc069f610b9f 100644 (file)
@@ -77,10 +77,10 @@ lp_disassemble(const void* func)
    while (ud_disassemble(&ud_obj)) {
 
 #ifdef PIPE_ARCH_X86
-      debug_printf("%08lx: ", (unsigned long)ud_insn_off(&ud_obj));
+      debug_printf("0x%08lx:\t", (unsigned long)ud_insn_off(&ud_obj));
 #endif
 #ifdef PIPE_ARCH_X86_64
-      debug_printf("%016llx: ", (unsigned long long)ud_insn_off(&ud_obj));
+      debug_printf("0x%016llx:\t", (unsigned long long)ud_insn_off(&ud_obj));
 #endif
 
 #if 0
@@ -115,9 +115,16 @@ lp_disassemble(const void* func)
          }
       }
 
-      if (ud_insn_off(&ud_obj) >= max_jmp_pc && ud_obj.mnemonic == UD_Iret)
+      if ((ud_insn_off(&ud_obj) >= max_jmp_pc && ud_obj.mnemonic == UD_Iret) ||
+           ud_obj.mnemonic == UD_Iinvalid)
          break;
    }
+
+#if 0
+   /* Print GDB command, useful to verify udis86 output */
+   debug_printf("disassemble %p %p\n", func, (void*)(uintptr_t)ud_obj.pc);
+#endif
+
    debug_printf("\n");
 #else
    (void)func;
index 98ec1cb1b9dca428118830ae8e8f70173a4d9648..d438c0e63d7d43695775d89ed9ca137c8cf4d9db 100644 (file)
@@ -90,7 +90,7 @@ lp_depth_type(const struct util_format_description *format_desc,
 
    if(format_desc->channel[swizzle].type == UTIL_FORMAT_TYPE_FLOAT) {
       type.floating = TRUE;
-      assert(swizzle = 0);
+      assert(swizzle == 0);
       assert(format_desc->channel[swizzle].size == format_desc->block.bits);
    }
    else if(format_desc->channel[swizzle].type == UTIL_FORMAT_TYPE_UNSIGNED) {
index 818c0e943e3dacbf4722111f3ccd40662b4ca4b9..49dab8ab61eed907c5d465e1579a29c910fa47a5 100644 (file)
@@ -303,8 +303,8 @@ lp_build_interp_soa_init(struct lp_build_interp_soa_context *bld,
             unsigned first, last, mask;
             unsigned attrib;
 
-            first = decl->DeclarationRange.First;
-            last = decl->DeclarationRange.Last;
+            first = decl->Range.First;
+            last = decl->Range.Last;
             mask = decl->Declaration.UsageMask;
 
             for( attrib = first; attrib <= last; ++attrib ) {
index 4d272bea87e7add07363045510a35a1a3ee4ddf4..af70ddc6ab98ddd47d1261184a1080f7e18aa5b3 100644 (file)
@@ -59,9 +59,9 @@ lp_sampler_static_state(struct lp_sampler_static_state *state,
 
    state->format            = texture->format;
    state->target            = texture->target;
-   state->pot_width         = util_is_pot(texture->width[0]);
-   state->pot_height        = util_is_pot(texture->height[0]);
-   state->pot_depth         = util_is_pot(texture->depth[0]);
+   state->pot_width         = util_is_pot(texture->width0);
+   state->pot_height        = util_is_pot(texture->height0);
+   state->pot_depth         = util_is_pot(texture->depth0);
 
    state->wrap_s            = sampler->wrap_s;
    state->wrap_t            = sampler->wrap_t;
index 94ce4ae83186e38076249a7a306af0274ab4145c..a67c70ff25a277a394476e9d916e2443aa1b7692 100644 (file)
@@ -64,7 +64,7 @@
    for (CHAN = 0; CHAN < NUM_CHANNELS; CHAN++)
 
 #define IS_DST0_CHANNEL_ENABLED( INST, CHAN )\
-   ((INST)->FullDstRegisters[0].DstRegister.WriteMask & (1 << (CHAN)))
+   ((INST)->Dst[0].Register.WriteMask & (1 << (CHAN)))
 
 #define IF_IS_DST0_CHANNEL_ENABLED( INST, CHAN )\
    if (IS_DST0_CHANNEL_ENABLED( INST, CHAN ))
@@ -157,7 +157,7 @@ emit_fetch(
    unsigned index,
    const unsigned chan_index )
 {
-   const struct tgsi_full_src_register *reg = &inst->FullSrcRegisters[index];
+   const struct tgsi_full_src_register *reg = &inst->Src[index];
    unsigned swizzle = tgsi_util_get_full_src_register_swizzle( reg, chan_index );
    LLVMValueRef res;
 
@@ -167,9 +167,9 @@ emit_fetch(
    case TGSI_SWIZZLE_Z:
    case TGSI_SWIZZLE_W:
 
-      switch (reg->SrcRegister.File) {
+      switch (reg->Register.File) {
       case TGSI_FILE_CONSTANT: {
-         LLVMValueRef index = LLVMConstInt(LLVMInt32Type(), reg->SrcRegister.Index*4 + swizzle, 0);
+         LLVMValueRef index = LLVMConstInt(LLVMInt32Type(), reg->Register.Index*4 + swizzle, 0);
          LLVMValueRef scalar_ptr = LLVMBuildGEP(bld->base.builder, bld->consts_ptr, &index, 1, "");
          LLVMValueRef scalar = LLVMBuildLoad(bld->base.builder, scalar_ptr, "");
          res = lp_build_broadcast_scalar(&bld->base, scalar);
@@ -177,17 +177,17 @@ emit_fetch(
       }
 
       case TGSI_FILE_IMMEDIATE:
-         res = bld->immediates[reg->SrcRegister.Index][swizzle];
+         res = bld->immediates[reg->Register.Index][swizzle];
          assert(res);
          break;
 
       case TGSI_FILE_INPUT:
-         res = bld->inputs[reg->SrcRegister.Index][swizzle];
+         res = bld->inputs[reg->Register.Index][swizzle];
          assert(res);
          break;
 
       case TGSI_FILE_TEMPORARY:
-         res = bld->temps[reg->SrcRegister.Index][swizzle];
+         res = bld->temps[reg->Register.Index][swizzle];
          if(!res)
             return bld->base.undef;
          break;
@@ -267,7 +267,7 @@ emit_store(
    unsigned chan_index,
    LLVMValueRef value)
 {
-   const struct tgsi_full_dst_register *reg = &inst->FullDstRegisters[index];
+   const struct tgsi_full_dst_register *reg = &inst->Dst[index];
 
    switch( inst->Instruction.Saturate ) {
    case TGSI_SAT_NONE:
@@ -287,13 +287,13 @@ emit_store(
       assert(0);
    }
 
-   switch( reg->DstRegister.File ) {
+   switch( reg->Register.File ) {
    case TGSI_FILE_OUTPUT:
-      bld->outputs[reg->DstRegister.Index][chan_index] = value;
+      bld->outputs[reg->Register.Index][chan_index] = value;
       break;
 
    case TGSI_FILE_TEMPORARY:
-      bld->temps[reg->DstRegister.Index][chan_index] = value;
+      bld->temps[reg->Register.Index][chan_index] = value;
       break;
 
    case TGSI_FILE_ADDRESS:
@@ -319,14 +319,14 @@ emit_tex( struct lp_build_tgsi_soa_context *bld,
           boolean projected,
           LLVMValueRef *texel)
 {
-   const uint unit = inst->FullSrcRegisters[1].SrcRegister.Index;
+   const uint unit = inst->Src[1].Register.Index;
    LLVMValueRef lodbias;
    LLVMValueRef oow;
    LLVMValueRef coords[3];
    unsigned num_coords;
    unsigned i;
 
-   switch (inst->InstructionExtTexture.Texture) {
+   switch (inst->Texture.Texture) {
    case TGSI_TEXTURE_1D:
       num_coords = 1;
       break;
@@ -375,7 +375,7 @@ emit_kil(
    struct lp_build_tgsi_soa_context *bld,
    const struct tgsi_full_instruction *inst )
 {
-   const struct tgsi_full_src_register *reg = &inst->FullSrcRegisters[0];
+   const struct tgsi_full_src_register *reg = &inst->Src[0];
    LLVMValueRef terms[NUM_CHANNELS];
    LLVMValueRef mask;
    unsigned chan_index;
@@ -423,15 +423,15 @@ indirect_temp_reference(const struct tgsi_full_instruction *inst)
 {
    uint i;
    for (i = 0; i < inst->Instruction.NumSrcRegs; i++) {
-      const struct tgsi_full_src_register *reg = &inst->FullSrcRegisters[i];
-      if (reg->SrcRegister.File == TGSI_FILE_TEMPORARY &&
-          reg->SrcRegister.Indirect)
+      const struct tgsi_full_src_register *reg = &inst->Src[i];
+      if (reg->Register.File == TGSI_FILE_TEMPORARY &&
+          reg->Register.Indirect)
          return TRUE;
    }
    for (i = 0; i < inst->Instruction.NumDstRegs; i++) {
-      const struct tgsi_full_dst_register *reg = &inst->FullDstRegisters[i];
-      if (reg->DstRegister.File == TGSI_FILE_TEMPORARY &&
-          reg->DstRegister.Indirect)
+      const struct tgsi_full_dst_register *reg = &inst->Dst[i];
+      if (reg->Register.File == TGSI_FILE_TEMPORARY &&
+          reg->Register.Indirect)
          return TRUE;
    }
    return FALSE;
@@ -763,7 +763,7 @@ emit_instruction(
       FOR_EACH_DST0_ENABLED_CHANNEL( inst, chan_index ) {
          src0 = emit_fetch( bld, inst, 0, chan_index );
          tmp0 = lp_build_floor(&bld->base, src0);
-         tmp0 = lp_build_sub(&bld->base, tmp0, src0);
+         tmp0 = lp_build_sub(&bld->base, src0, tmp0);
          dst0[chan_index] = tmp0;
       }
       break;
index bdcff94b9bf1fe06b4800361a63d9720285078c8..08d9f2e2735e3f7eb2ce7eb5f20b3981dd6fa472 100644 (file)
@@ -50,6 +50,7 @@ llvmpipe_clear(struct pipe_context *pipe, unsigned buffers, const float *rgba,
                double depth, unsigned stencil)
 {
    struct llvmpipe_context *llvmpipe = llvmpipe_context(pipe);
+   union util_color uc;
    unsigned cv;
    uint i;
 
@@ -64,8 +65,8 @@ llvmpipe_clear(struct pipe_context *pipe, unsigned buffers, const float *rgba,
       for (i = 0; i < llvmpipe->framebuffer.nr_cbufs; i++) {
          struct pipe_surface *ps = llvmpipe->framebuffer.cbufs[i];
 
-         util_pack_color(rgba, ps->format, &cv);
-         lp_tile_cache_clear(llvmpipe->cbuf_cache[i], rgba, cv);
+         util_pack_color(rgba, ps->format, &uc);
+         lp_tile_cache_clear(llvmpipe->cbuf_cache[i], rgba, uc.ui);
       }
       llvmpipe->dirty_render_cache = TRUE;
    }
index 57e71f3e98644ec878cd302c588df59d3d090ebd..37587d4f792c2af5bc6c48d4e3253ce6c77fc4e8 100644 (file)
@@ -118,6 +118,11 @@ static void llvmpipe_destroy( struct pipe_context *pipe )
       pipe_texture_reference(&llvmpipe->texture[i], NULL);
    }
 
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      lp_destroy_tex_tile_cache(llvmpipe->vertex_tex_cache[i]);
+      pipe_texture_reference(&llvmpipe->vertex_textures[i], NULL);
+   }
+
    for (i = 0; i < Elements(llvmpipe->constants); i++) {
       if (llvmpipe->constants[i].buffer) {
          pipe_buffer_reference(&llvmpipe->constants[i].buffer, NULL);
@@ -135,6 +140,7 @@ llvmpipe_is_texture_referenced( struct pipe_context *pipe,
    struct llvmpipe_context *llvmpipe = llvmpipe_context( pipe );
    unsigned i;
 
+   /* check if any of the bound drawing surfaces are this texture */
    if(llvmpipe->dirty_render_cache) {
       for (i = 0; i < llvmpipe->framebuffer.nr_cbufs; i++) {
          if(llvmpipe->framebuffer.cbufs[i] && 
@@ -145,6 +151,18 @@ llvmpipe_is_texture_referenced( struct pipe_context *pipe,
          llvmpipe->framebuffer.zsbuf->texture == texture)
          return PIPE_REFERENCED_FOR_WRITE;
    }
+
+   /* check if any of the tex_cache textures are this texture */
+   for (i = 0; i < PIPE_MAX_SAMPLERS; i++) {
+      if (llvmpipe->tex_cache[i] &&
+            llvmpipe->tex_cache[i]->texture == texture)
+         return PIPE_REFERENCED_FOR_READ;
+   }
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      if (llvmpipe->vertex_tex_cache[i] &&
+          llvmpipe->vertex_tex_cache[i]->texture == texture)
+         return PIPE_REFERENCED_FOR_READ;
+   }
    
    return PIPE_UNREFERENCED;
 }
@@ -180,7 +198,8 @@ llvmpipe_create( struct pipe_screen *screen )
    llvmpipe->pipe.delete_blend_state = llvmpipe_delete_blend_state;
 
    llvmpipe->pipe.create_sampler_state = llvmpipe_create_sampler_state;
-   llvmpipe->pipe.bind_sampler_states  = llvmpipe_bind_sampler_states;
+   llvmpipe->pipe.bind_fragment_sampler_states  = llvmpipe_bind_sampler_states;
+   llvmpipe->pipe.bind_vertex_sampler_states  = llvmpipe_bind_vertex_sampler_states;
    llvmpipe->pipe.delete_sampler_state = llvmpipe_delete_sampler_state;
 
    llvmpipe->pipe.create_depth_stencil_alpha_state = llvmpipe_create_depth_stencil_state;
@@ -205,7 +224,8 @@ llvmpipe_create( struct pipe_screen *screen )
    llvmpipe->pipe.set_framebuffer_state = llvmpipe_set_framebuffer_state;
    llvmpipe->pipe.set_polygon_stipple = llvmpipe_set_polygon_stipple;
    llvmpipe->pipe.set_scissor_state = llvmpipe_set_scissor_state;
-   llvmpipe->pipe.set_sampler_textures = llvmpipe_set_sampler_textures;
+   llvmpipe->pipe.set_fragment_sampler_textures = llvmpipe_set_sampler_textures;
+   llvmpipe->pipe.set_vertex_sampler_textures = llvmpipe_set_vertex_sampler_textures;
    llvmpipe->pipe.set_viewport_state = llvmpipe_set_viewport_state;
 
    llvmpipe->pipe.set_vertex_buffers = llvmpipe_set_vertex_buffers;
@@ -214,8 +234,6 @@ llvmpipe_create( struct pipe_screen *screen )
    llvmpipe->pipe.draw_arrays = llvmpipe_draw_arrays;
    llvmpipe->pipe.draw_elements = llvmpipe_draw_elements;
    llvmpipe->pipe.draw_range_elements = llvmpipe_draw_range_elements;
-   llvmpipe->pipe.set_edgeflags = llvmpipe_set_edgeflags;
-
 
    llvmpipe->pipe.clear = llvmpipe_clear;
    llvmpipe->pipe.flush = llvmpipe_flush;
@@ -234,13 +252,15 @@ llvmpipe_create( struct pipe_screen *screen )
 
    for (i = 0; i < PIPE_MAX_SAMPLERS; i++)
       llvmpipe->tex_cache[i] = lp_create_tex_tile_cache( screen );
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++)
+      llvmpipe->vertex_tex_cache[i] = lp_create_tex_tile_cache(screen);
 
 
    /* vertex shader samplers */
-   for (i = 0; i < PIPE_MAX_SAMPLERS; i++) {
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
       llvmpipe->tgsi.vert_samplers[i].base.get_samples = lp_get_samples;
       llvmpipe->tgsi.vert_samplers[i].processor = TGSI_PROCESSOR_VERTEX;
-      llvmpipe->tgsi.vert_samplers[i].cache = llvmpipe->tex_cache[i];
+      llvmpipe->tgsi.vert_samplers[i].cache = llvmpipe->vertex_tex_cache[i];
       llvmpipe->tgsi.vert_samplers_list[i] = &llvmpipe->tgsi.vert_samplers[i];
    }
 
@@ -260,7 +280,7 @@ llvmpipe_create( struct pipe_screen *screen )
       goto fail;
 
    draw_texture_samplers(llvmpipe->draw,
-                         PIPE_MAX_SAMPLERS,
+                         PIPE_MAX_VERTEX_SAMPLERS,
                          (struct tgsi_sampler **)
                             llvmpipe->tgsi.vert_samplers_list);
 
index 3ad95d0bfc2f0eeffa489b8a983ce3b7e3271287..cc4d5ad5fd9423301db47ce674466e226d9e80d4 100644 (file)
@@ -55,6 +55,7 @@ struct llvmpipe_context {
    /** Constant state objects */
    const struct pipe_blend_state *blend;
    const struct pipe_sampler_state *sampler[PIPE_MAX_SAMPLERS];
+   struct pipe_sampler_state *vertex_samplers[PIPE_MAX_VERTEX_SAMPLERS];
    const struct pipe_depth_stencil_alpha_state *depth_stencil;
    const struct pipe_rasterizer_state *rasterizer;
    struct lp_fragment_shader *fs;
@@ -68,12 +69,15 @@ struct llvmpipe_context {
    struct pipe_poly_stipple poly_stipple;
    struct pipe_scissor_state scissor;
    struct pipe_texture *texture[PIPE_MAX_SAMPLERS];
+   struct pipe_texture *vertex_textures[PIPE_MAX_VERTEX_SAMPLERS];
    struct pipe_viewport_state viewport;
    struct pipe_vertex_buffer vertex_buffer[PIPE_MAX_ATTRIBS];
    struct pipe_vertex_element vertex_element[PIPE_MAX_ATTRIBS];
 
    unsigned num_samplers;
    unsigned num_textures;
+   unsigned num_vertex_samplers;
+   unsigned num_vertex_textures;
    unsigned num_vertex_elements;
    unsigned num_vertex_buffers;
 
@@ -136,6 +140,7 @@ struct llvmpipe_context {
 
    unsigned tex_timestamp;
    struct llvmpipe_tex_tile_cache *tex_cache[PIPE_MAX_SAMPLERS];
+   struct llvmpipe_tex_tile_cache *vertex_tex_cache[PIPE_MAX_VERTEX_SAMPLERS];
 
    unsigned no_rast : 1;
 
diff --git a/src/gallium/drivers/llvmpipe/lp_debug.h b/src/gallium/drivers/llvmpipe/lp_debug.h
new file mode 100644 (file)
index 0000000..74b2757
--- /dev/null
@@ -0,0 +1,71 @@
+/**************************************************************************
+ * 
+ * Copyright 2007 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+
+#ifndef LP_DEBUG_H
+#define LP_DEBUG_H
+
+#include "pipe/p_compiler.h"
+#include "util/u_debug.h"
+
+extern void
+st_print_current(void);
+
+
+#define DEBUG_PIPE      0x1
+#define DEBUG_TGSI      0x2
+#define DEBUG_TEX       0x4
+#define DEBUG_ASM       0x8
+#define DEBUG_SETUP     0x10
+#define DEBUG_RAST      0x20
+#define DEBUG_QUERY     0x40
+#define DEBUG_SCREEN    0x80
+#define DEBUG_JIT       0x100
+
+#ifdef DEBUG
+extern int LP_DEBUG;
+#else
+#define LP_DEBUG 0
+#endif
+
+void st_debug_init( void );
+
+static INLINE void
+LP_DBG( unsigned flag, const char *fmt, ... )
+{
+    if (LP_DEBUG & flag)
+    {
+        va_list args;
+
+        va_start( args, fmt );
+        debug_vprintf( fmt, args );
+        va_end( args );
+    }
+}
+
+
+#endif /* LP_DEBUG_H */
index 0aa13a1fc679d282e7606a42007d36855a7cf005..2299566c665ea78d21da369f57a6665b9e95055c 100644 (file)
@@ -133,10 +133,3 @@ llvmpipe_draw_elements(struct pipe_context *pipe,
                                         mode, start, count );
 }
 
-
-void
-llvmpipe_set_edgeflags(struct pipe_context *pipe, const unsigned *edgeflags)
-{
-   struct llvmpipe_context *lp = llvmpipe_context(pipe);
-   draw_set_edgeflags(lp->draw, edgeflags);
-}
index 13535dd638e6146ba375e61d4f4abc1f04c2d4d1..bce3baec1643637f70dfbc8c2d1988aa0c5282ab 100644 (file)
@@ -154,6 +154,7 @@ lp_jit_screen_init(struct llvmpipe_screen *screen)
 #if 0
    /* For simulating less capable machines */
    util_cpu_caps.has_sse3 = 0;
+   util_cpu_caps.has_ssse3 = 0;
    util_cpu_caps.has_sse4_1 = 0;
 #endif
 
@@ -167,7 +168,7 @@ lp_jit_screen_init(struct llvmpipe_screen *screen)
    if (LLVMCreateJITCompiler(&screen->engine, screen->provider, 1, &error)) {
       _debug_printf("%s\n", error);
       LLVMDisposeMessage(error);
-      abort();
+      assert(0);
    }
 
    screen->target = LLVMGetExecutionEngineTargetData(screen->engine);
index 05189274589608a2cd95d8965e45e2c00295df59..9b47415f003350c7592894fb6338f50c7943e2b6 100644 (file)
@@ -27,6 +27,7 @@
 
 
 #include "util/u_memory.h"
+#include "util/u_format.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_screen.h"
 
 #include "lp_winsys.h"
 #include "lp_jit.h"
 #include "lp_screen.h"
+#include "lp_debug.h"
+
+#ifdef DEBUG
+int LP_DEBUG = 0;
+
+static const struct debug_named_value lp_debug_flags[] = {
+   { "pipe",   DEBUG_PIPE },
+   { "tgsi",   DEBUG_TGSI },
+   { "tex",    DEBUG_TEX },
+   { "asm",    DEBUG_ASM },
+   { "setup",  DEBUG_SETUP },
+   { "rast",   DEBUG_RAST },
+   { "query",  DEBUG_QUERY },
+   { "screen", DEBUG_SCREEN },
+   { "jit",    DEBUG_JIT },
+   {NULL, 0}
+};
+#endif
 
 
 static const char *
@@ -58,7 +77,9 @@ llvmpipe_get_param(struct pipe_screen *screen, int param)
    case PIPE_CAP_MAX_TEXTURE_IMAGE_UNITS:
       return PIPE_MAX_SAMPLERS;
    case PIPE_CAP_MAX_VERTEX_TEXTURE_UNITS:
-      return PIPE_MAX_SAMPLERS;
+      return PIPE_MAX_VERTEX_SAMPLERS;
+   case PIPE_CAP_MAX_COMBINED_SAMPLERS:
+      return PIPE_MAX_SAMPLERS + PIPE_MAX_VERTEX_SAMPLERS;
    case PIPE_CAP_NPOT_TEXTURES:
       return 1;
    case PIPE_CAP_TWO_SIDED_STENCIL:
@@ -131,17 +152,17 @@ llvmpipe_is_format_supported( struct pipe_screen *_screen,
 {
    struct llvmpipe_screen *screen = llvmpipe_screen(_screen);
    struct llvmpipe_winsys *winsys = screen->winsys;
+   const struct util_format_description *format_desc;
+
+   format_desc = util_format_description(format);
+   if(!format_desc)
+      return FALSE;
 
    assert(target == PIPE_TEXTURE_1D ||
           target == PIPE_TEXTURE_2D ||
           target == PIPE_TEXTURE_3D ||
           target == PIPE_TEXTURE_CUBE);
 
-   if(format == PIPE_FORMAT_Z16_UNORM)
-      return FALSE;
-   if(format == PIPE_FORMAT_S8_UNORM)
-      return FALSE;
-
    switch(format) {
    case PIPE_FORMAT_DXT1_RGB:
    case PIPE_FORMAT_DXT1_RGBA:
@@ -152,8 +173,51 @@ llvmpipe_is_format_supported( struct pipe_screen *_screen,
       break;
    }
 
-   if(tex_usage & PIPE_TEXTURE_USAGE_DISPLAY_TARGET)
-      return winsys->is_displaytarget_format_supported(winsys, format);
+   if(tex_usage & PIPE_TEXTURE_USAGE_RENDER_TARGET) {
+      if(format_desc->block.width != 1 ||
+         format_desc->block.height != 1)
+         return FALSE;
+
+      if(format_desc->layout != UTIL_FORMAT_LAYOUT_SCALAR &&
+         format_desc->layout != UTIL_FORMAT_LAYOUT_ARITH &&
+         format_desc->layout != UTIL_FORMAT_LAYOUT_ARRAY)
+         return FALSE;
+
+      if(format_desc->colorspace != UTIL_FORMAT_COLORSPACE_RGB &&
+         format_desc->colorspace != UTIL_FORMAT_COLORSPACE_SRGB)
+         return FALSE;
+   }
+
+   if(tex_usage & PIPE_TEXTURE_USAGE_DISPLAY_TARGET) {
+      if(!winsys->is_displaytarget_format_supported(winsys, format))
+         return FALSE;
+   }
+
+   if(tex_usage & PIPE_TEXTURE_USAGE_DEPTH_STENCIL) {
+      if(format_desc->colorspace != UTIL_FORMAT_COLORSPACE_ZS)
+         return FALSE;
+
+      /* FIXME: Temporary restriction. See lp_state_fs.c. */
+      if(format_desc->block.bits != 32)
+         return FALSE;
+   }
+
+   /* FIXME: Temporary restrictions. See lp_bld_sample_soa.c */
+   if(tex_usage & PIPE_TEXTURE_USAGE_SAMPLER) {
+      if(format_desc->block.width != 1 ||
+         format_desc->block.height != 1)
+         return FALSE;
+
+      if(format_desc->layout != UTIL_FORMAT_LAYOUT_SCALAR &&
+         format_desc->layout != UTIL_FORMAT_LAYOUT_ARITH &&
+         format_desc->layout != UTIL_FORMAT_LAYOUT_ARRAY)
+         return FALSE;
+
+      if(format_desc->colorspace != UTIL_FORMAT_COLORSPACE_RGB &&
+         format_desc->colorspace != UTIL_FORMAT_COLORSPACE_SRGB &&
+         format_desc->colorspace != UTIL_FORMAT_COLORSPACE_ZS)
+         return FALSE;
+   }
 
    return TRUE;
 }
@@ -213,6 +277,10 @@ llvmpipe_create_screen(struct llvmpipe_winsys *winsys)
 {
    struct llvmpipe_screen *screen = CALLOC_STRUCT(llvmpipe_screen);
 
+#ifdef DEBUG
+   LP_DEBUG = debug_get_flags_option("LP_DEBUG", lp_debug_flags, 0 );
+#endif
+
    if (!screen)
       return NULL;
 
index ffcbc9a379f08b562f12f233a76ac50afe5adb75..b18f17c0cd34a2ef76f1b01e360d0f58b9d612d3 100644 (file)
@@ -41,6 +41,7 @@
 #include "draw/draw_vertex.h"
 #include "pipe/p_shader_tokens.h"
 #include "pipe/p_thread.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 #include "lp_bld_debug.h"
@@ -166,7 +167,7 @@ shade_quads(struct llvmpipe_context *llvmpipe,
       assert((y % 2) == 0);
       depth = llvmpipe->zsbuf_map +
               y*llvmpipe->zsbuf_transfer->stride +
-              2*x*llvmpipe->zsbuf_transfer->block.size;
+              2*x*util_format_get_blocksize(llvmpipe->zsbuf_transfer->texture->format);
    }
    else
       depth = NULL;
index 7b26ce61a380748d1c41882921d5987712835f51..5cee7bf74b6aa5a1d4d6053286bc0b818a4b7a81 100644 (file)
@@ -126,6 +126,10 @@ void *
 llvmpipe_create_sampler_state(struct pipe_context *,
                               const struct pipe_sampler_state *);
 void llvmpipe_bind_sampler_states(struct pipe_context *, unsigned, void **);
+void
+llvmpipe_bind_vertex_sampler_states(struct pipe_context *,
+                                    unsigned num_samplers,
+                                    void **samplers);
 void llvmpipe_delete_sampler_state(struct pipe_context *, void *);
 
 void *
@@ -172,6 +176,11 @@ void llvmpipe_set_sampler_textures( struct pipe_context *,
                                     unsigned num,
                                     struct pipe_texture ** );
 
+void
+llvmpipe_set_vertex_sampler_textures(struct pipe_context *,
+                                     unsigned num_textures,
+                                     struct pipe_texture **);
+
 void llvmpipe_set_viewport_state( struct pipe_context *,
                                   const struct pipe_viewport_state * );
 
@@ -203,10 +212,6 @@ llvmpipe_draw_range_elements(struct pipe_context *pipe,
                              unsigned max_index,
                              unsigned mode, unsigned start, unsigned count);
 
-void
-llvmpipe_set_edgeflags(struct pipe_context *pipe, const unsigned *edgeflags);
-
-
 void
 llvmpipe_map_transfers(struct llvmpipe_context *lp);
 
index b2e75d3b14e30b1a4efa902b5e13ba47043a77d2..a94cd05ef20c7a9616dc4dbb6c985c2ce58f30fe 100644 (file)
@@ -34,6 +34,7 @@
 #include "util/u_memory.h"
 #include "util/u_math.h"
 #include "util/u_debug_dump.h"
+#include "draw/draw_context.h"
 #include "lp_screen.h"
 #include "lp_context.h"
 #include "lp_state.h"
@@ -51,6 +52,11 @@ void llvmpipe_bind_blend_state( struct pipe_context *pipe,
 {
    struct llvmpipe_context *llvmpipe = llvmpipe_context(pipe);
 
+   if (llvmpipe->blend == blend)
+      return;
+
+   draw_flush(llvmpipe->draw);
+
    llvmpipe->blend = blend;
 
    llvmpipe->dirty |= LP_NEW_BLEND;
@@ -69,6 +75,11 @@ void llvmpipe_set_blend_color( struct pipe_context *pipe,
    struct llvmpipe_context *llvmpipe = llvmpipe_context(pipe);
    unsigned i, j;
 
+   if(memcmp(&llvmpipe->blend_color, blend_color, sizeof *blend_color) == 0)
+      return;
+
+   draw_flush(llvmpipe->draw);
+
    memcpy(&llvmpipe->blend_color, blend_color, sizeof *blend_color);
 
    if(!llvmpipe->jit_context.blend_color)
@@ -99,7 +110,12 @@ llvmpipe_bind_depth_stencil_state(struct pipe_context *pipe,
 {
    struct llvmpipe_context *llvmpipe = llvmpipe_context(pipe);
 
-   llvmpipe->depth_stencil = (const struct pipe_depth_stencil_alpha_state *)depth_stencil;
+   if (llvmpipe->depth_stencil == depth_stencil)
+      return;
+
+   draw_flush(llvmpipe->draw);
+
+   llvmpipe->depth_stencil = depth_stencil;
 
    if(llvmpipe->depth_stencil)
       llvmpipe->jit_context.alpha_ref_value = llvmpipe->depth_stencil->alpha.ref_value;
index c753b183c0ca36f0c4ff7d956c18e39eb7bc107d..acfd7be5f74732f7ab637609594254f9948dc1ef 100644 (file)
@@ -66,7 +66,7 @@ llvmpipe_get_vertex_info(struct llvmpipe_context *llvmpipe)
       /* compute vertex layout now */
       const struct lp_fragment_shader *lpfs = llvmpipe->fs;
       struct vertex_info *vinfo_vbuf = &llvmpipe->vertex_info_vbuf;
-      const uint num = draw_num_vs_outputs(llvmpipe->draw);
+      const uint num = draw_current_shader_outputs(llvmpipe->draw);
       uint i;
 
       /* Tell draw_vbuf to simply emit the whole post-xform vertex
@@ -116,13 +116,13 @@ llvmpipe_get_vertex_info(struct llvmpipe_context *llvmpipe)
          }
 
          /* this includes texcoords and varying vars */
-         src = draw_find_vs_output(llvmpipe->draw,
+         src = draw_find_shader_output(llvmpipe->draw,
                                    lpfs->info.input_semantic_name[i],
                                    lpfs->info.input_semantic_index[i]);
          draw_emit_vertex_attr(vinfo, EMIT_4F, interp, src);
       }
 
-      llvmpipe->psize_slot = draw_find_vs_output(llvmpipe->draw,
+      llvmpipe->psize_slot = draw_find_shader_output(llvmpipe->draw,
                                                  TGSI_SEMANTIC_PSIZE, 0);
       if (llvmpipe->psize_slot > 0) {
          draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_CONSTANT,
@@ -198,10 +198,14 @@ update_tgsi_samplers( struct llvmpipe_context *llvmpipe )
    unsigned i;
 
    /* vertex shader samplers */
-   for (i = 0; i < PIPE_MAX_SAMPLERS; i++) {
-      llvmpipe->tgsi.vert_samplers[i].sampler = llvmpipe->sampler[i];
-      llvmpipe->tgsi.vert_samplers[i].texture = llvmpipe->texture[i];
-      llvmpipe->tgsi.frag_samplers[i].base.get_samples = lp_get_samples;
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      llvmpipe->tgsi.vert_samplers[i].sampler = llvmpipe->vertex_samplers[i];
+      llvmpipe->tgsi.vert_samplers[i].texture = llvmpipe->vertex_textures[i];
+      llvmpipe->tgsi.vert_samplers[i].base.get_samples = lp_get_samples;
+   }
+
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      lp_tex_tile_cache_validate_texture( llvmpipe->vertex_tex_cache[i] );
    }
 
    /* fragment shader samplers */
index 2e9aa9fffe377ac9abc972eef7b6c41761dea364..3e8f595e994026ae92768f4d2516c94e856fb229 100644 (file)
@@ -87,6 +87,7 @@
 #include "lp_state.h"
 #include "lp_quad.h"
 #include "lp_tex_sample.h"
+#include "lp_debug.h"
 
 
 static const unsigned char quad_offset_x[4] = {0, 1, 0, 1};
@@ -148,6 +149,20 @@ generate_depth(LLVMBuilderRef builder,
    format_desc = util_format_description(key->zsbuf_format);
    assert(format_desc);
 
+   /*
+    * Depths are expected to be between 0 and 1, even if they are stored in
+    * floats. Setting these bits here will ensure that the lp_build_conv() call
+    * below won't try to unnecessarily clamp the incoming values.
+    */
+   if(src_type.floating) {
+      src_type.sign = FALSE;
+      src_type.norm = TRUE;
+   }
+   else {
+      assert(!src_type.sign);
+      assert(src_type.norm);
+   }
+
    /* Pick the depth type. */
    dst_type = lp_depth_type(format_desc, src_type.width*src_type.length);
 
@@ -155,14 +170,11 @@ generate_depth(LLVMBuilderRef builder,
    assert(dst_type.width == src_type.width);
    assert(dst_type.length == src_type.length);
 
-#if 1
-   src = lp_build_clamped_float_to_unsigned_norm(builder,
-                                                 src_type,
-                                                 dst_type.width,
-                                                 src);
-#else
    lp_build_conv(builder, src_type, dst_type, &src, 1, &src, 1);
-#endif
+
+   dst_ptr = LLVMBuildBitCast(builder,
+                              dst_ptr,
+                              LLVMPointerType(lp_build_vec_type(dst_type), 0), "");
 
    lp_build_depth_test(builder,
                        &key->depth,
@@ -397,59 +409,58 @@ generate_fragment(struct llvmpipe_context *lp,
    unsigned i;
    unsigned chan;
 
-#ifdef DEBUG
-   tgsi_dump(shader->base.tokens, 0);
-   if(key->depth.enabled) {
-      debug_printf("depth.format = %s\n", pf_name(key->zsbuf_format));
-      debug_printf("depth.func = %s\n", debug_dump_func(key->depth.func, TRUE));
-      debug_printf("depth.writemask = %u\n", key->depth.writemask);
-   }
-   if(key->alpha.enabled) {
-      debug_printf("alpha.func = %s\n", debug_dump_func(key->alpha.func, TRUE));
-      debug_printf("alpha.ref_value = %f\n", key->alpha.ref_value);
-   }
-   if(key->blend.logicop_enable) {
-      debug_printf("blend.logicop_func = %u\n", key->blend.logicop_func);
-   }
-   else if(key->blend.blend_enable) {
-      debug_printf("blend.rgb_func = %s\n",   debug_dump_blend_func  (key->blend.rgb_func, TRUE));
-      debug_printf("rgb_src_factor = %s\n",   debug_dump_blend_factor(key->blend.rgb_src_factor, TRUE));
-      debug_printf("rgb_dst_factor = %s\n",   debug_dump_blend_factor(key->blend.rgb_dst_factor, TRUE));
-      debug_printf("alpha_func = %s\n",       debug_dump_blend_func  (key->blend.alpha_func, TRUE));
-      debug_printf("alpha_src_factor = %s\n", debug_dump_blend_factor(key->blend.alpha_src_factor, TRUE));
-      debug_printf("alpha_dst_factor = %s\n", debug_dump_blend_factor(key->blend.alpha_dst_factor, TRUE));
-   }
-   debug_printf("blend.colormask = 0x%x\n", key->blend.colormask);
-   for(i = 0; i < PIPE_MAX_SAMPLERS; ++i) {
-      if(key->sampler[i].format) {
-         debug_printf("sampler[%u] = \n", i);
-         debug_printf("  .format = %s\n",
-                      pf_name(key->sampler[i].format));
-         debug_printf("  .target = %s\n",
-                      debug_dump_tex_target(key->sampler[i].target, TRUE));
-         debug_printf("  .pot = %u %u %u\n",
-                      key->sampler[i].pot_width,
-                      key->sampler[i].pot_height,
-                      key->sampler[i].pot_depth);
-         debug_printf("  .wrap = %s %s %s\n",
-                      debug_dump_tex_wrap(key->sampler[i].wrap_s, TRUE),
-                      debug_dump_tex_wrap(key->sampler[i].wrap_t, TRUE),
-                      debug_dump_tex_wrap(key->sampler[i].wrap_r, TRUE));
-         debug_printf("  .min_img_filter = %s\n",
-                      debug_dump_tex_filter(key->sampler[i].min_img_filter, TRUE));
-         debug_printf("  .min_mip_filter = %s\n",
-                      debug_dump_tex_mipfilter(key->sampler[i].min_mip_filter, TRUE));
-         debug_printf("  .mag_img_filter = %s\n",
-                      debug_dump_tex_filter(key->sampler[i].mag_img_filter, TRUE));
-         if(key->sampler[i].compare_mode)
-            debug_printf("  .compare_mode = %s\n", debug_dump_func(key->sampler[i].compare_func, TRUE));
-         debug_printf("  .normalized_coords = %u\n", key->sampler[i].normalized_coords);
-         debug_printf("  .prefilter = %u\n", key->sampler[i].prefilter);
+   if (LP_DEBUG & DEBUG_JIT) {
+      tgsi_dump(shader->base.tokens, 0);
+      if(key->depth.enabled) {
+         debug_printf("depth.format = %s\n", pf_name(key->zsbuf_format));
+         debug_printf("depth.func = %s\n", debug_dump_func(key->depth.func, TRUE));
+         debug_printf("depth.writemask = %u\n", key->depth.writemask);
+      }
+      if(key->alpha.enabled) {
+         debug_printf("alpha.func = %s\n", debug_dump_func(key->alpha.func, TRUE));
+         debug_printf("alpha.ref_value = %f\n", key->alpha.ref_value);
+      }
+      if(key->blend.logicop_enable) {
+         debug_printf("blend.logicop_func = %u\n", key->blend.logicop_func);
+      }
+      else if(key->blend.blend_enable) {
+         debug_printf("blend.rgb_func = %s\n",   debug_dump_blend_func  (key->blend.rgb_func, TRUE));
+         debug_printf("rgb_src_factor = %s\n",   debug_dump_blend_factor(key->blend.rgb_src_factor, TRUE));
+         debug_printf("rgb_dst_factor = %s\n",   debug_dump_blend_factor(key->blend.rgb_dst_factor, TRUE));
+         debug_printf("alpha_func = %s\n",       debug_dump_blend_func  (key->blend.alpha_func, TRUE));
+         debug_printf("alpha_src_factor = %s\n", debug_dump_blend_factor(key->blend.alpha_src_factor, TRUE));
+         debug_printf("alpha_dst_factor = %s\n", debug_dump_blend_factor(key->blend.alpha_dst_factor, TRUE));
+      }
+      debug_printf("blend.colormask = 0x%x\n", key->blend.colormask);
+      for(i = 0; i < PIPE_MAX_SAMPLERS; ++i) {
+         if(key->sampler[i].format) {
+            debug_printf("sampler[%u] = \n", i);
+            debug_printf("  .format = %s\n",
+                         pf_name(key->sampler[i].format));
+            debug_printf("  .target = %s\n",
+                         debug_dump_tex_target(key->sampler[i].target, TRUE));
+            debug_printf("  .pot = %u %u %u\n",
+                         key->sampler[i].pot_width,
+                         key->sampler[i].pot_height,
+                         key->sampler[i].pot_depth);
+            debug_printf("  .wrap = %s %s %s\n",
+                         debug_dump_tex_wrap(key->sampler[i].wrap_s, TRUE),
+                         debug_dump_tex_wrap(key->sampler[i].wrap_t, TRUE),
+                         debug_dump_tex_wrap(key->sampler[i].wrap_r, TRUE));
+            debug_printf("  .min_img_filter = %s\n",
+                         debug_dump_tex_filter(key->sampler[i].min_img_filter, TRUE));
+            debug_printf("  .min_mip_filter = %s\n",
+                         debug_dump_tex_mipfilter(key->sampler[i].min_mip_filter, TRUE));
+            debug_printf("  .mag_img_filter = %s\n",
+                         debug_dump_tex_filter(key->sampler[i].mag_img_filter, TRUE));
+            if(key->sampler[i].compare_mode)
+               debug_printf("  .compare_mode = %s\n", debug_dump_func(key->sampler[i].compare_func, TRUE));
+            debug_printf("  .normalized_coords = %u\n", key->sampler[i].normalized_coords);
+            debug_printf("  .prefilter = %u\n", key->sampler[i].prefilter);
+         }
       }
    }
 
-#endif
-
    variant = CALLOC_STRUCT(lp_fragment_shader_variant);
    if(!variant)
       return NULL;
@@ -588,8 +599,8 @@ generate_fragment(struct llvmpipe_context *lp,
    }
 
    lp_build_conv_mask(builder, fs_type, blend_type,
-                               fs_mask, num_fs,
-                               &blend_mask, 1);
+                      fs_mask, num_fs,
+                      &blend_mask, 1);
 
    /*
     * Blending.
@@ -611,23 +622,24 @@ generate_fragment(struct llvmpipe_context *lp,
     * Translate the LLVM IR into machine code.
     */
 
+#ifdef DEBUG
    if(LLVMVerifyFunction(variant->function, LLVMPrintMessageAction)) {
       LLVMDumpValue(variant->function);
-      abort();
+      assert(0);
    }
+#endif
 
    LLVMRunFunctionPassManager(screen->pass, variant->function);
 
-#ifdef DEBUG
-   LLVMDumpValue(variant->function);
-   debug_printf("\n");
-#endif
+   if (LP_DEBUG & DEBUG_JIT) {
+      LLVMDumpValue(variant->function);
+      debug_printf("\n");
+   }
 
    variant->jit_function = (lp_jit_frag_func)LLVMGetPointerToGlobal(screen->engine, variant->function);
 
-#ifdef DEBUG
-   lp_disassemble(variant->jit_function);
-#endif
+   if (LP_DEBUG & DEBUG_ASM)
+      lp_disassemble(variant->jit_function);
 
    variant->next = shader->variants;
    shader->variants = variant;
@@ -661,7 +673,12 @@ llvmpipe_bind_fs_state(struct pipe_context *pipe, void *fs)
 {
    struct llvmpipe_context *llvmpipe = llvmpipe_context(pipe);
 
-   llvmpipe->fs = (struct lp_fragment_shader *) fs;
+   if (llvmpipe->fs == fs)
+      return;
+
+   draw_flush(llvmpipe->draw);
+
+   llvmpipe->fs = fs;
 
    llvmpipe->dirty |= LP_NEW_FS;
 }
@@ -711,8 +728,7 @@ llvmpipe_set_constant_buffer(struct pipe_context *pipe,
    assert(shader < PIPE_SHADER_TYPES);
    assert(index == 0);
 
-   if(shader == PIPE_SHADER_VERTEX)
-      draw_flush(llvmpipe->draw);
+   draw_flush(llvmpipe->draw);
 
    /* note: reference counting */
    pipe_buffer_reference(&llvmpipe->constants[shader].buffer, buffer);
@@ -722,7 +738,8 @@ llvmpipe_set_constant_buffer(struct pipe_context *pipe,
    }
 
    if(shader == PIPE_SHADER_VERTEX) {
-      draw_set_mapped_constant_buffer(llvmpipe->draw, data, size);
+      draw_set_mapped_constant_buffer(llvmpipe->draw, PIPE_SHADER_VERTEX,
+                                      data, size);
    }
 
    llvmpipe->dirty |= LP_NEW_CONSTANTS;
index 4561c6b84560a40b12bc4d595d42121d476c3603..aa3b5a3f91ef1ce7971eccd40f166813109ba27f 100644 (file)
@@ -41,14 +41,17 @@ llvmpipe_create_rasterizer_state(struct pipe_context *pipe,
 }
 
 void llvmpipe_bind_rasterizer_state(struct pipe_context *pipe,
-                                    void *setup)
+                                    void *rasterizer)
 {
    struct llvmpipe_context *llvmpipe = llvmpipe_context(pipe);
 
+   if (llvmpipe->rasterizer == rasterizer)
+      return;
+
    /* pass-through to draw module */
-   draw_set_rasterizer_state(llvmpipe->draw, setup);
+   draw_set_rasterizer_state(llvmpipe->draw, rasterizer);
 
-   llvmpipe->rasterizer = (struct pipe_rasterizer_state *)setup;
+   llvmpipe->rasterizer = rasterizer;
 
    llvmpipe->dirty |= LP_NEW_RASTERIZER;
 }
index c69d90c723ac74ff1cbdf2b1e1d65a70210750bd..d382f9ca87e1cea4563785e21c423cda04bad305 100644 (file)
@@ -77,6 +77,34 @@ llvmpipe_bind_sampler_states(struct pipe_context *pipe,
 }
 
 
+void
+llvmpipe_bind_vertex_sampler_states(struct pipe_context *pipe,
+                                    unsigned num_samplers,
+                                    void **samplers)
+{
+   struct llvmpipe_context *llvmpipe = llvmpipe_context(pipe);
+   unsigned i;
+
+   assert(num_samplers <= PIPE_MAX_VERTEX_SAMPLERS);
+
+   /* Check for no-op */
+   if (num_samplers == llvmpipe->num_vertex_samplers &&
+       !memcmp(llvmpipe->vertex_samplers, samplers, num_samplers * sizeof(void *)))
+      return;
+
+   draw_flush(llvmpipe->draw);
+
+   for (i = 0; i < num_samplers; ++i)
+      llvmpipe->vertex_samplers[i] = samplers[i];
+   for (i = num_samplers; i < PIPE_MAX_VERTEX_SAMPLERS; ++i)
+      llvmpipe->vertex_samplers[i] = NULL;
+
+   llvmpipe->num_vertex_samplers = num_samplers;
+
+   llvmpipe->dirty |= LP_NEW_SAMPLER;
+}
+
+
 void
 llvmpipe_set_sampler_textures(struct pipe_context *pipe,
                               unsigned num, struct pipe_texture **texture)
@@ -102,8 +130,8 @@ llvmpipe_set_sampler_textures(struct pipe_context *pipe,
       if(tex) {
          struct llvmpipe_texture *lp_tex = llvmpipe_texture(tex);
          struct lp_jit_texture *jit_tex = &llvmpipe->jit_context.textures[i];
-         jit_tex->width = tex->width[0];
-         jit_tex->height = tex->height[0];
+         jit_tex->width = tex->width0;
+         jit_tex->height = tex->height0;
          jit_tex->stride = lp_tex->stride[0];
          if(!lp_tex->dt)
             jit_tex->data = lp_tex->data;
@@ -116,6 +144,37 @@ llvmpipe_set_sampler_textures(struct pipe_context *pipe,
 }
 
 
+void
+llvmpipe_set_vertex_sampler_textures(struct pipe_context *pipe,
+                                     unsigned num_textures,
+                                     struct pipe_texture **textures)
+{
+   struct llvmpipe_context *llvmpipe = llvmpipe_context(pipe);
+   uint i;
+
+   assert(num_textures <= PIPE_MAX_VERTEX_SAMPLERS);
+
+   /* Check for no-op */
+   if (num_textures == llvmpipe->num_vertex_textures &&
+       !memcmp(llvmpipe->vertex_textures, textures, num_textures * sizeof(struct pipe_texture *))) {
+      return;
+   }
+
+   draw_flush(llvmpipe->draw);
+
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      struct pipe_texture *tex = i < num_textures ? textures[i] : NULL;
+
+      pipe_texture_reference(&llvmpipe->vertex_textures[i], tex);
+      lp_tex_tile_cache_set_texture(llvmpipe->vertex_tex_cache[i], tex);
+   }
+
+   llvmpipe->num_vertex_textures = num_textures;
+
+   llvmpipe->dirty |= LP_NEW_TEXTURE;
+}
+
+
 void
 llvmpipe_delete_sampler_state(struct pipe_context *pipe,
                               void *sampler)
index c06ce8b75c1b21a0f8e00fe0bf3e125d69a50b8c..e37ff04f3df172aab3e3c764a645f2124c5f715f 100644 (file)
@@ -35,6 +35,8 @@
 
 #include "draw/draw_context.h"
 
+#include "util/u_format.h"
+
 
 /**
  * XXX this might get moved someday
@@ -49,6 +51,8 @@ llvmpipe_set_framebuffer_state(struct pipe_context *pipe,
    struct llvmpipe_context *lp = llvmpipe_context(pipe);
    uint i;
 
+   draw_flush(lp->draw);
+
    for (i = 0; i < PIPE_MAX_COLOR_BUFS; i++) {
       /* check if changing cbuf */
       if (lp->framebuffer.cbufs[i] != fb->cbufs[i]) {
@@ -88,8 +92,9 @@ llvmpipe_set_framebuffer_state(struct pipe_context *pipe,
       if (lp->framebuffer.zsbuf) {
          int depth_bits;
          double mrd;
-         depth_bits = pf_get_component_bits(lp->framebuffer.zsbuf->format,
-                                            PIPE_FORMAT_COMP_Z);
+         depth_bits = util_format_get_component_bits(lp->framebuffer.zsbuf->format,
+                                                     UTIL_FORMAT_COLORSPACE_ZS,
+                                                     0);
          if (depth_bits > 16) {
             mrd = 0.0000001;
          }
index 15c30296144117462f3b9f9bad4d3be86f4af877..884e3878e621af5fea2eb76220b4f6c39549ab7e 100644 (file)
@@ -70,14 +70,18 @@ fail:
 
 
 void
-llvmpipe_bind_vs_state(struct pipe_context *pipe, void *vs)
+llvmpipe_bind_vs_state(struct pipe_context *pipe, void *_vs)
 {
    struct llvmpipe_context *llvmpipe = llvmpipe_context(pipe);
+   const struct lp_vertex_shader *vs = (const struct lp_vertex_shader *)_vs;
 
-   llvmpipe->vs = (const struct lp_vertex_shader *)vs;
+   if (llvmpipe->vs == vs)
+      return;
 
-   draw_bind_vertex_shader(llvmpipe->draw,
-                           (llvmpipe->vs ? llvmpipe->vs->draw_data : NULL));
+   draw_bind_vertex_shader(llvmpipe->draw, 
+                           vs ? vs->draw_data : NULL);
+
+   llvmpipe->vs = vs;
 
    llvmpipe->dirty |= LP_NEW_VS;
 }
@@ -92,5 +96,6 @@ llvmpipe_delete_vs_state(struct pipe_context *pipe, void *vs)
       (struct lp_vertex_shader *)vs;
 
    draw_delete_vertex_shader(llvmpipe->draw, state->draw_data);
+   FREE( (void *)state->shader.tokens );
    FREE( state );
 }
index 773e84824256642d254ccedc604a1e2c07a7cc36..a6d9a2c1acfc724056542bcca14df24cb63c5641 100644 (file)
@@ -36,6 +36,7 @@
 #include "util/u_memory.h"
 #include "util/u_tile.h"
 #include "util/u_format.h"
+#include "util/u_math.h"
 #include "lp_context.h"
 #include "lp_surface.h"
 #include "lp_texture.h"
@@ -154,7 +155,6 @@ lp_tex_tile_cache_validate_texture(struct llvmpipe_tex_tile_cache *tc)
       if (lpt->timestamp != tc->timestamp) {
          /* texture was modified, invalidate all cached tiles */
          uint i;
-         debug_printf("INV %d %d\n", tc->timestamp, lpt->timestamp);
          for (i = 0; i < NUM_ENTRIES; i++) {
             tc->entries[i].addr.bits.invalid = 1;
          }
@@ -270,8 +270,8 @@ lp_find_cached_tex_tile(struct llvmpipe_tex_tile_cache *tc,
                                      addr.bits.level, 
                                      addr.bits.z, 
                                      PIPE_TRANSFER_READ, 0, 0,
-                                     tc->texture->width[addr.bits.level],
-                                     tc->texture->height[addr.bits.level]);
+                                     u_minify(tc->texture->width0, addr.bits.level),
+                                     u_minify(tc->texture->height0, addr.bits.level));
 
          tc->tex_trans_map = screen->transfer_map(screen, tc->tex_trans);
 
@@ -290,7 +290,7 @@ lp_find_cached_tex_tile(struct llvmpipe_tex_tile_cache *tc,
             assert(0);
          }
 
-         util_format_read_4ub(tc->tex_trans->format,
+         util_format_read_4ub(tc->tex_trans->texture->format,
                               (uint8_t *)tile->color, sizeof tile->color[0],
                               tc->tex_trans_map, tc->tex_trans->stride,
                               x, y, w, h);
index a1365a045f178db64eb36a66c560a02e1d334d53..0d01c07fb5e3f38d6a87b2cea03d47c45ce9a3cc 100644 (file)
@@ -544,7 +544,7 @@ compute_lambda(struct tgsi_sampler *tgsi_sampler,
       float dsdy = s[QUAD_TOP_LEFT]     - s[QUAD_BOTTOM_LEFT];
       dsdx = fabsf(dsdx);
       dsdy = fabsf(dsdy);
-      rho = MAX2(dsdx, dsdy) * texture->width[0];
+      rho = MAX2(dsdx, dsdy) * texture->width0;
    }
    if (t) {
       float dtdx = t[QUAD_BOTTOM_RIGHT] - t[QUAD_BOTTOM_LEFT];
@@ -552,7 +552,7 @@ compute_lambda(struct tgsi_sampler *tgsi_sampler,
       float max;
       dtdx = fabsf(dtdx);
       dtdy = fabsf(dtdy);
-      max = MAX2(dtdx, dtdy) * texture->height[0];
+      max = MAX2(dtdx, dtdy) * texture->height0;
       rho = MAX2(rho, max);
    }
    if (p) {
@@ -561,7 +561,7 @@ compute_lambda(struct tgsi_sampler *tgsi_sampler,
       float max;
       dpdx = fabsf(dpdx);
       dpdy = fabsf(dpdy);
-      max = MAX2(dpdx, dpdy) * texture->depth[0];
+      max = MAX2(dpdx, dpdy) * texture->depth0;
       rho = MAX2(rho, max);
    }
 
@@ -726,9 +726,9 @@ get_texel(const struct tgsi_sampler *tgsi_sampler,
    const struct pipe_texture *texture = samp->texture;
    const struct pipe_sampler_state *sampler = samp->sampler;
 
-   if (x < 0 || x >= (int) texture->width[level] ||
-       y < 0 || y >= (int) texture->height[level] ||
-       z < 0 || z >= (int) texture->depth[level]) {
+   if (x < 0 || x >= (int) u_minify(texture->width0, level) ||
+       y < 0 || y >= (int) u_minify(texture->height0, level) ||
+       z < 0 || z >= (int) u_minify(texture->depth0, level)) {
       rgba[0][j] = sampler->border_color[0];
       rgba[1][j] = sampler->border_color[1];
       rgba[2][j] = sampler->border_color[2];
@@ -1093,8 +1093,8 @@ lp_get_samples_2d_common(struct tgsi_sampler *tgsi_sampler,
 
    assert(sampler->normalized_coords);
 
-   width = texture->width[level0];
-   height = texture->height[level0];
+   width = u_minify(texture->width0, level0);
+   height = u_minify(texture->height0, level0);
 
    assert(width > 0);
 
@@ -1250,9 +1250,9 @@ lp_get_samples_3d(struct tgsi_sampler *tgsi_sampler,
 
    assert(sampler->normalized_coords);
 
-   width = texture->width[level0];
-   height = texture->height[level0];
-   depth = texture->depth[level0];
+   width = u_minify(texture->width0, level0);
+   height = u_minify(texture->height0, level0);
+   depth = u_minify(texture->depth0, level0);
 
    assert(width > 0);
    assert(height > 0);
@@ -1394,8 +1394,8 @@ lp_get_samples_rect(struct tgsi_sampler *tgsi_sampler,
    /* texture RECTS cannot be mipmapped */
    assert(level0 == level1);
 
-   width = texture->width[level0];
-   height = texture->height[level0];
+   width = u_minify(texture->width0, level0);
+   height = u_minify(texture->height0, level0);
 
    assert(width > 0);
 
@@ -1513,8 +1513,8 @@ lp_get_samples(struct tgsi_sampler *tgsi_sampler,
 
    /* Do this elsewhere: 
     */
-   samp->xpot = util_unsigned_logbase2( samp->texture->width[0] );
-   samp->ypot = util_unsigned_logbase2( samp->texture->height[0] );
+   samp->xpot = util_unsigned_logbase2( samp->texture->width0 );
+   samp->ypot = util_unsigned_logbase2( samp->texture->height0 );
 
    /* Try to hook in a faster sampler.  Ultimately we'll have to
     * code-generate these.  Luckily most of this looks like it is
index a00f2495dfc2dab39269f9f22d97c7a824a7b5e3..2c135029ea2d1b5bd817c6f6d8ca5485bfe19fe3 100644 (file)
@@ -34,6 +34,8 @@
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
 #include "pipe/internal/p_winsys_screen.h"
+
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -48,7 +50,6 @@
 /* Simple, maximally packed layout.
  */
 
-
 /* Conventional allocation path for non-display textures:
  */
 static boolean
@@ -57,29 +58,21 @@ llvmpipe_texture_layout(struct llvmpipe_screen *screen,
 {
    struct pipe_texture *pt = &lpt->base;
    unsigned level;
-   unsigned width = pt->width[0];
-   unsigned height = pt->height[0];
-   unsigned depth = pt->depth[0];
+   unsigned width = pt->width0;
+   unsigned height = pt->height0;
+   unsigned depth = pt->depth0;
 
    unsigned buffer_size = 0;
 
-   pf_get_block(lpt->base.format, &lpt->base.block);
-
    for (level = 0; level <= pt->last_level; level++) {
       unsigned nblocksx, nblocksy;
 
-      pt->width[level] = width;
-      pt->height[level] = height;
-      pt->depth[level] = depth;
-      pt->nblocksx[level] = pf_get_nblocksx(&pt->block, width);  
-      pt->nblocksy[level] = pf_get_nblocksy(&pt->block, height);
-
       /* Allocate storage for whole quads. This is particularly important
        * for depth surfaces, which are currently stored in a swizzled format. */
-      nblocksx = pf_get_nblocksx(&pt->block, align(width, 2));
-      nblocksy = pf_get_nblocksy(&pt->block, align(height, 2));
+      nblocksx = util_format_get_nblocksx(pt->format, align(width, 2));
+      nblocksy = util_format_get_nblocksy(pt->format, align(height, 2));
 
-      lpt->stride[level] = align(nblocksx*pt->block.size, 16);
+      lpt->stride[level] = align(nblocksx * util_format_get_blocksize(pt->format), 16);
 
       lpt->level_offset[level] = buffer_size;
 
@@ -87,9 +80,9 @@ llvmpipe_texture_layout(struct llvmpipe_screen *screen,
                       ((pt->target == PIPE_TEXTURE_CUBE) ? 6 : depth) *
                       lpt->stride[level]);
 
-      width  = minify(width);
-      height = minify(height);
-      depth = minify(depth);
+      width  = u_minify(width, 1);
+      height = u_minify(height, 1);
+      depth = u_minify(depth, 1);
    }
 
    lpt->data = align_malloc(buffer_size, 16);
@@ -103,14 +96,10 @@ llvmpipe_displaytarget_layout(struct llvmpipe_screen *screen,
 {
    struct llvmpipe_winsys *winsys = screen->winsys;
 
-   pf_get_block(lpt->base.format, &lpt->base.block);
-   lpt->base.nblocksx[0] = pf_get_nblocksx(&lpt->base.block, lpt->base.width[0]);  
-   lpt->base.nblocksy[0] = pf_get_nblocksy(&lpt->base.block, lpt->base.height[0]);  
-
    lpt->dt = winsys->displaytarget_create(winsys,
                                           lpt->base.format,
-                                          lpt->base.width[0],
-                                          lpt->base.height[0],
+                                          lpt->base.width0,
+                                          lpt->base.height0,
                                           16,
                                           &lpt->stride[0] );
 
@@ -172,7 +161,7 @@ llvmpipe_texture_blanket(struct pipe_screen * screen,
    /* Only supports one type */
    if (base->target != PIPE_TEXTURE_2D ||
        base->last_level != 0 ||
-       base->depth[0] != 1) {
+       base->depth0 != 1) {
       return NULL;
    }
 
@@ -183,8 +172,6 @@ llvmpipe_texture_blanket(struct pipe_screen * screen,
    lpt->base = *base;
    pipe_reference_init(&lpt->base.reference, 1);
    lpt->base.screen = screen;
-   lpt->base.nblocksx[0] = pf_get_nblocksx(&lpt->base.block, lpt->base.width[0]);  
-   lpt->base.nblocksy[0] = pf_get_nblocksy(&lpt->base.block, lpt->base.height[0]);  
    lpt->stride[0] = stride[0];
 
    pipe_buffer_reference(&lpt->buffer, buffer);
@@ -229,8 +216,8 @@ llvmpipe_get_tex_surface(struct pipe_screen *screen,
       pipe_reference_init(&ps->reference, 1);
       pipe_texture_reference(&ps->texture, pt);
       ps->format = pt->format;
-      ps->width = pt->width[level];
-      ps->height = pt->height[level];
+      ps->width = u_minify(pt->width0, level);
+      ps->height = u_minify(pt->height0, level);
       ps->offset = lpt->level_offset[level];
       ps->usage = usage;
 
@@ -258,11 +245,17 @@ llvmpipe_get_tex_surface(struct pipe_screen *screen,
       ps->level = level;
       ps->zslice = zslice;
 
+      /* XXX shouldn't that rather be
+         tex_height = align(ps->height, 2);
+         to account for alignment done in llvmpipe_texture_layout ?
+      */
       if (pt->target == PIPE_TEXTURE_CUBE) {
-         ps->offset += face * pt->nblocksy[level] * lpt->stride[level];
+         unsigned tex_height = ps->height;
+         ps->offset += face * util_format_get_nblocksy(pt->format, tex_height) * lpt->stride[level];
       }
       else if (pt->target == PIPE_TEXTURE_3D) {
-         ps->offset += zslice * pt->nblocksy[level] * lpt->stride[level];
+         unsigned tex_height = ps->height;
+         ps->offset += zslice * util_format_get_nblocksy(pt->format, tex_height) * lpt->stride[level];
       }
       else {
          assert(face == 0);
@@ -303,14 +296,10 @@ llvmpipe_get_tex_transfer(struct pipe_screen *screen,
    if (lpt) {
       struct pipe_transfer *pt = &lpt->base;
       pipe_texture_reference(&pt->texture, texture);
-      pt->format = texture->format;
-      pt->block = texture->block;
       pt->x = x;
       pt->y = y;
       pt->width = w;
       pt->height = h;
-      pt->nblocksx = texture->nblocksx[level];
-      pt->nblocksy = texture->nblocksy[level];
       pt->stride = lptex->stride[level];
       pt->usage = usage;
       pt->face = face;
@@ -319,11 +308,17 @@ llvmpipe_get_tex_transfer(struct pipe_screen *screen,
 
       lpt->offset = lptex->level_offset[level];
 
+      /* XXX shouldn't that rather be
+         tex_height = align(u_minify(texture->height0, level), 2)
+         to account for alignment done in llvmpipe_texture_layout ?
+      */
       if (texture->target == PIPE_TEXTURE_CUBE) {
-         lpt->offset += face * pt->nblocksy * pt->stride;
+         unsigned tex_height = u_minify(texture->height0, level);
+         lpt->offset += face *  util_format_get_nblocksy(texture->format, tex_height) * pt->stride;
       }
       else if (texture->target == PIPE_TEXTURE_3D) {
-         lpt->offset += zslice * pt->nblocksy * pt->stride;
+         unsigned tex_height = u_minify(texture->height0, level);
+         lpt->offset += zslice * util_format_get_nblocksy(texture->format, tex_height) * pt->stride;
       }
       else {
          assert(face == 0);
@@ -355,9 +350,11 @@ llvmpipe_transfer_map( struct pipe_screen *_screen,
    struct llvmpipe_screen *screen = llvmpipe_screen(_screen);
    ubyte *map, *xfer_map;
    struct llvmpipe_texture *lpt;
+   enum pipe_format format;
 
    assert(transfer->texture);
    lpt = llvmpipe_texture(transfer->texture);
+   format = lpt->base.format;
 
    if(lpt->dt) {
       struct llvmpipe_winsys *winsys = screen->winsys;
@@ -382,8 +379,8 @@ llvmpipe_transfer_map( struct pipe_screen *_screen,
    }
    
    xfer_map = map + llvmpipe_transfer(transfer)->offset +
-      transfer->y / transfer->block.height * transfer->stride +
-      transfer->x / transfer->block.width * transfer->block.size;
+      transfer->y / util_format_get_blockheight(format) * transfer->stride +
+      transfer->x / util_format_get_blockwidth(format) * util_format_get_blocksize(format);
    /*printf("map = %p  xfer map = %p\n", map, xfer_map);*/
    return xfer_map;
 }
index ec3e002d6282697f2bba27c45581773cbc1516db..7a1ecf5107b218e8844876ef3a9c85fa961c55bc 100644 (file)
@@ -252,13 +252,13 @@ lp_flush_tile_cache(struct llvmpipe_tile_cache *tc)
                case LP_TILE_STATUS_CLEAR:
                   /* Actually clear the tiles which were flagged as being in a
                    * clear state. */
-                  util_fill_rect(tc->transfer_map, &pt->block, pt->stride,
+                  util_fill_rect(tc->transfer_map, pt->texture->format, pt->stride,
                                  x, y, w, h,
                                  tc->clear_val);
                   break;
 
                case LP_TILE_STATUS_DEFINED:
-                  lp_tile_write_4ub(pt->format,
+                  lp_tile_write_4ub(pt->texture->format,
                                     tile->color,
                                     tc->transfer_map, pt->stride,
                                     x, y, w, h);
@@ -291,6 +291,11 @@ lp_get_cached_tile(struct llvmpipe_tile_cache *tc,
    assert(tc->surface);
    assert(tc->transfer);
 
+   if(!tc->transfer_map)
+      lp_tile_cache_map_transfers(tc);
+
+   assert(tc->transfer_map);
+
    switch(tile->status) {
    case LP_TILE_STATUS_CLEAR:
       /* don't get tile from framebuffer, just clear it */
@@ -306,7 +311,7 @@ lp_get_cached_tile(struct llvmpipe_tile_cache *tc,
       y &= ~(TILE_SIZE - 1);
 
       if (!pipe_clip_tile(x, y, &w, &h, tc->transfer))
-         lp_tile_read_4ub(pt->format,
+         lp_tile_read_4ub(pt->texture->format,
                           tile->color,
                           tc->transfer_map, tc->transfer->stride,
                           x, y, w, h);
index b595405357f0044255261993c4887f1acdcfb8c5..9aee9e495664e3c8b3caf97669106e2b729ff256 100644 (file)
@@ -48,13 +48,14 @@ so_ref(struct nouveau_stateobj *ref, struct nouveau_stateobj **pso)
        struct nouveau_stateobj *so = *pso;
        int i;
 
-        if (pipe_reference((struct pipe_reference**)pso, &ref->reference)) {
+        if (pipe_reference(&(*pso)->reference, &ref->reference)) {
                free(so->push);
                for (i = 0; i < so->cur_reloc; i++)
                        nouveau_bo_ref(NULL, &so->reloc[i].bo);
                free(so->reloc);
                free(so);
        }
+       *pso = ref;
 }
 
 static INLINE void
@@ -111,20 +112,30 @@ so_emit(struct nouveau_channel *chan, struct nouveau_stateobj *so)
 {
        struct nouveau_pushbuf *pb = chan->pushbuf;
        unsigned nr, i;
+       int ret = 0;
 
        nr = so->cur - so->push;
-       if (pb->remaining < nr)
-               nouveau_pushbuf_flush(chan, nr);
+       /* This will flush if we need space.
+        * We don't actually need the marker.
+        */
+       if ((ret = nouveau_pushbuf_marker_emit(chan, nr, so->cur_reloc))) {
+               debug_printf("so_emit failed marker emit with error %d\n", ret);
+               return;
+       }
        pb->remaining -= nr;
 
        memcpy(pb->cur, so->push, nr * 4);
        for (i = 0; i < so->cur_reloc; i++) {
                struct nouveau_stateobj_reloc *r = &so->reloc[i];
 
-               nouveau_pushbuf_emit_reloc(chan, pb->cur + r->offset,
+               if ((ret = nouveau_pushbuf_emit_reloc(chan, pb->cur + r->offset,
                                           r->bo, r->data, 0, r->flags,
-                                          r->vor, r->tor);
+                                          r->vor, r->tor))) {
+                       debug_printf("so_emit failed reloc with error %d\n", ret);
+                       goto out;
+               }
        }
+out:
        pb->cur += nr;
 }
 
@@ -133,26 +144,45 @@ so_emit_reloc_markers(struct nouveau_channel *chan, struct nouveau_stateobj *so)
 {
        struct nouveau_pushbuf *pb = chan->pushbuf;
        unsigned i;
+       int ret = 0;
 
        if (!so)
                return;
 
        i = so->cur_reloc << 1;
-       if (pb->remaining < i)
-               nouveau_pushbuf_flush(chan, i);
+       /* This will flush if we need space.
+        * We don't actually need the marker.
+        */
+       if ((ret = nouveau_pushbuf_marker_emit(chan, i, i))) {
+               debug_printf("so_emit_reloc_markers failed marker emit with" \
+                       "error %d\n", ret);
+               return;
+       }
        pb->remaining -= i;
 
        for (i = 0; i < so->cur_reloc; i++) {
                struct nouveau_stateobj_reloc *r = &so->reloc[i];
 
-               nouveau_pushbuf_emit_reloc(chan, pb->cur++, r->bo, r->packet, 0,
+               if ((ret = nouveau_pushbuf_emit_reloc(chan, pb->cur++, r->bo,
+                                          r->packet, 0,
                                           (r->flags & (NOUVEAU_BO_VRAM |
                                                        NOUVEAU_BO_GART |
                                                        NOUVEAU_BO_RDWR)) |
-                                          NOUVEAU_BO_DUMMY, 0, 0);
-               nouveau_pushbuf_emit_reloc(chan, pb->cur++, r->bo, r->data, 0,
+                                          NOUVEAU_BO_DUMMY, 0, 0))) {
+                       debug_printf("so_emit_reloc_markers failed reloc" \
+                                               "with error %d\n", ret);
+                       pb->remaining += ((so->cur_reloc - i) << 1);
+                       return;
+               }
+               if ((ret = nouveau_pushbuf_emit_reloc(chan, pb->cur++, r->bo,
+                                          r->data, 0,
                                           r->flags | NOUVEAU_BO_DUMMY,
-                                          r->vor, r->tor);
+                                          r->vor, r->tor))) {
+                       debug_printf("so_emit_reloc_markers failed reloc" \
+                                               "with error %d\n", ret);
+                       pb->remaining += ((so->cur_reloc - i) << 1) - 1;
+                       return;
+               }
        }
 }
 
index 10d984ace9b6a556d3be85a1cd19932e3a6b7123..4b33636b2eb5b258e474bdbd20191d7740b63153 100644 (file)
@@ -27,11 +27,6 @@ nv04_destroy(struct pipe_context *pipe)
        FREE(nv04);
 }
 
-static void
-nv04_set_edgeflags(struct pipe_context *pipe, const unsigned *bitfield)
-{
-}
-
 static boolean
 nv04_init_hwctx(struct nv04_context *nv04)
 {
@@ -83,7 +78,6 @@ nv04_create(struct pipe_screen *pscreen, unsigned pctx_id)
        nv04->pipe.winsys = ws;
        nv04->pipe.screen = pscreen;
        nv04->pipe.destroy = nv04_destroy;
-       nv04->pipe.set_edgeflags = nv04_set_edgeflags;
        nv04->pipe.draw_arrays = nv04_draw_arrays;
        nv04->pipe.draw_elements = nv04_draw_elements;
        nv04->pipe.clear = nv04_clear;
index 21f990fd536deb3a4a4fe32319c9d775c2c1f649..0cce71ad1de4328d6e99a5d47dc979fb8fa64618 100644 (file)
@@ -57,8 +57,8 @@ nv04_fragtex_build(struct nv04_context *nv04, int unit)
                | NV04_DX5_TEXTURED_TRIANGLE_FORMAT_ORIGIN_FOH_CORNER
                | nv04_fragtex_format(pt->format)
                | ( (pt->last_level + 1) << NV04_DX5_TEXTURED_TRIANGLE_FORMAT_MIPMAP_LEVELS_SHIFT )
-               | ( log2i(pt->width[0]) << NV04_DX5_TEXTURED_TRIANGLE_FORMAT_BASE_SIZE_U_SHIFT )
-               | ( log2i(pt->height[0]) << NV04_DX5_TEXTURED_TRIANGLE_FORMAT_BASE_SIZE_V_SHIFT )
+               | ( log2i(pt->width0) << NV04_DX5_TEXTURED_TRIANGLE_FORMAT_BASE_SIZE_U_SHIFT )
+               | ( log2i(pt->height0) << NV04_DX5_TEXTURED_TRIANGLE_FORMAT_BASE_SIZE_V_SHIFT )
                | NV04_DX5_TEXTURED_TRIANGLE_FORMAT_ADDRESSU_CLAMP_TO_EDGE
                | NV04_DX5_TEXTURED_TRIANGLE_FORMAT_ADDRESSV_CLAMP_TO_EDGE
                ;
index 93f752faec90c135e7b4e38f313b9a2bee4b3197..e0a6948aeb45eef0017ca1e0d87c1716048646c8 100644 (file)
@@ -1,6 +1,7 @@
 #include "pipe/p_state.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
+#include "util/u_math.h"
 
 #include "nv04_context.h"
 #include "nv04_screen.h"
@@ -9,31 +10,22 @@ static void
 nv04_miptree_layout(struct nv04_miptree *nv04mt)
 {
        struct pipe_texture *pt = &nv04mt->base;
-       uint width = pt->width[0], height = pt->height[0];
        uint offset = 0;
        int nr_faces, l;
 
        nr_faces = 1;
 
        for (l = 0; l <= pt->last_level; l++) {
-               pt->width[l] = width;
-               pt->height[l] = height;
-
-               pt->nblocksx[l] = pf_get_nblocksx(&pt->block, width);
-               pt->nblocksy[l] = pf_get_nblocksy(&pt->block, height);
-               
-               nv04mt->level[l].pitch = pt->width[0];
+               nv04mt->level[l].pitch = pt->width0;
                nv04mt->level[l].pitch = (nv04mt->level[l].pitch + 63) & ~63;
-
-               width  = MAX2(1, width  >> 1);
-               height = MAX2(1, height >> 1);
        }
 
        for (l = 0; l <= pt->last_level; l++) {
-
                nv04mt->level[l].image_offset = 
                        CALLOC(nr_faces, sizeof(unsigned));
-               offset += nv04mt->level[l].pitch * pt->height[l];
+               /* XXX guess was obviously missing */
+               nv04mt->level[l].image_offset[0] = offset;
+               offset += nv04mt->level[l].pitch * u_minify(pt->height0, l);
        }
 
        nv04mt->total_size = offset;
@@ -63,7 +55,7 @@ nv04_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *pt)
                FREE(mt);
                return NULL;
        }
-       
+       mt->bo = nouveau_bo(mt->buffer);
        return &mt->base;
 }
 
@@ -75,7 +67,7 @@ nv04_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
 
        /* Only supports 2D, non-mipmapped textures for the moment */
        if (pt->target != PIPE_TEXTURE_2D || pt->last_level != 0 ||
-           pt->depth[0] != 1)
+           pt->depth0 != 1)
                return NULL;
 
        mt = CALLOC_STRUCT(nv04_miptree);
@@ -89,6 +81,7 @@ nv04_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
        mt->level[0].image_offset = CALLOC(1, sizeof(unsigned));
 
        pipe_buffer_reference(&mt->buffer, pb);
+       mt->bo = nouveau_bo(mt->buffer);
        return &mt->base;
 }
 
@@ -120,8 +113,8 @@ nv04_miptree_surface_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
                return NULL;
        pipe_texture_reference(&ns->base.texture, pt);
        ns->base.format = pt->format;
-       ns->base.width = pt->width[level];
-       ns->base.height = pt->height[level];
+       ns->base.width = u_minify(pt->width0, level);
+       ns->base.height = u_minify(pt->height0, level);
        ns->base.usage = flags;
        pipe_reference_init(&ns->base.reference, 1);
        ns->base.face = face;
@@ -129,7 +122,7 @@ nv04_miptree_surface_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
        ns->base.zslice = zslice;
        ns->pitch = nv04mt->level[level].pitch;
 
-       ns->base.offset = nv04mt->level[level].image_offset;
+       ns->base.offset = nv04mt->level[level].image_offset[0];
 
        return &ns->base;
 }
index d356ebd8b3688217740b49d490b644342f3253e3..ef3005db5fc11e261f83dd25ee356dda8cabcd1f 100644 (file)
@@ -425,9 +425,9 @@ nv04_init_state_functions(struct nv04_context *nv04)
        nv04->pipe.delete_blend_state = nv04_blend_state_delete;
 
        nv04->pipe.create_sampler_state = nv04_sampler_state_create;
-       nv04->pipe.bind_sampler_states = nv04_sampler_state_bind;
+       nv04->pipe.bind_fragment_sampler_states = nv04_sampler_state_bind;
        nv04->pipe.delete_sampler_state = nv04_sampler_state_delete;
-       nv04->pipe.set_sampler_textures = nv04_set_sampler_texture;
+       nv04->pipe.set_fragment_sampler_textures = nv04_set_sampler_texture;
 
        nv04->pipe.create_rasterizer_state = nv04_rasterizer_state_create;
        nv04->pipe.bind_rasterizer_state = nv04_rasterizer_state_bind;
index 399f750dbe7caf4de1d33036b2704aa297fc81d7..81d1d2ebaa94a034c9dba060527403b765a29e22 100644 (file)
@@ -31,6 +31,7 @@ struct nv04_rasterizer_state {
 
 struct nv04_miptree {
        struct pipe_texture base;
+       struct nouveau_bo *bo;
 
        struct pipe_buffer *buffer;
        uint total_size;
index 8be134b83dd8898dc8ef00f30c54c9f382c19a55..12df7fd1997b4c0eb07389e7f8cf37dc89cf2bef 100644 (file)
@@ -1,5 +1,6 @@
 #include "pipe/p_context.h"
 #include "pipe/p_format.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -133,6 +134,9 @@ nv04_surface_copy_swizzle(struct nv04_surface_2d *ctx,
        assert(sub_w == w || util_is_pot(sub_w));
        assert(sub_h == h || util_is_pot(sub_h));
 
+       MARK_RING (chan, 8 + ((w+sub_w)/sub_w)*((h+sub_h)/sub_h)*17, 2 +
+                        ((w+sub_w)/sub_w)*((h+sub_h)/sub_h)*2);
+
        BEGIN_RING(chan, swzsurf, NV04_SWIZZLED_SURFACE_DMA_IMAGE, 1);
        OUT_RELOCo(chan, dst_bo,
                         NOUVEAU_BO_GART | NOUVEAU_BO_VRAM | NOUVEAU_BO_WR);
@@ -155,10 +159,10 @@ nv04_surface_copy_swizzle(struct nv04_surface_2d *ctx,
            sub_w = MIN2(sub_w, w - x);
 
            /* Must be 64-byte aligned */
-           assert(!((dst->offset + nv04_swizzle_bits(dx+x, dy+y) * dst->texture->block.size) & 63));
+           assert(!((dst->offset + nv04_swizzle_bits(dx+x, dy+y) * util_format_get_blocksize(dst->texture->format)) & 63));
 
            BEGIN_RING(chan, swzsurf, NV04_SWIZZLED_SURFACE_OFFSET, 1);
-           OUT_RELOCl(chan, dst_bo, dst->offset + nv04_swizzle_bits(dx+x, dy+y) * dst->texture->block.size,
+           OUT_RELOCl(chan, dst_bo, dst->offset + nv04_swizzle_bits(dx+x, dy+y) * util_format_get_blocksize(dst->texture->format),
                              NOUVEAU_BO_GART | NOUVEAU_BO_VRAM | NOUVEAU_BO_WR);
 
            BEGIN_RING(chan, sifm, NV04_SCALED_IMAGE_FROM_MEMORY_COLOR_CONVERSION, 9);
@@ -177,7 +181,7 @@ nv04_surface_copy_swizzle(struct nv04_surface_2d *ctx,
            OUT_RING  (chan, src_pitch |
                             NV04_SCALED_IMAGE_FROM_MEMORY_FORMAT_ORIGIN_CENTER |
                             NV04_SCALED_IMAGE_FROM_MEMORY_FORMAT_FILTER_POINT_SAMPLE);
-           OUT_RELOCl(chan, src_bo, src->offset + (sy+y) * src_pitch + (sx+x) * src->texture->block.size,
+           OUT_RELOCl(chan, src_bo, src->offset + (sy+y) * src_pitch + (sx+x) * util_format_get_blocksize(src->texture->format),
                              NOUVEAU_BO_GART | NOUVEAU_BO_VRAM | NOUVEAU_BO_RD);
            OUT_RING  (chan, 0);
          }
@@ -198,11 +202,11 @@ nv04_surface_copy_m2mf(struct nv04_surface_2d *ctx,
        unsigned src_pitch = ((struct nv04_surface *)src)->pitch;
        unsigned dst_pitch = ((struct nv04_surface *)dst)->pitch;
        unsigned dst_offset = dst->offset + dy * dst_pitch +
-                             dx * dst->texture->block.size;
+                             dx * util_format_get_blocksize(dst->texture->format);
        unsigned src_offset = src->offset + sy * src_pitch +
-                             sx * src->texture->block.size;
+                             sx * util_format_get_blocksize(src->texture->format);
 
-       WAIT_RING (chan, 3 + ((h / 2047) + 1) * 9);
+       MARK_RING (chan, 3 + ((h / 2047) + 1) * 9, 2 + ((h / 2047) + 1) * 2);
        BEGIN_RING(chan, m2mf, NV04_MEMORY_TO_MEMORY_FORMAT_DMA_BUFFER_IN, 2);
        OUT_RELOCo(chan, src_bo,
                   NOUVEAU_BO_GART | NOUVEAU_BO_VRAM | NOUVEAU_BO_RD);
@@ -219,7 +223,7 @@ nv04_surface_copy_m2mf(struct nv04_surface_2d *ctx,
                           NOUVEAU_BO_VRAM | NOUVEAU_BO_GART | NOUVEAU_BO_WR);
                OUT_RING  (chan, src_pitch);
                OUT_RING  (chan, dst_pitch);
-               OUT_RING  (chan, w * src->texture->block.size);
+               OUT_RING  (chan, w * util_format_get_blocksize(src->texture->format));
                OUT_RING  (chan, count);
                OUT_RING  (chan, 0x0101);
                OUT_RING  (chan, 0);
@@ -250,7 +254,7 @@ nv04_surface_copy_blit(struct nv04_surface_2d *ctx, struct pipe_surface *dst,
        if (format < 0)
                return 1;
 
-       WAIT_RING (chan, 12);
+       MARK_RING (chan, 12, 4);
        BEGIN_RING(chan, surf2d, NV04_CONTEXT_SURFACES_2D_DMA_IMAGE_SOURCE, 2);
        OUT_RELOCo(chan, src_bo, NOUVEAU_BO_VRAM | NOUVEAU_BO_RD);
        OUT_RELOCo(chan, dst_bo, NOUVEAU_BO_VRAM | NOUVEAU_BO_WR);
@@ -315,7 +319,7 @@ nv04_surface_fill(struct nv04_surface_2d *ctx, struct pipe_surface *dst,
        gdirect_format = nv04_rect_format(dst->format);
        assert(gdirect_format >= 0);
 
-       WAIT_RING (chan, 16);
+       MARK_RING (chan, 16, 4);
        BEGIN_RING(chan, surf2d, NV04_CONTEXT_SURFACES_2D_DMA_IMAGE_SOURCE, 2);
        OUT_RELOCo(chan, dst_bo, NOUVEAU_BO_VRAM | NOUVEAU_BO_WR);
        OUT_RELOCo(chan, dst_bo, NOUVEAU_BO_VRAM | NOUVEAU_BO_WR);
index 6618660743d48388048912c1bb45d4014f37ced6..8446073ae8023dd9bc9417e6ae9a07a6223f1c16 100644 (file)
@@ -1,7 +1,9 @@
 #include <pipe/p_state.h>
 #include <pipe/p_defines.h>
 #include <pipe/p_inlines.h>
+#include <util/u_format.h>
 #include <util/u_memory.h>
+#include <util/u_math.h>
 #include <nouveau/nouveau_winsys.h>
 #include "nv04_context.h"
 #include "nv04_screen.h"
@@ -10,7 +12,7 @@
 struct nv04_transfer {
        struct pipe_transfer base;
        struct pipe_surface *surface;
-       bool direct;
+       boolean direct;
 };
 
 static void
@@ -20,12 +22,9 @@ nv04_compatible_transfer_tex(struct pipe_texture *pt, unsigned level,
        memset(template, 0, sizeof(struct pipe_texture));
        template->target = pt->target;
        template->format = pt->format;
-       template->width[0] = pt->width[level];
-       template->height[0] = pt->height[level];
-       template->depth[0] = 1;
-       template->block = pt->block;
-       template->nblocksx[0] = pt->nblocksx[level];
-       template->nblocksy[0] = pt->nblocksx[level];
+       template->width0 = u_minify(pt->width0, level);
+       template->height0 = u_minify(pt->height0, level);
+       template->depth0 = 1;
        template->last_level = 0;
        template->nr_samples = pt->nr_samples;
 
@@ -48,14 +47,10 @@ nv04_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
                return NULL;
 
        pipe_texture_reference(&tx->base.texture, pt);
-       tx->base.format = pt->format;
        tx->base.x = x;
        tx->base.y = y;
        tx->base.width = w;
        tx->base.height = h;
-       tx->base.block = pt->block;
-       tx->base.nblocksx = pt->nblocksx[level];
-       tx->base.nblocksy = pt->nblocksy[level];
        tx->base.stride = mt->level[level].pitch;
        tx->base.usage = usage;
        tx->base.face = face;
@@ -157,7 +152,7 @@ nv04_transfer_map(struct pipe_screen *pscreen, struct pipe_transfer *ptx)
                                    pipe_transfer_buffer_flags(ptx));
 
        return map + ns->base.offset +
-              ptx->y * ns->pitch + ptx->x * ptx->block.size;
+              ptx->y * ns->pitch + ptx->x * util_format_get_blocksize(ptx->texture->format);
 }
 
 static void
index e3167814f2bcb39d5ec92280c3391f8b1fea9c40..099ab100433e175767bc1433d39d0402d17bb3fb 100644 (file)
@@ -45,7 +45,7 @@ boolean nv04_draw_elements( struct pipe_context *pipe,
                draw_set_mapped_element_buffer(draw, 0, NULL);
        }
 
-       draw_set_mapped_constant_buffer(draw,
+       draw_set_mapped_constant_buffer(draw, PIPE_SHADER_VERTEX,
                                        nv04->constbuf[PIPE_SHADER_VERTEX],
                                        nv04->constbuf_nr[PIPE_SHADER_VERTEX]);
 
index 65a22b175e11ee373970c984aacf131cb91c6ede..0dadeb03dd3a752993c615fd460c51c3dfc4c9ea 100644 (file)
@@ -252,11 +252,6 @@ static void nv10_init_hwctx(struct nv10_context *nv10)
        FIRE_RING (NULL);
 }
 
-static void
-nv10_set_edgeflags(struct pipe_context *pipe, const unsigned *bitfield)
-{
-}
-
 struct pipe_context *
 nv10_create(struct pipe_screen *pscreen, unsigned pctx_id)
 {
@@ -276,7 +271,6 @@ nv10_create(struct pipe_screen *pscreen, unsigned pctx_id)
        nv10->pipe.winsys = ws;
        nv10->pipe.screen = pscreen;
        nv10->pipe.destroy = nv10_destroy;
-       nv10->pipe.set_edgeflags = nv10_set_edgeflags;
        nv10->pipe.draw_arrays = nv10_draw_arrays;
        nv10->pipe.draw_elements = nv10_draw_elements;
        nv10->pipe.clear = nv10_clear;
index 27f2f8758471064b7b689be1f0241de6e3279437..906fdfeeb93d39d0eacecda354e584aef8a0c445 100644 (file)
@@ -62,9 +62,9 @@ nv10_fragtex_build(struct nv10_context *nv10, int unit)
 
        txf  = tf->format << 8;
        txf |= (pt->last_level + 1) << 16;
-       txf |= log2i(pt->width[0]) << 20;
-       txf |= log2i(pt->height[0]) << 24;
-       txf |= log2i(pt->depth[0]) << 28;
+       txf |= log2i(pt->width0) << 20;
+       txf |= log2i(pt->height0) << 24;
+       txf |= log2i(pt->depth0) << 28;
        txf |= 8;
 
        switch (pt->target) {
@@ -89,7 +89,7 @@ nv10_fragtex_build(struct nv10_context *nv10, int unit)
        OUT_RING  (0x40000000); /* enable */
        OUT_RING  (txs);
        OUT_RING  (ps->filt | 0x2000 /* magic */);
-       OUT_RING  ((pt->width[0] << 16) | pt->height[0]);
+       OUT_RING  ((pt->width0 << 16) | pt->height0);
        OUT_RING  (ps->bcol);
 #endif
 }
index 34e3c2ebd770e86d2456049a2e34e6f32b3f19fe..908482ad854890eaee16fa33006f395714292de6 100644 (file)
@@ -1,6 +1,8 @@
 #include "pipe/p_state.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
 
 #include "nv10_context.h"
 #include "nv10_screen.h"
@@ -10,7 +12,7 @@ nv10_miptree_layout(struct nv10_miptree *nv10mt)
 {
        struct pipe_texture *pt = &nv10mt->base;
        boolean swizzled = FALSE;
-       uint width = pt->width[0], height = pt->height[0];
+       uint width = pt->width0;
        uint offset = 0;
        int nr_faces, l, f;
 
@@ -21,29 +23,23 @@ nv10_miptree_layout(struct nv10_miptree *nv10mt)
        }
        
        for (l = 0; l <= pt->last_level; l++) {
-               pt->width[l] = width;
-               pt->height[l] = height;
-               pt->nblocksx[l] = pf_get_nblocksx(&pt->block, width);
-               pt->nblocksy[l] = pf_get_nblocksy(&pt->block, height);
-
                if (swizzled)
-                       nv10mt->level[l].pitch = pt->nblocksx[l] * pt->block.size;
+                       nv10mt->level[l].pitch = util_format_get_stride(pt->format, width);
                else
-                       nv10mt->level[l].pitch = pt->nblocksx[0] * pt->block.size;
+                       nv10mt->level[l].pitch = util_format_get_stride(pt->format, pt->width0);
                nv10mt->level[l].pitch = (nv10mt->level[l].pitch + 63) & ~63;
 
                nv10mt->level[l].image_offset =
                        CALLOC(nr_faces, sizeof(unsigned));
 
-               width  = MAX2(1, width  >> 1);
-               height = MAX2(1, height >> 1);
+               width  = u_minify(width, 1);
 
        }
 
        for (f = 0; f < nr_faces; f++) {
                for (l = 0; l <= pt->last_level; l++) {
                        nv10mt->level[l].image_offset[f] = offset;
-                       offset += nv10mt->level[l].pitch * pt->height[l];
+                       offset += nv10mt->level[l].pitch * u_minify(pt->height0, l);
                }
        }
 
@@ -58,7 +54,7 @@ nv10_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
 
        /* Only supports 2D, non-mipmapped textures for the moment */
        if (pt->target != PIPE_TEXTURE_2D || pt->last_level != 0 ||
-           pt->depth[0] != 1)
+           pt->depth0 != 1)
                return NULL;
 
        mt = CALLOC_STRUCT(nv10_miptree);
@@ -72,6 +68,7 @@ nv10_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
        mt->level[0].image_offset = CALLOC(1, sizeof(unsigned));
 
        pipe_buffer_reference(&mt->buffer, pb);
+       mt->bo = nouveau_bo(mt->buffer);
        return &mt->base;
 }
 
@@ -95,6 +92,7 @@ nv10_miptree_create(struct pipe_screen *screen, const struct pipe_texture *pt)
                FREE(mt);
                return NULL;
        }
+       mt->bo = nouveau_bo(mt->buffer);
        
        return &mt->base;
 }
@@ -133,8 +131,8 @@ nv10_miptree_surface_get(struct pipe_screen *screen, struct pipe_texture *pt,
                return NULL;
        pipe_texture_reference(&ns->base.texture, pt);
        ns->base.format = pt->format;
-       ns->base.width = pt->width[level];
-       ns->base.height = pt->height[level];
+       ns->base.width = u_minify(pt->width0, level);
+       ns->base.height = u_minify(pt->height0, level);
        ns->base.usage = flags;
        pipe_reference_init(&ns->base.reference, 1);
        ns->base.face = face;
index 9b38219b996d70cbc1d607c11e35586b8fe6db1f..ffc6be3c401d68740f162bb86d15ab001d4d2299 100644 (file)
@@ -553,9 +553,9 @@ nv10_init_state_functions(struct nv10_context *nv10)
        nv10->pipe.delete_blend_state = nv10_blend_state_delete;
 
        nv10->pipe.create_sampler_state = nv10_sampler_state_create;
-       nv10->pipe.bind_sampler_states = nv10_sampler_state_bind;
+       nv10->pipe.bind_fragment_sampler_states = nv10_sampler_state_bind;
        nv10->pipe.delete_sampler_state = nv10_sampler_state_delete;
-       nv10->pipe.set_sampler_textures = nv10_set_sampler_texture;
+       nv10->pipe.set_fragment_sampler_textures = nv10_set_sampler_texture;
 
        nv10->pipe.create_rasterizer_state = nv10_rasterizer_state_create;
        nv10->pipe.bind_rasterizer_state = nv10_rasterizer_state_bind;
index 3a3fd0d4f4f01f151ceaa00d975c7813d4bf02f0..2524ac02e29d95bb39cbc4879eb894a84e18456c 100644 (file)
@@ -126,6 +126,7 @@ struct nv10_depth_stencil_alpha_state {
 
 struct nv10_miptree {
        struct pipe_texture base;
+       struct nouveau_bo *bo;
 
        struct pipe_buffer *buffer;
        uint total_size;
index 8feb85e4bda22eb33837d7002fd0777cf9edd0ca..c664973e90463c8cfaeab56c28f7310980c37192 100644 (file)
@@ -1,7 +1,9 @@
 #include <pipe/p_state.h>
 #include <pipe/p_defines.h>
 #include <pipe/p_inlines.h>
+#include <util/u_format.h>
 #include <util/u_memory.h>
+#include <util/u_math.h>
 #include <nouveau/nouveau_winsys.h>
 #include "nv10_context.h"
 #include "nv10_screen.h"
@@ -10,7 +12,7 @@
 struct nv10_transfer {
        struct pipe_transfer base;
        struct pipe_surface *surface;
-       bool direct;
+       boolean direct;
 };
 
 static void
@@ -20,12 +22,9 @@ nv10_compatible_transfer_tex(struct pipe_texture *pt, unsigned level,
        memset(template, 0, sizeof(struct pipe_texture));
        template->target = pt->target;
        template->format = pt->format;
-       template->width[0] = pt->width[level];
-       template->height[0] = pt->height[level];
-       template->depth[0] = 1;
-       template->block = pt->block;
-       template->nblocksx[0] = pt->nblocksx[level];
-       template->nblocksy[0] = pt->nblocksx[level];
+       template->width0 = u_minify(pt->width0, level);
+       template->height0 = u_minify(pt->height0, level);
+       template->depth0 = 1;
        template->last_level = 0;
        template->nr_samples = pt->nr_samples;
 
@@ -48,14 +47,10 @@ nv10_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
                return NULL;
 
        pipe_texture_reference(&tx->base.texture, pt);
-       tx->base.format = pt->format;
        tx->base.x = x;
        tx->base.y = y;
        tx->base.width = w;
        tx->base.height = h;
-       tx->base.block = pt->block;
-       tx->base.nblocksx = pt->nblocksx[level];
-       tx->base.nblocksy = pt->nblocksy[level];
        tx->base.stride = mt->level[level].pitch;
        tx->base.usage = usage;
        tx->base.face = face;
@@ -157,7 +152,7 @@ nv10_transfer_map(struct pipe_screen *pscreen, struct pipe_transfer *ptx)
                                    pipe_transfer_buffer_flags(ptx));
 
        return map + ns->base.offset +
-              ptx->y * ns->pitch + ptx->x * ptx->block.size;
+              ptx->y * ns->pitch + ptx->x * util_format_get_blocksize(ptx->texture->format);
 }
 
 static void
index 441a4f75f3f84ddfde01116b4ed01d44e1b16b0a..0d2614124853375a64bd2261580694ca202cc460 100644 (file)
@@ -45,6 +45,7 @@ boolean nv10_draw_elements( struct pipe_context *pipe,
        }
 
        draw_set_mapped_constant_buffer(draw,
+                                        PIPE_SHADER_VERTEX,
                                        nv10->constbuf[PIPE_SHADER_VERTEX],
                                        nv10->constbuf_nr[PIPE_SHADER_VERTEX]);
 
index 276db8b57b695886cd74373edd5eb16fc1c275ae..6a147a4159cebc38685a1d1389871e92901dba01 100644 (file)
@@ -375,11 +375,6 @@ static void nv20_init_hwctx(struct nv20_context *nv20)
        FIRE_RING (NULL);
 }
 
-static void
-nv20_set_edgeflags(struct pipe_context *pipe, const unsigned *bitfield)
-{
-}
-
 struct pipe_context *
 nv20_create(struct pipe_screen *pscreen, unsigned pctx_id)
 {
@@ -399,7 +394,6 @@ nv20_create(struct pipe_screen *pscreen, unsigned pctx_id)
        nv20->pipe.winsys = ws;
        nv20->pipe.screen = pscreen;
        nv20->pipe.destroy = nv20_destroy;
-       nv20->pipe.set_edgeflags = nv20_set_edgeflags;
        nv20->pipe.draw_arrays = nv20_draw_arrays;
        nv20->pipe.draw_elements = nv20_draw_elements;
        nv20->pipe.clear = nv20_clear;
index 495a7be9127f85301a593cfec1e14c7ddeaed2aa..2db4a4015a69c4dbef85e7c3ea322432d7396c10 100644 (file)
@@ -62,9 +62,9 @@ nv20_fragtex_build(struct nv20_context *nv20, int unit)
 
        txf  = tf->format << 8;
        txf |= (pt->last_level + 1) << 16;
-       txf |= log2i(pt->width[0]) << 20;
-       txf |= log2i(pt->height[0]) << 24;
-       txf |= log2i(pt->depth[0]) << 28;
+       txf |= log2i(pt->width0) << 20;
+       txf |= log2i(pt->height0) << 24;
+       txf |= log2i(pt->depth0) << 28;
        txf |= 8;
 
        switch (pt->target) {
@@ -89,7 +89,7 @@ nv20_fragtex_build(struct nv20_context *nv20, int unit)
        OUT_RING  (0x40000000); /* enable */
        OUT_RING  (txs);
        OUT_RING  (ps->filt | 0x2000 /* magic */);
-       OUT_RING  ((pt->width[0] << 16) | pt->height[0]);
+       OUT_RING  ((pt->width0 << 16) | pt->height0);
        OUT_RING  (ps->bcol);
 #endif
 }
index 185fbf53e0f8fb7419fc9bd0939c0fbe9b2433b1..d1291a92e0ab1405c6e8bb18860ae27b2a905caf 100644 (file)
@@ -1,6 +1,8 @@
 #include "pipe/p_state.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
 
 #include "nv20_context.h"
 #include "nv20_screen.h"
@@ -9,7 +11,7 @@ static void
 nv20_miptree_layout(struct nv20_miptree *nv20mt)
 {
        struct pipe_texture *pt = &nv20mt->base;
-       uint width = pt->width[0], height = pt->height[0];
+       uint width = pt->width0;
        uint offset = 0;
        int nr_faces, l, f;
        uint wide_pitch = pt->tex_usage & (PIPE_TEXTURE_USAGE_SAMPLER |
@@ -25,21 +27,15 @@ nv20_miptree_layout(struct nv20_miptree *nv20mt)
        }
        
        for (l = 0; l <= pt->last_level; l++) {
-               pt->width[l] = width;
-               pt->height[l] = height;
-               pt->nblocksx[l] = pf_get_nblocksx(&pt->block, width);
-               pt->nblocksy[l] = pf_get_nblocksy(&pt->block, height);
-
                if (wide_pitch && (pt->tex_usage & NOUVEAU_TEXTURE_USAGE_LINEAR))
-                       nv20mt->level[l].pitch = align(pt->width[0] * pt->block.size, 64);
+                       nv20mt->level[l].pitch = align(util_format_get_stride(pt->format, pt->width0), 64);
                else
-                       nv20mt->level[l].pitch = pt->width[l] * pt->block.size;
+                       nv20mt->level[l].pitch = util_format_get_stride(pt->format, width);
 
                nv20mt->level[l].image_offset =
                        CALLOC(nr_faces, sizeof(unsigned));
 
-               width  = MAX2(1, width  >> 1);
-               height = MAX2(1, height >> 1);
+               width  = u_minify(width, 1);
        }
 
        for (f = 0; f < nr_faces; f++) {
@@ -47,14 +43,14 @@ nv20_miptree_layout(struct nv20_miptree *nv20mt)
                        nv20mt->level[l].image_offset[f] = offset;
 
                        if (!(pt->tex_usage & NOUVEAU_TEXTURE_USAGE_LINEAR) &&
-                           pt->width[l + 1] > 1 && pt->height[l + 1] > 1)
-                               offset += align(nv20mt->level[l].pitch * pt->height[l], 64);
+                           u_minify(pt->width0, l + 1) > 1 && u_minify(pt->height0, l + 1) > 1)
+                               offset += align(nv20mt->level[l].pitch * u_minify(pt->height0, l), 64);
                        else
-                               offset += nv20mt->level[l].pitch * pt->height[l];
+                               offset += nv20mt->level[l].pitch * u_minify(pt->height0, l);
                }
 
                nv20mt->level[l].image_offset[f] = offset;
-               offset += nv20mt->level[l].pitch * pt->height[l];
+               offset += nv20mt->level[l].pitch * u_minify(pt->height0, l);
        }
 
        nv20mt->total_size = offset;
@@ -68,7 +64,7 @@ nv20_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
 
        /* Only supports 2D, non-mipmapped textures for the moment */
        if (pt->target != PIPE_TEXTURE_2D || pt->last_level != 0 ||
-           pt->depth[0] != 1)
+           pt->depth0 != 1)
                return NULL;
 
        mt = CALLOC_STRUCT(nv20_miptree);
@@ -82,6 +78,7 @@ nv20_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
        mt->level[0].image_offset = CALLOC(1, sizeof(unsigned));
 
        pipe_buffer_reference(&mt->buffer, pb);
+       mt->bo = nouveau_bo(mt->buffer);
        return &mt->base;
 }
 
@@ -100,8 +97,8 @@ nv20_miptree_create(struct pipe_screen *screen, const struct pipe_texture *pt)
        mt->base.screen = screen;
 
        /* Swizzled textures must be POT */
-       if (pt->width[0] & (pt->width[0] - 1) ||
-           pt->height[0] & (pt->height[0] - 1))
+       if (pt->width0 & (pt->width0 - 1) ||
+           pt->height0 & (pt->height0 - 1))
                mt->base.tex_usage |= NOUVEAU_TEXTURE_USAGE_LINEAR;
        else
        if (pt->tex_usage & (PIPE_TEXTURE_USAGE_PRIMARY |
@@ -137,6 +134,7 @@ nv20_miptree_create(struct pipe_screen *screen, const struct pipe_texture *pt)
                FREE(mt);
                return NULL;
        }
+       mt->bo = nouveau_bo(mt->buffer);
        
        return &mt->base;
 }
@@ -167,8 +165,8 @@ nv20_miptree_surface_get(struct pipe_screen *screen, struct pipe_texture *pt,
                return NULL;
        pipe_texture_reference(&ns->base.texture, pt);
        ns->base.format = pt->format;
-       ns->base.width = pt->width[level];
-       ns->base.height = pt->height[level];
+       ns->base.width = u_minify(pt->width0, level);
+       ns->base.height = u_minify(pt->height0, level);
        ns->base.usage = flags;
        pipe_reference_init(&ns->base.reference, 1);
        ns->base.face = face;
index ed4084980f20148af52ae8aec449511eff03d531..3a82e63423dcea9b595a573c25479f9af90c452b 100644 (file)
@@ -546,9 +546,9 @@ nv20_init_state_functions(struct nv20_context *nv20)
        nv20->pipe.delete_blend_state = nv20_blend_state_delete;
 
        nv20->pipe.create_sampler_state = nv20_sampler_state_create;
-       nv20->pipe.bind_sampler_states = nv20_sampler_state_bind;
+       nv20->pipe.bind_fragment_sampler_states = nv20_sampler_state_bind;
        nv20->pipe.delete_sampler_state = nv20_sampler_state_delete;
-       nv20->pipe.set_sampler_textures = nv20_set_sampler_texture;
+       nv20->pipe.set_fragment_sampler_textures = nv20_set_sampler_texture;
 
        nv20->pipe.create_rasterizer_state = nv20_rasterizer_state_create;
        nv20->pipe.bind_rasterizer_state = nv20_rasterizer_state_bind;
index 34f402fdcbf4140c74bb335109218287bf08533c..dde41065685d6186a0d1644ee30787116d1de4c8 100644 (file)
@@ -126,6 +126,7 @@ struct nv20_depth_stencil_alpha_state {
 
 struct nv20_miptree {
        struct pipe_texture base;
+       struct nouveau_bo *bo;
 
        struct pipe_buffer *buffer;
        uint total_size;
index 0122b1c2cdbd70eeefb3b05635f58dd117204e53..63cba1f4122138dfa29f40aace11c170818dea80 100644 (file)
@@ -228,7 +228,7 @@ static void nv20_vertex_layout(struct nv20_context *nv20)
        }
 
        /* always do position */ {
-               src = draw_find_vs_output(dc, TGSI_SEMANTIC_POSITION, 0);
+               src = draw_find_shader_output(dc, TGSI_SEMANTIC_POSITION, 0);
                draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_LINEAR, src);
                vinfo->hwfmt[0] |= (1 << 0);
        }
@@ -237,19 +237,19 @@ static void nv20_vertex_layout(struct nv20_context *nv20)
        for (i = 4; i < 6; i++) {
                if (!generics[i])
                        continue;
-               src = draw_find_vs_output(dc, TGSI_SEMANTIC_GENERIC, i);
+               src = draw_find_shader_output(dc, TGSI_SEMANTIC_GENERIC, i);
                draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_PERSPECTIVE, src);
                vinfo->hwfmt[0] |= (1 << (i - 3));
        }
 
        if (colors[0]) {
-               src = draw_find_vs_output(dc, TGSI_SEMANTIC_COLOR, 0);
+               src = draw_find_shader_output(dc, TGSI_SEMANTIC_COLOR, 0);
                draw_emit_vertex_attr(vinfo, EMIT_4F, colorInterp, src);
                vinfo->hwfmt[0] |= (1 << 3);
        }
 
        if (colors[1]) {
-               src = draw_find_vs_output(dc, TGSI_SEMANTIC_COLOR, 1);
+               src = draw_find_shader_output(dc, TGSI_SEMANTIC_COLOR, 1);
                draw_emit_vertex_attr(vinfo, EMIT_4F, colorInterp, src);
                vinfo->hwfmt[0] |= (1 << 4);
        }
@@ -258,7 +258,7 @@ static void nv20_vertex_layout(struct nv20_context *nv20)
        for (i = 6; i < 10; i++) {
                if (!generics[i])
                        continue;
-               src = draw_find_vs_output(dc, TGSI_SEMANTIC_GENERIC, i);
+               src = draw_find_shader_output(dc, TGSI_SEMANTIC_GENERIC, i);
                draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_PERSPECTIVE, src);
                vinfo->hwfmt[0] |= (1 << (i - 1));
        }
@@ -267,7 +267,7 @@ static void nv20_vertex_layout(struct nv20_context *nv20)
        for (i = 0; i < 4; i++) {
                if (!generics[i])
                        continue;
-               src = draw_find_vs_output(dc, TGSI_SEMANTIC_GENERIC, i);
+               src = draw_find_shader_output(dc, TGSI_SEMANTIC_GENERIC, i);
                draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_PERSPECTIVE, src);
                vinfo->hwfmt[0] |= (1 << (i + 9));
        }
@@ -276,13 +276,13 @@ static void nv20_vertex_layout(struct nv20_context *nv20)
        for (i = 10; i < 12; i++) {
                if (!generics[i])
                        continue;
-               src = draw_find_vs_output(dc, TGSI_SEMANTIC_GENERIC, i);
+               src = draw_find_shader_output(dc, TGSI_SEMANTIC_GENERIC, i);
                draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_PERSPECTIVE, src);
                vinfo->hwfmt[0] |= (1 << (i + 3));
        }
 
        if (fog) {
-               src = draw_find_vs_output(dc, TGSI_SEMANTIC_FOG, 0);
+               src = draw_find_shader_output(dc, TGSI_SEMANTIC_FOG, 0);
                draw_emit_vertex_attr(vinfo, EMIT_1F, INTERP_PERSPECTIVE, src);
                vinfo->hwfmt[0] |= (1 << 15);
        }
index 81b4f1a9177f47beb75e7c138abb17ca3b975a7e..69b79c809f470ff4209950bd2f0dcf83852c06b6 100644 (file)
@@ -1,7 +1,9 @@
 #include <pipe/p_state.h>
 #include <pipe/p_defines.h>
 #include <pipe/p_inlines.h>
+#include <util/u_format.h>
 #include <util/u_memory.h>
+#include <util/u_math.h>
 #include <nouveau/nouveau_winsys.h>
 #include "nv20_context.h"
 #include "nv20_screen.h"
@@ -10,7 +12,7 @@
 struct nv20_transfer {
        struct pipe_transfer base;
        struct pipe_surface *surface;
-       bool direct;
+       boolean direct;
 };
 
 static void
@@ -20,12 +22,9 @@ nv20_compatible_transfer_tex(struct pipe_texture *pt, unsigned level,
        memset(template, 0, sizeof(struct pipe_texture));
        template->target = pt->target;
        template->format = pt->format;
-       template->width[0] = pt->width[level];
-       template->height[0] = pt->height[level];
-       template->depth[0] = 1;
-       template->block = pt->block;
-       template->nblocksx[0] = pt->nblocksx[level];
-       template->nblocksy[0] = pt->nblocksx[level];
+       template->width0 = u_minify(pt->width0, level);
+       template->height0 = u_minify(pt->height0, level);
+       template->depth0 = 1;
        template->last_level = 0;
        template->nr_samples = pt->nr_samples;
 
@@ -48,14 +47,10 @@ nv20_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
                return NULL;
 
        pipe_texture_reference(&tx->base.texture, pt);
-       tx->base.format = pt->format;
        tx->base.x = x;
        tx->base.y = y;
        tx->base.width = w;
        tx->base.height = h;
-       tx->base.block = pt->block;
-       tx->base.nblocksx = pt->nblocksx[level];
-       tx->base.nblocksy = pt->nblocksy[level];
        tx->base.stride = mt->level[level].pitch;
        tx->base.usage = usage;
        tx->base.face = face;
@@ -157,7 +152,7 @@ nv20_transfer_map(struct pipe_screen *pscreen, struct pipe_transfer *ptx)
                                    pipe_transfer_buffer_flags(ptx));
 
        return map + ns->base.offset +
-              ptx->y * ns->pitch + ptx->x * ptx->block.size;
+              ptx->y * ns->pitch + ptx->x * util_format_get_blocksize(ptx->texture->format);
 }
 
 static void
index 84d7db6c5e27c3274eea2f40340e06311db8a04e..4bf461eba923d2b4c52a4b101bd99cff30ce3f37 100644 (file)
@@ -45,7 +45,7 @@ boolean nv20_draw_elements( struct pipe_context *pipe,
                draw_set_mapped_element_buffer(draw, 0, NULL);
        }
 
-       draw_set_mapped_constant_buffer(draw,
+       draw_set_mapped_constant_buffer(draw, PIPE_SHADER_VERTEX,
                                        nv20->constbuf[PIPE_SHADER_VERTEX],
                                        nv20->constbuf_nr[PIPE_SHADER_VERTEX]);
 
index 388245ecb042a430ed800588a497ec88423726e1..7886c2af7e62d75d66deb3d61824d013358c7ddf 100644 (file)
@@ -253,32 +253,32 @@ static INLINE struct nv20_sreg
 tgsi_src(struct nv20_vpc *vpc, const struct tgsi_full_src_register *fsrc) {
        struct nv20_sreg src;
 
-       switch (fsrc->SrcRegister.File) {
+       switch (fsrc->Register.File) {
        case TGSI_FILE_INPUT:
-               src = nv20_sr(NV30SR_INPUT, fsrc->SrcRegister.Index);
+               src = nv20_sr(NV30SR_INPUT, fsrc->Register.Index);
                break;
        case TGSI_FILE_CONSTANT:
-               src = constant(vpc, fsrc->SrcRegister.Index, 0, 0, 0, 0);
+               src = constant(vpc, fsrc->Register.Index, 0, 0, 0, 0);
                break;
        case TGSI_FILE_IMMEDIATE:
-               src = vpc->imm[fsrc->SrcRegister.Index];
+               src = vpc->imm[fsrc->Register.Index];
                break;
        case TGSI_FILE_TEMPORARY:
-               if (vpc->high_temp < fsrc->SrcRegister.Index)
-                       vpc->high_temp = fsrc->SrcRegister.Index;
-               src = nv20_sr(NV30SR_TEMP, fsrc->SrcRegister.Index);
+               if (vpc->high_temp < fsrc->Register.Index)
+                       vpc->high_temp = fsrc->Register.Index;
+               src = nv20_sr(NV30SR_TEMP, fsrc->Register.Index);
                break;
        default:
                NOUVEAU_ERR("bad src file\n");
                break;
        }
 
-       src.abs = fsrc->SrcRegisterExtMod.Absolute;
-       src.negate = fsrc->SrcRegister.Negate;
-       src.swz[0] = fsrc->SrcRegister.SwizzleX;
-       src.swz[1] = fsrc->SrcRegister.SwizzleY;
-       src.swz[2] = fsrc->SrcRegister.SwizzleZ;
-       src.swz[3] = fsrc->SrcRegister.SwizzleW;
+       src.abs = fsrc->Register.Absolute;
+       src.negate = fsrc->Register.Negate;
+       src.swz[0] = fsrc->Register.SwizzleX;
+       src.swz[1] = fsrc->Register.SwizzleY;
+       src.swz[2] = fsrc->Register.SwizzleZ;
+       src.swz[3] = fsrc->Register.SwizzleW;
        return src;
 }
 
@@ -286,14 +286,14 @@ static INLINE struct nv20_sreg
 tgsi_dst(struct nv20_vpc *vpc, const struct tgsi_full_dst_register *fdst) {
        struct nv20_sreg dst;
 
-       switch (fdst->DstRegister.File) {
+       switch (fdst->Register.File) {
        case TGSI_FILE_OUTPUT:
                dst = nv20_sr(NV30SR_OUTPUT,
-                             vpc->output_map[fdst->DstRegister.Index]);
+                             vpc->output_map[fdst->Register.Index]);
 
                break;
        case TGSI_FILE_TEMPORARY:
-               dst = nv20_sr(NV30SR_TEMP, fdst->DstRegister.Index);
+               dst = nv20_sr(NV30SR_TEMP, fdst->Register.Index);
                if (vpc->high_temp < dst.index)
                        vpc->high_temp = dst.index;
                break;
@@ -334,8 +334,8 @@ nv20_vertprog_parse_instruction(struct nv20_vpc *vpc,
        for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
                const struct tgsi_full_src_register *fsrc;
 
-               fsrc = &finst->FullSrcRegisters[i];
-               if (fsrc->SrcRegister.File == TGSI_FILE_TEMPORARY) {
+               fsrc = &finst->Src[i];
+               if (fsrc->Register.File == TGSI_FILE_TEMPORARY) {
                        src[i] = tgsi_src(vpc, fsrc);
                }
        }
@@ -343,11 +343,11 @@ nv20_vertprog_parse_instruction(struct nv20_vpc *vpc,
        for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
                const struct tgsi_full_src_register *fsrc;
 
-               fsrc = &finst->FullSrcRegisters[i];
-               switch (fsrc->SrcRegister.File) {
+               fsrc = &finst->Src[i];
+               switch (fsrc->Register.File) {
                case TGSI_FILE_INPUT:
-                       if (ai == -1 || ai == fsrc->SrcRegister.Index) {
-                               ai = fsrc->SrcRegister.Index;
+                       if (ai == -1 || ai == fsrc->Register.Index) {
+                               ai = fsrc->Register.Index;
                                src[i] = tgsi_src(vpc, fsrc);
                        } else {
                                src[i] = temp(vpc);
@@ -360,8 +360,8 @@ nv20_vertprog_parse_instruction(struct nv20_vpc *vpc,
                 */
                case TGSI_FILE_CONSTANT:
                case TGSI_FILE_IMMEDIATE:
-                       if (ci == -1 || ci == fsrc->SrcRegister.Index) {
-                               ci = fsrc->SrcRegister.Index;
+                       if (ci == -1 || ci == fsrc->Register.Index) {
+                               ci = fsrc->Register.Index;
                                src[i] = tgsi_src(vpc, fsrc);
                        } else {
                                src[i] = temp(vpc);
@@ -378,8 +378,8 @@ nv20_vertprog_parse_instruction(struct nv20_vpc *vpc,
                }
        }
 
-       dst  = tgsi_dst(vpc, &finst->FullDstRegisters[0]);
-       mask = tgsi_mask(finst->FullDstRegisters[0].DstRegister.WriteMask);
+       dst  = tgsi_dst(vpc, &finst->Dst[0]);
+       mask = tgsi_mask(finst->Dst[0].Register.WriteMask);
 
        switch (finst->Instruction.Opcode) {
        case TGSI_OPCODE_ABS:
@@ -490,15 +490,15 @@ nv20_vertprog_parse_decl_output(struct nv20_vpc *vpc,
 {
        int hw;
 
-       switch (fdec->Semantic.SemanticName) {
+       switch (fdec->Semantic.Name) {
        case TGSI_SEMANTIC_POSITION:
                hw = NV30_VP_INST_DEST_POS;
                break;
        case TGSI_SEMANTIC_COLOR:
-               if (fdec->Semantic.SemanticIndex == 0) {
+               if (fdec->Semantic.Index == 0) {
                        hw = NV30_VP_INST_DEST_COL0;
                } else
-               if (fdec->Semantic.SemanticIndex == 1) {
+               if (fdec->Semantic.Index == 1) {
                        hw = NV30_VP_INST_DEST_COL1;
                } else {
                        NOUVEAU_ERR("bad colour semantic index\n");
@@ -506,10 +506,10 @@ nv20_vertprog_parse_decl_output(struct nv20_vpc *vpc,
                }
                break;
        case TGSI_SEMANTIC_BCOLOR:
-               if (fdec->Semantic.SemanticIndex == 0) {
+               if (fdec->Semantic.Index == 0) {
                        hw = NV30_VP_INST_DEST_BFC0;
                } else
-               if (fdec->Semantic.SemanticIndex == 1) {
+               if (fdec->Semantic.Index == 1) {
                        hw = NV30_VP_INST_DEST_BFC1;
                } else {
                        NOUVEAU_ERR("bad bcolour semantic index\n");
@@ -523,19 +523,22 @@ nv20_vertprog_parse_decl_output(struct nv20_vpc *vpc,
                hw = NV30_VP_INST_DEST_PSZ;
                break;
        case TGSI_SEMANTIC_GENERIC:
-               if (fdec->Semantic.SemanticIndex <= 7) {
-                       hw = NV30_VP_INST_DEST_TC(fdec->Semantic.SemanticIndex);
+               if (fdec->Semantic.Index <= 7) {
+                       hw = NV30_VP_INST_DEST_TC(fdec->Semantic.Index);
                } else {
                        NOUVEAU_ERR("bad generic semantic index\n");
                        return FALSE;
                }
                break;
+       case TGSI_SEMANTIC_EDGEFLAG:
+               NOUVEAU_ERR("cannot handle edgeflag output\n");
+               return FALSE;
        default:
                NOUVEAU_ERR("bad output semantic\n");
                return FALSE;
        }
 
-       vpc->output_map[fdec->DeclarationRange.First] = hw;
+       vpc->output_map[fdec->Range.First] = hw;
        return TRUE;
 }
 
index d8300fd69f647a95ed00ce0581deed5012707a3b..46a821a48b1f03b22e071120083e425d946918f8 100644 (file)
@@ -58,6 +58,9 @@ nv30_create(struct pipe_screen *pscreen, unsigned pctx_id)
        nv30->pipe.is_texture_referenced = nouveau_is_texture_referenced;
        nv30->pipe.is_buffer_referenced = nouveau_is_buffer_referenced;
 
+       screen->base.channel->user_private = nv30;
+       screen->base.channel->flush_notify = nv30_state_flush_notify;
+
        nv30_init_query_functions(nv30);
        nv30_init_surface_functions(nv30);
        nv30_init_state_functions(nv30);
index 8d49366dfcb813dd91e54812593e52caeb760466..864ddaeb598eefb937c2bc7e1ddb482ecba99098 100644 (file)
@@ -144,7 +144,6 @@ struct nv30_context {
        unsigned vtxbuf_nr;
        struct pipe_vertex_element vtxelt[PIPE_MAX_ATTRIBS];
        unsigned vtxelt_nr;
-       const unsigned *edgeflags;
 };
 
 static INLINE struct nv30_context *
@@ -184,6 +183,7 @@ extern void nv30_fragtex_bind(struct nv30_context *);
 /* nv30_state.c and friends */
 extern boolean nv30_state_validate(struct nv30_context *nv30);
 extern void nv30_state_emit(struct nv30_context *nv30);
+extern void nv30_state_flush_notify(struct nouveau_channel *chan);
 extern struct nv30_state_entry nv30_state_rasterizer;
 extern struct nv30_state_entry nv30_state_scissor;
 extern struct nv30_state_entry nv30_state_stipple;
index 0ce702d6f8449d6ca8399d03ce0d7f0943df7b48..40965a97723b5c2cfe2927ab60c4d3c7b3b0d079 100644 (file)
@@ -237,20 +237,20 @@ tgsi_src(struct nv30_fpc *fpc, const struct tgsi_full_src_register *fsrc)
 {
        struct nv30_sreg src;
 
-       switch (fsrc->SrcRegister.File) {
+       switch (fsrc->Register.File) {
        case TGSI_FILE_INPUT:
                src = nv30_sr(NV30SR_INPUT,
-                             fpc->attrib_map[fsrc->SrcRegister.Index]);
+                             fpc->attrib_map[fsrc->Register.Index]);
                break;
        case TGSI_FILE_CONSTANT:
-               src = constant(fpc, fsrc->SrcRegister.Index, NULL);
+               src = constant(fpc, fsrc->Register.Index, NULL);
                break;
        case TGSI_FILE_IMMEDIATE:
-               assert(fsrc->SrcRegister.Index < fpc->nr_imm);
-               src = fpc->imm[fsrc->SrcRegister.Index];
+               assert(fsrc->Register.Index < fpc->nr_imm);
+               src = fpc->imm[fsrc->Register.Index];
                break;
        case TGSI_FILE_TEMPORARY:
-               src = nv30_sr(NV30SR_TEMP, fsrc->SrcRegister.Index + 1);
+               src = nv30_sr(NV30SR_TEMP, fsrc->Register.Index + 1);
                if (fpc->high_temp < src.index)
                        fpc->high_temp = src.index;
                break;
@@ -258,7 +258,7 @@ tgsi_src(struct nv30_fpc *fpc, const struct tgsi_full_src_register *fsrc)
         * Luckily fragprog results are just temp regs..
         */
        case TGSI_FILE_OUTPUT:
-               if (fsrc->SrcRegister.Index == fpc->colour_id)
+               if (fsrc->Register.Index == fpc->colour_id)
                        return nv30_sr(NV30SR_OUTPUT, 0);
                else
                        return nv30_sr(NV30SR_OUTPUT, 1);
@@ -268,12 +268,12 @@ tgsi_src(struct nv30_fpc *fpc, const struct tgsi_full_src_register *fsrc)
                break;
        }
 
-       src.abs = fsrc->SrcRegisterExtMod.Absolute;
-       src.negate = fsrc->SrcRegister.Negate;
-       src.swz[0] = fsrc->SrcRegister.SwizzleX;
-       src.swz[1] = fsrc->SrcRegister.SwizzleY;
-       src.swz[2] = fsrc->SrcRegister.SwizzleZ;
-       src.swz[3] = fsrc->SrcRegister.SwizzleW;
+       src.abs = fsrc->Register.Absolute;
+       src.negate = fsrc->Register.Negate;
+       src.swz[0] = fsrc->Register.SwizzleX;
+       src.swz[1] = fsrc->Register.SwizzleY;
+       src.swz[2] = fsrc->Register.SwizzleZ;
+       src.swz[3] = fsrc->Register.SwizzleW;
        return src;
 }
 
@@ -281,22 +281,22 @@ static INLINE struct nv30_sreg
 tgsi_dst(struct nv30_fpc *fpc, const struct tgsi_full_dst_register *fdst) {
        int idx;
 
-       switch (fdst->DstRegister.File) {
+       switch (fdst->Register.File) {
        case TGSI_FILE_OUTPUT:
-               if (fdst->DstRegister.Index == fpc->colour_id)
+               if (fdst->Register.Index == fpc->colour_id)
                        return nv30_sr(NV30SR_OUTPUT, 0);
                else
                        return nv30_sr(NV30SR_OUTPUT, 1);
                break;
        case TGSI_FILE_TEMPORARY:
-               idx = fdst->DstRegister.Index + 1;
+               idx = fdst->Register.Index + 1;
                if (fpc->high_temp < idx)
                        fpc->high_temp = idx;
                return nv30_sr(NV30SR_TEMP, idx);
        case TGSI_FILE_NULL:
                return nv30_sr(NV30SR_NONE, 0);
        default:
-               NOUVEAU_ERR("bad dst file %d\n", fdst->DstRegister.File);
+               NOUVEAU_ERR("bad dst file %d\n", fdst->Register.File);
                return nv30_sr(NV30SR_NONE, 0);
        }
 }
@@ -363,8 +363,8 @@ nv30_fragprog_parse_instruction(struct nv30_fpc *fpc,
        for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
                const struct tgsi_full_src_register *fsrc;
 
-               fsrc = &finst->FullSrcRegisters[i];
-               if (fsrc->SrcRegister.File == TGSI_FILE_TEMPORARY) {
+               fsrc = &finst->Src[i];
+               if (fsrc->Register.File == TGSI_FILE_TEMPORARY) {
                        src[i] = tgsi_src(fpc, fsrc);
                }
        }
@@ -372,9 +372,9 @@ nv30_fragprog_parse_instruction(struct nv30_fpc *fpc,
        for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
                const struct tgsi_full_src_register *fsrc;
 
-               fsrc = &finst->FullSrcRegisters[i];
+               fsrc = &finst->Src[i];
 
-               switch (fsrc->SrcRegister.File) {
+               switch (fsrc->Register.File) {
                case TGSI_FILE_INPUT:
                case TGSI_FILE_CONSTANT:
                case TGSI_FILE_TEMPORARY:
@@ -385,14 +385,14 @@ nv30_fragprog_parse_instruction(struct nv30_fpc *fpc,
                        break;
                }
 
-               switch (fsrc->SrcRegister.File) {
+               switch (fsrc->Register.File) {
                case TGSI_FILE_INPUT:
-                       if (ai == -1 || ai == fsrc->SrcRegister.Index) {
-                               ai = fsrc->SrcRegister.Index;
+                       if (ai == -1 || ai == fsrc->Register.Index) {
+                               ai = fsrc->Register.Index;
                                src[i] = tgsi_src(fpc, fsrc);
                        } else {
                                NOUVEAU_MSG("extra src attr %d\n",
-                                        fsrc->SrcRegister.Index);
+                                        fsrc->Register.Index);
                                src[i] = temp(fpc);
                                arith(fpc, 0, MOV, src[i], MASK_ALL,
                                      tgsi_src(fpc, fsrc), none, none);
@@ -400,8 +400,8 @@ nv30_fragprog_parse_instruction(struct nv30_fpc *fpc,
                        break;
                case TGSI_FILE_CONSTANT:
                case TGSI_FILE_IMMEDIATE:
-                       if (ci == -1 || ci == fsrc->SrcRegister.Index) {
-                               ci = fsrc->SrcRegister.Index;
+                       if (ci == -1 || ci == fsrc->Register.Index) {
+                               ci = fsrc->Register.Index;
                                src[i] = tgsi_src(fpc, fsrc);
                        } else {
                                src[i] = temp(fpc);
@@ -413,7 +413,7 @@ nv30_fragprog_parse_instruction(struct nv30_fpc *fpc,
                        /* handled above */
                        break;
                case TGSI_FILE_SAMPLER:
-                       unit = fsrc->SrcRegister.Index;
+                       unit = fsrc->Register.Index;
                        break;
                case TGSI_FILE_OUTPUT:
                        break;
@@ -423,8 +423,8 @@ nv30_fragprog_parse_instruction(struct nv30_fpc *fpc,
                }
        }
 
-       dst  = tgsi_dst(fpc, &finst->FullDstRegisters[0]);
-       mask = tgsi_mask(finst->FullDstRegisters[0].DstRegister.WriteMask);
+       dst  = tgsi_dst(fpc, &finst->Dst[0]);
+       mask = tgsi_mask(finst->Dst[0].Register.WriteMask);
        sat  = (finst->Instruction.Saturate == TGSI_SAT_ZERO_ONE);
 
        switch (finst->Instruction.Opcode) {
@@ -572,15 +572,15 @@ nv30_fragprog_parse_decl_attrib(struct nv30_fpc *fpc,
 {
        int hw;
 
-       switch (fdec->Semantic.SemanticName) {
+       switch (fdec->Semantic.Name) {
        case TGSI_SEMANTIC_POSITION:
                hw = NV30_FP_OP_INPUT_SRC_POSITION;
                break;
        case TGSI_SEMANTIC_COLOR:
-               if (fdec->Semantic.SemanticIndex == 0) {
+               if (fdec->Semantic.Index == 0) {
                        hw = NV30_FP_OP_INPUT_SRC_COL0;
                } else
-               if (fdec->Semantic.SemanticIndex == 1) {
+               if (fdec->Semantic.Index == 1) {
                        hw = NV30_FP_OP_INPUT_SRC_COL1;
                } else {
                        NOUVEAU_ERR("bad colour semantic index\n");
@@ -591,9 +591,9 @@ nv30_fragprog_parse_decl_attrib(struct nv30_fpc *fpc,
                hw = NV30_FP_OP_INPUT_SRC_FOGC;
                break;
        case TGSI_SEMANTIC_GENERIC:
-               if (fdec->Semantic.SemanticIndex <= 7) {
+               if (fdec->Semantic.Index <= 7) {
                        hw = NV30_FP_OP_INPUT_SRC_TC(fdec->Semantic.
-                                                    SemanticIndex);
+                                                    Index);
                } else {
                        NOUVEAU_ERR("bad generic semantic index\n");
                        return FALSE;
@@ -604,7 +604,7 @@ nv30_fragprog_parse_decl_attrib(struct nv30_fpc *fpc,
                return FALSE;
        }
 
-       fpc->attrib_map[fdec->DeclarationRange.First] = hw;
+       fpc->attrib_map[fdec->Range.First] = hw;
        return TRUE;
 }
 
@@ -612,12 +612,12 @@ static boolean
 nv30_fragprog_parse_decl_output(struct nv30_fpc *fpc,
                                const struct tgsi_full_declaration *fdec)
 {
-       switch (fdec->Semantic.SemanticName) {
+       switch (fdec->Semantic.Name) {
        case TGSI_SEMANTIC_POSITION:
-               fpc->depth_id = fdec->DeclarationRange.First;
+               fpc->depth_id = fdec->Range.First;
                break;
        case TGSI_SEMANTIC_COLOR:
-               fpc->colour_id = fdec->DeclarationRange.First;
+               fpc->colour_id = fdec->Range.First;
                break;
        default:
                NOUVEAU_ERR("bad output semantic\n");
@@ -653,9 +653,9 @@ nv30_fragprog_prepare(struct nv30_fpc *fpc)
                                        goto out_err;
                                break;
                        /*case TGSI_FILE_TEMPORARY:
-                               if (fdec->DeclarationRange.Last > high_temp) {
+                               if (fdec->Range.Last > high_temp) {
                                        high_temp =
-                                               fdec->DeclarationRange.Last;
+                                               fdec->Range.Last;
                                }
                                break;*/
                        default:
index dca760cae62c40c121d88fa87f4d485895be8b72..b3293ee700dc7ab023bb1939d699115b0557cbc1 100644 (file)
@@ -74,9 +74,9 @@ nv30_fragtex_build(struct nv30_context *nv30, int unit)
 
        txf  = tf->format;
        txf |= ((pt->last_level>0) ? NV34TCL_TX_FORMAT_MIPMAP : 0);
-       txf |= log2i(pt->width[0]) << NV34TCL_TX_FORMAT_BASE_SIZE_U_SHIFT;
-       txf |= log2i(pt->height[0]) << NV34TCL_TX_FORMAT_BASE_SIZE_V_SHIFT;
-       txf |= log2i(pt->depth[0]) << NV34TCL_TX_FORMAT_BASE_SIZE_W_SHIFT;
+       txf |= log2i(pt->width0) << NV34TCL_TX_FORMAT_BASE_SIZE_U_SHIFT;
+       txf |= log2i(pt->height0) << NV34TCL_TX_FORMAT_BASE_SIZE_V_SHIFT;
+       txf |= log2i(pt->depth0) << NV34TCL_TX_FORMAT_BASE_SIZE_W_SHIFT;
        txf |= NV34TCL_TX_FORMAT_NO_BORDER | 0x10000;
 
        switch (pt->target) {
@@ -115,8 +115,8 @@ nv30_fragtex_build(struct nv30_context *nv30, int unit)
        so_data  (so, NV34TCL_TX_ENABLE_ENABLE | ps->en);
        so_data  (so, txs);
        so_data  (so, ps->filt | 0x2000 /*voodoo*/);
-       so_data  (so, (pt->width[0] << NV34TCL_TX_NPOT_SIZE_W_SHIFT) |
-                      pt->height[0]);
+       so_data  (so, (pt->width0 << NV34TCL_TX_NPOT_SIZE_W_SHIFT) |
+                      pt->height0);
        so_data  (so, ps->bcol);
 
        return so;
index 280696d45039724faa5d80bf78f067b0824c79a6..ce95d9700f60702ad7ff6a1aa377d49e64bf2c55 100644 (file)
@@ -1,6 +1,8 @@
 #include "pipe/p_state.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
 
 #include "nv30_context.h"
 
@@ -8,7 +10,7 @@ static void
 nv30_miptree_layout(struct nv30_miptree *nv30mt)
 {
        struct pipe_texture *pt = &nv30mt->base;
-       uint width = pt->width[0], height = pt->height[0], depth = pt->depth[0];
+       uint width = pt->width0;
        uint offset = 0;
        int nr_faces, l, f;
        uint wide_pitch = pt->tex_usage & (PIPE_TEXTURE_USAGE_SAMPLER |
@@ -21,29 +23,21 @@ nv30_miptree_layout(struct nv30_miptree *nv30mt)
                nr_faces = 6;
        } else
        if (pt->target == PIPE_TEXTURE_3D) {
-               nr_faces = pt->depth[0];
+               nr_faces = pt->depth0;
        } else {
                nr_faces = 1;
        }
 
        for (l = 0; l <= pt->last_level; l++) {
-               pt->width[l] = width;
-               pt->height[l] = height;
-               pt->depth[l] = depth;
-               pt->nblocksx[l] = pf_get_nblocksx(&pt->block, width);
-               pt->nblocksy[l] = pf_get_nblocksy(&pt->block, height);
-
                if (wide_pitch && (pt->tex_usage & NOUVEAU_TEXTURE_USAGE_LINEAR))
-                       nv30mt->level[l].pitch = align(pt->width[0] * pt->block.size, 64);
+                       nv30mt->level[l].pitch = align(util_format_get_stride(pt->format, pt->width0), 64);
                else
-                       nv30mt->level[l].pitch = pt->width[l] * pt->block.size;
+                       nv30mt->level[l].pitch = util_format_get_stride(pt->format, width);
 
                nv30mt->level[l].image_offset =
                        CALLOC(nr_faces, sizeof(unsigned));
 
-               width  = MAX2(1, width  >> 1);
-               height = MAX2(1, height >> 1);
-               depth  = MAX2(1, depth  >> 1);
+               width  = u_minify(width, 1);
        }
 
        for (f = 0; f < nr_faces; f++) {
@@ -51,14 +45,14 @@ nv30_miptree_layout(struct nv30_miptree *nv30mt)
                        nv30mt->level[l].image_offset[f] = offset;
 
                        if (!(pt->tex_usage & NOUVEAU_TEXTURE_USAGE_LINEAR) &&
-                           pt->width[l + 1] > 1 && pt->height[l + 1] > 1)
-                               offset += align(nv30mt->level[l].pitch * pt->height[l], 64);
+                           u_minify(pt->width0, l + 1) > 1 && u_minify(pt->height0, l + 1) > 1)
+                               offset += align(nv30mt->level[l].pitch * u_minify(pt->height0, l), 64);
                        else
-                               offset += nv30mt->level[l].pitch * pt->height[l];
+                               offset += nv30mt->level[l].pitch * u_minify(pt->height0, l);
                }
 
                nv30mt->level[l].image_offset[f] = offset;
-               offset += nv30mt->level[l].pitch * pt->height[l];
+               offset += nv30mt->level[l].pitch * u_minify(pt->height0, l);
        }
 
        nv30mt->total_size = offset;
@@ -79,8 +73,8 @@ nv30_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *pt)
        mt->base.screen = pscreen;
 
        /* Swizzled textures must be POT */
-       if (pt->width[0] & (pt->width[0] - 1) ||
-           pt->height[0] & (pt->height[0] - 1))
+       if (pt->width0 & (pt->width0 - 1) ||
+           pt->height0 & (pt->height0 - 1))
                mt->base.tex_usage |= NOUVEAU_TEXTURE_USAGE_LINEAR;
        else
        if (pt->tex_usage & (PIPE_TEXTURE_USAGE_PRIMARY |
@@ -122,6 +116,7 @@ nv30_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *pt)
                FREE(mt);
                return NULL;
        }
+       mt->bo = nouveau_bo(mt->buffer);
 
        return &mt->base;
 }
@@ -134,7 +129,7 @@ nv30_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
 
        /* Only supports 2D, non-mipmapped textures for the moment */
        if (pt->target != PIPE_TEXTURE_2D || pt->last_level != 0 ||
-           pt->depth[0] != 1)
+           pt->depth0 != 1)
                return NULL;
 
        mt = CALLOC_STRUCT(nv30_miptree);
@@ -151,6 +146,7 @@ nv30_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
        mt->base.tex_usage |= NOUVEAU_TEXTURE_USAGE_LINEAR;
 
        pipe_buffer_reference(&mt->buffer, pb);
+       mt->bo = nouveau_bo(mt->buffer);
        return &mt->base;
 }
 
@@ -182,8 +178,8 @@ nv30_miptree_surface_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
                return NULL;
        pipe_texture_reference(&ns->base.texture, pt);
        ns->base.format = pt->format;
-       ns->base.width = pt->width[level];
-       ns->base.height = pt->height[level];
+       ns->base.width = u_minify(pt->width0, level);
+       ns->base.height = u_minify(pt->height0, level);
        ns->base.usage = flags;
        pipe_reference_init(&ns->base.reference, 1);
        ns->base.face = face;
index b91e972c12336783abd0910564f87e011eaf7b31..e6321b480f6f55cc96d3082da5d40493404cf3a3 100644 (file)
@@ -672,16 +672,6 @@ nv30_set_vertex_elements(struct pipe_context *pipe, unsigned count,
        /*nv30->draw_dirty |= NV30_NEW_ARRAYS;*/
 }
 
-static void
-nv30_set_edgeflags(struct pipe_context *pipe, const unsigned *bitfield)
-{
-       struct nv30_context *nv30 = nv30_context(pipe);
-
-       nv30->edgeflags = bitfield;
-       nv30->dirty |= NV30_NEW_ARRAYS;
-       /*nv30->draw_dirty |= NV30_NEW_ARRAYS;*/
-}
-
 void
 nv30_init_state_functions(struct nv30_context *nv30)
 {
@@ -690,9 +680,9 @@ nv30_init_state_functions(struct nv30_context *nv30)
        nv30->pipe.delete_blend_state = nv30_blend_state_delete;
 
        nv30->pipe.create_sampler_state = nv30_sampler_state_create;
-       nv30->pipe.bind_sampler_states = nv30_sampler_state_bind;
+       nv30->pipe.bind_fragment_sampler_states = nv30_sampler_state_bind;
        nv30->pipe.delete_sampler_state = nv30_sampler_state_delete;
-       nv30->pipe.set_sampler_textures = nv30_set_sampler_texture;
+       nv30->pipe.set_fragment_sampler_textures = nv30_set_sampler_texture;
 
        nv30->pipe.create_rasterizer_state = nv30_rasterizer_state_create;
        nv30->pipe.bind_rasterizer_state = nv30_rasterizer_state_bind;
@@ -721,7 +711,6 @@ nv30_init_state_functions(struct nv30_context *nv30)
        nv30->pipe.set_scissor_state = nv30_set_scissor_state;
        nv30->pipe.set_viewport_state = nv30_set_viewport_state;
 
-       nv30->pipe.set_edgeflags = nv30_set_edgeflags;
        nv30->pipe.set_vertex_buffers = nv30_set_vertex_buffers;
        nv30->pipe.set_vertex_elements = nv30_set_vertex_elements;
 }
index e6f23bf166715c2fbcaf4e03d994212d6d74f2b9..e42e872de757257f3c2998b8e25e09b24d87a95e 100644 (file)
@@ -72,6 +72,7 @@ struct nv30_fragment_program {
 
 struct nv30_miptree {
        struct pipe_texture base;
+       struct nouveau_bo *bo;
 
        struct pipe_buffer *buffer;
        uint total_size;
index 621b8846c8ed3835c97fd809ddfbe0dcba208e26..ac52d946f021232356e8912b47041080fe19d049 100644 (file)
@@ -41,7 +41,7 @@ nv30_state_emit(struct nv30_context *nv30)
        struct nouveau_channel *chan = nv30->screen->base.channel;
        struct nv30_state *state = &nv30->state;
        struct nv30_screen *screen = nv30->screen;
-       unsigned i, samplers;
+       unsigned i;
        uint64_t states;
 
        if (nv30->pctx_id != screen->cur_pctx) {
@@ -63,6 +63,14 @@ nv30_state_emit(struct nv30_context *nv30)
        }
 
        state->dirty = 0;
+}
+
+void
+nv30_state_flush_notify(struct nouveau_channel *chan)
+{
+       struct nv30_context *nv30 = chan->user_private;
+       struct nv30_state *state = &nv30->state;
+       unsigned i, samplers;
 
        so_emit_reloc_markers(chan, state->hw[NV30_STATE_FB]);
        for (i = 0, samplers = state->fp_samplers; i < 16 && samplers; i++) {
index 98011decf7cac2c3f1e2a4cab30fe2be12f0f5d9..2255a02caedb059e5add52eb7f44413b64b073be 100644 (file)
@@ -1,7 +1,9 @@
 #include <pipe/p_state.h>
 #include <pipe/p_defines.h>
 #include <pipe/p_inlines.h>
+#include <util/u_format.h>
 #include <util/u_memory.h>
+#include <util/u_math.h>
 #include <nouveau/nouveau_winsys.h>
 #include "nv30_context.h"
 #include "nv30_screen.h"
@@ -10,7 +12,7 @@
 struct nv30_transfer {
        struct pipe_transfer base;
        struct pipe_surface *surface;
-       bool direct;
+       boolean direct;
 };
 
 static void
@@ -20,12 +22,9 @@ nv30_compatible_transfer_tex(struct pipe_texture *pt, unsigned level,
        memset(template, 0, sizeof(struct pipe_texture));
        template->target = pt->target;
        template->format = pt->format;
-       template->width[0] = pt->width[level];
-       template->height[0] = pt->height[level];
-       template->depth[0] = 1;
-       template->block = pt->block;
-       template->nblocksx[0] = pt->nblocksx[level];
-       template->nblocksy[0] = pt->nblocksx[level];
+       template->width0 = u_minify(pt->width0, level);
+       template->height0 = u_minify(pt->height0, level);
+       template->depth0 = 1;
        template->last_level = 0;
        template->nr_samples = pt->nr_samples;
 
@@ -48,14 +47,10 @@ nv30_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
                return NULL;
 
        pipe_texture_reference(&tx->base.texture, pt);
-       tx->base.format = pt->format;
        tx->base.x = x;
        tx->base.y = y;
        tx->base.width = w;
        tx->base.height = h;
-       tx->base.block = pt->block;
-       tx->base.nblocksx = pt->nblocksx[level];
-       tx->base.nblocksy = pt->nblocksy[level];
        tx->base.stride = mt->level[level].pitch;
        tx->base.usage = usage;
        tx->base.face = face;
@@ -157,7 +152,7 @@ nv30_transfer_map(struct pipe_screen *pscreen, struct pipe_transfer *ptx)
                                    pipe_transfer_buffer_flags(ptx));
 
        return map + ns->base.offset +
-              ptx->y * ns->pitch + ptx->x * ptx->block.size;
+              ptx->y * ns->pitch + ptx->x * util_format_get_blocksize(ptx->texture->format);
 }
 
 static void
index 189656ec817aadb637c83b47fe31614400ab3d45..e32b8141af85022abd998a1f89dbb580203684c9 100644 (file)
@@ -485,11 +485,6 @@ nv30_vbo_validate(struct nv30_context *nv30)
        unsigned vb_flags = NOUVEAU_BO_VRAM | NOUVEAU_BO_GART | NOUVEAU_BO_RD;
        int hw;
 
-       if (nv30->edgeflags) {
-               /*nv30->fallback_swtnl |= NV30_NEW_ARRAYS;*/
-               return FALSE;
-       }
-
        vtxbuf = so_new(20, 18);
        so_method(vtxbuf, rankine, NV34TCL_VTXBUF_ADDRESS(0), nv30->vtxelt_nr);
        vtxfmt = so_new(17, 0);
index 14a5c0260d073a6c139a0ac537bbcdcc2f4beeb7..5d6098462295f3b5cb52f3b75c8b895cb18c2fed 100644 (file)
@@ -253,32 +253,32 @@ static INLINE struct nv30_sreg
 tgsi_src(struct nv30_vpc *vpc, const struct tgsi_full_src_register *fsrc) {
        struct nv30_sreg src;
 
-       switch (fsrc->SrcRegister.File) {
+       switch (fsrc->Register.File) {
        case TGSI_FILE_INPUT:
-               src = nv30_sr(NV30SR_INPUT, fsrc->SrcRegister.Index);
+               src = nv30_sr(NV30SR_INPUT, fsrc->Register.Index);
                break;
        case TGSI_FILE_CONSTANT:
-               src = constant(vpc, fsrc->SrcRegister.Index, 0, 0, 0, 0);
+               src = constant(vpc, fsrc->Register.Index, 0, 0, 0, 0);
                break;
        case TGSI_FILE_IMMEDIATE:
-               src = vpc->imm[fsrc->SrcRegister.Index];
+               src = vpc->imm[fsrc->Register.Index];
                break;
        case TGSI_FILE_TEMPORARY:
-               if (vpc->high_temp < fsrc->SrcRegister.Index)
-                       vpc->high_temp = fsrc->SrcRegister.Index;
-               src = nv30_sr(NV30SR_TEMP, fsrc->SrcRegister.Index);
+               if (vpc->high_temp < fsrc->Register.Index)
+                       vpc->high_temp = fsrc->Register.Index;
+               src = nv30_sr(NV30SR_TEMP, fsrc->Register.Index);
                break;
        default:
                NOUVEAU_ERR("bad src file\n");
                break;
        }
 
-       src.abs = fsrc->SrcRegisterExtMod.Absolute;
-       src.negate = fsrc->SrcRegister.Negate;
-       src.swz[0] = fsrc->SrcRegister.SwizzleX;
-       src.swz[1] = fsrc->SrcRegister.SwizzleY;
-       src.swz[2] = fsrc->SrcRegister.SwizzleZ;
-       src.swz[3] = fsrc->SrcRegister.SwizzleW;
+       src.abs = fsrc->Register.Absolute;
+       src.negate = fsrc->Register.Negate;
+       src.swz[0] = fsrc->Register.SwizzleX;
+       src.swz[1] = fsrc->Register.SwizzleY;
+       src.swz[2] = fsrc->Register.SwizzleZ;
+       src.swz[3] = fsrc->Register.SwizzleW;
        return src;
 }
 
@@ -286,14 +286,14 @@ static INLINE struct nv30_sreg
 tgsi_dst(struct nv30_vpc *vpc, const struct tgsi_full_dst_register *fdst) {
        struct nv30_sreg dst;
 
-       switch (fdst->DstRegister.File) {
+       switch (fdst->Register.File) {
        case TGSI_FILE_OUTPUT:
                dst = nv30_sr(NV30SR_OUTPUT,
-                             vpc->output_map[fdst->DstRegister.Index]);
+                             vpc->output_map[fdst->Register.Index]);
 
                break;
        case TGSI_FILE_TEMPORARY:
-               dst = nv30_sr(NV30SR_TEMP, fdst->DstRegister.Index);
+               dst = nv30_sr(NV30SR_TEMP, fdst->Register.Index);
                if (vpc->high_temp < dst.index)
                        vpc->high_temp = dst.index;
                break;
@@ -334,8 +334,8 @@ nv30_vertprog_parse_instruction(struct nv30_vpc *vpc,
        for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
                const struct tgsi_full_src_register *fsrc;
 
-               fsrc = &finst->FullSrcRegisters[i];
-               if (fsrc->SrcRegister.File == TGSI_FILE_TEMPORARY) {
+               fsrc = &finst->Src[i];
+               if (fsrc->Register.File == TGSI_FILE_TEMPORARY) {
                        src[i] = tgsi_src(vpc, fsrc);
                }
        }
@@ -343,11 +343,11 @@ nv30_vertprog_parse_instruction(struct nv30_vpc *vpc,
        for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
                const struct tgsi_full_src_register *fsrc;
 
-               fsrc = &finst->FullSrcRegisters[i];
-               switch (fsrc->SrcRegister.File) {
+               fsrc = &finst->Src[i];
+               switch (fsrc->Register.File) {
                case TGSI_FILE_INPUT:
-                       if (ai == -1 || ai == fsrc->SrcRegister.Index) {
-                               ai = fsrc->SrcRegister.Index;
+                       if (ai == -1 || ai == fsrc->Register.Index) {
+                               ai = fsrc->Register.Index;
                                src[i] = tgsi_src(vpc, fsrc);
                        } else {
                                src[i] = temp(vpc);
@@ -360,8 +360,8 @@ nv30_vertprog_parse_instruction(struct nv30_vpc *vpc,
                 */
                case TGSI_FILE_CONSTANT:
                case TGSI_FILE_IMMEDIATE:
-                       if (ci == -1 || ci == fsrc->SrcRegister.Index) {
-                               ci = fsrc->SrcRegister.Index;
+                       if (ci == -1 || ci == fsrc->Register.Index) {
+                               ci = fsrc->Register.Index;
                                src[i] = tgsi_src(vpc, fsrc);
                        } else {
                                src[i] = temp(vpc);
@@ -378,8 +378,8 @@ nv30_vertprog_parse_instruction(struct nv30_vpc *vpc,
                }
        }
 
-       dst  = tgsi_dst(vpc, &finst->FullDstRegisters[0]);
-       mask = tgsi_mask(finst->FullDstRegisters[0].DstRegister.WriteMask);
+       dst  = tgsi_dst(vpc, &finst->Dst[0]);
+       mask = tgsi_mask(finst->Dst[0].Register.WriteMask);
 
        switch (finst->Instruction.Opcode) {
        case TGSI_OPCODE_ABS:
@@ -490,15 +490,15 @@ nv30_vertprog_parse_decl_output(struct nv30_vpc *vpc,
 {
        int hw;
 
-       switch (fdec->Semantic.SemanticName) {
+       switch (fdec->Semantic.Name) {
        case TGSI_SEMANTIC_POSITION:
                hw = NV30_VP_INST_DEST_POS;
                break;
        case TGSI_SEMANTIC_COLOR:
-               if (fdec->Semantic.SemanticIndex == 0) {
+               if (fdec->Semantic.Index == 0) {
                        hw = NV30_VP_INST_DEST_COL0;
                } else
-               if (fdec->Semantic.SemanticIndex == 1) {
+               if (fdec->Semantic.Index == 1) {
                        hw = NV30_VP_INST_DEST_COL1;
                } else {
                        NOUVEAU_ERR("bad colour semantic index\n");
@@ -506,10 +506,10 @@ nv30_vertprog_parse_decl_output(struct nv30_vpc *vpc,
                }
                break;
        case TGSI_SEMANTIC_BCOLOR:
-               if (fdec->Semantic.SemanticIndex == 0) {
+               if (fdec->Semantic.Index == 0) {
                        hw = NV30_VP_INST_DEST_BFC0;
                } else
-               if (fdec->Semantic.SemanticIndex == 1) {
+               if (fdec->Semantic.Index == 1) {
                        hw = NV30_VP_INST_DEST_BFC1;
                } else {
                        NOUVEAU_ERR("bad bcolour semantic index\n");
@@ -523,19 +523,22 @@ nv30_vertprog_parse_decl_output(struct nv30_vpc *vpc,
                hw = NV30_VP_INST_DEST_PSZ;
                break;
        case TGSI_SEMANTIC_GENERIC:
-               if (fdec->Semantic.SemanticIndex <= 7) {
-                       hw = NV30_VP_INST_DEST_TC(fdec->Semantic.SemanticIndex);
+               if (fdec->Semantic.Index <= 7) {
+                       hw = NV30_VP_INST_DEST_TC(fdec->Semantic.Index);
                } else {
                        NOUVEAU_ERR("bad generic semantic index\n");
                        return FALSE;
                }
                break;
+       case TGSI_SEMANTIC_EDGEFLAG:
+               NOUVEAU_ERR("cannot handle edgeflag output\n");
+               return FALSE;
        default:
                NOUVEAU_ERR("bad output semantic\n");
                return FALSE;
        }
 
-       vpc->output_map[fdec->DeclarationRange.First] = hw;
+       vpc->output_map[fdec->Range.First] = hw;
        return TRUE;
 }
 
index 7f008274a4e301d5621234853afbfb036e51bb3e..eb9cce4c786718b45708de45a47eb83e2d8538a3 100644 (file)
@@ -58,6 +58,9 @@ nv40_create(struct pipe_screen *pscreen, unsigned pctx_id)
        nv40->pipe.is_texture_referenced = nouveau_is_texture_referenced;
        nv40->pipe.is_buffer_referenced = nouveau_is_buffer_referenced;
 
+       screen->base.channel->user_private = nv40;
+       screen->base.channel->flush_notify = nv40_state_flush_notify;
+
        nv40_init_query_functions(nv40);
        nv40_init_surface_functions(nv40);
        nv40_init_state_functions(nv40);
index a3d594167aad1c668c93ff74a85996220397a658..83fcf1785d9f0187655d415b8a3914701df00f3f 100644 (file)
@@ -159,7 +159,6 @@ struct nv40_context {
        unsigned vtxbuf_nr;
        struct pipe_vertex_element vtxelt[PIPE_MAX_ATTRIBS];
        unsigned vtxelt_nr;
-       const unsigned *edgeflags;
 };
 
 static INLINE struct nv40_context *
@@ -204,6 +203,7 @@ extern void nv40_fragtex_bind(struct nv40_context *);
 extern boolean nv40_state_validate(struct nv40_context *nv40);
 extern boolean nv40_state_validate_swtnl(struct nv40_context *nv40);
 extern void nv40_state_emit(struct nv40_context *nv40);
+extern void nv40_state_flush_notify(struct nouveau_channel *chan);
 extern struct nv40_state_entry nv40_state_rasterizer;
 extern struct nv40_state_entry nv40_state_scissor;
 extern struct nv40_state_entry nv40_state_stipple;
index b2f19ecb699826bccd729640a6834b7dca64cf6e..3875bc35457359ae105ff80f78cdd543b4be7614 100644 (file)
@@ -261,7 +261,8 @@ nv40_draw_elements_swtnl(struct pipe_context *pipe,
                map = pipe_buffer_map(pscreen,
                                      nv40->constbuf[PIPE_SHADER_VERTEX],
                                      PIPE_BUFFER_USAGE_CPU_READ);
-               draw_set_mapped_constant_buffer(nv40->draw, map, nr);
+               draw_set_mapped_constant_buffer(nv40->draw, PIPE_SHADER_VERTEX,
+                                                map, nr);
        }
 
        draw_arrays(nv40->draw, mode, start, count);
@@ -285,7 +286,7 @@ static INLINE void
 emit_attrib(struct nv40_context *nv40, unsigned hw, unsigned emit,
            unsigned semantic, unsigned index)
 {
-       unsigned draw_out = draw_find_vs_output(nv40->draw, semantic, index);
+       unsigned draw_out = draw_find_shader_output(nv40->draw, semantic, index);
        unsigned a = nv40->swtnl.nr_attribs++;
 
        nv40->swtnl.hw[a] = hw;
index 99277506fc207dddcfa6e89279269d35be76a6e8..1bf16726d105e1314a20b84290a94bf58e8d9aab 100644 (file)
@@ -255,50 +255,50 @@ tgsi_src(struct nv40_fpc *fpc, const struct tgsi_full_src_register *fsrc)
 {
        struct nv40_sreg src;
 
-       switch (fsrc->SrcRegister.File) {
+       switch (fsrc->Register.File) {
        case TGSI_FILE_INPUT:
                src = nv40_sr(NV40SR_INPUT,
-                             fpc->attrib_map[fsrc->SrcRegister.Index]);
+                             fpc->attrib_map[fsrc->Register.Index]);
                break;
        case TGSI_FILE_CONSTANT:
-               src = constant(fpc, fsrc->SrcRegister.Index, NULL);
+               src = constant(fpc, fsrc->Register.Index, NULL);
                break;
        case TGSI_FILE_IMMEDIATE:
-               assert(fsrc->SrcRegister.Index < fpc->nr_imm);
-               src = fpc->imm[fsrc->SrcRegister.Index];
+               assert(fsrc->Register.Index < fpc->nr_imm);
+               src = fpc->imm[fsrc->Register.Index];
                break;
        case TGSI_FILE_TEMPORARY:
-               src = fpc->r_temp[fsrc->SrcRegister.Index];
+               src = fpc->r_temp[fsrc->Register.Index];
                break;
        /* NV40 fragprog result regs are just temps, so this is simple */
        case TGSI_FILE_OUTPUT:
-               src = fpc->r_result[fsrc->SrcRegister.Index];
+               src = fpc->r_result[fsrc->Register.Index];
                break;
        default:
                NOUVEAU_ERR("bad src file\n");
                break;
        }
 
-       src.abs = fsrc->SrcRegisterExtMod.Absolute;
-       src.negate = fsrc->SrcRegister.Negate;
-       src.swz[0] = fsrc->SrcRegister.SwizzleX;
-       src.swz[1] = fsrc->SrcRegister.SwizzleY;
-       src.swz[2] = fsrc->SrcRegister.SwizzleZ;
-       src.swz[3] = fsrc->SrcRegister.SwizzleW;
+       src.abs = fsrc->Register.Absolute;
+       src.negate = fsrc->Register.Negate;
+       src.swz[0] = fsrc->Register.SwizzleX;
+       src.swz[1] = fsrc->Register.SwizzleY;
+       src.swz[2] = fsrc->Register.SwizzleZ;
+       src.swz[3] = fsrc->Register.SwizzleW;
        return src;
 }
 
 static INLINE struct nv40_sreg
 tgsi_dst(struct nv40_fpc *fpc, const struct tgsi_full_dst_register *fdst) {
-       switch (fdst->DstRegister.File) {
+       switch (fdst->Register.File) {
        case TGSI_FILE_OUTPUT:
-               return fpc->r_result[fdst->DstRegister.Index];
+               return fpc->r_result[fdst->Register.Index];
        case TGSI_FILE_TEMPORARY:
-               return fpc->r_temp[fdst->DstRegister.Index];
+               return fpc->r_temp[fdst->Register.Index];
        case TGSI_FILE_NULL:
                return nv40_sr(NV40SR_NONE, 0);
        default:
-               NOUVEAU_ERR("bad dst file %d\n", fdst->DstRegister.File);
+               NOUVEAU_ERR("bad dst file %d\n", fdst->Register.File);
                return nv40_sr(NV40SR_NONE, 0);
        }
 }
@@ -364,8 +364,8 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
        for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
                const struct tgsi_full_src_register *fsrc;
 
-               fsrc = &finst->FullSrcRegisters[i];
-               if (fsrc->SrcRegister.File == TGSI_FILE_TEMPORARY) {
+               fsrc = &finst->Src[i];
+               if (fsrc->Register.File == TGSI_FILE_TEMPORARY) {
                        src[i] = tgsi_src(fpc, fsrc);
                }
        }
@@ -373,9 +373,9 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
        for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
                const struct tgsi_full_src_register *fsrc;
 
-               fsrc = &finst->FullSrcRegisters[i];
+               fsrc = &finst->Src[i];
 
-               switch (fsrc->SrcRegister.File) {
+               switch (fsrc->Register.File) {
                case TGSI_FILE_INPUT:
                case TGSI_FILE_CONSTANT:
                case TGSI_FILE_TEMPORARY:
@@ -386,10 +386,10 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
                        break;
                }
 
-               switch (fsrc->SrcRegister.File) {
+               switch (fsrc->Register.File) {
                case TGSI_FILE_INPUT:
-                       if (ai == -1 || ai == fsrc->SrcRegister.Index) {
-                               ai = fsrc->SrcRegister.Index;
+                       if (ai == -1 || ai == fsrc->Register.Index) {
+                               ai = fsrc->Register.Index;
                                src[i] = tgsi_src(fpc, fsrc);
                        } else {
                                src[i] = temp(fpc);
@@ -399,8 +399,8 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
                        break;
                case TGSI_FILE_CONSTANT:
                        if ((ci == -1 && ii == -1) ||
-                           ci == fsrc->SrcRegister.Index) {
-                               ci = fsrc->SrcRegister.Index;
+                           ci == fsrc->Register.Index) {
+                               ci = fsrc->Register.Index;
                                src[i] = tgsi_src(fpc, fsrc);
                        } else {
                                src[i] = temp(fpc);
@@ -410,8 +410,8 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
                        break;
                case TGSI_FILE_IMMEDIATE:
                        if ((ci == -1 && ii == -1) ||
-                           ii == fsrc->SrcRegister.Index) {
-                               ii = fsrc->SrcRegister.Index;
+                           ii == fsrc->Register.Index) {
+                               ii = fsrc->Register.Index;
                                src[i] = tgsi_src(fpc, fsrc);
                        } else {
                                src[i] = temp(fpc);
@@ -423,7 +423,7 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
                        /* handled above */
                        break;
                case TGSI_FILE_SAMPLER:
-                       unit = fsrc->SrcRegister.Index;
+                       unit = fsrc->Register.Index;
                        break;
                case TGSI_FILE_OUTPUT:
                        break;
@@ -433,8 +433,8 @@ nv40_fragprog_parse_instruction(struct nv40_fpc *fpc,
                }
        }
 
-       dst  = tgsi_dst(fpc, &finst->FullDstRegisters[0]);
-       mask = tgsi_mask(finst->FullDstRegisters[0].DstRegister.WriteMask);
+       dst  = tgsi_dst(fpc, &finst->Dst[0]);
+       mask = tgsi_mask(finst->Dst[0].Register.WriteMask);
        sat  = (finst->Instruction.Saturate == TGSI_SAT_ZERO_ONE);
 
        switch (finst->Instruction.Opcode) {
@@ -644,15 +644,15 @@ nv40_fragprog_parse_decl_attrib(struct nv40_fpc *fpc,
 {
        int hw;
 
-       switch (fdec->Semantic.SemanticName) {
+       switch (fdec->Semantic.Name) {
        case TGSI_SEMANTIC_POSITION:
                hw = NV40_FP_OP_INPUT_SRC_POSITION;
                break;
        case TGSI_SEMANTIC_COLOR:
-               if (fdec->Semantic.SemanticIndex == 0) {
+               if (fdec->Semantic.Index == 0) {
                        hw = NV40_FP_OP_INPUT_SRC_COL0;
                } else
-               if (fdec->Semantic.SemanticIndex == 1) {
+               if (fdec->Semantic.Index == 1) {
                        hw = NV40_FP_OP_INPUT_SRC_COL1;
                } else {
                        NOUVEAU_ERR("bad colour semantic index\n");
@@ -663,9 +663,9 @@ nv40_fragprog_parse_decl_attrib(struct nv40_fpc *fpc,
                hw = NV40_FP_OP_INPUT_SRC_FOGC;
                break;
        case TGSI_SEMANTIC_GENERIC:
-               if (fdec->Semantic.SemanticIndex <= 7) {
+               if (fdec->Semantic.Index <= 7) {
                        hw = NV40_FP_OP_INPUT_SRC_TC(fdec->Semantic.
-                                                    SemanticIndex);
+                                                    Index);
                } else {
                        NOUVEAU_ERR("bad generic semantic index\n");
                        return FALSE;
@@ -676,7 +676,7 @@ nv40_fragprog_parse_decl_attrib(struct nv40_fpc *fpc,
                return FALSE;
        }
 
-       fpc->attrib_map[fdec->DeclarationRange.First] = hw;
+       fpc->attrib_map[fdec->Range.First] = hw;
        return TRUE;
 }
 
@@ -684,15 +684,15 @@ static boolean
 nv40_fragprog_parse_decl_output(struct nv40_fpc *fpc,
                                const struct tgsi_full_declaration *fdec)
 {
-       unsigned idx = fdec->DeclarationRange.First;
+       unsigned idx = fdec->Range.First;
        unsigned hw;
 
-       switch (fdec->Semantic.SemanticName) {
+       switch (fdec->Semantic.Name) {
        case TGSI_SEMANTIC_POSITION:
                hw = 1;
                break;
        case TGSI_SEMANTIC_COLOR:
-               switch (fdec->Semantic.SemanticIndex) {
+               switch (fdec->Semantic.Index) {
                case 0: hw = 0; break;
                case 1: hw = 2; break;
                case 2: hw = 3; break;
@@ -738,9 +738,9 @@ nv40_fragprog_prepare(struct nv40_fpc *fpc)
                                        goto out_err;
                                break;
                        case TGSI_FILE_TEMPORARY:
-                               if (fdec->DeclarationRange.Last > high_temp) {
+                               if (fdec->Range.Last > high_temp) {
                                        high_temp =
-                                               fdec->DeclarationRange.Last;
+                                               fdec->Range.Last;
                                }
                                break;
                        default:
index e2ec57564d14a142c6b3692f7655a4f6f9178bd8..44abc845969f4b1fe208a27abd0420408b04896a 100644 (file)
@@ -117,11 +117,11 @@ nv40_fragtex_build(struct nv40_context *nv40, int unit)
        so_data  (so, NV40TCL_TEX_ENABLE_ENABLE | ps->en);
        so_data  (so, txs);
        so_data  (so, ps->filt | tf->sign | 0x2000 /*voodoo*/);
-       so_data  (so, (pt->width[0] << NV40TCL_TEX_SIZE0_W_SHIFT) |
-                      pt->height[0]);
+       so_data  (so, (pt->width0 << NV40TCL_TEX_SIZE0_W_SHIFT) |
+                      pt->height0);
        so_data  (so, ps->bcol);
        so_method(so, nv40->screen->curie, NV40TCL_TEX_SIZE1(unit), 1);
-       so_data  (so, (pt->depth[0] << NV40TCL_TEX_SIZE1_DEPTH_SHIFT) | txp);
+       so_data  (so, (pt->depth0 << NV40TCL_TEX_SIZE1_DEPTH_SHIFT) | txp);
 
        return so;
 }
index 465dd3b069399e8b1e0a2d31842bbd5b4d57ce38..b974e68a077643c6857da233fea18dd0a7c4cdee 100644 (file)
@@ -1,14 +1,18 @@
 #include "pipe/p_state.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
 
 #include "nv40_context.h"
 
+
+
 static void
 nv40_miptree_layout(struct nv40_miptree *mt)
 {
        struct pipe_texture *pt = &mt->base;
-       uint width = pt->width[0], height = pt->height[0], depth = pt->depth[0];
+       uint width = pt->width0;
        uint offset = 0;
        int nr_faces, l, f;
        uint wide_pitch = pt->tex_usage & (PIPE_TEXTURE_USAGE_SAMPLER |
@@ -21,29 +25,21 @@ nv40_miptree_layout(struct nv40_miptree *mt)
                nr_faces = 6;
        } else
        if (pt->target == PIPE_TEXTURE_3D) {
-               nr_faces = pt->depth[0];
+               nr_faces = pt->depth0;
        } else {
                nr_faces = 1;
        }
 
        for (l = 0; l <= pt->last_level; l++) {
-               pt->width[l] = width;
-               pt->height[l] = height;
-               pt->depth[l] = depth;
-               pt->nblocksx[l] = pf_get_nblocksx(&pt->block, width);
-               pt->nblocksy[l] = pf_get_nblocksy(&pt->block, height);
-
                if (wide_pitch && (pt->tex_usage & NOUVEAU_TEXTURE_USAGE_LINEAR))
-                       mt->level[l].pitch = align(pt->width[0] * pt->block.size, 64);
+                       mt->level[l].pitch = align(util_format_get_stride(pt->format, pt->width0), 64);
                else
-                       mt->level[l].pitch = pt->width[l] * pt->block.size;
+                       mt->level[l].pitch = util_format_get_stride(pt->format, width);
 
                mt->level[l].image_offset =
                        CALLOC(nr_faces, sizeof(unsigned));
 
-               width  = MAX2(1, width  >> 1);
-               height = MAX2(1, height >> 1);
-               depth  = MAX2(1, depth  >> 1);
+               width  = u_minify(width, 1);
        }
 
        for (f = 0; f < nr_faces; f++) {
@@ -51,14 +47,14 @@ nv40_miptree_layout(struct nv40_miptree *mt)
                        mt->level[l].image_offset[f] = offset;
 
                        if (!(pt->tex_usage & NOUVEAU_TEXTURE_USAGE_LINEAR) &&
-                           pt->width[l + 1] > 1 && pt->height[l + 1] > 1)
-                               offset += align(mt->level[l].pitch * pt->height[l], 64);
+                           u_minify(pt->width0, l + 1) > 1 && u_minify(pt->height0, l + 1) > 1)
+                               offset += align(mt->level[l].pitch * u_minify(pt->height0, l), 64);
                        else
-                               offset += mt->level[l].pitch * pt->height[l];
+                               offset += mt->level[l].pitch * u_minify(pt->height0, l);
                }
 
                mt->level[l].image_offset[f] = offset;
-               offset += mt->level[l].pitch * pt->height[l];
+               offset += mt->level[l].pitch * u_minify(pt->height0, l);
        }
 
        mt->total_size = offset;
@@ -79,8 +75,8 @@ nv40_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *pt)
        mt->base.screen = pscreen;
 
        /* Swizzled textures must be POT */
-       if (pt->width[0] & (pt->width[0] - 1) ||
-           pt->height[0] & (pt->height[0] - 1))
+       if (pt->width0 & (pt->width0 - 1) ||
+           pt->height0 & (pt->height0 - 1))
                mt->base.tex_usage |= NOUVEAU_TEXTURE_USAGE_LINEAR;
        else
        if (pt->tex_usage & (PIPE_TEXTURE_USAGE_PRIMARY |
@@ -116,7 +112,7 @@ nv40_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *pt)
                FREE(mt);
                return NULL;
        }
-
+       mt->bo = nouveau_bo(mt->buffer);
        return &mt->base;
 }
 
@@ -128,7 +124,7 @@ nv40_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
 
        /* Only supports 2D, non-mipmapped textures for the moment */
        if (pt->target != PIPE_TEXTURE_2D || pt->last_level != 0 ||
-           pt->depth[0] != 1)
+           pt->depth0 != 1)
                return NULL;
 
        mt = CALLOC_STRUCT(nv40_miptree);
@@ -145,6 +141,7 @@ nv40_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
        mt->base.tex_usage |= NOUVEAU_TEXTURE_USAGE_LINEAR;
 
        pipe_buffer_reference(&mt->buffer, pb);
+       mt->bo = nouveau_bo(mt->buffer);
        return &mt->base;
 }
 
@@ -176,8 +173,8 @@ nv40_miptree_surface_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
                return NULL;
        pipe_texture_reference(&ns->base.texture, pt);
        ns->base.format = pt->format;
-       ns->base.width = pt->width[level];
-       ns->base.height = pt->height[level];
+       ns->base.width = u_minify(pt->width0, level);
+       ns->base.height = u_minify(pt->height0, level);
        ns->base.usage = flags;
        pipe_reference_init(&ns->base.reference, 1);
        ns->base.face = face;
index c3ee4d234537d1dc317649dff573316899c46f94..ed55d29affde8565541fcb077400424a98a2418f 100644 (file)
@@ -687,16 +687,6 @@ nv40_set_vertex_elements(struct pipe_context *pipe, unsigned count,
        nv40->draw_dirty |= NV40_NEW_ARRAYS;
 }
 
-static void
-nv40_set_edgeflags(struct pipe_context *pipe, const unsigned *bitfield)
-{
-       struct nv40_context *nv40 = nv40_context(pipe);
-
-       nv40->edgeflags = bitfield;
-       nv40->dirty |= NV40_NEW_ARRAYS;
-       nv40->draw_dirty |= NV40_NEW_ARRAYS;
-}
-
 void
 nv40_init_state_functions(struct nv40_context *nv40)
 {
@@ -705,9 +695,9 @@ nv40_init_state_functions(struct nv40_context *nv40)
        nv40->pipe.delete_blend_state = nv40_blend_state_delete;
 
        nv40->pipe.create_sampler_state = nv40_sampler_state_create;
-       nv40->pipe.bind_sampler_states = nv40_sampler_state_bind;
+       nv40->pipe.bind_fragment_sampler_states = nv40_sampler_state_bind;
        nv40->pipe.delete_sampler_state = nv40_sampler_state_delete;
-       nv40->pipe.set_sampler_textures = nv40_set_sampler_texture;
+       nv40->pipe.set_fragment_sampler_textures = nv40_set_sampler_texture;
 
        nv40->pipe.create_rasterizer_state = nv40_rasterizer_state_create;
        nv40->pipe.bind_rasterizer_state = nv40_rasterizer_state_bind;
@@ -736,7 +726,6 @@ nv40_init_state_functions(struct nv40_context *nv40)
        nv40->pipe.set_scissor_state = nv40_set_scissor_state;
        nv40->pipe.set_viewport_state = nv40_set_viewport_state;
 
-       nv40->pipe.set_edgeflags = nv40_set_edgeflags;
        nv40->pipe.set_vertex_buffers = nv40_set_vertex_buffers;
        nv40->pipe.set_vertex_elements = nv40_set_vertex_elements;
 }
index 8a9d8c8fdf608bc409f3b157f53a2edc38ba9a18..192074e7471c8fe382e8b5f1f5f037870628cd5a 100644 (file)
@@ -75,6 +75,7 @@ struct nv40_fragment_program {
 
 struct nv40_miptree {
        struct pipe_texture base;
+       struct nouveau_bo *bo;
 
        struct pipe_buffer *buffer;
        uint total_size;
index 198692965dc4ad3d482c2826161af03c04457524..789ed16126a330499100a580bcdfe57284ef5611 100644 (file)
@@ -57,7 +57,7 @@ nv40_state_emit(struct nv40_context *nv40)
        struct nouveau_channel *chan = nv40->screen->base.channel;
        struct nv40_state *state = &nv40->state;
        struct nv40_screen *screen = nv40->screen;
-       unsigned i, samplers;
+       unsigned i;
        uint64_t states;
 
        if (nv40->pctx_id != screen->cur_pctx) {
@@ -87,6 +87,14 @@ nv40_state_emit(struct nv40_context *nv40)
        }
 
        state->dirty = 0;
+}
+
+void
+nv40_state_flush_notify(struct nouveau_channel *chan)
+{
+       struct nv40_context *nv40 = chan->user_private;
+       struct nv40_state *state = &nv40->state;
+       unsigned i, samplers;
 
        so_emit_reloc_markers(chan, state->hw[NV40_STATE_FB]);
        for (i = 0, samplers = state->fp_samplers; i < 16 && samplers; i++) {
@@ -160,7 +168,6 @@ nv40_state_validate_swtnl(struct nv40_context *nv40)
                draw_set_viewport_state(draw, &nv40->viewport);
 
        if (nv40->draw_dirty & NV40_NEW_ARRAYS) {
-               draw_set_edgeflags(draw, nv40->edgeflags);
                draw_set_vertex_buffers(draw, nv40->vtxbuf_nr, nv40->vtxbuf);
                draw_set_vertex_elements(draw, nv40->vtxelt_nr, nv40->vtxelt);  
        }
index 92caee6f3825dda7324b3326501e37cee9d11504..b084a38b48273cbb7ebb42f2d0f36f8ffbffa9a8 100644 (file)
@@ -1,7 +1,9 @@
 #include <pipe/p_state.h>
 #include <pipe/p_defines.h>
 #include <pipe/p_inlines.h>
+#include <util/u_format.h>
 #include <util/u_memory.h>
+#include <util/u_math.h>
 #include <nouveau/nouveau_winsys.h>
 #include "nv40_context.h"
 #include "nv40_screen.h"
@@ -10,7 +12,7 @@
 struct nv40_transfer {
        struct pipe_transfer base;
        struct pipe_surface *surface;
-       bool direct;
+       boolean direct;
 };
 
 static void
@@ -20,12 +22,9 @@ nv40_compatible_transfer_tex(struct pipe_texture *pt, unsigned level,
        memset(template, 0, sizeof(struct pipe_texture));
        template->target = pt->target;
        template->format = pt->format;
-       template->width[0] = pt->width[level];
-       template->height[0] = pt->height[level];
-       template->depth[0] = 1;
-       template->block = pt->block;
-       template->nblocksx[0] = pt->nblocksx[level];
-       template->nblocksy[0] = pt->nblocksx[level];
+       template->width0 = u_minify(pt->width0, level);
+       template->height0 = u_minify(pt->height0, level);
+       template->depth0 = 1;
        template->last_level = 0;
        template->nr_samples = pt->nr_samples;
 
@@ -48,14 +47,10 @@ nv40_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
                return NULL;
 
        pipe_texture_reference(&tx->base.texture, pt);
-       tx->base.format = pt->format;
        tx->base.x = x;
        tx->base.y = y;
        tx->base.width = w;
        tx->base.height = h;
-       tx->base.block = pt->block;
-       tx->base.nblocksx = pt->nblocksx[level];
-       tx->base.nblocksy = pt->nblocksy[level];
        tx->base.stride = mt->level[level].pitch;
        tx->base.usage = usage;
        tx->base.face = face;
@@ -157,7 +152,7 @@ nv40_transfer_map(struct pipe_screen *pscreen, struct pipe_transfer *ptx)
                                    pipe_transfer_buffer_flags(ptx));
 
        return map + ns->base.offset +
-              ptx->y * ns->pitch + ptx->x * ptx->block.size;
+              ptx->y * ns->pitch + ptx->x * util_format_get_blocksize(ptx->texture->format);
 }
 
 static void
index b2753b8e2e052a655a7f89c28e5d3243ec5c84ed..af3fcf6a3487adc109bb7bc7e3cd8596afd03802 100644 (file)
@@ -484,11 +484,6 @@ nv40_vbo_validate(struct nv40_context *nv40)
        unsigned vb_flags = NOUVEAU_BO_VRAM | NOUVEAU_BO_GART | NOUVEAU_BO_RD;
        int hw;
 
-       if (nv40->edgeflags) {
-               nv40->fallback_swtnl |= NV40_NEW_ARRAYS;
-               return FALSE;
-       }
-
        vtxbuf = so_new(20, 18);
        so_method(vtxbuf, curie, NV40TCL_VTXBUF_ADDRESS(0), nv40->vtxelt_nr);
        vtxfmt = so_new(17, 0);
index 31dae2457fdcdf3ab52646773b7fb9f398b85641..d9fc31006f5aff7562c4204a828a3efa4e2cb07f 100644 (file)
@@ -295,30 +295,30 @@ static INLINE struct nv40_sreg
 tgsi_src(struct nv40_vpc *vpc, const struct tgsi_full_src_register *fsrc) {
        struct nv40_sreg src;
 
-       switch (fsrc->SrcRegister.File) {
+       switch (fsrc->Register.File) {
        case TGSI_FILE_INPUT:
-               src = nv40_sr(NV40SR_INPUT, fsrc->SrcRegister.Index);
+               src = nv40_sr(NV40SR_INPUT, fsrc->Register.Index);
                break;
        case TGSI_FILE_CONSTANT:
-               src = constant(vpc, fsrc->SrcRegister.Index, 0, 0, 0, 0);
+               src = constant(vpc, fsrc->Register.Index, 0, 0, 0, 0);
                break;
        case TGSI_FILE_IMMEDIATE:
-               src = vpc->imm[fsrc->SrcRegister.Index];
+               src = vpc->imm[fsrc->Register.Index];
                break;
        case TGSI_FILE_TEMPORARY:
-               src = vpc->r_temp[fsrc->SrcRegister.Index];
+               src = vpc->r_temp[fsrc->Register.Index];
                break;
        default:
                NOUVEAU_ERR("bad src file\n");
                break;
        }
 
-       src.abs = fsrc->SrcRegisterExtMod.Absolute;
-       src.negate = fsrc->SrcRegister.Negate;
-       src.swz[0] = fsrc->SrcRegister.SwizzleX;
-       src.swz[1] = fsrc->SrcRegister.SwizzleY;
-       src.swz[2] = fsrc->SrcRegister.SwizzleZ;
-       src.swz[3] = fsrc->SrcRegister.SwizzleW;
+       src.abs = fsrc->Register.Absolute;
+       src.negate = fsrc->Register.Negate;
+       src.swz[0] = fsrc->Register.SwizzleX;
+       src.swz[1] = fsrc->Register.SwizzleY;
+       src.swz[2] = fsrc->Register.SwizzleZ;
+       src.swz[3] = fsrc->Register.SwizzleW;
        return src;
 }
 
@@ -326,15 +326,15 @@ static INLINE struct nv40_sreg
 tgsi_dst(struct nv40_vpc *vpc, const struct tgsi_full_dst_register *fdst) {
        struct nv40_sreg dst;
 
-       switch (fdst->DstRegister.File) {
+       switch (fdst->Register.File) {
        case TGSI_FILE_OUTPUT:
-               dst = vpc->r_result[fdst->DstRegister.Index];
+               dst = vpc->r_result[fdst->Register.Index];
                break;
        case TGSI_FILE_TEMPORARY:
-               dst = vpc->r_temp[fdst->DstRegister.Index];
+               dst = vpc->r_temp[fdst->Register.Index];
                break;
        case TGSI_FILE_ADDRESS:
-               dst = vpc->r_address[fdst->DstRegister.Index];
+               dst = vpc->r_address[fdst->Register.Index];
                break;
        default:
                NOUVEAU_ERR("bad dst file\n");
@@ -405,8 +405,8 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
        for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
                const struct tgsi_full_src_register *fsrc;
 
-               fsrc = &finst->FullSrcRegisters[i];
-               if (fsrc->SrcRegister.File == TGSI_FILE_TEMPORARY) {
+               fsrc = &finst->Src[i];
+               if (fsrc->Register.File == TGSI_FILE_TEMPORARY) {
                        src[i] = tgsi_src(vpc, fsrc);
                }
        }
@@ -414,9 +414,9 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
        for (i = 0; i < finst->Instruction.NumSrcRegs; i++) {
                const struct tgsi_full_src_register *fsrc;
 
-               fsrc = &finst->FullSrcRegisters[i];
+               fsrc = &finst->Src[i];
 
-               switch (fsrc->SrcRegister.File) {
+               switch (fsrc->Register.File) {
                case TGSI_FILE_INPUT:
                case TGSI_FILE_CONSTANT:
                case TGSI_FILE_TEMPORARY:
@@ -427,10 +427,10 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
                        break;
                }
 
-               switch (fsrc->SrcRegister.File) {
+               switch (fsrc->Register.File) {
                case TGSI_FILE_INPUT:
-                       if (ai == -1 || ai == fsrc->SrcRegister.Index) {
-                               ai = fsrc->SrcRegister.Index;
+                       if (ai == -1 || ai == fsrc->Register.Index) {
+                               ai = fsrc->Register.Index;
                                src[i] = tgsi_src(vpc, fsrc);
                        } else {
                                src[i] = temp(vpc);
@@ -440,8 +440,8 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
                        break;
                case TGSI_FILE_CONSTANT:
                        if ((ci == -1 && ii == -1) ||
-                           ci == fsrc->SrcRegister.Index) {
-                               ci = fsrc->SrcRegister.Index;
+                           ci == fsrc->Register.Index) {
+                               ci = fsrc->Register.Index;
                                src[i] = tgsi_src(vpc, fsrc);
                        } else {
                                src[i] = temp(vpc);
@@ -451,8 +451,8 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
                        break;
                case TGSI_FILE_IMMEDIATE:
                        if ((ci == -1 && ii == -1) ||
-                           ii == fsrc->SrcRegister.Index) {
-                               ii = fsrc->SrcRegister.Index;
+                           ii == fsrc->Register.Index) {
+                               ii = fsrc->Register.Index;
                                src[i] = tgsi_src(vpc, fsrc);
                        } else {
                                src[i] = temp(vpc);
@@ -469,8 +469,8 @@ nv40_vertprog_parse_instruction(struct nv40_vpc *vpc,
                }
        }
 
-       dst  = tgsi_dst(vpc, &finst->FullDstRegisters[0]);
-       mask = tgsi_mask(finst->FullDstRegisters[0].DstRegister.WriteMask);
+       dst  = tgsi_dst(vpc, &finst->Dst[0]);
+       mask = tgsi_mask(finst->Dst[0].Register.WriteMask);
 
        switch (finst->Instruction.Opcode) {
        case TGSI_OPCODE_ABS:
@@ -577,19 +577,19 @@ static boolean
 nv40_vertprog_parse_decl_output(struct nv40_vpc *vpc,
                                const struct tgsi_full_declaration *fdec)
 {
-       unsigned idx = fdec->DeclarationRange.First;
+       unsigned idx = fdec->Range.First;
        int hw;
 
-       switch (fdec->Semantic.SemanticName) {
+       switch (fdec->Semantic.Name) {
        case TGSI_SEMANTIC_POSITION:
                hw = NV40_VP_INST_DEST_POS;
                vpc->hpos_idx = idx;
                break;
        case TGSI_SEMANTIC_COLOR:
-               if (fdec->Semantic.SemanticIndex == 0) {
+               if (fdec->Semantic.Index == 0) {
                        hw = NV40_VP_INST_DEST_COL0;
                } else
-               if (fdec->Semantic.SemanticIndex == 1) {
+               if (fdec->Semantic.Index == 1) {
                        hw = NV40_VP_INST_DEST_COL1;
                } else {
                        NOUVEAU_ERR("bad colour semantic index\n");
@@ -597,10 +597,10 @@ nv40_vertprog_parse_decl_output(struct nv40_vpc *vpc,
                }
                break;
        case TGSI_SEMANTIC_BCOLOR:
-               if (fdec->Semantic.SemanticIndex == 0) {
+               if (fdec->Semantic.Index == 0) {
                        hw = NV40_VP_INST_DEST_BFC0;
                } else
-               if (fdec->Semantic.SemanticIndex == 1) {
+               if (fdec->Semantic.Index == 1) {
                        hw = NV40_VP_INST_DEST_BFC1;
                } else {
                        NOUVEAU_ERR("bad bcolour semantic index\n");
@@ -614,13 +614,17 @@ nv40_vertprog_parse_decl_output(struct nv40_vpc *vpc,
                hw = NV40_VP_INST_DEST_PSZ;
                break;
        case TGSI_SEMANTIC_GENERIC:
-               if (fdec->Semantic.SemanticIndex <= 7) {
-                       hw = NV40_VP_INST_DEST_TC(fdec->Semantic.SemanticIndex);
+               if (fdec->Semantic.Index <= 7) {
+                       hw = NV40_VP_INST_DEST_TC(fdec->Semantic.Index);
                } else {
                        NOUVEAU_ERR("bad generic semantic index\n");
                        return FALSE;
                }
                break;
+       case TGSI_SEMANTIC_EDGEFLAG:
+               /* not really an error just a fallback */
+               NOUVEAU_ERR("cannot handle edgeflag output\n");
+               return FALSE;
        default:
                NOUVEAU_ERR("bad output semantic\n");
                return FALSE;
@@ -652,16 +656,16 @@ nv40_vertprog_prepare(struct nv40_vpc *vpc)
                        fdec = &p.FullToken.FullDeclaration;
                        switch (fdec->Declaration.File) {
                        case TGSI_FILE_TEMPORARY:
-                               if (fdec->DeclarationRange.Last > high_temp) {
+                               if (fdec->Range.Last > high_temp) {
                                        high_temp =
-                                               fdec->DeclarationRange.Last;
+                                               fdec->Range.Last;
                                }
                                break;
 #if 0 /* this would be nice.. except gallium doesn't track it */
                        case TGSI_FILE_ADDRESS:
-                               if (fdec->DeclarationRange.Last > high_addr) {
+                               if (fdec->Range.Last > high_addr) {
                                        high_addr =
-                                               fdec->DeclarationRange.Last;
+                                               fdec->Range.Last;
                                }
                                break;
 #endif
@@ -681,11 +685,11 @@ nv40_vertprog_prepare(struct nv40_vpc *vpc)
                        const struct tgsi_full_dst_register *fdst;
 
                        finst = &p.FullToken.FullInstruction;
-                       fdst = &finst->FullDstRegisters[0];
+                       fdst = &finst->Dst[0];
 
-                       if (fdst->DstRegister.File == TGSI_FILE_ADDRESS) {
-                               if (fdst->DstRegister.Index > high_addr)
-                                       high_addr = fdst->DstRegister.Index;
+                       if (fdst->Register.File == TGSI_FILE_ADDRESS) {
+                               if (fdst->Register.Index > high_addr)
+                                       high_addr = fdst->Register.Index;
                        }
                
                }
index 219e7a7862321f2ede5eeabbcbce52785b3d4b30..d21b80eab8d476ea14679673d180bc5b76d7176e 100644 (file)
@@ -48,11 +48,6 @@ nv50_destroy(struct pipe_context *pipe)
 }
 
 
-static void
-nv50_set_edgeflags(struct pipe_context *pipe, const unsigned *bitfield)
-{
-}
-
 struct pipe_context *
 nv50_create(struct pipe_screen *pscreen, unsigned pctx_id)
 {
@@ -71,7 +66,6 @@ nv50_create(struct pipe_screen *pscreen, unsigned pctx_id)
 
        nv50->pipe.destroy = nv50_destroy;
 
-       nv50->pipe.set_edgeflags = nv50_set_edgeflags;
        nv50->pipe.draw_arrays = nv50_draw_arrays;
        nv50->pipe.draw_elements = nv50_draw_elements;
        nv50->pipe.clear = nv50_clear;
index 4b0f062295338eb8125071ad0af0b1aabd33e0e5..5578a5838fb4e32254129907b7550c6c44112c71 100644 (file)
@@ -65,7 +65,7 @@ struct nv50_rasterizer_stateobj {
 };
 
 struct nv50_sampler_stateobj {
-       bool normalized;
+       boolean normalized;
        unsigned tsc[8];
 };
 
@@ -126,7 +126,7 @@ struct nv50_state {
        unsigned viewport_bypass;
        struct nouveau_stateobj *tsc_upload;
        struct nouveau_stateobj *tic_upload;
-       unsigned miptree_nr;
+       unsigned miptree_nr[PIPE_SHADER_TYPES];
        struct nouveau_stateobj *vertprog;
        struct nouveau_stateobj *fragprog;
        struct nouveau_stateobj *programs;
@@ -162,10 +162,10 @@ struct nv50_context {
        unsigned vtxbuf_nr;
        struct pipe_vertex_element vtxelt[PIPE_MAX_ATTRIBS];
        unsigned vtxelt_nr;
-       struct nv50_sampler_stateobj *sampler[PIPE_MAX_SAMPLERS];
-       unsigned sampler_nr;
-       struct nv50_miptree *miptree[PIPE_MAX_SAMPLERS];
-       unsigned miptree_nr;
+       struct nv50_sampler_stateobj *sampler[PIPE_SHADER_TYPES][PIPE_MAX_SAMPLERS];
+       unsigned sampler_nr[PIPE_SHADER_TYPES];
+       struct nv50_miptree *miptree[PIPE_SHADER_TYPES][PIPE_MAX_SAMPLERS];
+       unsigned miptree_nr[PIPE_SHADER_TYPES];
 
        uint16_t vbo_fifo;
 };
@@ -218,7 +218,7 @@ extern void nv50_state_flush_notify(struct nouveau_channel *chan);
 extern void nv50_so_init_sifc(struct nv50_context *nv50,
                              struct nouveau_stateobj *so,
                              struct nouveau_bo *bo, unsigned reloc,
-                             unsigned size);
+                             unsigned offset, unsigned size);
 
 /* nv50_tex.c */
 extern void nv50_tex_validate(struct nv50_context *);
index 9c20c5cc28283741e5f42a9377c5518d39ce79c9..3f1edf0a13970cba730556e2b3eee69f97d44239 100644 (file)
@@ -23,6 +23,7 @@
 #include "pipe/p_state.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 
 #include "nv50_context.h"
 
@@ -55,14 +56,28 @@ get_tile_mode(unsigned ny, unsigned d)
        return tile_mode | 0x10;
 }
 
+static INLINE unsigned
+get_zslice_offset(unsigned tile_mode, unsigned z, unsigned pitch, unsigned nb_h)
+{
+       unsigned tile_h = get_tile_height(tile_mode);
+       unsigned tile_d = get_tile_depth(tile_mode);
+
+       /* pitch_2d == to next slice within this volume-tile */
+       /* pitch_3d == size (in bytes) of a volume-tile */
+       unsigned pitch_2d = tile_h * 64;
+       unsigned pitch_3d = tile_d * align(nb_h, tile_h) * pitch;
+
+       return (z % tile_d) * pitch_2d + (z / tile_d) * pitch_3d;
+}
+
 static struct pipe_texture *
 nv50_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *tmp)
 {
        struct nouveau_device *dev = nouveau_screen(pscreen)->device;
        struct nv50_miptree *mt = CALLOC_STRUCT(nv50_miptree);
        struct pipe_texture *pt = &mt->base.base;
-       unsigned width = tmp->width[0], height = tmp->height[0];
-       unsigned depth = tmp->depth[0], image_alignment;
+       unsigned width = tmp->width0, height = tmp->height0;
+       unsigned depth = tmp->depth0, image_alignment;
        uint32_t tile_flags;
        int ret, i, l;
 
@@ -91,20 +106,15 @@ nv50_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *tmp)
 
        for (l = 0; l <= pt->last_level; l++) {
                struct nv50_miptree_level *lvl = &mt->level[l];
-
-               pt->width[l] = width;
-               pt->height[l] = height;
-               pt->depth[l] = depth;
-               pt->nblocksx[l] = pf_get_nblocksx(&pt->block, width);
-               pt->nblocksy[l] = pf_get_nblocksy(&pt->block, height);
+               unsigned nblocksy = util_format_get_nblocksy(pt->format, height);
 
                lvl->image_offset = CALLOC(mt->image_nr, sizeof(int));
-               lvl->pitch = align(pt->nblocksx[l] * pt->block.size, 64);
-               lvl->tile_mode = get_tile_mode(pt->nblocksy[l], depth);
+               lvl->pitch = align(util_format_get_stride(pt->format, width), 64);
+               lvl->tile_mode = get_tile_mode(nblocksy, depth);
 
-               width = MAX2(1, width >> 1);
-               height = MAX2(1, height >> 1);
-               depth = MAX2(1, depth >> 1);
+               width = u_minify(width, 1);
+               height = u_minify(height, 1);
+               depth = u_minify(depth, 1);
        }
 
        image_alignment  = get_tile_height(mt->level[0].tile_mode) * 64;
@@ -121,8 +131,8 @@ nv50_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *tmp)
                        unsigned tile_d = get_tile_depth(lvl->tile_mode);
 
                        size  = lvl->pitch;
-                       size *= align(pt->nblocksy[l], tile_h);
-                       size *= align(pt->depth[l], tile_d);
+                       size *= align(util_format_get_nblocksy(pt->format, u_minify(pt->height0, l)), tile_h);
+                       size *= align(u_minify(pt->depth0, l), tile_d);
 
                        lvl->image_offset[i] = mt->total_size;
 
@@ -135,6 +145,8 @@ nv50_miptree_create(struct pipe_screen *pscreen, const struct pipe_texture *tmp)
                                  mt->level[0].tile_mode, tile_flags,
                                  &mt->base.bo);
        if (ret) {
+               for (l = 0; l < pt->last_level; ++l)
+                       FREE(mt->level[l].image_offset);
                FREE(mt);
                return NULL;
        }
@@ -151,7 +163,7 @@ nv50_miptree_blanket(struct pipe_screen *pscreen, const struct pipe_texture *pt,
 
        /* Only supports 2D, non-mipmapped textures for the moment */
        if (pt->target != PIPE_TEXTURE_2D || pt->last_level != 0 ||
-           pt->depth[0] != 1)
+           pt->depth0 != 1)
                return NULL;
 
        mt = CALLOC_STRUCT(nv50_miptree);
@@ -174,6 +186,10 @@ static void
 nv50_miptree_destroy(struct pipe_texture *pt)
 {
        struct nv50_miptree *mt = nv50_miptree(pt);
+       unsigned l;
+
+       for (l = 0; l < pt->last_level; ++l)
+               FREE(mt->level[l].image_offset);
 
        nouveau_bo_ref(NULL, &mt->base.bo);
        FREE(mt);
@@ -187,23 +203,18 @@ nv50_miptree_surface_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
        struct nv50_miptree *mt = nv50_miptree(pt);
        struct nv50_miptree_level *lvl = &mt->level[level];
        struct pipe_surface *ps;
-       int img;
+       unsigned img = 0;
 
        if (pt->target == PIPE_TEXTURE_CUBE)
                img = face;
-       else
-       if (pt->target == PIPE_TEXTURE_3D)
-               img = zslice;
-       else
-               img = 0;
 
        ps = CALLOC_STRUCT(pipe_surface);
        if (!ps)
                return NULL;
        pipe_texture_reference(&ps->texture, pt);
        ps->format = pt->format;
-       ps->width = pt->width[level];
-       ps->height = pt->height[level];
+       ps->width = u_minify(pt->width0, level);
+       ps->height = u_minify(pt->height0, level);
        ps->usage = flags;
        pipe_reference_init(&ps->reference, 1);
        ps->face = face;
@@ -211,6 +222,12 @@ nv50_miptree_surface_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
        ps->zslice = zslice;
        ps->offset = lvl->image_offset[img];
 
+       if (pt->target == PIPE_TEXTURE_3D) {
+               unsigned nb_h = util_format_get_nblocksy(pt->format, ps->height);
+               ps->offset += get_zslice_offset(lvl->tile_mode, zslice,
+                                               lvl->pitch, nb_h);
+       }
+
        return ps;
 }
 
index bf50982dd162c3278fcddf3105e537965f47a863..a101ac095c270991d9ebdbc2979eed427a232717 100644 (file)
@@ -98,9 +98,17 @@ struct nv50_reg {
 #define NV50_MOD_ABS 2
 #define NV50_MOD_SAT 4
 
-/* arbitrary limits */
-#define MAX_IF_DEPTH 4
-#define MAX_LOOP_DEPTH 4
+/* STACK: Conditionals and loops have to use the (per warp) stack.
+ * Stack entries consist of an entry type (divergent path, join at),
+ * a mask indicating the active threads of the warp, and an address.
+ * MPs can store 12 stack entries internally, if we need more (and
+ * we probably do), we have to create a stack buffer in VRAM.
+ */
+/* impose low limits for now */
+#define NV50_MAX_COND_NESTING 4
+#define NV50_MAX_LOOP_NESTING 3
+
+#define JOIN_ON(e) e; pc->p->exec_tail->inst[1] |= 2
 
 struct nv50_pc {
        struct nv50_program *p;
@@ -119,10 +127,11 @@ struct nv50_pc {
        struct nv50_reg *param;
        int param_nr;
        struct nv50_reg *immd;
-       float *immd_buf;
+       uint32_t *immd_buf;
        int immd_nr;
        struct nv50_reg **addr;
        int addr_nr;
+       uint8_t addr_alloc; /* set bit indicates used for TGSI_FILE_ADDRESS */
 
        struct nv50_reg *temp_temp[16];
        unsigned temp_temp_nr;
@@ -131,25 +140,43 @@ struct nv50_pc {
        struct nv50_reg *r_brdc;
        struct nv50_reg *r_dst[4];
 
+       struct nv50_reg reg_instances[16];
+       unsigned reg_instance_nr;
+
        unsigned interp_mode[32];
        /* perspective interpolation registers */
        struct nv50_reg *iv_p;
        struct nv50_reg *iv_c;
 
-       struct nv50_program_exec *if_cond;
-       struct nv50_program_exec *if_insn[MAX_IF_DEPTH];
-       struct nv50_program_exec *br_join[MAX_IF_DEPTH];
-       struct nv50_program_exec *br_loop[MAX_LOOP_DEPTH]; /* for BRK branch */
+       struct nv50_program_exec *if_insn[NV50_MAX_COND_NESTING];
+       struct nv50_program_exec *if_join[NV50_MAX_COND_NESTING];
+       struct nv50_program_exec *loop_brka[NV50_MAX_LOOP_NESTING];
        int if_lvl, loop_lvl;
-       unsigned loop_pos[MAX_LOOP_DEPTH];
+       unsigned loop_pos[NV50_MAX_LOOP_NESTING];
 
        /* current instruction and total number of insns */
        unsigned insn_cur;
        unsigned insn_nr;
 
        boolean allow32;
+
+       uint8_t edgeflag_out;
 };
 
+static INLINE struct nv50_reg *
+reg_instance(struct nv50_pc *pc, struct nv50_reg *reg)
+{
+       struct nv50_reg *ri;
+
+       assert(pc->reg_instance_nr < 16);
+       ri = &pc->reg_instances[pc->reg_instance_nr++];
+       if (reg) {
+               *ri = *reg;
+               reg->mod = 0;
+       }
+       return ri;
+}
+
 static INLINE void
 ctor_reg(struct nv50_reg *reg, unsigned type, int index, int hw)
 {
@@ -176,8 +203,7 @@ terminate_mbb(struct nv50_pc *pc)
 
        /* remove records of temporary address register values */
        for (i = 0; i < NV50_SU_MAX_ADDR; ++i)
-               if (pc->r_addr[i].index < 0)
-                       pc->r_addr[i].rhw = -1;
+               pc->r_addr[i].rhw = -1;
 }
 
 static void
@@ -342,25 +368,34 @@ static void
 kill_temp_temp(struct nv50_pc *pc)
 {
        int i;
-       
+
        for (i = 0; i < pc->temp_temp_nr; i++)
                free_temp(pc, pc->temp_temp[i]);
        pc->temp_temp_nr = 0;
 }
 
 static int
-ctor_immd(struct nv50_pc *pc, float x, float y, float z, float w)
+ctor_immd_4u32(struct nv50_pc *pc,
+              uint32_t x, uint32_t y, uint32_t z, uint32_t w)
 {
-       pc->immd_buf = REALLOC(pc->immd_buf, (pc->immd_nr * 4 * sizeof(float)),
-                              (pc->immd_nr + 1) * 4 * sizeof(float));
+       unsigned size = pc->immd_nr * 4 * sizeof(uint32_t);
+
+       pc->immd_buf = REALLOC(pc->immd_buf, size, size + 4 * sizeof(uint32_t));
+
        pc->immd_buf[(pc->immd_nr * 4) + 0] = x;
        pc->immd_buf[(pc->immd_nr * 4) + 1] = y;
        pc->immd_buf[(pc->immd_nr * 4) + 2] = z;
        pc->immd_buf[(pc->immd_nr * 4) + 3] = w;
-       
+
        return pc->immd_nr++;
 }
 
+static INLINE int
+ctor_immd_4f32(struct nv50_pc *pc, float x, float y, float z, float w)
+{
+       return ctor_immd_4u32(pc, fui(x), fui(y), fui(z), fui(w));
+}
+
 static struct nv50_reg *
 alloc_immd(struct nv50_pc *pc, float f)
 {
@@ -368,11 +403,11 @@ alloc_immd(struct nv50_pc *pc, float f)
        unsigned hw;
 
        for (hw = 0; hw < pc->immd_nr * 4; hw++)
-               if (pc->immd_buf[hw] == f)
+               if (pc->immd_buf[hw] == fui(f))
                        break;
 
        if (hw == pc->immd_nr * 4)
-               hw = ctor_immd(pc, f, -f, 0.5 * f, 0) * 4;
+               hw = ctor_immd_4f32(pc, f, -f, 0.5 * f, 0) * 4;
 
        ctor_reg(r, P_IMMD, -1, hw);
        return r;
@@ -464,22 +499,24 @@ set_dst(struct nv50_pc *pc, struct nv50_reg *dst, struct nv50_program_exec *e)
 static INLINE void
 set_immd(struct nv50_pc *pc, struct nv50_reg *imm, struct nv50_program_exec *e)
 {
-       unsigned val;
-       float f = pc->immd_buf[imm->hw];
+       union {
+               float f;
+               uint32_t ui;
+       } u;
+       u.ui = pc->immd_buf[imm->hw];
 
-       if (imm->mod & NV50_MOD_ABS)
-               f = fabsf(f);
-       val = fui((imm->mod & NV50_MOD_NEG) ? -f : f);
+       u.f = (imm->mod & NV50_MOD_ABS) ? fabsf(u.f) : u.f;
+       u.f = (imm->mod & NV50_MOD_NEG) ? -u.f : u.f;
 
        set_long(pc, e);
-       /*XXX: can't be predicated - bits overlap.. catch cases where both
-        *     are required and avoid them. */
+       /* XXX: can't be predicated - bits overlap; cases where both
+        * are required should be avoided by using pc->allow32 */
        set_pred(pc, 0, 0, e);
        set_pred_wr(pc, 0, 0, e);
 
        e->inst[1] |= 0x00000002 | 0x00000001;
-       e->inst[0] |= (val & 0x3f) << 16;
-       e->inst[1] |= (val >> 6) << 2;
+       e->inst[0] |= (u.ui & 0x3f) << 16;
+       e->inst[1] |= (u.ui >> 6) << 2;
 }
 
 static INLINE void
@@ -511,21 +548,24 @@ emit_add_addr_imm(struct nv50_pc *pc, struct nv50_reg *dst,
 static struct nv50_reg *
 alloc_addr(struct nv50_pc *pc, struct nv50_reg *ref)
 {
-       int i;
        struct nv50_reg *a_tgsi = NULL, *a = NULL;
+       int i;
+       uint8_t avail = ~pc->addr_alloc;
 
        if (!ref) {
-               /* allocate for TGSI address reg */
-               for (i = 0; i < NV50_SU_MAX_ADDR; ++i) {
-                       if (pc->r_addr[i].index >= 0)
-                               continue;
-                       if (pc->r_addr[i].rhw >= 0 &&
-                           pc->r_addr[i].acc == pc->insn_cur)
-                               continue;
+               /* allocate for TGSI_FILE_ADDRESS */
+               while (avail) {
+                       i = ffs(avail) - 1;
 
-                       pc->r_addr[i].rhw = -1;
-                       pc->r_addr[i].index = i;
-                       return &pc->r_addr[i];
+                       if (pc->r_addr[i].rhw < 0 ||
+                           pc->r_addr[i].acc != pc->insn_cur) {
+                               pc->addr_alloc |= (1 << i);
+
+                               pc->r_addr[i].rhw = -1;
+                               pc->r_addr[i].index = i;
+                               return &pc->r_addr[i];
+                       }
+                       avail &= ~(1 << i);
                }
                assert(0);
                return NULL;
@@ -533,15 +573,16 @@ alloc_addr(struct nv50_pc *pc, struct nv50_reg *ref)
 
        /* Allocate and set an address reg so we can access 'ref'.
         *
-        * If and r_addr has index < 0, it is not reserved for TGSI,
-        * and index will be the negative of the TGSI addr index the
-        * value in rhw is relative to, or -256 if rhw is an offset
-        * from 0. If rhw < 0, the reg has not been initialized.
+        * If and r_addr->index will be -1 or the hw index the value
+        * value in rhw is relative to. If rhw < 0, the reg has not
+        * been initialized or is in use for TGSI_FILE_ADDRESS.
         */
-       for (i = NV50_SU_MAX_ADDR - 1; i >= 0; --i) {
-               if (pc->r_addr[i].index >= 0) /* occupied for TGSI */
-                       continue;
-               if (pc->r_addr[i].rhw < 0) { /* unused */
+       while (avail) { /* only consider regs that are not TGSI */
+               i = ffs(avail) - 1;
+               avail &= ~(1 << i);
+
+               if ((!a || a->rhw >= 0) && pc->r_addr[i].rhw < 0) {
+                       /* prefer an usused reg with low hw index */
                        a = &pc->r_addr[i];
                        continue;
                }
@@ -551,8 +592,8 @@ alloc_addr(struct nv50_pc *pc, struct nv50_reg *ref)
                if (ref->hw - pc->r_addr[i].rhw >= 128)
                        continue;
 
-               if ((ref->acc >= 0 && pc->r_addr[i].index == -256) ||
-                   (ref->acc < 0 && -pc->r_addr[i].index == ref->index)) {
+               if ((ref->acc >= 0 && pc->r_addr[i].index < 0) ||
+                   (ref->acc < 0 && pc->r_addr[i].index == ref->index)) {
                        pc->r_addr[i].acc = pc->insn_cur;
                        return &pc->r_addr[i];
                }
@@ -566,7 +607,7 @@ alloc_addr(struct nv50_pc *pc, struct nv50_reg *ref)
 
        a->rhw = ref->hw & ~0x7f;
        a->acc = pc->insn_cur;
-       a->index = a_tgsi ? -ref->index : -256;
+       a->index = a_tgsi ? ref->index : -1;
        return a;
 }
 
@@ -644,7 +685,7 @@ emit_mov(struct nv50_pc *pc, struct nv50_reg *dst, struct nv50_reg *src)
        if (src->type == P_IMMD || src->type == P_CONST) {
                set_long(pc, e);
                set_data(pc, src, 0x7f, 9, e);
-               e->inst[1] |= 0x20000000; /* src0 const? */
+               e->inst[1] |= 0x20000000; /* mov from c[] */
        } else {
                if (src->type == P_ATTR) {
                        set_long(pc, e);
@@ -659,9 +700,9 @@ emit_mov(struct nv50_pc *pc, struct nv50_reg *dst, struct nv50_reg *src)
 
        if (is_long(e) && !is_immd(e)) {
                e->inst[1] |= 0x04000000; /* 32-bit */
-               e->inst[1] |= 0x0000c000; /* "subsubop" 0x3 */
+               e->inst[1] |= 0x0000c000; /* 32-bit c[] load / lane mask 0:1 */
                if (!(e->inst[1] & 0x20000000))
-                       e->inst[1] |= 0x00030000; /* "subsubop" 0xf */
+                       e->inst[1] |= 0x00030000; /* lane mask 2:3 */
        } else
                e->inst[0] |= 0x00008000;
 
@@ -676,6 +717,17 @@ emit_mov_immdval(struct nv50_pc *pc, struct nv50_reg *dst, float f)
        FREE(imm);
 }
 
+static void
+emit_nop(struct nv50_pc *pc)
+{
+       struct nv50_program_exec *e = exec(pc);
+
+       e->inst[0] = 0xf0000000;
+       set_long(pc, e);
+       e->inst[1] = 0xe0000000;
+       emit(pc, e);
+}
+
 static boolean
 check_swap_src_0_1(struct nv50_pc *pc,
                   struct nv50_reg **s0, struct nv50_reg **s1)
@@ -794,6 +846,33 @@ set_src_2(struct nv50_pc *pc, struct nv50_reg *src, struct nv50_program_exec *e)
        e->inst[1] |= ((src->hw & 127) << 14);
 }
 
+static void
+emit_mov_from_pred(struct nv50_pc *pc, struct nv50_reg *dst, int pred)
+{
+       struct nv50_program_exec *e = exec(pc);
+
+       assert(dst->type == P_TEMP);
+       e->inst[1] = 0x20000000 | (pred << 12);
+       set_long(pc, e);
+       set_dst(pc, dst, e);
+
+       emit(pc, e);
+}
+
+static void
+emit_mov_to_pred(struct nv50_pc *pc, int pred, struct nv50_reg *src)
+{
+       struct nv50_program_exec *e = exec(pc);
+
+       e->inst[0] = 0x000001fc;
+       e->inst[1] = 0xa0000008;
+       set_long(pc, e);
+       set_pred_wr(pc, 1, pred, e);
+       set_src_0_restricted(pc, src, e);
+
+       emit(pc, e);
+}
+
 static void
 emit_mul(struct nv50_pc *pc, struct nv50_reg *dst, struct nv50_reg *src0,
         struct nv50_reg *src1)
@@ -898,7 +977,6 @@ static INLINE void
 emit_sub(struct nv50_pc *pc, struct nv50_reg *dst, struct nv50_reg *src0,
         struct nv50_reg *src1)
 {
-       assert(src0 != src1);
        src1->mod ^= NV50_MOD_NEG;
        emit_add(pc, dst, src0, src1);
        src1->mod ^= NV50_MOD_NEG;
@@ -967,7 +1045,6 @@ static INLINE void
 emit_msb(struct nv50_pc *pc, struct nv50_reg *dst, struct nv50_reg *src0,
         struct nv50_reg *src1, struct nv50_reg *src2)
 {
-       assert(src2 != src0 && src2 != src1);
        src2->mod ^= NV50_MOD_NEG;
        emit_mad(pc, dst, src0, src1, src2);
        src2->mod ^= NV50_MOD_NEG;
@@ -1120,7 +1197,6 @@ emit_set(struct nv50_pc *pc, unsigned ccode, struct nv50_reg *dst, int wp,
        set_src_1(pc, src1, e);
 
        emit(pc, e);
-       pc->if_cond = pc->p->exec_tail; /* record for OPCODE_IF */
 
        /* cvt.f32.u32/s32 (?) if we didn't only write the predicate */
        if (rdst)
@@ -1242,24 +1318,128 @@ emit_kil(struct nv50_pc *pc, struct nv50_reg *src)
 {
        struct nv50_program_exec *e;
        const int r_pred = 1;
-       unsigned cvn = CVT_F32_F32;
 
-       if (src->mod & NV50_MOD_NEG)
-               cvn |= CVT_NEG;
-       /* write predicate reg */
-       emit_cvt(pc, NULL, src, r_pred, CVTOP_RN, cvn);
-
-       /* conditional discard */
        e = exec(pc);
-       e->inst[0] = 0x00000002;
+       e->inst[0] = 0x00000002; /* discard */
+       set_long(pc, e); /* sets cond code to ALWAYS */
+
+       if (src) {
+               unsigned cvn = CVT_F32_F32;
+
+               set_pred(pc, 0x1 /* cc = LT */, r_pred, e);
+
+               if (src->mod & NV50_MOD_NEG)
+                       cvn |= CVT_NEG;
+               /* write predicate reg */
+               emit_cvt(pc, NULL, src, r_pred, CVTOP_RN, cvn);
+       }
+
+       emit(pc, e);
+}
+
+static struct nv50_program_exec *
+emit_breakaddr(struct nv50_pc *pc)
+{
+       struct nv50_program_exec *e = exec(pc);
+
+       e->inst[0] = 0x40000002;
+       set_long(pc, e);
+
+       emit(pc, e);
+       return e;
+}
+
+static void
+emit_break(struct nv50_pc *pc, int pred, unsigned cc)
+{
+       struct nv50_program_exec *e = exec(pc);
+
+       e->inst[0] = 0x50000002;
+       set_long(pc, e);
+       if (pred >= 0)
+               set_pred(pc, cc, pred, e);
+
+       emit(pc, e);
+}
+
+static struct nv50_program_exec *
+emit_joinat(struct nv50_pc *pc)
+{
+       struct nv50_program_exec *e = exec(pc);
+
+       e->inst[0] = 0xa0000002;
        set_long(pc, e);
-       set_pred(pc, 0x1 /* LT */, r_pred, e);
+
+       emit(pc, e);
+       return e;
+}
+
+static struct nv50_program_exec *
+emit_branch(struct nv50_pc *pc, int pred, unsigned cc)
+{
+       struct nv50_program_exec *e = exec(pc);
+
+       e->inst[0] = 0x10000002;
+       set_long(pc, e);
+       if (pred >= 0)
+               set_pred(pc, cc, pred, e);
        emit(pc, e);
+       return pc->p->exec_tail;
+}
+
+static void
+emit_ret(struct nv50_pc *pc, int pred, unsigned cc)
+{
+       struct nv50_program_exec *e = exec(pc);
+
+       e->inst[0] = 0x30000002;
+       set_long(pc, e);
+       if (pred >= 0)
+               set_pred(pc, cc, pred, e);
+
+       emit(pc, e);
+}
+
+#define QOP_ADD 0
+#define QOP_SUBR 1
+#define QOP_SUB 2
+#define QOP_MOV_SRC1 3
+
+/* For a quad of threads / top left, top right, bottom left, bottom right
+ * pixels, do a different operation, and take src0 from a specific thread.
+ */
+static void
+emit_quadop(struct nv50_pc *pc, struct nv50_reg *dst, int wp, int lane_src0,
+           struct nv50_reg *src0, struct nv50_reg *src1, ubyte qop)
+{
+       struct nv50_program_exec *e = exec(pc);
+
+       e->inst[0] = 0xc0000000;
+       e->inst[1] = 0x80000000;
+       set_long(pc, e);
+       e->inst[0] |= lane_src0 << 16;
+       set_src_0(pc, src0, e);
+       set_src_2(pc, src1, e);
+
+       if (wp >= 0)
+              set_pred_wr(pc, 1, wp, e);
+
+       if (dst)
+              set_dst(pc, dst, e);
+       else {
+              e->inst[0] |= 0x000001fc;
+              e->inst[1] |= 0x00000008;
+       }
+
+       e->inst[0] |= (qop & 3) << 20;
+       e->inst[1] |= (qop >> 2) << 22;
+
+       emit(pc, e);
 }
 
 static void
 load_cube_tex_coords(struct nv50_pc *pc, struct nv50_reg *t[4],
-                    struct nv50_reg **src, boolean proj)
+                    struct nv50_reg **src, unsigned arg, boolean proj)
 {
        int mod[3] = { src[0]->mod, src[1]->mod, src[2]->mod };
 
@@ -1276,6 +1456,10 @@ load_cube_tex_coords(struct nv50_pc *pc, struct nv50_reg *t[4],
 
        if (proj && 0 /* looks more correct without this */)
                emit_mul(pc, t[2], t[2], src[3]);
+       else
+       if (arg == 4) /* there is no textureProj(samplerCubeShadow) */
+               emit_mov(pc, t[3], src[3]);
+
        emit_flop(pc, 0, t[2], t[2]);
 
        emit_mul(pc, t[0], src[0], t[2]);
@@ -1284,89 +1468,214 @@ load_cube_tex_coords(struct nv50_pc *pc, struct nv50_reg *t[4],
 }
 
 static void
-emit_tex(struct nv50_pc *pc, struct nv50_reg **dst, unsigned mask,
-        struct nv50_reg **src, unsigned unit, unsigned type, boolean proj)
+load_proj_tex_coords(struct nv50_pc *pc, struct nv50_reg *t[4],
+                    struct nv50_reg **src, unsigned dim, unsigned arg)
 {
-       struct nv50_reg *t[4];
-       struct nv50_program_exec *e;
+       unsigned c, mode;
+
+       if (src[0]->type == P_TEMP && src[0]->rhw != -1) {
+               mode = pc->interp_mode[src[0]->index] | INTERP_PERSPECTIVE;
 
-       unsigned c, mode, dim;
+               t[3]->rhw = src[3]->rhw;
+               emit_interp(pc, t[3], NULL, (mode & INTERP_CENTROID));
+               emit_flop(pc, 0, t[3], t[3]);
 
+               for (c = 0; c < dim; ++c) {
+                       t[c]->rhw = src[c]->rhw;
+                       emit_interp(pc, t[c], t[3], mode);
+               }
+               if (arg != dim) { /* depth reference value */
+                       t[dim]->rhw = src[2]->rhw;
+                       emit_interp(pc, t[dim], t[3], mode);
+               }
+       } else {
+               /* XXX: for some reason the blob sometimes uses MAD
+                * (mad f32 $rX $rY $rZ neg $r63)
+                */
+               emit_flop(pc, 0, t[3], src[3]);
+               for (c = 0; c < dim; ++c)
+                       emit_mul(pc, t[c], src[c], t[3]);
+               if (arg != dim) /* depth reference value */
+                       emit_mul(pc, t[dim], src[2], t[3]);
+       }
+}
+
+static INLINE void
+get_tex_dim(unsigned type, unsigned *dim, unsigned *arg)
+{
        switch (type) {
        case TGSI_TEXTURE_1D:
-               dim = 1;
+               *arg = *dim = 1;
+               break;
+       case TGSI_TEXTURE_SHADOW1D:
+               *dim = 1;
+               *arg = 2;
                break;
        case TGSI_TEXTURE_UNKNOWN:
        case TGSI_TEXTURE_2D:
-       case TGSI_TEXTURE_SHADOW1D: /* XXX: x, z */
        case TGSI_TEXTURE_RECT:
-               dim = 2;
+               *arg = *dim = 2;
+               break;
+       case TGSI_TEXTURE_SHADOW2D:
+       case TGSI_TEXTURE_SHADOWRECT:
+               *dim = 2;
+               *arg = 3;
                break;
        case TGSI_TEXTURE_3D:
        case TGSI_TEXTURE_CUBE:
-       case TGSI_TEXTURE_SHADOW2D:
-       case TGSI_TEXTURE_SHADOWRECT: /* XXX */
-               dim = 3;
+               *dim = *arg = 3;
                break;
        default:
                assert(0);
                break;
        }
+}
 
-       /* some cards need t[0]'s hw index to be a multiple of 4 */
-       alloc_temp4(pc, t, 0);
+/* We shouldn't execute TEXLOD if any of the pixels in a quad have
+ * different LOD values, so branch off groups of equal LOD.
+ */
+static void
+emit_texlod_sequence(struct nv50_pc *pc, struct nv50_reg *tlod,
+                    struct nv50_reg *src, struct nv50_program_exec *tex)
+{
+       struct nv50_program_exec *join_at;
+       unsigned i, target = pc->p->exec_size + 7 * 2;
 
-       if (type == TGSI_TEXTURE_CUBE) {
-               load_cube_tex_coords(pc, t, src, proj);
-       } else
-       if (proj) {
-               if (src[0]->type == P_TEMP && src[0]->rhw != -1) {
-                       mode = pc->interp_mode[src[0]->index];
-
-                       t[3]->rhw = src[3]->rhw;
-                       emit_interp(pc, t[3], NULL, (mode & INTERP_CENTROID));
-                       emit_flop(pc, 0, t[3], t[3]);
-
-                       for (c = 0; c < dim; c++) {
-                               t[c]->rhw = src[c]->rhw;
-                               emit_interp(pc, t[c], t[3],
-                                           (mode | INTERP_PERSPECTIVE));
-                       }
-               } else {
-                       emit_flop(pc, 0, t[3], src[3]);
-                       for (c = 0; c < dim; c++)
-                               emit_mul(pc, t[c], src[c], t[3]);
+       /* Subtract lod of each pixel from lod of top left pixel, jump
+        * texlod insn if result is 0, then repeat for 2 other pixels.
+        */
+       join_at = emit_joinat(pc);
+       emit_quadop(pc, NULL, 0, 0, tlod, tlod, 0x55);
+       emit_branch(pc, 0, 2)->param.index = target;
 
-                       /* XXX: for some reason the blob sometimes uses MAD:
-                        * emit_mad(pc, t[c], src[0][c], t[3], t[3])
-                        * pc->p->exec_tail->inst[1] |= 0x080fc000;
-                        */
+       for (i = 1; i < 4; ++i) {
+               emit_quadop(pc, NULL, 0, i, tlod, tlod, 0x55);
+               emit_branch(pc, 0, 2)->param.index = target;
+       }
+
+       emit_mov(pc, tlod, src); /* target */
+       emit(pc, tex); /* texlod */
+
+       join_at->param.index = target + 2 * 2;
+       JOIN_ON(emit_nop(pc)); /* join _after_ tex */
+}
+
+static void
+emit_texbias_sequence(struct nv50_pc *pc, struct nv50_reg *t[4], unsigned arg,
+                     struct nv50_program_exec *tex)
+{
+       struct nv50_program_exec *e;
+       struct nv50_reg imm_1248, *t123[4][4], *r_bits = alloc_temp(pc, NULL);
+       int r_pred = 0;
+       unsigned n, c, i, cc[4] = { 0x0a, 0x13, 0x11, 0x10 };
+
+       pc->allow32 = FALSE;
+       ctor_reg(&imm_1248, P_IMMD, -1, ctor_immd_4u32(pc, 1, 2, 4, 8) * 4);
+
+       /* Subtract bias value of thread i from bias values of each thread,
+        * store result in r_pred, and set bit i in r_bits if result was 0.
+        */
+       assert(arg < 4);
+       for (i = 0; i < 4; ++i, ++imm_1248.hw) {
+               emit_quadop(pc, NULL, r_pred, i, t[arg], t[arg], 0x55);
+               emit_mov(pc, r_bits, &imm_1248);
+               set_pred(pc, 2, r_pred, pc->p->exec_tail);
+       }
+       emit_mov_to_pred(pc, r_pred, r_bits);
+
+       /* The lanes of a quad are now grouped by the bit in r_pred they have
+        * set. Put the input values for TEX into a new register set for each
+        * group and execute TEX only for a specific group.
+        * We cannot use the same register set for each group because we need
+        * the derivatives, which are implicitly calculated, to be correct.
+        */
+       for (i = 1; i < 4; ++i) {
+               alloc_temp4(pc, t123[i], 0);
+
+               for (c = 0; c <= arg; ++c)
+                       emit_mov(pc, t123[i][c], t[c]);
+
+               *(e = exec(pc)) = *(tex);
+               e->inst[0] &= ~0x01fc;
+               set_dst(pc, t123[i][0], e);
+               set_pred(pc, cc[i], r_pred, e);
+               emit(pc, e);
+       }
+       /* finally TEX on the original regs (where we kept the input) */
+       set_pred(pc, cc[0], r_pred, tex);
+       emit(pc, tex);
+
+       /* put the 3 * n other results into regs for lane 0 */
+       n = popcnt4(((e->inst[0] >> 25) & 0x3) | ((e->inst[1] >> 12) & 0xc));
+       for (i = 1; i < 4; ++i) {
+               for (c = 0; c < n; ++c) {
+                       emit_mov(pc, t[c], t123[i][c]);
+                       set_pred(pc, cc[i], r_pred, pc->p->exec_tail);
                }
-       } else {
-               for (c = 0; c < dim; c++)
-                       emit_mov(pc, t[c], src[c]);
+               free_temp4(pc, t123[i]);
        }
 
+       emit_nop(pc);
+       free_temp(pc, r_bits);
+}
+
+static void
+emit_tex(struct nv50_pc *pc, struct nv50_reg **dst, unsigned mask,
+        struct nv50_reg **src, unsigned unit, unsigned type,
+        boolean proj, int bias_lod)
+{
+       struct nv50_reg *t[4];
+       struct nv50_program_exec *e;
+       unsigned c, dim, arg;
+
+       /* t[i] must be within a single 128 bit super-reg */
+       alloc_temp4(pc, t, 0);
+
        e = exec(pc);
+       e->inst[0] = 0xf0000000;
        set_long(pc, e);
-       e->inst[0] |= 0xf0000000;
-       e->inst[1] |= 0x00000004;
        set_dst(pc, t[0], e);
-       e->inst[0] |= (unit << 9);
 
-       if (dim == 2)
-               e->inst[0] |= 0x00400000;
-       else
-       if (dim == 3) {
-               e->inst[0] |= 0x00800000;
-               if (type == TGSI_TEXTURE_CUBE)
-                       e->inst[0] |= 0x08000000;
+       /* TIC and TSC binding indices (TSC is ignored as TSC_LINKED = TRUE): */
+       e->inst[0] |= (unit << 9) /* | (unit << 17) */;
+
+       /* live flag (don't set if TEX results affect input to another TEX): */
+       /* e->inst[0] |= 0x00000004; */
+
+       get_tex_dim(type, &dim, &arg);
+
+       if (type == TGSI_TEXTURE_CUBE) {
+               e->inst[0] |= 0x08000000;
+               load_cube_tex_coords(pc, t, src, arg, proj);
+       } else
+       if (proj)
+               load_proj_tex_coords(pc, t, src, dim, arg);
+       else {
+               for (c = 0; c < dim; c++)
+                       emit_mov(pc, t[c], src[c]);
+               if (arg != dim) /* depth reference value (always src.z here) */
+                       emit_mov(pc, t[dim], src[2]);
        }
 
        e->inst[0] |= (mask & 0x3) << 25;
        e->inst[1] |= (mask & 0xc) << 12;
 
-       emit(pc, e);
+       if (!bias_lod) {
+               e->inst[0] |= (arg - 1) << 22;
+               emit(pc, e);
+       } else
+       if (bias_lod < 0) {
+               e->inst[0] |= arg << 22;
+               e->inst[1] |= 0x20000000; /* texbias */
+               emit_mov(pc, t[arg], src[3]);
+               emit_texbias_sequence(pc, t, arg, e);
+       } else {
+               e->inst[0] |= arg << 22;
+               e->inst[1] |= 0x40000000; /* texlod */
+               emit_mov(pc, t[arg], src[3]);
+               emit_texlod_sequence(pc, t[arg], src[3], e);
+       }
+
 #if 1
        c = 0;
        if (mask & 1) emit_mov(pc, dst[0], t[c++]);
@@ -1388,38 +1697,6 @@ emit_tex(struct nv50_pc *pc, struct nv50_reg **dst, unsigned mask,
 #endif
 }
 
-static void
-emit_branch(struct nv50_pc *pc, int pred, unsigned cc,
-           struct nv50_program_exec **join)
-{
-       struct nv50_program_exec *e = exec(pc);
-
-       if (join) {
-               set_long(pc, e);
-               e->inst[0] |= 0xa0000002;
-               emit(pc, e);
-               *join = e;
-               e = exec(pc);
-       }
-
-       set_long(pc, e);
-       e->inst[0] |= 0x10000002;
-       if (pred >= 0)
-               set_pred(pc, cc, pred, e);
-       emit(pc, e);
-}
-
-static void
-emit_nop(struct nv50_pc *pc)
-{
-       struct nv50_program_exec *e = exec(pc);
-
-       e->inst[0] = 0xf0000000;
-       set_long(pc, e);
-       e->inst[1] = 0xe0000000;
-       emit(pc, e);
-}
-
 static void
 emit_ddx(struct nv50_pc *pc, struct nv50_reg *dst, struct nv50_reg *src)
 {
@@ -1427,8 +1704,8 @@ emit_ddx(struct nv50_pc *pc, struct nv50_reg *dst, struct nv50_reg *src)
 
        assert(src->type == P_TEMP);
 
-       e->inst[0] = 0xc0140000;
-       e->inst[1] = 0x89800000;
+       e->inst[0] = (src->mod & NV50_MOD_NEG) ? 0xc0240000 : 0xc0140000;
+       e->inst[1] = (src->mod & NV50_MOD_NEG) ? 0x86400000 : 0x89800000;
        set_long(pc, e);
        set_dst(pc, dst, e);
        set_src_0(pc, src, e);
@@ -1444,11 +1721,8 @@ emit_ddy(struct nv50_pc *pc, struct nv50_reg *dst, struct nv50_reg *src)
 
        assert(src->type == P_TEMP);
 
-       if (!(src->mod & NV50_MOD_NEG)) /* ! double negation */
-               emit_neg(pc, src, src);
-
-       e->inst[0] = 0xc0150000;
-       e->inst[1] = 0x8a400000;
+       e->inst[0] = (src->mod & NV50_MOD_NEG) ? 0xc0250000 : 0xc0150000;
+       e->inst[1] = (src->mod & NV50_MOD_NEG) ? 0x85800000 : 0x8a400000;
        set_long(pc, e);
        set_dst(pc, dst, e);
        set_src_0(pc, src, e);
@@ -1509,9 +1783,8 @@ convert_to_long(struct nv50_pc *pc, struct nv50_program_exec *e)
 static boolean
 negate_supported(const struct tgsi_full_instruction *insn, int i)
 {
-       int s;
-
        switch (insn->Instruction.Opcode) {
+       case TGSI_OPCODE_DDX:
        case TGSI_OPCODE_DDY:
        case TGSI_OPCODE_DP3:
        case TGSI_OPCODE_DP4:
@@ -1520,36 +1793,21 @@ negate_supported(const struct tgsi_full_instruction *insn, int i)
        case TGSI_OPCODE_ADD:
        case TGSI_OPCODE_SUB:
        case TGSI_OPCODE_MAD:
-               break;
+               return TRUE;
        case TGSI_OPCODE_POW:
                if (i == 1)
-                       break;
+                       return TRUE;
                return FALSE;
        default:
                return FALSE;
        }
-
-       /* Watch out for possible multiple uses of an nv50_reg, we
-        * can't use nv50_reg::neg in these cases.
-        */
-       for (s = 0; s < insn->Instruction.NumSrcRegs; ++s) {
-               if (s == i)
-                       continue;
-               if ((insn->FullSrcRegisters[s].SrcRegister.Index ==
-                    insn->FullSrcRegisters[i].SrcRegister.Index) &&
-                   (insn->FullSrcRegisters[s].SrcRegister.File ==
-                    insn->FullSrcRegisters[i].SrcRegister.File))
-                       return FALSE;
-       }
-
-       return TRUE;
 }
 
 /* Return a read mask for source registers deduced from opcode & write mask. */
 static unsigned
 nv50_tgsi_src_mask(const struct tgsi_full_instruction *insn, int c)
 {
-       unsigned x, mask = insn->FullDstRegisters[0].DstRegister.WriteMask;
+       unsigned x, mask = insn->Dst[0].Register.WriteMask;
 
        switch (insn->Instruction.Opcode) {
        case TGSI_OPCODE_COS:
@@ -1570,24 +1828,32 @@ nv50_tgsi_src_mask(const struct tgsi_full_instruction *insn, int c)
        case TGSI_OPCODE_RSQ:
        case TGSI_OPCODE_SCS:
                return 0x1;
+       case TGSI_OPCODE_IF:
+               return 0x1;
        case TGSI_OPCODE_LIT:
                return 0xb;
        case TGSI_OPCODE_TEX:
+       case TGSI_OPCODE_TXB:
+       case TGSI_OPCODE_TXL:
        case TGSI_OPCODE_TXP:
        {
-               const struct tgsi_instruction_ext_texture *tex;
+               const struct tgsi_instruction_texture *tex;
 
-               assert(insn->Instruction.Extended);
-               tex = &insn->InstructionExtTexture;
+               assert(insn->Instruction.Texture);
+               tex = &insn->Texture;
 
                mask = 0x7;
-               if (insn->Instruction.Opcode == TGSI_OPCODE_TXP)
-                       mask |= 0x8;
+               if (insn->Instruction.Opcode != TGSI_OPCODE_TEX &&
+                   insn->Instruction.Opcode != TGSI_OPCODE_TXD)
+                       mask |= 0x8; /* bias, lod or proj */
 
                switch (tex->Texture) {
                case TGSI_TEXTURE_1D:
                        mask &= 0x9;
                        break;
+               case TGSI_TEXTURE_SHADOW1D:
+                       mask &= 0x5;
+                       break;
                case TGSI_TEXTURE_2D:
                        mask &= 0xb;
                        break;
@@ -1612,17 +1878,17 @@ nv50_tgsi_src_mask(const struct tgsi_full_instruction *insn, int c)
 static struct nv50_reg *
 tgsi_dst(struct nv50_pc *pc, int c, const struct tgsi_full_dst_register *dst)
 {
-       switch (dst->DstRegister.File) {
+       switch (dst->Register.File) {
        case TGSI_FILE_TEMPORARY:
-               return &pc->temp[dst->DstRegister.Index * 4 + c];
+               return &pc->temp[dst->Register.Index * 4 + c];
        case TGSI_FILE_OUTPUT:
-               return &pc->result[dst->DstRegister.Index * 4 + c];
+               return &pc->result[dst->Register.Index * 4 + c];
        case TGSI_FILE_ADDRESS:
        {
-               struct nv50_reg *r = pc->addr[dst->DstRegister.Index * 4 + c];
+               struct nv50_reg *r = pc->addr[dst->Register.Index * 4 + c];
                if (!r) {
                        r = alloc_addr(pc, NULL);
-                       pc->addr[dst->DstRegister.Index * 4 + c] = r;
+                       pc->addr[dst->Register.Index * 4 + c] = r;
                }
                assert(r);
                return r;
@@ -1644,8 +1910,8 @@ tgsi_src(struct nv50_pc *pc, int chan, const struct tgsi_full_src_register *src,
        struct nv50_reg *temp;
        unsigned sgn, c, swz;
 
-       if (src->SrcRegister.File != TGSI_FILE_CONSTANT)
-               assert(!src->SrcRegister.Indirect);
+       if (src->Register.File != TGSI_FILE_CONSTANT)
+               assert(!src->Register.Indirect);
 
        sgn = tgsi_util_get_full_src_register_sign_mode(src, chan);
 
@@ -1655,36 +1921,36 @@ tgsi_src(struct nv50_pc *pc, int chan, const struct tgsi_full_src_register *src,
        case TGSI_SWIZZLE_Y:
        case TGSI_SWIZZLE_Z:
        case TGSI_SWIZZLE_W:
-               switch (src->SrcRegister.File) {
+               switch (src->Register.File) {
                case TGSI_FILE_INPUT:
-                       r = &pc->attr[src->SrcRegister.Index * 4 + c];
+                       r = &pc->attr[src->Register.Index * 4 + c];
                        break;
                case TGSI_FILE_TEMPORARY:
-                       r = &pc->temp[src->SrcRegister.Index * 4 + c];
+                       r = &pc->temp[src->Register.Index * 4 + c];
                        break;
                case TGSI_FILE_CONSTANT:
-                       if (!src->SrcRegister.Indirect) {
-                               r = &pc->param[src->SrcRegister.Index * 4 + c];
+                       if (!src->Register.Indirect) {
+                               r = &pc->param[src->Register.Index * 4 + c];
                                break;
                        }
                        /* Indicate indirection by setting r->acc < 0 and
                         * use the index field to select the address reg.
                         */
-                       r = MALLOC_STRUCT(nv50_reg);
+                       r = reg_instance(pc, NULL);
                        swz = tgsi_util_get_src_register_swizzle(
-                                                &src->SrcRegisterInd, 0);
+                                                &src->Indirect, 0);
                        ctor_reg(r, P_CONST,
-                                src->SrcRegisterInd.Index * 4 + swz,
-                                src->SrcRegister.Index * 4 + c);
+                                src->Indirect.Index * 4 + swz,
+                                src->Register.Index * 4 + c);
                        r->acc = -1;
                        break;
                case TGSI_FILE_IMMEDIATE:
-                       r = &pc->immd[src->SrcRegister.Index * 4 + c];
+                       r = &pc->immd[src->Register.Index * 4 + c];
                        break;
                case TGSI_FILE_SAMPLER:
                        break;
                case TGSI_FILE_ADDRESS:
-                       r = pc->addr[src->SrcRegister.Index * 4 + c];
+                       r = pc->addr[src->Register.Index * 4 + c];
                        assert(r);
                        break;
                default:
@@ -1724,6 +1990,8 @@ tgsi_src(struct nv50_pc *pc, int chan, const struct tgsi_full_src_register *src,
                break;
        }
 
+       if (r && r->acc >= 0 && r != temp)
+               return reg_instance(pc, r);
        return r;
 }
 
@@ -1779,6 +2047,8 @@ nv50_tgsi_dst_revdep(unsigned op, int s, int c)
        case TGSI_OPCODE_LIT:
        case TGSI_OPCODE_SCS:
        case TGSI_OPCODE_TEX:
+       case TGSI_OPCODE_TXB:
+       case TGSI_OPCODE_TXL:
        case TGSI_OPCODE_TXP:
                /* these take care of dangerous swizzles themselves */
                return 0x0;
@@ -1820,28 +2090,43 @@ nv50_kill_branch(struct nv50_pc *pc)
         */
        if (has_pred(pc->if_insn[lvl], 0xf))
                return FALSE;
-       assert(pc->if_insn[lvl] && pc->br_join[lvl]);
+       assert(pc->if_insn[lvl] && pc->if_join[lvl]);
 
-       /* We'll use the exec allocated for JOIN_AT (as we can't easily
-        * update prev's next); if exec_tail is BRK, update the pointer.
+       /* We'll use the exec allocated for JOIN_AT (we can't easily
+        * access nv50_program_exec's prev).
         */
-       if (pc->loop_lvl && pc->br_loop[pc->loop_lvl - 1] == pc->p->exec_tail)
-               pc->br_loop[pc->loop_lvl - 1] = pc->br_join[lvl];
-
        pc->p->exec_size -= 4; /* remove JOIN_AT and BRA */
 
-       *pc->br_join[lvl] = *pc->p->exec_tail;
+       *pc->if_join[lvl] = *pc->p->exec_tail;
 
        FREE(pc->if_insn[lvl]);
        FREE(pc->p->exec_tail);
 
-       pc->p->exec_tail = pc->br_join[lvl];
+       pc->p->exec_tail = pc->if_join[lvl];
        pc->p->exec_tail->next = NULL;
        set_pred(pc, 0xd, 0, pc->p->exec_tail);
 
        return TRUE;
 }
 
+static void
+nv50_fp_move_results(struct nv50_pc *pc)
+{
+       struct nv50_reg reg;
+       unsigned i;
+
+       ctor_reg(&reg, P_TEMP, -1, -1);
+
+       for (i = 0; i < pc->result_nr * 4; ++i) {
+               if (pc->result[i].rhw < 0 || pc->result[i].hw < 0)
+                       continue;
+               if (pc->result[i].rhw != pc->result[i].hw) {
+                       reg.hw = pc->result[i].rhw;
+                       emit_mov(pc, &reg, &pc->result[i]);
+               }
+       }
+}
+
 static boolean
 nv50_program_tx_insn(struct nv50_pc *pc,
                     const struct tgsi_full_instruction *inst)
@@ -1850,29 +2135,29 @@ nv50_program_tx_insn(struct nv50_pc *pc,
        unsigned mask, sat, unit;
        int i, c;
 
-       mask = inst->FullDstRegisters[0].DstRegister.WriteMask;
+       mask = inst->Dst[0].Register.WriteMask;
        sat = inst->Instruction.Saturate == TGSI_SAT_ZERO_ONE;
 
        memset(src, 0, sizeof(src));
 
        for (c = 0; c < 4; c++) {
                if ((mask & (1 << c)) && !pc->r_dst[c])
-                       dst[c] = tgsi_dst(pc, c, &inst->FullDstRegisters[0]);
+                       dst[c] = tgsi_dst(pc, c, &inst->Dst[0]);
                else
                        dst[c] = pc->r_dst[c];
                rdst[c] = dst[c];
        }
 
        for (i = 0; i < inst->Instruction.NumSrcRegs; i++) {
-               const struct tgsi_full_src_register *fs = &inst->FullSrcRegisters[i];
+               const struct tgsi_full_src_register *fs = &inst->Src[i];
                unsigned src_mask;
                boolean neg_supp;
 
                src_mask = nv50_tgsi_src_mask(inst, i);
                neg_supp = negate_supported(inst, i);
 
-               if (fs->SrcRegister.File == TGSI_FILE_SAMPLER)
-                       unit = fs->SrcRegister.Index;
+               if (fs->Register.File == TGSI_FILE_SAMPLER)
+                       unit = fs->Register.Index;
 
                for (c = 0; c < 4; c++)
                        if (src_mask & (1 << c))
@@ -1928,13 +2213,13 @@ nv50_program_tx_insn(struct nv50_pc *pc,
                emit_arl(pc, dst[0], temp, 4);
                break;
        case TGSI_OPCODE_BGNLOOP:
+               pc->loop_brka[pc->loop_lvl] = emit_breakaddr(pc);
                pc->loop_pos[pc->loop_lvl++] = pc->p->exec_size;
                terminate_mbb(pc);
                break;
        case TGSI_OPCODE_BRK:
-               emit_branch(pc, -1, 0, NULL);
                assert(pc->loop_lvl > 0);
-               pc->br_loop[pc->loop_lvl - 1] = pc->p->exec_tail;
+               emit_break(pc, -1, 0);
                break;
        case TGSI_OPCODE_CEIL:
                for (c = 0; c < 4; c++) {
@@ -1956,6 +2241,11 @@ nv50_program_tx_insn(struct nv50_pc *pc,
                        set_pred(pc, 0x6, 1, pc->p->exec_tail); /* @NSF */
                }
                break;
+       case TGSI_OPCODE_CONT:
+               assert(pc->loop_lvl > 0);
+               emit_branch(pc, -1, 0)->param.index =
+                       pc->loop_pos[pc->loop_lvl - 1];
+               break;
        case TGSI_OPCODE_COS:
                if (mask & 8) {
                        emit_precossin(pc, temp, src[0][3]);
@@ -2010,7 +2300,7 @@ nv50_program_tx_insn(struct nv50_pc *pc,
                        emit_mov_immdval(pc, dst[0], 1.0f);
                break;
        case TGSI_OPCODE_ELSE:
-               emit_branch(pc, -1, 0, NULL);
+               emit_branch(pc, -1, 0);
                pc->if_insn[--pc->if_lvl]->param.index = pc->p->exec_size;
                pc->if_insn[pc->if_lvl++] = pc->p->exec_tail;
                terminate_mbb(pc);
@@ -2022,21 +2312,20 @@ nv50_program_tx_insn(struct nv50_pc *pc,
                if (nv50_kill_branch(pc) == TRUE)
                        break;
 
-               if (pc->br_join[pc->if_lvl]) {
-                       pc->br_join[pc->if_lvl]->param.index = pc->p->exec_size;
-                       pc->br_join[pc->if_lvl] = NULL;
+               if (pc->if_join[pc->if_lvl]) {
+                       pc->if_join[pc->if_lvl]->param.index = pc->p->exec_size;
+                       pc->if_join[pc->if_lvl] = NULL;
                }
                terminate_mbb(pc);
                /* emit a NOP as join point, we could set it on the next
                 * one, but would have to make sure it is long and !immd
                 */
-               emit_nop(pc);
-               pc->p->exec_tail->inst[1] |= 2;
+               JOIN_ON(emit_nop(pc));
                break;
        case TGSI_OPCODE_ENDLOOP:
-               emit_branch(pc, -1, 0, NULL);
-               pc->p->exec_tail->param.index = pc->loop_pos[--pc->loop_lvl];
-               pc->br_loop[pc->loop_lvl]->param.index = pc->p->exec_size;
+               emit_branch(pc, -1, 0)->param.index =
+                       pc->loop_pos[--pc->loop_lvl];
+               pc->loop_brka[pc->loop_lvl]->param.index = pc->p->exec_size;
                terminate_mbb(pc);
                break;
        case TGSI_OPCODE_EX2:
@@ -2060,21 +2349,23 @@ nv50_program_tx_insn(struct nv50_pc *pc,
                }
                break;
        case TGSI_OPCODE_IF:
-               /* emitting a join_at may not be necessary */
-               assert(pc->if_lvl < MAX_IF_DEPTH);
-               /* set_pred_wr(pc, 1, 0, pc->if_cond); */
+               assert(pc->if_lvl < NV50_MAX_COND_NESTING);
                emit_cvt(pc, NULL, src[0][0], 0, CVTOP_ABS | CVTOP_RN,
                         CVT_F32_F32);
-               emit_branch(pc, 0, 2, &pc->br_join[pc->if_lvl]);
-               pc->if_insn[pc->if_lvl++] = pc->p->exec_tail;
+               pc->if_join[pc->if_lvl] = emit_joinat(pc);
+               pc->if_insn[pc->if_lvl++] = emit_branch(pc, 0, 2);;
                terminate_mbb(pc);
                break;
        case TGSI_OPCODE_KIL:
+               assert(src[0][0] && src[0][1] && src[0][2] && src[0][3]);
                emit_kil(pc, src[0][0]);
                emit_kil(pc, src[0][1]);
                emit_kil(pc, src[0][2]);
                emit_kil(pc, src[0][3]);
                break;
+       case TGSI_OPCODE_KILP:
+               emit_kil(pc, NULL);
+               break;
        case TGSI_OPCODE_LIT:
                emit_lit(pc, &dst[0], mask, &src[0][0]);
                break;
@@ -2131,6 +2422,11 @@ nv50_program_tx_insn(struct nv50_pc *pc,
        case TGSI_OPCODE_RCP:
                emit_flop(pc, 0, brdc, src[0][0]);
                break;
+       case TGSI_OPCODE_RET:
+               if (pc->p->type == PIPE_SHADER_FRAGMENT)
+                       nv50_fp_move_results(pc);
+               emit_ret(pc, -1, 0);
+               break;
        case TGSI_OPCODE_RSQ:
                emit_flop(pc, 2, brdc, src[0][0]);
                break;
@@ -2181,11 +2477,19 @@ nv50_program_tx_insn(struct nv50_pc *pc,
                break;
        case TGSI_OPCODE_TEX:
                emit_tex(pc, dst, mask, src[0], unit,
-                        inst->InstructionExtTexture.Texture, FALSE);
+                        inst->Texture.Texture, FALSE, 0);
+               break;
+       case TGSI_OPCODE_TXB:
+               emit_tex(pc, dst, mask, src[0], unit,
+                        inst->Texture.Texture, FALSE, -1);
+               break;
+       case TGSI_OPCODE_TXL:
+               emit_tex(pc, dst, mask, src[0], unit,
+                        inst->Texture.Texture, FALSE, 1);
                break;
        case TGSI_OPCODE_TXP:
                emit_tex(pc, dst, mask, src[0], unit,
-                        inst->InstructionExtTexture.Texture, TRUE);
+                        inst->Texture.Texture, TRUE, 0);
                break;
        case TGSI_OPCODE_TRUNC:
                for (c = 0; c < 4; c++) {
@@ -2239,20 +2543,9 @@ nv50_program_tx_insn(struct nv50_pc *pc,
                }
        }
 
-       for (i = 0; i < inst->Instruction.NumSrcRegs; i++) {
-               for (c = 0; c < 4; c++) {
-                       if (!src[i][c])
-                               continue;
-                       src[i][c]->mod = 0;
-                       if (src[i][c]->index == -1 && src[i][c]->type == P_IMMD)
-                               FREE(src[i][c]);
-                       else
-                       if (src[i][c]->acc < 0 && src[i][c]->type == P_CONST)
-                               FREE(src[i][c]); /* indirect constant */
-               }
-       }
-
        kill_temp_temp(pc);
+       pc->reg_instance_nr = 0;
+
        return TRUE;
 }
 
@@ -2264,14 +2557,20 @@ prep_inspect_insn(struct nv50_pc *pc, const struct tgsi_full_instruction *insn)
        const struct tgsi_dst_register *dst;
        unsigned i, c, k, mask;
 
-       dst = &insn->FullDstRegisters[0].DstRegister;
+       dst = &insn->Dst[0].Register;
        mask = dst->WriteMask;
 
         if (dst->File == TGSI_FILE_TEMPORARY)
-                reg = pc->temp;
+               reg = pc->temp;
         else
-        if (dst->File == TGSI_FILE_OUTPUT)
-                reg = pc->result;
+       if (dst->File == TGSI_FILE_OUTPUT) {
+               reg = pc->result;
+
+               if (insn->Instruction.Opcode == TGSI_OPCODE_MOV &&
+                   dst->Index == pc->edgeflag_out &&
+                   insn->Src[0].Register.File == TGSI_FILE_INPUT)
+                       pc->p->cfg.edgeflag_in = insn->Src[0].Register.Index;
+       }
 
        if (reg) {
                for (c = 0; c < 4; c++) {
@@ -2282,12 +2581,12 @@ prep_inspect_insn(struct nv50_pc *pc, const struct tgsi_full_instruction *insn)
        }
 
        for (i = 0; i < insn->Instruction.NumSrcRegs; i++) {
-               src = &insn->FullSrcRegisters[i];
+               src = &insn->Src[i];
 
-               if (src->SrcRegister.File == TGSI_FILE_TEMPORARY)
+               if (src->Register.File == TGSI_FILE_TEMPORARY)
                        reg = pc->temp;
                else
-               if (src->SrcRegister.File == TGSI_FILE_INPUT)
+               if (src->Register.File == TGSI_FILE_INPUT)
                        reg = pc->attr;
                else
                        continue;
@@ -2299,7 +2598,7 @@ prep_inspect_insn(struct nv50_pc *pc, const struct tgsi_full_instruction *insn)
                                continue;
                        k = tgsi_util_get_full_src_register_swizzle(src, c);
 
-                       reg[src->SrcRegister.Index * 4 + k].acc = pc->insn_nr;
+                       reg[src->Register.Index * 4 + k].acc = pc->insn_nr;
                }
        }
 }
@@ -2359,13 +2658,13 @@ static struct nv50_reg *
 tgsi_broadcast_dst(struct nv50_pc *pc,
                   const struct tgsi_full_dst_register *fd, unsigned mask)
 {
-       if (fd->DstRegister.File == TGSI_FILE_TEMPORARY) {
-               int c = ffs(~mask & fd->DstRegister.WriteMask);
+       if (fd->Register.File == TGSI_FILE_TEMPORARY) {
+               int c = ffs(~mask & fd->Register.WriteMask);
                if (c)
                        return tgsi_dst(pc, c - 1, fd);
        } else {
-               int c = ffs(fd->DstRegister.WriteMask) - 1;
-               if ((1 << c) == fd->DstRegister.WriteMask)
+               int c = ffs(fd->Register.WriteMask) - 1;
+               if ((1 << c) == fd->Register.WriteMask)
                        return tgsi_dst(pc, c, fd);
        }
 
@@ -2379,7 +2678,7 @@ static unsigned
 nv50_tgsi_scan_swizzle(const struct tgsi_full_instruction *insn,
                       unsigned rdep[4])
 {
-       const struct tgsi_full_dst_register *fd = &insn->FullDstRegisters[0];
+       const struct tgsi_full_dst_register *fd = &insn->Dst[0];
        const struct tgsi_full_src_register *fs;
        unsigned i, deqs = 0;
 
@@ -2390,9 +2689,9 @@ nv50_tgsi_scan_swizzle(const struct tgsi_full_instruction *insn,
                unsigned chn, mask = nv50_tgsi_src_mask(insn, i);
                boolean neg_supp = negate_supported(insn, i);
 
-               fs = &insn->FullSrcRegisters[i];
-               if (fs->SrcRegister.File != fd->DstRegister.File ||
-                   fs->SrcRegister.Index != fd->DstRegister.Index)
+               fs = &insn->Src[i];
+               if (fs->Register.File != fd->Register.File ||
+                   fs->Register.Index != fd->Register.Index)
                        continue;
 
                for (chn = 0; chn < 4; ++chn) {
@@ -2403,7 +2702,7 @@ nv50_tgsi_scan_swizzle(const struct tgsi_full_instruction *insn,
                        c = tgsi_util_get_full_src_register_swizzle(fs, chn);
                        s = tgsi_util_get_full_src_register_sign_mode(fs, chn);
 
-                       if (!(fd->DstRegister.WriteMask & (1 << c)))
+                       if (!(fd->Register.WriteMask & (1 << c)))
                                continue;
 
                        /* no danger if src is copied to TEMP first */
@@ -2427,7 +2726,7 @@ nv50_tgsi_insn(struct nv50_pc *pc, const union tgsi_full_token *tok)
        const struct tgsi_full_dst_register *fd;
        unsigned i, deqs, rdep[4], m[4];
 
-       fd = &tok->FullInstruction.FullDstRegisters[0];
+       fd = &tok->FullInstruction.Dst[0];
        deqs = nv50_tgsi_scan_swizzle(&insn, rdep);
 
        if (is_scalar_op(insn.Instruction.Opcode)) {
@@ -2446,10 +2745,10 @@ nv50_tgsi_insn(struct nv50_pc *pc, const union tgsi_full_token *tok)
        for (i = 0; i < 4; ++i) {
                assert(pc->r_dst[m[i]] == NULL);
 
-               insn.FullDstRegisters[0].DstRegister.WriteMask =
-                       fd->DstRegister.WriteMask & (1 << m[i]);
+               insn.Dst[0].Register.WriteMask =
+                       fd->Register.WriteMask & (1 << m[i]);
 
-               if (!insn.FullDstRegisters[0].DstRegister.WriteMask)
+               if (!insn.Dst[0].Register.WriteMask)
                        continue;
 
                if (deqs & (1 << i))
@@ -2535,10 +2834,10 @@ nv50_program_tx_prep(struct nv50_pc *pc)
                        const struct tgsi_full_immediate *imm =
                                &tp.FullToken.FullImmediate;
 
-                       ctor_immd(pc, imm->u[0].Float,
-                                     imm->u[1].Float,
-                                     imm->u[2].Float,
-                                     imm->u[3].Float);
+                       ctor_immd_4f32(pc, imm->u[0].Float,
+                                      imm->u[1].Float,
+                                      imm->u[2].Float,
+                                      imm->u[3].Float);
                }
                        break;
                case TGSI_TOKEN_TYPE_DECLARATION:
@@ -2547,8 +2846,8 @@ nv50_program_tx_prep(struct nv50_pc *pc)
                        unsigned si, last, first, mode;
 
                        d = &tp.FullToken.FullDeclaration;
-                       first = d->DeclarationRange.First;
-                       last = d->DeclarationRange.Last;
+                       first = d->Range.First;
+                       last = d->Range.Last;
 
                        switch (d->Declaration.File) {
                        case TGSI_FILE_TEMPORARY:
@@ -2558,8 +2857,8 @@ nv50_program_tx_prep(struct nv50_pc *pc)
                                    p->type == PIPE_SHADER_FRAGMENT)
                                        break;
 
-                               si = d->Semantic.SemanticIndex;
-                               switch (d->Semantic.SemanticName) {
+                               si = d->Semantic.Index;
+                               switch (d->Semantic.Name) {
                                case TGSI_SEMANTIC_BCOLOR:
                                        p->cfg.two_side[si].hw = first;
                                        if (p->cfg.io_nr > first)
@@ -2570,6 +2869,9 @@ nv50_program_tx_prep(struct nv50_pc *pc)
                                        if (p->cfg.io_nr > first)
                                                p->cfg.io_nr = first;
                                        break;
+                               case TGSI_SEMANTIC_EDGEFLAG:
+                                       pc->edgeflag_out = first;
+                                       break;
                                        /*
                                case TGSI_SEMANTIC_CLIP_DISTANCE:
                                        p->cfg.clpd = MIN2(p->cfg.clpd, first);
@@ -2637,7 +2939,7 @@ nv50_program_tx_prep(struct nv50_pc *pc)
 
                for (i = 0, rid = 0; i < pc->result_nr; ++i) {
                        p->cfg.io[i].hw = rid;
-                       p->cfg.io[i].id_vp = i;
+                       p->cfg.io[i].id = i;
 
                        for (c = 0; c < 4; ++c) {
                                int n = i * 4 + c;
@@ -2669,14 +2971,12 @@ nv50_program_tx_prep(struct nv50_pc *pc)
                 * the lower hardware IDs, so sort them:
                 */
                for (i = 0; i < pc->attr_nr; i++) {
-                       if (pc->interp_mode[i] == INTERP_FLAT) {
-                               p->cfg.io[m].id_vp = i + base;
-                               p->cfg.io[m++].id_fp = i;
-                       } else {
+                       if (pc->interp_mode[i] == INTERP_FLAT)
+                               p->cfg.io[m++].id = i;
+                       else {
                                if (!(pc->interp_mode[i] & INTERP_PERSPECTIVE))
                                        p->cfg.io[n].linear = TRUE;
-                               p->cfg.io[n].id_vp = i + base;
-                               p->cfg.io[n++].id_fp = i;
+                               p->cfg.io[n++].id = i;
                        }
                }
 
@@ -2688,7 +2988,7 @@ nv50_program_tx_prep(struct nv50_pc *pc)
 
                for (n = 0; n < pc->attr_nr; ++n) {
                        p->cfg.io[n].hw = rid = aid;
-                       i = p->cfg.io[n].id_fp;
+                       i = p->cfg.io[n].id;
 
                        if (p->info.input_semantic_name[n] ==
                            TGSI_SEMANTIC_FACE) {
@@ -2728,8 +3028,8 @@ nv50_program_tx_prep(struct nv50_pc *pc)
                for (i = 0; i < pc->attr_nr; i++) {
                        ubyte si, sn;
 
-                       sn = p->info.input_semantic_name[p->cfg.io[i].id_fp];
-                       si = p->info.input_semantic_index[p->cfg.io[i].id_fp];
+                       sn = p->info.input_semantic_name[p->cfg.io[i].id];
+                       si = p->info.input_semantic_index[p->cfg.io[i].id];
 
                        if (sn == TGSI_SEMANTIC_COLOR) {
                                p->cfg.two_side[si] = p->cfg.io[i];
@@ -2820,6 +3120,8 @@ ctor_nv50_pc(struct nv50_pc *pc, struct nv50_program *p)
        p->cfg.two_side[0].hw = 0x40;
        p->cfg.two_side[1].hw = 0x40;
 
+       p->cfg.edgeflag_in = pc->edgeflag_out = 0xff;
+
        switch (p->type) {
        case PIPE_SHADER_VERTEX:
                p->cfg.psiz = 0x40;
@@ -2893,28 +3195,10 @@ ctor_nv50_pc(struct nv50_pc *pc, struct nv50_program *p)
        return TRUE;
 }
 
-static void
-nv50_fp_move_results(struct nv50_pc *pc)
-{
-       struct nv50_reg reg;
-       unsigned i;
-
-       ctor_reg(&reg, P_TEMP, -1, -1);
-
-       for (i = 0; i < pc->result_nr * 4; ++i) {
-               if (pc->result[i].rhw < 0 || pc->result[i].hw < 0)
-                       continue;
-               if (pc->result[i].rhw != pc->result[i].hw) {
-                       reg.hw = pc->result[i].rhw;
-                       emit_mov(pc, &reg, &pc->result[i]);
-               }
-       }
-}
-
 static void
 nv50_program_fixup_insns(struct nv50_pc *pc)
 {
-       struct nv50_program_exec *e, *prev = NULL, **bra_list;
+       struct nv50_program_exec *e, **bra_list;
        unsigned i, n, pos;
 
        bra_list = CALLOC(pc->p->exec_size, sizeof(struct nv50_program_exec *));
@@ -2926,6 +3210,16 @@ nv50_program_fixup_insns(struct nv50_pc *pc)
                if (e->param.index >= 0 && !e->param.mask)
                        bra_list[n++] = e;
 
+       /* last instruction must be long so it can have the exit bit set */
+       if (!is_long(pc->p->exec_tail))
+               convert_to_long(pc, pc->p->exec_tail);
+       /* set exit bit */
+       pc->p->exec_tail->inst[1] |= 1;
+
+       /* !immd on exit insn simultaneously means !join */
+       assert(!is_immd(pc->p->exec_head));
+       assert(!is_immd(pc->p->exec_tail));
+
        /* Make sure we don't have any single 32 bit instructions. */
        for (e = pc->p->exec_head, pos = 0; e; e = e->next) {
                pos += is_long(e) ? 2 : 1;
@@ -2937,22 +3231,7 @@ nv50_program_fixup_insns(struct nv50_pc *pc)
                        convert_to_long(pc, e);
                        ++pos;
                }
-               if (e->next)
-                       prev = e;
-       }
-
-       assert(!is_immd(pc->p->exec_head));
-       assert(!is_immd(pc->p->exec_tail));
-
-       /* last instruction must be long so it can have the end bit set */
-       if (!is_long(pc->p->exec_tail)) {
-               convert_to_long(pc, pc->p->exec_tail);
-               if (prev)
-                       convert_to_long(pc, prev);
        }
-       assert(!(pc->p->exec_tail->inst[1] & 2));
-       /* set the end-bit */
-       pc->p->exec_tail->inst[1] |= 1;
 
        FREE(bra_list);
 }
@@ -3025,7 +3304,7 @@ nv50_program_validate(struct nv50_context *nv50, struct nv50_program *p)
 }
 
 static void
-nv50_program_upload_data(struct nv50_context *nv50, float *map,
+nv50_program_upload_data(struct nv50_context *nv50, uint32_t *map,
                        unsigned start, unsigned count, unsigned cbuf)
 {
        struct nouveau_channel *chan = nv50->screen->base.channel;
@@ -3073,8 +3352,8 @@ nv50_program_validate_data(struct nv50_context *nv50, struct nv50_program *p)
 
        if (p->param_nr) {
                unsigned cb;
-               float *map = pipe_buffer_map(pscreen, nv50->constbuf[p->type],
-                                            PIPE_BUFFER_USAGE_CPU_READ);
+               uint32_t *map = pipe_buffer_map(pscreen, nv50->constbuf[p->type],
+                                               PIPE_BUFFER_USAGE_CPU_READ);
 
                if (p->type == PIPE_SHADER_VERTEX)
                        cb = NV50_CB_PVP;
@@ -3236,15 +3515,15 @@ nv50_pntc_replace(struct nv50_context *nv50, uint32_t pntc[8], unsigned base)
        struct nv50_program *vp = nv50->vertprog;
        unsigned i, c, m = base;
 
-       /* XXX: This can't work correctly in all cases yet, we either
-        * have to create TGSI_SEMANTIC_PNTC or sprite_coord_mode has
-        * to be per FP input instead of per VP output
+       /* XXX: this might not work correctly in all cases yet - we'll
+        * just assume that an FP generic input that is not written in
+        * the VP is PointCoord.
         */
        memset(pntc, 0, 8 * sizeof(uint32_t));
 
        for (i = 0; i < fp->cfg.io_nr; i++) {
                uint8_t sn, si;
-               uint8_t j = fp->cfg.io[i].id_vp, k = fp->cfg.io[i].id_fp;
+               uint8_t j, k = fp->cfg.io[i].id;
                unsigned n = popcnt4(fp->cfg.io[i].mask);
 
                if (fp->info.input_semantic_name[k] != TGSI_SEMANTIC_GENERIC) {
@@ -3252,10 +3531,16 @@ nv50_pntc_replace(struct nv50_context *nv50, uint32_t pntc[8], unsigned base)
                        continue;
                }
 
-               sn = vp->info.input_semantic_name[j];
-               si = vp->info.input_semantic_index[j];
+               for (j = 0; j < vp->info.num_outputs; ++j) {
+                       sn = vp->info.output_semantic_name[j];
+                       si = vp->info.output_semantic_index[j];
+
+                       if (sn == fp->info.input_semantic_name[k] &&
+                           si == fp->info.input_semantic_index[k])
+                               break;
+               }
 
-               if (j < fp->cfg.io_nr && sn == TGSI_SEMANTIC_GENERIC) {
+               if (j < vp->info.num_outputs) {
                        ubyte mode =
                                nv50->rasterizer->pipe.sprite_coord_mode[si];
 
@@ -3343,20 +3628,24 @@ nv50_linkage_validate(struct nv50_context *nv50)
        reg[0] += m - 4; /* adjust FFC0 id */
        reg[4] |= m << 8; /* set mid where 'normal' FP inputs start */
 
-       i = 0;
-       if (fp->info.input_semantic_name[0] == TGSI_SEMANTIC_POSITION)
-               i = 1;
-       for (; i < fp->cfg.io_nr; i++) {
-               ubyte sn = fp->info.input_semantic_name[fp->cfg.io[i].id_fp];
-               ubyte si = fp->info.input_semantic_index[fp->cfg.io[i].id_fp];
-
-               n = fp->cfg.io[i].id_vp;
-               if (n >= vp->cfg.io_nr ||
-                   vp->info.output_semantic_name[n] != sn ||
-                   vp->info.output_semantic_index[n] != si)
-                       vpo = &dummy;
-               else
-                       vpo = &vp->cfg.io[n];
+       for (i = 0; i < fp->cfg.io_nr; i++) {
+               ubyte sn = fp->info.input_semantic_name[fp->cfg.io[i].id];
+               ubyte si = fp->info.input_semantic_index[fp->cfg.io[i].id];
+
+               /* position must be mapped first */
+               assert(i == 0 || sn != TGSI_SEMANTIC_POSITION);
+
+               /* maybe even remove these from cfg.io */
+               if (sn == TGSI_SEMANTIC_POSITION || sn == TGSI_SEMANTIC_FACE)
+                       continue;
+
+               /* VP outputs and vp->cfg.io are in the same order */
+               for (n = 0; n < vp->info.num_outputs; ++n) {
+                       if (vp->info.output_semantic_name[n] == sn &&
+                           vp->info.output_semantic_index[n] == si)
+                               break;
+               }
+               vpo = (n < vp->info.num_outputs) ? &vp->cfg.io[n] : &dummy;
 
                m = nv50_sreg4_map(map, m, lin, &fp->cfg.io[i], vpo);
        }
index d78dee083f1452cf6ed725c48718ae7364278078..461fec1d89c323663602f52e89e4d23358bebdd9 100644 (file)
@@ -17,8 +17,7 @@ struct nv50_program_exec {
 
 struct nv50_sreg4 {
        uint8_t hw;
-       uint8_t id_vp;
-       uint8_t id_fp;
+       uint8_t id; /* tgsi index, nv50 needs them sorted: flat ones last */
 
        uint8_t mask;
        boolean linear;
@@ -38,7 +37,7 @@ struct nv50_program {
 
        struct nouveau_bo *bo;
 
-       float *immd;
+       uint32_t *immd;
        unsigned immd_nr;
        unsigned param_nr;
 
@@ -59,6 +58,7 @@ struct nv50_program {
                /* VP only */
                uint8_t clpd, clpd_nr;
                uint8_t psiz;
+               uint8_t edgeflag_in;
        } cfg;
 };
 
index 5305c93d59ab15de76094c8a14e667d8d2b335a2..268c9823f7d8887ceab9b3022662d61a6c372da8 100644 (file)
@@ -93,7 +93,7 @@ nv50_query_end(struct pipe_context *pipe, struct pipe_query *pq)
        struct nouveau_grobj *tesla = nv50->screen->tesla;
        struct nv50_query *q = nv50_query(pq);
 
-       WAIT_RING (chan, 5);
+       MARK_RING (chan, 5, 2); /* flush on lack of space or relocs */
        BEGIN_RING(chan, tesla, NV50TCL_QUERY_ADDRESS_HIGH, 4);
        OUT_RELOCh(chan, q->bo, 0, NOUVEAU_BO_VRAM | NOUVEAU_BO_WR);
        OUT_RELOCl(chan, q->bo, 0, NOUVEAU_BO_VRAM | NOUVEAU_BO_WR);
index e1b2f11239a4e1cc3859c98d112d3b59818878ce..5a1efd3998bc2fed34237ae18bc98c9df67576b2 100644 (file)
@@ -76,6 +76,7 @@ nv50_screen_is_format_supported(struct pipe_screen *pscreen,
                case PIPE_FORMAT_DXT3_RGBA:
                case PIPE_FORMAT_DXT5_RGBA:
                case PIPE_FORMAT_Z24S8_UNORM:
+               case PIPE_FORMAT_S8Z24_UNORM:
                case PIPE_FORMAT_Z32_FLOAT:
                case PIPE_FORMAT_R16G16B16A16_SNORM:
                case PIPE_FORMAT_R16G16B16A16_UNORM:
@@ -97,6 +98,10 @@ nv50_screen_get_param(struct pipe_screen *pscreen, int param)
        switch (param) {
        case PIPE_CAP_MAX_TEXTURE_IMAGE_UNITS:
                return 32;
+       case PIPE_CAP_MAX_VERTEX_TEXTURE_UNITS:
+               return 32;
+       case PIPE_CAP_MAX_COMBINED_SAMPLERS:
+               return 64;
        case PIPE_CAP_NPOT_TEXTURES:
                return 1;
        case PIPE_CAP_TWO_SIDED_STENCIL:
@@ -122,10 +127,8 @@ nv50_screen_get_param(struct pipe_screen *pscreen, int param)
        case PIPE_CAP_TEXTURE_MIRROR_CLAMP:
        case PIPE_CAP_TEXTURE_MIRROR_REPEAT:
                return 1;
-       case PIPE_CAP_MAX_VERTEX_TEXTURE_UNITS:
-               return 0;
        case PIPE_CAP_TGSI_CONT_SUPPORTED:
-               return 0;
+               return 1;
        case PIPE_CAP_BLEND_EQUATION_SEPARATE:
                return 1;
        case NOUVEAU_CAP_HW_VTXBUF:
@@ -315,6 +318,9 @@ nv50_screen_create(struct pipe_winsys *ws, struct nouveau_device *dev)
        so_method(so, screen->tesla, 0x1400, 1);
        so_data  (so, 0xf);
 
+       /* max TIC (bits 4:8) & TSC (ignored) bindings, per program type */
+       so_method(so, screen->tesla, 0x13b4, 1);
+       so_data  (so, 0x54);
        so_method(so, screen->tesla, 0x13bc, 1);
        so_data  (so, 0x54);
        /* origin is top left (set to 1 for bottom left) */
@@ -387,7 +393,8 @@ nv50_screen_create(struct pipe_winsys *ws, struct nouveau_device *dev)
        so_method(so, screen->tesla, NV50TCL_SET_PROGRAM_CB, 1);
        so_data  (so, 0x00000131 | (NV50_CB_PFP << 12));
 
-       ret = nouveau_bo_new(dev, NOUVEAU_BO_VRAM, 0, 64*8*4, &screen->tic);
+       ret = nouveau_bo_new(dev, NOUVEAU_BO_VRAM, 0, PIPE_SHADER_TYPES*32*32,
+                            &screen->tic);
        if (ret) {
                nv50_screen_destroy(pscreen);
                return NULL;
@@ -398,9 +405,10 @@ nv50_screen_create(struct pipe_winsys *ws, struct nouveau_device *dev)
                  NOUVEAU_BO_RD | NOUVEAU_BO_HIGH, 0, 0);
        so_reloc (so, screen->tic, 0, NOUVEAU_BO_VRAM |
                  NOUVEAU_BO_RD | NOUVEAU_BO_LOW, 0, 0);
-       so_data  (so, 0x000007ff);
+       so_data  (so, PIPE_SHADER_TYPES * 32 - 1);
 
-       ret = nouveau_bo_new(dev, NOUVEAU_BO_VRAM, 0, 64*8*4, &screen->tsc);
+       ret = nouveau_bo_new(dev, NOUVEAU_BO_VRAM, 0, PIPE_SHADER_TYPES*32*32,
+                            &screen->tsc);
        if (ret) {
                nv50_screen_destroy(pscreen);
                return NULL;
@@ -411,7 +419,7 @@ nv50_screen_create(struct pipe_winsys *ws, struct nouveau_device *dev)
                  NOUVEAU_BO_RD | NOUVEAU_BO_HIGH, 0, 0);
        so_reloc (so, screen->tsc, 0, NOUVEAU_BO_VRAM |
                  NOUVEAU_BO_RD | NOUVEAU_BO_LOW, 0, 0);
-       so_data  (so, 0x00000000);
+       so_data  (so, 0x00000000); /* ignored if TSC_LINKED (0x1234) = 1 */
 
 
        /* Vertex array limits - max them out */
@@ -425,6 +433,7 @@ nv50_screen_create(struct pipe_winsys *ws, struct nouveau_device *dev)
        so_data  (so, fui(0.0));
        so_data  (so, fui(1.0));
 
+       /* no dynamic combination of TIC & TSC entries => only BIND_TIC used */
        so_method(so, screen->tesla, 0x1234, 1);
        so_data  (so, 1);
 
@@ -432,6 +441,9 @@ nv50_screen_create(struct pipe_winsys *ws, struct nouveau_device *dev)
        so_method(so, screen->tesla, NV50TCL_SCISSOR_ENABLE, 1);
        so_data  (so, 1);
 
+       so_method(so, screen->tesla, 0x15e4, 1);
+       so_data  (so, 1); /* default edgeflag to TRUE */
+
        so_emit(chan, so);
        so_ref (so, &screen->static_init);
        so_ref (NULL, &so);
index ffaa5e29d1c22286a422efa1afa3179cfd65e6f3..88aef52d08c6bb57218cea85a52d9af073581b7c 100644 (file)
@@ -196,8 +196,9 @@ nv50_sampler_state_create(struct pipe_context *pipe,
        }
 
        if (cso->compare_mode == PIPE_TEX_COMPARE_R_TO_TEXTURE) {
-               tsc[0] |= (1 << 8);
-               tsc[0] |= (nvgl_comparison_op(cso->compare_func) & 0x7);
+               /* XXX: must be deactivated for non-shadow textures */
+               tsc[0] |= (1 << 9);
+               tsc[0] |= (nvgl_comparison_op(cso->compare_func) & 0x7) << 10;
        }
 
        limit = CLAMP(cso->lod_bias, -16.0, 15.0);
@@ -215,41 +216,66 @@ nv50_sampler_state_create(struct pipe_context *pipe,
        return (void *)sso;
 }
 
-static void
-nv50_sampler_state_bind(struct pipe_context *pipe, unsigned nr, void **sampler)
+static INLINE void
+nv50_sampler_state_bind(struct pipe_context *pipe, unsigned type,
+                       unsigned nr, void **sampler)
 {
        struct nv50_context *nv50 = nv50_context(pipe);
-       int i;
 
-       nv50->sampler_nr = nr;
-       for (i = 0; i < nv50->sampler_nr; i++)
-               nv50->sampler[i] = sampler[i];
+       memcpy(nv50->sampler[type], sampler, nr * sizeof(void *));
 
+       nv50->sampler_nr[type] = nr;
        nv50->dirty |= NV50_NEW_SAMPLER;
 }
 
+static void
+nv50_vp_sampler_state_bind(struct pipe_context *pipe, unsigned nr, void **s)
+{
+       nv50_sampler_state_bind(pipe, PIPE_SHADER_VERTEX, nr, s);
+}
+
+static void
+nv50_fp_sampler_state_bind(struct pipe_context *pipe, unsigned nr, void **s)
+{
+       nv50_sampler_state_bind(pipe, PIPE_SHADER_FRAGMENT, nr, s);
+}
+
 static void
 nv50_sampler_state_delete(struct pipe_context *pipe, void *hwcso)
 {
        FREE(hwcso);
 }
 
-static void
-nv50_set_sampler_texture(struct pipe_context *pipe, unsigned nr,
-                        struct pipe_texture **pt)
+static INLINE void
+nv50_set_sampler_texture(struct pipe_context *pipe, unsigned type,
+                        unsigned nr, struct pipe_texture **pt)
 {
        struct nv50_context *nv50 = nv50_context(pipe);
-       int i;
+       unsigned i;
 
        for (i = 0; i < nr; i++)
-               pipe_texture_reference((void *)&nv50->miptree[i], pt[i]);
-       for (i = nr; i < nv50->miptree_nr; i++)
-               pipe_texture_reference((void *)&nv50->miptree[i], NULL);
+               pipe_texture_reference((void *)&nv50->miptree[type][i], pt[i]);
+       for (i = nr; i < nv50->miptree_nr[type]; i++)
+               pipe_texture_reference((void *)&nv50->miptree[type][i], NULL);
 
-       nv50->miptree_nr = nr;
+       nv50->miptree_nr[type] = nr;
        nv50->dirty |= NV50_NEW_TEXTURE;
 }
 
+static void
+nv50_set_vp_sampler_textures(struct pipe_context *pipe,
+                            unsigned nr, struct pipe_texture **pt)
+{
+       nv50_set_sampler_texture(pipe, PIPE_SHADER_VERTEX, nr, pt);
+}
+
+static void
+nv50_set_fp_sampler_textures(struct pipe_context *pipe,
+                            unsigned nr, struct pipe_texture **pt)
+{
+       nv50_set_sampler_texture(pipe, PIPE_SHADER_FRAGMENT, nr, pt);
+}
+
 static void *
 nv50_rasterizer_state_create(struct pipe_context *pipe,
                             const struct pipe_rasterizer_state *cso)
@@ -648,9 +674,11 @@ nv50_init_state_functions(struct nv50_context *nv50)
        nv50->pipe.delete_blend_state = nv50_blend_state_delete;
 
        nv50->pipe.create_sampler_state = nv50_sampler_state_create;
-       nv50->pipe.bind_sampler_states = nv50_sampler_state_bind;
        nv50->pipe.delete_sampler_state = nv50_sampler_state_delete;
-       nv50->pipe.set_sampler_textures = nv50_set_sampler_texture;
+       nv50->pipe.bind_fragment_sampler_states = nv50_fp_sampler_state_bind;
+       nv50->pipe.bind_vertex_sampler_states   = nv50_vp_sampler_state_bind;
+       nv50->pipe.set_fragment_sampler_textures = nv50_set_fp_sampler_textures;
+       nv50->pipe.set_vertex_sampler_textures   = nv50_set_vp_sampler_textures;
 
        nv50->pipe.create_rasterizer_state = nv50_rasterizer_state_create;
        nv50->pipe.bind_rasterizer_state = nv50_rasterizer_state_bind;
index 799d2758fee17ee6468136f1c4e35acb816c66eb..871e8097b65a9650485c512946b4e6c1adad7dfa 100644 (file)
@@ -155,6 +155,30 @@ nv50_state_validate_fb(struct nv50_context *nv50)
        so_ref(NULL, &so);
 }
 
+static void
+nv50_validate_samplers(struct nv50_context *nv50, struct nouveau_stateobj *so,
+                      unsigned p)
+{
+       struct nouveau_grobj *eng2d = nv50->screen->eng2d;
+       unsigned i, j, dw = nv50->sampler_nr[p] * 8;
+
+       if (!dw)
+               return;
+       nv50_so_init_sifc(nv50, so, nv50->screen->tsc, NOUVEAU_BO_VRAM,
+                         p * (32 * 8 * 4), dw * 4);
+
+       so_method(so, eng2d, NV50_2D_SIFC_DATA | (2 << 29), dw);
+
+       for (i = 0; i < nv50->sampler_nr[p]; ++i) {
+               if (nv50->sampler[p][i])
+                       so_datap(so, nv50->sampler[p][i]->tsc, 8);
+               else {
+                       for (j = 0; j < 8; ++j) /* you get punished */
+                               so_data(so, 0); /* ... for leaving holes */
+               }
+       }
+}
+
 static void
 nv50_state_emit(struct nv50_context *nv50)
 {
@@ -201,7 +225,8 @@ nv50_state_emit(struct nv50_context *nv50)
                so_emit(chan, nv50->state.vertprog);
        if (nv50->state.dirty & NV50_NEW_FRAGPROG)
                so_emit(chan, nv50->state.fragprog);
-       if (nv50->state.dirty & (NV50_NEW_FRAGPROG | NV50_NEW_VERTPROG))
+       if (nv50->state.dirty & (NV50_NEW_FRAGPROG | NV50_NEW_VERTPROG |
+                                NV50_NEW_RASTERIZER))
                so_emit(chan, nv50->state.programs);
        if (nv50->state.dirty & NV50_NEW_RASTERIZER)
                so_emit(chan, nv50->state.rast);
@@ -245,7 +270,6 @@ boolean
 nv50_state_validate(struct nv50_context *nv50)
 {
        struct nouveau_grobj *tesla = nv50->screen->tesla;
-       struct nouveau_grobj *eng2d = nv50->screen->eng2d;
        struct nouveau_stateobj *so;
        unsigned i;
 
@@ -264,7 +288,8 @@ nv50_state_validate(struct nv50_context *nv50)
        if (nv50->dirty & (NV50_NEW_FRAGPROG | NV50_NEW_FRAGPROG_CB))
                nv50_fragprog_validate(nv50);
 
-       if (nv50->dirty & (NV50_NEW_FRAGPROG | NV50_NEW_VERTPROG))
+       if (nv50->dirty & (NV50_NEW_FRAGPROG | NV50_NEW_VERTPROG |
+                          NV50_NEW_RASTERIZER))
                nv50_linkage_validate(nv50);
 
        if (nv50->dirty & NV50_NEW_RASTERIZER)
@@ -285,7 +310,7 @@ nv50_state_validate(struct nv50_context *nv50)
                so = so_new(33, 0);
                so_method(so, tesla, NV50TCL_POLYGON_STIPPLE_PATTERN(0), 32);
                for (i = 0; i < 32; i++)
-                       so_data(so, nv50->stipple.stipple[i]);
+                       so_data(so, util_bswap32(nv50->stipple.stipple[i]));
                so_ref(so, &nv50->state.stipple);
                so_ref(NULL, &so);
        }
@@ -367,22 +392,16 @@ scissor_uptodate:
 viewport_uptodate:
 
        if (nv50->dirty & NV50_NEW_SAMPLER) {
-               unsigned i;
+               unsigned nr = 0;
 
-               so = so_new(nv50->sampler_nr * 9 + 23 + 4, 2);
+               for (i = 0; i < PIPE_SHADER_TYPES; ++i)
+                       nr += nv50->sampler_nr[i];
 
-               nv50_so_init_sifc(nv50, so, nv50->screen->tsc, NOUVEAU_BO_VRAM,
-                                 nv50->sampler_nr * 8 * 4);
+               so = so_new(nr * 8 + 24 * PIPE_SHADER_TYPES + 2, 4);
 
-               for (i = 0; i < nv50->sampler_nr; i++) {
-                       if (!nv50->sampler[i])
-                               continue;
-                       so_method(so, eng2d, NV50_2D_SIFC_DATA | (2 << 29), 8);
-                       so_datap (so, nv50->sampler[i]->tsc, 8);
-               }
+               nv50_validate_samplers(nv50, so, PIPE_SHADER_VERTEX);
+               nv50_validate_samplers(nv50, so, PIPE_SHADER_FRAGMENT);
 
-               so_method(so, tesla, 0x1440, 1); /* sync SIFC */
-               so_data  (so, 0);
                so_method(so, tesla, 0x1334, 1); /* flush TSC */
                so_data  (so, 0);
 
@@ -405,10 +424,13 @@ viewport_uptodate:
 
 void nv50_so_init_sifc(struct nv50_context *nv50,
                       struct nouveau_stateobj *so,
-                      struct nouveau_bo *bo, unsigned reloc, unsigned size)
+                      struct nouveau_bo *bo, unsigned reloc,
+                      unsigned offset, unsigned size)
 {
        struct nouveau_grobj *eng2d = nv50->screen->eng2d;
 
+       reloc |= NOUVEAU_BO_WR;
+
        so_method(so, eng2d, NV50_2D_DST_FORMAT, 2);
        so_data  (so, NV50_2D_DST_FORMAT_R8_UNORM);
        so_data  (so, 1);
@@ -416,8 +438,8 @@ void nv50_so_init_sifc(struct nv50_context *nv50,
        so_data  (so, 262144);
        so_data  (so, 65536);
        so_data  (so, 1);
-       so_reloc (so, bo, 0, reloc | NOUVEAU_BO_WR | NOUVEAU_BO_HIGH, 0, 0);
-       so_reloc (so, bo, 0, reloc | NOUVEAU_BO_WR | NOUVEAU_BO_LOW, 0, 0);
+       so_reloc (so, bo, offset, reloc | NOUVEAU_BO_HIGH, 0, 0);
+       so_reloc (so, bo, offset, reloc | NOUVEAU_BO_LOW, 0, 0);
        so_method(so, eng2d, NV50_2D_SIFC_UNK0800, 2);
        so_data  (so, 0);
        so_data  (so, NV50_2D_SIFC_FORMAT_R8_UNORM);
index 6bf6f773b0c26b075282175c1d3292a1e3ddd753..79655fc08d5a6b782b33eadb5916d0a6697fed19 100644 (file)
@@ -62,6 +62,7 @@ nv50_surface_set(struct nv50_screen *screen, struct pipe_surface *ps, int dst)
                return 1;
 
        if (!bo->tile_flags) {
+               MARK_RING (chan, 9, 2); /* flush on lack of space or relocs */
                BEGIN_RING(chan, eng2d, mthd, 2);
                OUT_RING  (chan, format);
                OUT_RING  (chan, 1);
@@ -72,6 +73,7 @@ nv50_surface_set(struct nv50_screen *screen, struct pipe_surface *ps, int dst)
                OUT_RELOCh(chan, bo, ps->offset, flags);
                OUT_RELOCl(chan, bo, ps->offset, flags);
        } else {
+               MARK_RING (chan, 11, 2); /* flush on lack of space or relocs */
                BEGIN_RING(chan, eng2d, mthd, 5);
                OUT_RING  (chan, format);
                OUT_RING  (chan, 0);
index 2813f5447704babcf196458c4a10f9930d9c8230..c4ca096d6ac5ef726c6f196d57f5a926250e64e7 100644 (file)
@@ -25,6 +25,8 @@
 
 #include "nouveau/nouveau_stateobj.h"
 
+#include "util/u_format.h"
+
 #define _MIXED(pf, t0, t1, t2, t3, cr, cg, cb, ca, f)          \
 {                                                              \
        PIPE_FORMAT_##pf,                                       \
@@ -68,6 +70,7 @@ static const struct nv50_texture_format nv50_tex_format_list[] =
        _(DXT5_RGBA, UNORM, C0, C1, C2, C3, DXT5),
 
        _MIXED(Z24S8_UNORM, UINT, UNORM, UINT, UINT, C1, C1, C1, ONE, 24_8),
+       _MIXED(S8Z24_UNORM, UNORM, UINT, UINT, UINT, C0, C0, C0, ONE, 8_24),
 
        _(R16G16B16A16_SNORM, UNORM, C0, C1, C2, C3, 16_16_16_16),
        _(R16G16B16A16_UNORM, SNORM, C0, C1, C2, C3, 16_16_16_16),
@@ -85,10 +88,11 @@ static const struct nv50_texture_format nv50_tex_format_list[] =
 
 static int
 nv50_tex_construct(struct nv50_context *nv50, struct nouveau_stateobj *so,
-                  struct nv50_miptree *mt, int unit)
+                  struct nv50_miptree *mt, int unit, unsigned p)
 {
        unsigned i;
        uint32_t mode;
+       const struct util_format_description *desc;
 
        for (i = 0; i < NV50_TEX_FORMAT_LIST_SIZE; i++)
                if (nv50_tex_format_list[i].pf == mt->base.base.format)
@@ -96,7 +100,7 @@ nv50_tex_construct(struct nv50_context *nv50, struct nouveau_stateobj *so,
        if (i == NV50_TEX_FORMAT_LIST_SIZE)
                 return 1;
 
-       if (nv50->sampler[unit]->normalized)
+       if (nv50->sampler[p][unit]->normalized)
                mode = 0x50001000 | (1 << 31);
        else {
                mode = 0x50001000 | (7 << 14);
@@ -106,7 +110,10 @@ nv50_tex_construct(struct nv50_context *nv50, struct nouveau_stateobj *so,
        mode |= ((mt->base.bo->tile_mode & 0x0f) << 22) |
                ((mt->base.bo->tile_mode & 0xf0) << 21);
 
-       if (pf_type(mt->base.base.format) == PIPE_FORMAT_TYPE_SRGB)
+       desc = util_format_description(mt->base.base.format);
+       assert(desc);
+
+       if (desc->colorspace == UTIL_FORMAT_COLORSPACE_SRGB)
                mode |= 0x0400;
 
        switch (mt->base.base.target) {
@@ -131,57 +138,87 @@ nv50_tex_construct(struct nv50_context *nv50, struct nouveau_stateobj *so,
                 NOUVEAU_BO_RD, 0, 0);
        so_data (so, mode);
        so_data (so, 0x00300000);
-       so_data (so, mt->base.base.width[0] | (1 << 31));
+       so_data (so, mt->base.base.width0 | (1 << 31));
        so_data (so, (mt->base.base.last_level << 28) |
-                (mt->base.base.depth[0] << 16) | mt->base.base.height[0]);
+                (mt->base.base.depth0 << 16) | mt->base.base.height0);
        so_data (so, 0x03000000);
        so_data (so, mt->base.base.last_level << 4);
 
        return 0;
 }
 
-void
-nv50_tex_validate(struct nv50_context *nv50)
+#ifndef NV50TCL_BIND_TIC
+#define NV50TCL_BIND_TIC(n) (0x1448 + 8 * n)
+#endif
+
+static boolean
+nv50_validate_textures(struct nv50_context *nv50, struct nouveau_stateobj *so,
+                      unsigned p)
 {
+       static const unsigned p_remap[PIPE_SHADER_TYPES] = { 0, 2 };
+
        struct nouveau_grobj *eng2d = nv50->screen->eng2d;
        struct nouveau_grobj *tesla = nv50->screen->tesla;
-       struct nouveau_stateobj *so;
-       unsigned i, unit, push;
-
-       push = MAX2(nv50->miptree_nr, nv50->state.miptree_nr) * 2 + 23 + 6;
-       so = so_new(nv50->miptree_nr * 9 + push, nv50->miptree_nr * 2 + 2);
+       unsigned unit, j, p_hw = p_remap[p];
 
        nv50_so_init_sifc(nv50, so, nv50->screen->tic, NOUVEAU_BO_VRAM,
-                         nv50->miptree_nr * 8 * 4);
-
-       for (i = 0, unit = 0; unit < nv50->miptree_nr; ++unit) {
-               struct nv50_miptree *mt = nv50->miptree[unit];
+                         p * (32 * 8 * 4), nv50->miptree_nr[p] * 8 * 4);
 
-               if (!mt)
-                       continue;
+       for (unit = 0; unit < nv50->miptree_nr[p]; ++unit) {
+               struct nv50_miptree *mt = nv50->miptree[p][unit];
 
                so_method(so, eng2d, NV50_2D_SIFC_DATA | (2 << 29), 8);
-               if (nv50_tex_construct(nv50, so, mt, unit)) {
-                       NOUVEAU_ERR("failed tex validate\n");
-                       so_ref(NULL, &so);
-                       return;
+               if (mt) {
+                       if (nv50_tex_construct(nv50, so, mt, unit, p))
+                               return FALSE;
+                       /* Set TEX insn $t src binding $unit in program type p
+                        * to TIC, TSC entry (32 * p + unit), mark valid (1).
+                        */
+                       so_method(so, tesla, NV50TCL_BIND_TIC(p_hw), 1);
+                       so_data  (so, ((32 * p + unit) << 9) | (unit << 1) | 1);
+               } else {
+                       for (j = 0; j < 8; ++j)
+                               so_data(so, 0);
+                       so_method(so, tesla, NV50TCL_BIND_TIC(p_hw), 1);
+                       so_data  (so, (unit << 1) | 0);
                }
+       }
+
+       for (; unit < nv50->state.miptree_nr[p]; unit++) {
+               /* Make other bindings invalid. */
+               so_method(so, tesla, NV50TCL_BIND_TIC(p_hw), 1);
+               so_data  (so, (unit << 1) | 0);
+       }
 
-               so_method(so, tesla, NV50TCL_SET_SAMPLER_TEX, 1);
-               so_data  (so, (i++ << NV50TCL_SET_SAMPLER_TEX_TIC_SHIFT) |
-                         (unit << NV50TCL_SET_SAMPLER_TEX_SAMPLER_SHIFT) |
-                         NV50TCL_SET_SAMPLER_TEX_VALID);
+       nv50->state.miptree_nr[p] = nv50->miptree_nr[p];
+       return TRUE;
+}
+
+void
+nv50_tex_validate(struct nv50_context *nv50)
+{
+       struct nouveau_stateobj *so;
+       struct nouveau_grobj *tesla = nv50->screen->tesla;
+       unsigned p, push, nrlc;
+
+       for (nrlc = 0, push = 0, p = 0; p < PIPE_SHADER_TYPES; ++p) {
+               push += MAX2(nv50->miptree_nr[p], nv50->state.miptree_nr[p]);
+               nrlc += nv50->miptree_nr[p];
        }
+       push = push * 11 + 23 * PIPE_SHADER_TYPES + 4;
+       nrlc = nrlc * 2 + 2 * PIPE_SHADER_TYPES;
+
+       so = so_new(push, nrlc);
 
-       for (; unit < nv50->state.miptree_nr; unit++) {
-               so_method(so, tesla, NV50TCL_SET_SAMPLER_TEX, 1);
-               so_data  (so,
-                         (unit << NV50TCL_SET_SAMPLER_TEX_SAMPLER_SHIFT) | 0);
+       if (nv50_validate_textures(nv50, so, PIPE_SHADER_VERTEX) == FALSE ||
+           nv50_validate_textures(nv50, so, PIPE_SHADER_FRAGMENT) == FALSE) {
+               so_ref(NULL, &so);
+
+               NOUVEAU_ERR("failed tex validate\n");
+               return;
        }
 
        /* not sure if the following really do what I think: */
-       so_method(so, tesla, 0x1440, 1); /* sync SIFC */
-       so_data  (so, 0);
        so_method(so, tesla, 0x1330, 1); /* flush TIC */
        so_data  (so, 0);
        so_method(so, tesla, 0x1338, 1); /* flush texture caches */
@@ -189,6 +226,4 @@ nv50_tex_validate(struct nv50_context *nv50)
 
        so_ref(so, &nv50->state.tic_upload);
        so_ref(NULL, &so);
-       nv50->state.miptree_nr = nv50->miptree_nr;
 }
-
index d531e6113278463b2f2915e72f2d86843ca6cfb9..b870302019a80359d63eba90e176d5e4045f2ee2 100644 (file)
@@ -82,6 +82,7 @@
 #define NV50TIC_0_0_FMT_RGTC1                                     0x00000027
 #define NV50TIC_0_0_FMT_RGTC2                                     0x00000028
 #define NV50TIC_0_0_FMT_24_8                                      0x00000029
+#define NV50TIC_0_0_FMT_8_24                                      0x0000002a
 #define NV50TIC_0_0_FMT_32_DEPTH                                  0x0000002f
 #define NV50TIC_0_0_FMT_32_8                                      0x00000030
 
index ea61357aaa6bb8faa61747ffa56a5810375e834e..4d9afa6fedc7342d7b53832ab2ecb166df295c2a 100644 (file)
@@ -1,6 +1,8 @@
 
 #include "pipe/p_context.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
 
 #include "nv50_context.h"
 
@@ -15,16 +17,19 @@ struct nv50_transfer {
        int level_depth;
        int level_x;
        int level_y;
+       int level_z;
+       unsigned nblocksx;
+       unsigned nblocksy;
 };
 
 static void
 nv50_transfer_rect_m2mf(struct pipe_screen *pscreen,
                        struct nouveau_bo *src_bo, unsigned src_offset,
                        int src_pitch, unsigned src_tile_mode,
-                       int sx, int sy, int sw, int sh, int sd,
+                       int sx, int sy, int sz, int sw, int sh, int sd,
                        struct nouveau_bo *dst_bo, unsigned dst_offset,
                        int dst_pitch, unsigned dst_tile_mode,
-                       int dx, int dy, int dw, int dh, int dd,
+                       int dx, int dy, int dz, int dw, int dh, int dd,
                        int cpp, int width, int height,
                        unsigned src_reloc, unsigned dst_reloc)
 {
@@ -53,7 +58,7 @@ nv50_transfer_rect_m2mf(struct pipe_screen *pscreen,
                OUT_RING  (chan, sw * cpp);
                OUT_RING  (chan, sh);
                OUT_RING  (chan, sd);
-               OUT_RING  (chan, 0);
+               OUT_RING  (chan, sz); /* copying only 1 zslice per call */
        }
 
        if (!dst_bo->tile_flags) {
@@ -72,13 +77,13 @@ nv50_transfer_rect_m2mf(struct pipe_screen *pscreen,
                OUT_RING  (chan, dw * cpp);
                OUT_RING  (chan, dh);
                OUT_RING  (chan, dd);
-               OUT_RING  (chan, 0);
+               OUT_RING  (chan, dz); /* copying only 1 zslice per call */
        }
 
        while (height) {
                int line_count = height > 2047 ? 2047 : height;
 
-               WAIT_RING (chan, 15);
+               MARK_RING (chan, 15, 4); /* flush on lack of space or relocs */
                BEGIN_RING(chan, m2mf,
                        NV50_MEMORY_TO_MEMORY_FORMAT_OFFSET_IN_HIGH, 2);
                OUT_RELOCh(chan, src_bo, src_offset, src_reloc);
@@ -115,20 +120,6 @@ nv50_transfer_rect_m2mf(struct pipe_screen *pscreen,
        }
 }
 
-static INLINE unsigned
-get_zslice_offset(unsigned tile_mode, unsigned z, unsigned pitch, unsigned ny)
-{
-       unsigned tile_h = get_tile_height(tile_mode);
-       unsigned tile_d = get_tile_depth(tile_mode);
-
-       /* pitch_2d == to next slice within this volume-tile */
-       /* pitch_3d == to next slice in next 2D array of blocks */
-       unsigned pitch_2d = tile_h * 64;
-       unsigned pitch_3d = tile_d * align(ny, tile_h) * pitch;
-
-       return (z % tile_d) * pitch_2d + (z / tile_d) * pitch_3d;
-}
-
 static struct pipe_transfer *
 nv50_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
                  unsigned face, unsigned level, unsigned zslice,
@@ -150,56 +141,43 @@ nv50_transfer_new(struct pipe_screen *pscreen, struct pipe_texture *pt,
                return NULL;
 
        pipe_texture_reference(&tx->base.texture, pt);
-       tx->base.format = pt->format;
+       tx->nblocksx = util_format_get_nblocksx(pt->format, u_minify(pt->width0, level));
+       tx->nblocksy = util_format_get_nblocksy(pt->format, u_minify(pt->height0, level));
        tx->base.width = w;
        tx->base.height = h;
-       tx->base.block = pt->block;
-       if (!pt->nblocksx[level]) {
-               tx->base.nblocksx = pf_get_nblocksx(&pt->block,
-                                                   pt->width[level]);
-               tx->base.nblocksy = pf_get_nblocksy(&pt->block,
-                                                   pt->height[level]);
-       } else {
-               tx->base.nblocksx = pt->nblocksx[level];
-               tx->base.nblocksy = pt->nblocksy[level];
-       }
-       tx->base.stride = tx->base.nblocksx * pt->block.size;
+       tx->base.stride = tx->nblocksx * util_format_get_blocksize(pt->format);
        tx->base.usage = usage;
 
        tx->level_pitch = lvl->pitch;
-       tx->level_width = mt->base.base.width[level];
-       tx->level_height = mt->base.base.height[level];
-       tx->level_depth = mt->base.base.depth[level];
+       tx->level_width = u_minify(mt->base.base.width0, level);
+       tx->level_height = u_minify(mt->base.base.height0, level);
+       tx->level_depth = u_minify(mt->base.base.depth0, level);
        tx->level_offset = lvl->image_offset[image];
        tx->level_tiling = lvl->tile_mode;
-       tx->level_x = pf_get_nblocksx(&tx->base.block, x);
-       tx->level_y = pf_get_nblocksy(&tx->base.block, y);
+       tx->level_z = zslice;
+       tx->level_x = util_format_get_nblocksx(pt->format, x);
+       tx->level_y = util_format_get_nblocksy(pt->format, y);
        ret = nouveau_bo_new(dev, NOUVEAU_BO_GART | NOUVEAU_BO_MAP, 0,
-                            tx->base.nblocksy * tx->base.stride, &tx->bo);
+                            tx->nblocksy * tx->base.stride, &tx->bo);
        if (ret) {
                FREE(tx);
                return NULL;
        }
 
-       if (pt->target == PIPE_TEXTURE_3D)
-               tx->level_offset += get_zslice_offset(lvl->tile_mode, zslice,
-                                                     lvl->pitch,
-                                                     tx->base.nblocksy);
-
        if (usage & PIPE_TRANSFER_READ) {
-               nx = pf_get_nblocksx(&tx->base.block, tx->base.width);
-               ny = pf_get_nblocksy(&tx->base.block, tx->base.height);
+               nx = util_format_get_nblocksx(pt->format, tx->base.width);
+               ny = util_format_get_nblocksy(pt->format, tx->base.height);
 
                nv50_transfer_rect_m2mf(pscreen, mt->base.bo, tx->level_offset,
                                        tx->level_pitch, tx->level_tiling,
-                                       x, y,
-                                       tx->base.nblocksx, tx->base.nblocksy,
+                                       x, y, zslice,
+                                       tx->nblocksx, tx->nblocksy,
                                        tx->level_depth,
                                        tx->bo, 0,
                                        tx->base.stride, tx->bo->tile_mode,
-                                       0, 0,
-                                       tx->base.nblocksx, tx->base.nblocksy, 1,
-                                       tx->base.block.size, nx, ny,
+                                       0, 0, 0,
+                                       tx->nblocksx, tx->nblocksy, 1,
+                                       util_format_get_blocksize(pt->format), nx, ny,
                                        NOUVEAU_BO_VRAM | NOUVEAU_BO_GART,
                                        NOUVEAU_BO_GART);
        }
@@ -212,23 +190,24 @@ nv50_transfer_del(struct pipe_transfer *ptx)
 {
        struct nv50_transfer *tx = (struct nv50_transfer *)ptx;
        struct nv50_miptree *mt = nv50_miptree(ptx->texture);
+       struct pipe_texture *pt = ptx->texture;
 
-       unsigned nx = pf_get_nblocksx(&tx->base.block, tx->base.width);
-       unsigned ny = pf_get_nblocksy(&tx->base.block, tx->base.height);
+       unsigned nx = util_format_get_nblocksx(pt->format, tx->base.width);
+       unsigned ny = util_format_get_nblocksy(pt->format, tx->base.height);
 
        if (ptx->usage & PIPE_TRANSFER_WRITE) {
-               struct pipe_screen *pscreen = ptx->texture->screen;
+               struct pipe_screen *pscreen = pt->screen;
 
                nv50_transfer_rect_m2mf(pscreen, tx->bo, 0,
                                        tx->base.stride, tx->bo->tile_mode,
-                                       0, 0,
-                                       tx->base.nblocksx, tx->base.nblocksy, 1,
+                                       0, 0, 0,
+                                       tx->nblocksx, tx->nblocksy, 1,
                                        mt->base.bo, tx->level_offset,
                                        tx->level_pitch, tx->level_tiling,
-                                       tx->level_x, tx->level_y,
-                                       tx->base.nblocksx, tx->base.nblocksy,
+                                       tx->level_x, tx->level_y, tx->level_z,
+                                       tx->nblocksx, tx->nblocksy,
                                        tx->level_depth,
-                                       tx->base.block.size, nx, ny,
+                                       util_format_get_blocksize(pt->format), nx, ny,
                                        NOUVEAU_BO_GART, NOUVEAU_BO_VRAM |
                                        NOUVEAU_BO_GART);
        }
@@ -287,7 +266,7 @@ nv50_upload_sifc(struct nv50_context *nv50,
 
        reloc |= NOUVEAU_BO_WR;
 
-       WAIT_RING (chan, 32);
+       MARK_RING (chan, 32, 2); /* flush on lack of space or relocs */
 
        if (bo->tile_flags) {
                BEGIN_RING(chan, eng2d, NV50_2D_DST_FORMAT, 5);
index db54380241f681afc474d563dec1a60e00cabd55..39324e30f6b86a3308b9159ea0bf73268e288f74 100644 (file)
@@ -24,6 +24,8 @@
 #include "pipe/p_state.h"
 #include "pipe/p_inlines.h"
 
+#include "util/u_format.h"
+
 #include "nv50_context.h"
 
 static boolean
@@ -62,18 +64,25 @@ nv50_prim(unsigned mode)
 }
 
 static INLINE uint32_t
-nv50_vbo_type_to_hw(unsigned type)
+nv50_vbo_type_to_hw(enum pipe_format format)
 {
-       switch (type) {
-       case PIPE_FORMAT_TYPE_FLOAT:
+       const struct util_format_description *desc;
+
+       desc = util_format_description(format);
+       assert(desc);
+
+       switch (desc->channel[0].type) {
+       case UTIL_FORMAT_TYPE_FLOAT:
                return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_FLOAT;
-       case PIPE_FORMAT_TYPE_UNORM:
-               return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_UNORM;
-       case PIPE_FORMAT_TYPE_SNORM:
-               return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_SNORM;
-       case PIPE_FORMAT_TYPE_USCALED:
+       case UTIL_FORMAT_TYPE_UNSIGNED:
+               if (desc->channel[0].normalized) {
+                       return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_UNORM;
+               }
                return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_USCALED;
-       case PIPE_FORMAT_TYPE_SSCALED:
+       case UTIL_FORMAT_TYPE_SIGNED:
+               if (desc->channel[0].normalized) {
+                       return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_SNORM;
+               }
                return NV50TCL_VERTEX_ARRAY_ATTRIB_TYPE_SSCALED;
        /*
        case PIPE_FORMAT_TYPE_UINT:
@@ -120,9 +129,15 @@ nv50_vbo_vtxelt_to_hw(struct pipe_vertex_element *ve)
 {
        uint32_t hw_type, hw_size;
        enum pipe_format pf = ve->src_format;
-       unsigned size = pf_size_x(pf) << pf_exp2(pf);
+       const struct util_format_description *desc;
+       unsigned size;
+
+       desc = util_format_description(pf);
+       assert(desc);
+
+       size = util_format_get_component_bits(pf, UTIL_FORMAT_COLORSPACE_RGB, 0);
 
-       hw_type = nv50_vbo_type_to_hw(pf_type(pf));
+       hw_type = nv50_vbo_type_to_hw(pf);
        hw_size = nv50_vbo_size_to_hw(size, ve->nr_components);
 
        if (!hw_type || !hw_size) {
@@ -131,7 +146,7 @@ nv50_vbo_vtxelt_to_hw(struct pipe_vertex_element *ve)
                return 0x24e80000;
        }
 
-       if (pf_swizzle_x(pf) == 2) /* BGRA */
+       if (desc->swizzle[0] == UTIL_FORMAT_SWIZZLE_Z) /* BGRA */
                hw_size |= (1 << 31); /* no real swizzle bits :-( */
 
        return (hw_type | hw_size);
@@ -319,9 +334,13 @@ nv50_vbo_static_attrib(struct nv50_context *nv50, unsigned attrib,
        float *v;
        int ret;
        enum pipe_format pf = ve->src_format;
+       const struct util_format_description *desc;
 
-       if ((pf_type(pf) != PIPE_FORMAT_TYPE_FLOAT) ||
-           (pf_size_x(pf) << pf_exp2(pf)) != 32)
+       desc = util_format_description(pf);
+       assert(desc);
+
+       if ((desc->channel[0].type != UTIL_FORMAT_TYPE_FLOAT) ||
+           util_format_get_component_bits(pf, UTIL_FORMAT_COLORSPACE_RGB, 0) != 32)
                return FALSE;
 
        ret = nouveau_bo_map(bo, NOUVEAU_BO_RD);
@@ -353,6 +372,10 @@ nv50_vbo_static_attrib(struct nv50_context *nv50, unsigned attrib,
                so_data  (so, fui(v[1]));
                break;
        case 1:
+               if (attrib == nv50->vertprog->cfg.edgeflag_in) {
+                       so_method(so, tesla, 0x15e4, 1);
+                       so_data  (so, v[0] ? 1 : 0);
+               }
                so_method(so, tesla, NV50TCL_VTX_ATTR_1F(attrib), 1);
                so_data  (so, fui(v[0]));
                break;
@@ -382,6 +405,9 @@ nv50_vbo_validate(struct nv50_context *nv50)
                    !(nv50->vtxbuf[i].buffer->usage & PIPE_BUFFER_USAGE_VERTEX))
                        nv50->vbo_fifo = 0xffff;
 
+       if (nv50->vertprog->cfg.edgeflag_in < 16)
+               nv50->vbo_fifo = 0xffff; /* vertprog can't set edgeflag */
+
        n_ve = MAX2(nv50->vtxelt_nr, nv50->state.vtxelt_nr);
 
        vtxattr = NULL;
@@ -460,6 +486,9 @@ struct nv50_vbo_emitctx
        unsigned nr_ve;
        unsigned vtx_dwords;
        unsigned vtx_max;
+
+       float edgeflag;
+       unsigned ve_edgeflag;
 };
 
 static INLINE void
@@ -603,13 +632,17 @@ emit_prepare(struct nv50_context *nv50, struct nv50_vbo_emitctx *emit,
        if (nv50_map_vbufs(nv50) == FALSE)
                return FALSE;
 
+       emit->ve_edgeflag = nv50->vertprog->cfg.edgeflag_in;
+
+       emit->edgeflag = 0.5f;
        emit->nr_ve = 0;
        emit->vtx_dwords = 0;
 
        for (i = 0; i < nv50->vtxelt_nr; ++i) {
                struct pipe_vertex_element *ve;
                struct pipe_vertex_buffer *vb;
-               unsigned n, type, size;
+               unsigned n, size;
+               const struct util_format_description *desc;
 
                ve = &nv50->vtxelt[i];
                vb = &nv50->vtxbuf[ve->vertex_buffer_index];
@@ -621,8 +654,11 @@ emit_prepare(struct nv50_context *nv50, struct nv50_vbo_emitctx *emit,
                emit->map[n] = nouveau_bo(vb->buffer)->map +
                        (start * vb->stride + ve->src_offset);
 
-               type = pf_type(ve->src_format);
-               size = pf_size_x(ve->src_format) << pf_exp2(ve->src_format);
+               desc = util_format_description(ve->src_format);
+               assert(desc);
+
+               size = util_format_get_component_bits(
+                       ve->src_format, UTIL_FORMAT_COLORSPACE_RGB, 0);
 
                assert(ve->nr_components > 0 && ve->nr_components <= 4);
 
@@ -664,10 +700,31 @@ emit_prepare(struct nv50_context *nv50, struct nv50_vbo_emitctx *emit,
        }
 
        emit->vtx_max = 512 / emit->vtx_dwords;
+       if (emit->ve_edgeflag < 16)
+               emit->vtx_max = 1;
 
        return TRUE;
 }
 
+static INLINE void
+set_edgeflag(struct nouveau_channel *chan,
+            struct nouveau_grobj *tesla,
+            struct nv50_vbo_emitctx *emit, uint32_t index)
+{
+       unsigned i = emit->ve_edgeflag;
+
+       if (i < 16) {
+               float f = *((float *)(emit->map[i] + index * emit->stride[i]));
+
+               if (emit->edgeflag != f) {
+                       emit->edgeflag = f;
+
+                       BEGIN_RING(chan, tesla, 0x15e4, 1);
+                       OUT_RING  (chan, f ? 1 : 0);
+               }
+       }
+}
+
 static boolean
 nv50_push_arrays(struct nv50_context *nv50, unsigned start, unsigned count)
 {
@@ -682,6 +739,8 @@ nv50_push_arrays(struct nv50_context *nv50, unsigned start, unsigned count)
                unsigned i, dw, nr = MIN2(count, emit.vtx_max);
                dw = nr * emit.vtx_dwords;
 
+               set_edgeflag(chan, tesla, &emit, 0); /* nr will be 1 */
+
                BEGIN_RING(chan, tesla, NV50TCL_VERTEX_DATA | 0x40000000, dw);
                for (i = 0; i < nr; ++i)
                        emit_vtx_next(chan, &emit);
@@ -707,6 +766,8 @@ nv50_push_elements_u32(struct nv50_context *nv50, uint32_t *map, unsigned count)
                unsigned i, dw, nr = MIN2(count, emit.vtx_max);
                dw = nr * emit.vtx_dwords;
 
+               set_edgeflag(chan, tesla, &emit, *map);
+
                BEGIN_RING(chan, tesla, NV50TCL_VERTEX_DATA | 0x40000000, dw);
                for (i = 0; i < nr; ++i)
                        emit_vtx(chan, &emit, *map++);
@@ -732,6 +793,8 @@ nv50_push_elements_u16(struct nv50_context *nv50, uint16_t *map, unsigned count)
                unsigned i, dw, nr = MIN2(count, emit.vtx_max);
                dw = nr * emit.vtx_dwords;
 
+               set_edgeflag(chan, tesla, &emit, *map);
+
                BEGIN_RING(chan, tesla, NV50TCL_VERTEX_DATA | 0x40000000, dw);
                for (i = 0; i < nr; ++i)
                        emit_vtx(chan, &emit, *map++);
@@ -757,6 +820,8 @@ nv50_push_elements_u08(struct nv50_context *nv50, uint8_t *map, unsigned count)
                unsigned i, dw, nr = MIN2(count, emit.vtx_max);
                dw = nr * emit.vtx_dwords;
 
+               set_edgeflag(chan, tesla, &emit, *map);
+
                BEGIN_RING(chan, tesla, NV50TCL_VERTEX_DATA | 0x40000000, dw);
                for (i = 0; i < nr; ++i)
                        emit_vtx(chan, &emit, *map++);
index 121b65063f33a718e63ae14b31efb9220999c77f..afddcb161fa76089a8cc1e2a8a04b9218323d15a 100644 (file)
@@ -4,8 +4,8 @@ include $(TOP)/configs/current
 LIBNAME = r300
 
 C_SOURCES = \
+       r300_blit.c \
        r300_chipset.c \
-       r300_clear.c \
        r300_context.c \
        r300_debug.c \
        r300_emit.c \
@@ -17,13 +17,13 @@ C_SOURCES = \
        r300_state.c \
        r300_state_derived.c \
        r300_state_invariant.c \
-       r300_vbo.c \
        r300_vs.c \
        r300_texture.c \
        r300_tgsi_to_rc.c
 
 LIBRARY_INCLUDES = \
-       -I$(TOP)/src/mesa/drivers/dri/r300/compiler
+       -I$(TOP)/src/mesa/drivers/dri/r300/compiler \
+       -I$(TOP)/src/gallium/winsys/drm/radeon/core
 
 COMPILER_ARCHIVE = $(TOP)/src/mesa/drivers/dri/r300/compiler/libr300compiler.a
 
index 97989040d2e06ee5d2be36bfedf360ea4315bb1d..0d2de17be93205b8c5375709b18010b04a242fa7 100644 (file)
@@ -9,8 +9,8 @@ env.Append(CPPPATH = ['#/src/mesa/drivers/dri/r300/compiler', '#/include', '#/sr
 r300 = env.ConvenienceLibrary(
     target = 'r300',
     source = [
+        'r300_blit.c',
         'r300_chipset.c',
-        'r300_clear.c',
         'r300_context.c',
         'r300_debug.c',
         'r300_emit.c',
diff --git a/src/gallium/drivers/r300/r300_blit.c b/src/gallium/drivers/r300/r300_blit.c
new file mode 100644 (file)
index 0000000..ffe066d
--- /dev/null
@@ -0,0 +1,130 @@
+/*
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * on the rights to use, copy, modify, merge, publish, distribute, sub
+ * license, and/or sell copies of the Software, and to permit persons to whom
+ * the Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE. */
+
+#include "r300_blit.h"
+#include "r300_context.h"
+
+#include "util/u_rect.h"
+
+static void r300_blitter_save_states(struct r300_context* r300)
+{
+    util_blitter_save_blend(r300->blitter, r300->blend_state);
+    util_blitter_save_depth_stencil_alpha(r300->blitter, r300->dsa_state);
+    util_blitter_save_rasterizer(r300->blitter, r300->rs_state);
+    util_blitter_save_fragment_shader(r300->blitter, r300->fs);
+    util_blitter_save_vertex_shader(r300->blitter, r300->vs);
+}
+
+/* Clear currently bound buffers. */
+void r300_clear(struct pipe_context* pipe,
+                unsigned buffers,
+                const float* rgba,
+                double depth,
+                unsigned stencil)
+{
+    /* XXX Implement fastfill.
+     *
+     * If fastfill is enabled, a few facts should be considered:
+     *
+     * 1) Zbuffer must be micro-tiled and whole microtiles must be
+     *    written.
+     *
+     * 2) ZB_DEPTHCLEARVALUE is used to clear a zbuffer and Z Mask must be
+     *    equal to 0.
+     *
+     * 3) RB3D_COLOR_CLEAR_VALUE is used to clear a colorbuffer and
+     *    RB3D_COLOR_CHANNEL_MASK must be equal to 0.
+     *
+     * 4) ZB_CB_CLEAR can be used to make the ZB units help in clearing
+     *    the colorbuffer. The color clear value is supplied through both
+     *    RB3D_COLOR_CLEAR_VALUE and ZB_DEPTHCLEARVALUE, and the colorbuffer
+     *    must be set in ZB_DEPTHOFFSET and ZB_DEPTHPITCH in addition to
+     *    RB3D_COLOROFFSET and RB3D_COLORPITCH. It's obvious that the zbuffer
+     *    will not be cleared and multiple render targets cannot be cleared
+     *    this way either.
+     *
+     * 5) For 16-bit integer buffering, compression causes a hung with one or
+     *    two samples and should not be used.
+     *
+     * 6) Fastfill must not be used if reading of compressed Z data is disabled
+     *    and writing of compressed Z data is enabled (RD/WR_COMP_ENABLE),
+     *    i.e. it cannot be used to compress the zbuffer.
+     *    (what the hell does that mean and how does it fit in clearing
+     *    the buffers?)
+     *
+     * - Marek
+     */
+
+    struct r300_context* r300 = r300_context(pipe);
+
+    r300_blitter_save_states(r300);
+
+    util_blitter_clear(r300->blitter,
+                       r300->framebuffer_state.width,
+                       r300->framebuffer_state.height,
+                       r300->framebuffer_state.nr_cbufs,
+                       buffers, rgba, depth, stencil);
+}
+
+/* Copy a block of pixels from one surface to another. */
+void r300_surface_copy(struct pipe_context* pipe,
+                       struct pipe_surface* dst,
+                       unsigned dstx, unsigned dsty,
+                       struct pipe_surface* src,
+                       unsigned srcx, unsigned srcy,
+                       unsigned width, unsigned height)
+{
+    struct r300_context* r300 = r300_context(pipe);
+
+    /* Yeah we have to save all those states to ensure this blitter operation
+     * is really transparent. The states will be restored by the blitter once
+     * copying is done. */
+    r300_blitter_save_states(r300);
+    util_blitter_save_framebuffer(r300->blitter, &r300->framebuffer_state);
+
+    util_blitter_save_fragment_sampler_states(
+        r300->blitter, r300->sampler_count, (void**)r300->sampler_states);
+
+    util_blitter_save_fragment_sampler_textures(
+        r300->blitter, r300->texture_count,
+        (struct pipe_texture**)r300->textures);
+
+    /* Do a copy */
+    util_blitter_copy(r300->blitter,
+                      dst, dstx, dsty, src, srcx, srcy, width, height, TRUE);
+}
+
+/* Fill a region of a surface with a constant value. */
+void r300_surface_fill(struct pipe_context* pipe,
+                       struct pipe_surface* dst,
+                       unsigned dstx, unsigned dsty,
+                       unsigned width, unsigned height,
+                       unsigned value)
+{
+    struct r300_context* r300 = r300_context(pipe);
+
+    r300_blitter_save_states(r300);
+    util_blitter_save_framebuffer(r300->blitter, &r300->framebuffer_state);
+
+    util_blitter_fill(r300->blitter,
+                      dst, dstx, dsty, width, height, value);
+}
diff --git a/src/gallium/drivers/r300/r300_blit.h b/src/gallium/drivers/r300/r300_blit.h
new file mode 100644 (file)
index 0000000..029e4f9
--- /dev/null
@@ -0,0 +1,48 @@
+/*
+ * Copyright 2008 Marek Olšák <maraeo@gmail.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * on the rights to use, copy, modify, merge, publish, distribute, sub
+ * license, and/or sell copies of the Software, and to permit persons to whom
+ * the Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE. */
+
+#ifndef R300_BLIT_H
+#define R300_BLIT_H
+
+struct pipe_context;
+struct pipe_surface;
+
+void r300_clear(struct pipe_context* pipe,
+                unsigned buffers,
+                const float* rgba,
+                double depth,
+                unsigned stencil);
+
+void r300_surface_copy(struct pipe_context* pipe,
+                       struct pipe_surface* dst,
+                       unsigned dstx, unsigned dsty,
+                       struct pipe_surface* src,
+                       unsigned srcx, unsigned srcy,
+                       unsigned width, unsigned height);
+
+void r300_surface_fill(struct pipe_context* pipe,
+                       struct pipe_surface* dst,
+                       unsigned dstx, unsigned dsty,
+                       unsigned width, unsigned height,
+                       unsigned value);
+
+#endif /* R300_BLIT_H */
diff --git a/src/gallium/drivers/r300/r300_clear.c b/src/gallium/drivers/r300/r300_clear.c
deleted file mode 100644 (file)
index 02d6d50..0000000
+++ /dev/null
@@ -1,38 +0,0 @@
-/*
- * Copyright 2008 Corbin Simpson <MostAwesomeDude@gmail.com>
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * on the rights to use, copy, modify, merge, publish, distribute, sub
- * license, and/or sell copies of the Software, and to permit persons to whom
- * the Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE. */
-
-#include "r300_clear.h"
-#include "r300_context.h"
-
-#include "util/u_clear.h"
-
-/* Clears currently bound buffers. */
-void r300_clear(struct pipe_context* pipe,
-                unsigned buffers,
-                const float* rgba,
-                double depth,
-                unsigned stencil)
-{
-    /* XXX we can and should do one clear if both color and zs are set */
-    util_clear(pipe, &r300_context(pipe)->framebuffer_state,
-            buffers, rgba, depth, stencil);
-}
diff --git a/src/gallium/drivers/r300/r300_clear.h b/src/gallium/drivers/r300/r300_clear.h
deleted file mode 100644 (file)
index b8fcdf2..0000000
+++ /dev/null
@@ -1,34 +0,0 @@
-/*
- * Copyright 2008 Corbin Simpson <MostAwesomeDude@gmail.com>
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * on the rights to use, copy, modify, merge, publish, distribute, sub
- * license, and/or sell copies of the Software, and to permit persons to whom
- * the Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE. */
-
-#ifndef R300_CLEAR_H
-#define R300_CLEAR_H
-
-struct pipe_context;
-
-void r300_clear(struct pipe_context* pipe,
-                unsigned buffers,
-                const float* rgba,
-                double depth,
-                unsigned stencil);
-
-#endif /* R300_CLEAR_H */
index ae23329b83f9c4d3d8c21d99914133c30bf2413f..d5c2d63d393babb83a5c18c6c400abc21b77b90b 100644 (file)
@@ -28,7 +28,7 @@
 #include "util/u_memory.h"
 #include "util/u_simple_list.h"
 
-#include "r300_clear.h"
+#include "r300_blit.h"
 #include "r300_context.h"
 #include "r300_flush.h"
 #include "r300_query.h"
@@ -36,6 +36,7 @@
 #include "r300_screen.h"
 #include "r300_state_derived.h"
 #include "r300_state_invariant.h"
+#include "r300_texture.h"
 #include "r300_winsys.h"
 
 static enum pipe_error r300_clear_hash_table(void* key, void* value,
@@ -51,6 +52,8 @@ static void r300_destroy_context(struct pipe_context* context)
     struct r300_context* r300 = r300_context(context);
     struct r300_query* query, * temp;
 
+    util_blitter_destroy(r300->blitter);
+
     util_hash_table_foreach(r300->shader_hash_table, r300_clear_hash_table,
         NULL);
     util_hash_table_destroy(r300->shader_hash_table);
@@ -69,6 +72,7 @@ static void r300_destroy_context(struct pipe_context* context)
     FREE(r300->blend_color_state);
     FREE(r300->rs_block);
     FREE(r300->scissor_state);
+    FREE(r300->vertex_info);
     FREE(r300->viewport_state);
     FREE(r300);
 }
@@ -104,7 +108,7 @@ static void r300_flush_cb(void *data)
 }
 
 struct pipe_context* r300_create_context(struct pipe_screen* screen,
-                                         struct r300_winsys* r300_winsys)
+                                         struct radeon_winsys* radeon_winsys)
 {
     struct r300_context* r300 = CALLOC_STRUCT(r300_context);
     struct r300_screen* r300screen = r300_screen(screen);
@@ -112,9 +116,9 @@ struct pipe_context* r300_create_context(struct pipe_screen* screen,
     if (!r300)
         return NULL;
 
-    r300->winsys = r300_winsys;
+    r300->winsys = radeon_winsys;
 
-    r300->context.winsys = (struct pipe_winsys*)r300_winsys;
+    r300->context.winsys = (struct pipe_winsys*)radeon_winsys;
     r300->context.screen = screen;
 
     r300_init_debug(r300);
@@ -122,16 +126,27 @@ struct pipe_context* r300_create_context(struct pipe_screen* screen,
     r300->context.destroy = r300_destroy_context;
 
     r300->context.clear = r300_clear;
+    r300->context.surface_copy = r300_surface_copy;
+    r300->context.surface_fill = r300_surface_fill;
 
-    if (r300screen->caps->has_tcl)
-    {
+    if (r300screen->caps->has_tcl) {
         r300->context.draw_arrays = r300_draw_arrays;
         r300->context.draw_elements = r300_draw_elements;
         r300->context.draw_range_elements = r300_draw_range_elements;
-    }
-    else
-    {
-        assert(0);
+    } else {
+        r300->context.draw_arrays = r300_swtcl_draw_arrays;
+        r300->context.draw_elements = r300_draw_elements;
+        r300->context.draw_range_elements = r300_swtcl_draw_range_elements;
+
+        /* Create a Draw. This is used for SW TCL. */
+        r300->draw = draw_create();
+        /* Enable our renderer. */
+        draw_set_rasterize_stage(r300->draw, r300_draw_stage(r300));
+        /* Enable Draw's clipping. */
+        draw_set_driver_clipping(r300->draw, FALSE);
+        /* Force Draw to never do viewport transform, since we can do
+         * transform in hardware, always. */
+        draw_set_viewport_state(r300->draw, &r300_viewport_identity);
     }
 
     r300->context.is_texture_referenced = r300_is_texture_referenced;
@@ -143,18 +158,9 @@ struct pipe_context* r300_create_context(struct pipe_screen* screen,
     r300->blend_color_state = CALLOC_STRUCT(r300_blend_color_state);
     r300->rs_block = CALLOC_STRUCT(r300_rs_block);
     r300->scissor_state = CALLOC_STRUCT(r300_scissor_state);
+    r300->vertex_info = CALLOC_STRUCT(r300_vertex_info);
     r300->viewport_state = CALLOC_STRUCT(r300_viewport_state);
 
-    /* Create a Draw. This is used for vert collation and SW TCL. */
-    r300->draw = draw_create();
-    /* Enable our renderer. */
-    draw_set_rasterize_stage(r300->draw, r300_draw_stage(r300));
-    /* Disable Draw's clipping if TCL is present. */
-    draw_set_driver_clipping(r300->draw, r300_screen(screen)->caps->has_tcl);
-    /* Force Draw to never do viewport transform, since (again) we can do
-     * transform in hardware, always. */
-    draw_set_viewport_state(r300->draw, &r300_viewport_identity);
-
     /* Open up the OQ BO. */
     r300->oqbo = screen->buffer_create(screen, 4096,
             PIPE_BUFFER_USAGE_VERTEX, 4096);
@@ -173,5 +179,8 @@ struct pipe_context* r300_create_context(struct pipe_screen* screen,
     r300->winsys->set_flush_cb(r300->winsys, r300_flush_cb, r300);
     r300->dirty_state = R300_NEW_KITCHEN_SINK;
     r300->dirty_hw++;
+
+    r300->blitter = util_blitter_create(&r300->context);
+
     return &r300->context;
 }
index f954ba7f9aab444f2f6622919fc30d67a4076b99..232530b7dc447b1840cadc9d1189ea4ca0e66186 100644 (file)
@@ -25,6 +25,8 @@
 
 #include "draw/draw_vertex.h"
 
+#include "util/u_blitter.h"
+
 #include "pipe/p_context.h"
 #include "pipe/p_inlines.h"
 
@@ -89,14 +91,28 @@ struct r300_rs_block {
 };
 
 struct r300_sampler_state {
+    struct pipe_sampler_state state;
+
     uint32_t filter0;      /* R300_TX_FILTER0: 0x4400 */
     uint32_t filter1;      /* R300_TX_FILTER1: 0x4440 */
     uint32_t border_color; /* R300_TX_BORDER_COLOR: 0x45c0 */
+
+    /* Min/max LOD must be clamped to [0, last_level], thus
+     * it's dependent on a currently bound texture */
+    unsigned min_lod, max_lod;
+};
+
+struct r300_scissor_regs {
+    uint32_t top_left;     /* R300_SC_SCISSORS_TL: 0x43e0 */
+    uint32_t bottom_right; /* R300_SC_SCISSORS_BR: 0x43e4 */
+
+    /* Whether everything is culled by scissoring. */
+    boolean empty_area;
 };
 
 struct r300_scissor_state {
-    uint32_t scissor_top_left;     /* R300_SC_SCISSORS_TL: 0x43e0 */
-    uint32_t scissor_bottom_right; /* R300_SC_SCISSORS_BR: 0x43e4 */
+    struct r300_scissor_regs framebuffer;
+    struct r300_scissor_regs scissor;
 };
 
 struct r300_texture_state {
@@ -219,11 +235,6 @@ struct r300_texture {
 struct r300_vertex_info {
     /* Parent class */
     struct vertex_info vinfo;
-    /* Map of vertex attributes into PVS memory for HW TCL,
-     * or GA memory for SW TCL. */
-    int vs_tab[16];
-    /* Map of rasterizer attributes from GB through RS to US. */
-    int fs_tab[16];
 
     /* R300_VAP_PROG_STREAK_CNTL_[0-7] */
     uint32_t vap_prog_stream_cntl[8];
@@ -238,9 +249,11 @@ struct r300_context {
     struct pipe_context context;
 
     /* The interface to the windowing system, etc. */
-    struct r300_winsys* winsys;
+    struct radeon_winsys* winsys;
     /* Draw module. Used mostly for SW TCL. */
     struct draw_context* draw;
+    /* Accelerated blit support. */
+    struct blitter_context* blitter;
 
     /* Vertex buffer for rendering. */
     struct pipe_buffer* vbo;
index 86ba91db52e46ffa4085cc188693fe854d7eed83..d142fee0502a058de93fdb94943a0cdcb3ad75c1 100644 (file)
@@ -26,7 +26,8 @@
 #include "util/u_math.h"
 
 #include "r300_reg.h"
-#include "r300_winsys.h"
+
+#include "radeon_winsys.h"
 
 /* Yes, I know macros are ugly. However, they are much prettier than the code
  * that they neatly hide away, and don't have the cost of function setup,so
 
 #define CS_LOCALS(context) \
     struct r300_context* const cs_context_copy = (context); \
-    struct r300_winsys* cs_winsys = cs_context_copy->winsys; \
+    struct radeon_winsys* cs_winsys = cs_context_copy->winsys; \
     int cs_count = 0;
 
 #define CHECK_CS(size) \
-    cs_winsys->check_cs(cs_winsys, (size))
+    assert(cs_winsys->check_cs(cs_winsys, (size)))
 
 #define BEGIN_CS(size) do { \
     CHECK_CS(size); \
     cs_count -= 3; \
 } while (0)
 
+#define OUT_CS_RELOC_NO_OFFSET(bo, rd, wd, flags) do { \
+    DBG(cs_context_copy, DBG_CS, "r300: writing relocation for buffer %p, " \
+            "domains (%d, %d, %d)\n", \
+        bo, rd, wd, flags); \
+    assert(bo); \
+    cs_winsys->write_cs_reloc(cs_winsys, bo, rd, wd, flags); \
+    cs_count -= 2; \
+} while (0)
+
 #define END_CS do { \
     if (VERY_VERBOSE_CS) { \
         DBG(cs_context_copy, DBG_CS, "r300: END_CS in %s (%s:%d)\n", __FUNCTION__, \
index eeb97a2d37053305878400a01fe3ab9a5561bb62..199ce3a945d509e9a7b0a75afd7707530e3d2b07 100644 (file)
@@ -1,5 +1,6 @@
 /*
  * Copyright 2008 Corbin Simpson <MostAwesomeDude@gmail.com>
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -22,6 +23,7 @@
 
 /* r300_emit: Functions for emitting state. */
 
+#include "util/u_format.h"
 #include "util/u_math.h"
 
 #include "r300_context.h"
@@ -40,9 +42,16 @@ void r300_emit_blend_state(struct r300_context* r300,
     CS_LOCALS(r300);
     BEGIN_CS(8);
     OUT_CS_REG_SEQ(R300_RB3D_CBLEND, 3);
-    OUT_CS(blend->blend_control);
-    OUT_CS(blend->alpha_blend_control);
-    OUT_CS(blend->color_channel_mask);
+    if (r300->framebuffer_state.nr_cbufs) {
+        OUT_CS(blend->blend_control);
+        OUT_CS(blend->alpha_blend_control);
+        OUT_CS(blend->color_channel_mask);
+    } else {
+        OUT_CS(0);
+        OUT_CS(0);
+        OUT_CS(0);
+        /* XXX also disable fastfill here once it's supported */
+    }
     OUT_CS_REG(R300_RB3D_ROPCNTL, blend->rop);
     OUT_CS_REG(R300_RB3D_DITHER_CTL, blend->dither);
     END_CS;
@@ -129,7 +138,9 @@ static const float * get_shader_constant(
     struct rc_constant * constant,
     struct r300_constant_buffer * externals)
 {
-    static const float zero[4] = { 0.0, 0.0, 0.0, 0.0 };
+    static float vec[4] = { 0.0, 0.0, 0.0, 1.0 };
+    struct pipe_texture *tex;
+
     switch(constant->Type) {
         case RC_CONSTANT_EXTERNAL:
             return externals->constants[constant->u.External];
@@ -137,11 +148,38 @@ static const float * get_shader_constant(
         case RC_CONSTANT_IMMEDIATE:
             return constant->u.Immediate;
 
+        case RC_CONSTANT_STATE:
+            switch (constant->u.State[0]) {
+                /* Factor for converting rectangle coords to
+                 * normalized coords. Should only show up on non-r500. */
+                case RC_STATE_R300_TEXRECT_FACTOR:
+                    tex = &r300->textures[constant->u.State[1]]->tex;
+                    vec[0] = 1.0 / tex->width0;
+                    vec[1] = 1.0 / tex->height0;
+                    break;
+
+                /* Texture compare-fail value. */
+                /* XXX Since Gallium doesn't support GL_ARB_shadow_ambient,
+                 * this is always (0,0,0,0). */
+                case RC_STATE_SHADOW_AMBIENT:
+                    vec[3] = 0;
+                    break;
+
+                default:
+                    debug_printf("r300: Implementation error: "
+                        "Unknown RC_CONSTANT type %d\n", constant->u.State[0]);
+            }
+            break;
+
         default:
-            debug_printf("r300: Implementation error: Unhandled constant type %i\n",
-                constant->Type);
-            return zero;
+            debug_printf("r300: Implementation error: "
+                "Unhandled constant type %d\n", constant->Type);
     }
+
+    /* This should either be (0, 0, 0, 1), which should be a relatively safe
+     * RGBA or STRQ value, or it could be one of the RC_CONSTANT_STATE
+     * state factors. */
+    return vec;
 }
 
 /* Convert a normal single-precision float into the 7.16 format
@@ -254,7 +292,7 @@ void r500_emit_fragment_program_code(struct r300_context* r300,
 
     BEGIN_CS(13 +
              ((code->inst_end + 1) * 6));
-    OUT_CS_REG(R500_US_CONFIG, 0);
+    OUT_CS_REG(R500_US_CONFIG, R500_ZERO_TIMES_ANYTHING_EQUALS_ZERO);
     OUT_CS_REG(R500_US_PIXSIZE, code->max_temp_idx);
     OUT_CS_REG(R500_US_CODE_RANGE,
                R500_US_CODE_RANGE_ADDR(0) | R500_US_CODE_RANGE_SIZE(code->inst_end));
@@ -308,7 +346,13 @@ void r300_emit_fb_state(struct r300_context* r300,
     int i;
     CS_LOCALS(r300);
 
-    BEGIN_CS((10 * fb->nr_cbufs) + (fb->zsbuf ? 10 : 0) + 4);
+    /* Shouldn't fail unless there is a bug in the state tracker. */
+    assert(fb->nr_cbufs <= 4);
+
+    BEGIN_CS((10 * fb->nr_cbufs) + (2 * (4 - fb->nr_cbufs)) +
+             (fb->zsbuf ? 10 : 0) + 6);
+
+    /* Flush and free renderbuffer caches. */
     OUT_CS_REG(R300_RB3D_DSTCACHE_CTLSTAT,
         R300_RB3D_DSTCACHE_CTLSTAT_DC_FREE_FREE_3D_TAGS |
         R300_RB3D_DSTCACHE_CTLSTAT_DC_FLUSH_FLUSH_DIRTY_3D);
@@ -316,6 +360,10 @@ void r300_emit_fb_state(struct r300_context* r300,
         R300_ZB_ZCACHE_CTLSTAT_ZC_FLUSH_FLUSH_AND_FREE |
         R300_ZB_ZCACHE_CTLSTAT_ZC_FREE_FREE);
 
+    /* Set the number of colorbuffers. */
+    OUT_CS_REG(R300_RB3D_CCTL, R300_RB3D_CCTL_NUM_MULTIWRITES(fb->nr_cbufs));
+
+    /* Set up colorbuffers. */
     for (i = 0; i < fb->nr_cbufs; i++) {
         surf = fb->cbufs[i];
         tex = (struct r300_texture*)surf->texture;
@@ -333,6 +381,12 @@ void r300_emit_fb_state(struct r300_context* r300,
             r300_translate_out_fmt(surf->format));
     }
 
+    /* Disable unused colorbuffers. */
+    for (; i < 4; i++) {
+        OUT_CS_REG(R300_US_OUT_FMT_0 + (4 * i), R300_US_OUT_FMT_UNUSED);
+    }
+
+    /* Set up a zbuffer. */
     if (fb->zsbuf) {
         surf = fb->zsbuf;
         tex = (struct r300_texture*)surf->texture;
@@ -360,8 +414,6 @@ static void r300_emit_query_start(struct r300_context *r300)
     if (!query)
        return;
 
-    /* XXX This will almost certainly not return good results
-     * for overlapping queries. */
     BEGIN_CS(4);
     if (caps->family == CHIP_FAMILY_RV530) {
         OUT_CS_REG(RV530_FG_ZBREG_DEST, RV530_FG_ZBREG_DEST_PIPE_SELECT_ALL);
@@ -543,24 +595,36 @@ void r300_emit_rs_block_state(struct r300_context* r300,
     END_CS;
 }
 
-void r300_emit_scissor_state(struct r300_context* r300,
-                             struct r300_scissor_state* scissor)
+static void r300_emit_scissor_regs(struct r300_context* r300,
+                                   struct r300_scissor_regs* scissor)
 {
     CS_LOCALS(r300);
 
     BEGIN_CS(3);
     OUT_CS_REG_SEQ(R300_SC_SCISSORS_TL, 2);
-    OUT_CS(scissor->scissor_top_left);
-    OUT_CS(scissor->scissor_bottom_right);
+    OUT_CS(scissor->top_left);
+    OUT_CS(scissor->bottom_right);
     END_CS;
 }
 
+void r300_emit_scissor_state(struct r300_context* r300,
+                             struct r300_scissor_state* scissor)
+{
+    if (r300->rs_state->rs.scissor) {
+        r300_emit_scissor_regs(r300, &scissor->scissor);
+    } else {
+        r300_emit_scissor_regs(r300, &scissor->framebuffer);
+    }
+}
+
 void r300_emit_texture(struct r300_context* r300,
                        struct r300_sampler_state* sampler,
                        struct r300_texture* tex,
                        unsigned offset)
 {
     uint32_t filter0 = sampler->filter0;
+    uint32_t format0 = tex->state.format0;
+    unsigned min_level, max_level;
     CS_LOCALS(r300);
 
     /* to emulate 1D textures through 2D ones correctly */
@@ -569,13 +633,20 @@ void r300_emit_texture(struct r300_context* r300,
         filter0 |= R300_TX_WRAP_T(R300_TX_CLAMP_TO_EDGE);
     }
 
+    /* determine min/max levels */
+    /* the MAX_MIP level is the largest (finest) one */
+    max_level = MIN2(sampler->max_lod, tex->tex.last_level);
+    min_level = MIN2(sampler->min_lod, max_level);
+    format0 |= R300_TX_NUM_LEVELS(max_level);
+    filter0 |= R300_TX_MAX_MIP_LEVEL(min_level);
+
     BEGIN_CS(16);
     OUT_CS_REG(R300_TX_FILTER0_0 + (offset * 4), filter0 |
         (offset << 28));
     OUT_CS_REG(R300_TX_FILTER1_0 + (offset * 4), sampler->filter1);
     OUT_CS_REG(R300_TX_BORDER_COLOR_0 + (offset * 4), sampler->border_color);
 
-    OUT_CS_REG(R300_TX_FORMAT0_0 + (offset * 4), tex->state.format0);
+    OUT_CS_REG(R300_TX_FORMAT0_0 + (offset * 4), format0);
     OUT_CS_REG(R300_TX_FORMAT1_0 + (offset * 4), tex->state.format1);
     OUT_CS_REG(R300_TX_FORMAT2_0 + (offset * 4), tex->state.format2);
     OUT_CS_REG_SEQ(R300_TX_OFFSET_0 + (offset * 4), 1);
@@ -584,50 +655,68 @@ void r300_emit_texture(struct r300_context* r300,
     END_CS;
 }
 
-/* XXX I can't read this and that's not good */
-void r300_emit_aos(struct r300_context* r300, unsigned offset)
+static boolean r300_validate_aos(struct r300_context *r300)
 {
     struct pipe_vertex_buffer *vbuf = r300->vertex_buffer;
     struct pipe_vertex_element *velem = r300->vertex_element;
-    CS_LOCALS(r300);
     int i;
-    unsigned aos_count = r300->vertex_element_count;
 
+    /* Check if formats and strides are aligned to the size of DWORD. */
+    for (i = 0; i < r300->vertex_element_count; i++) {
+        if (vbuf[velem[i].vertex_buffer_index].stride % 4 != 0 ||
+            util_format_get_blocksize(velem[i].src_format) % 4 != 0) {
+            return FALSE;
+        }
+    }
+    return TRUE;
+}
+
+void r300_emit_aos(struct r300_context* r300, unsigned offset)
+{
+    struct pipe_vertex_buffer *vb1, *vb2, *vbuf = r300->vertex_buffer;
+    struct pipe_vertex_element *velem = r300->vertex_element;
+    int i;
+    unsigned size1, size2, aos_count = r300->vertex_element_count;
     unsigned packet_size = (aos_count * 3 + 1) / 2;
+    CS_LOCALS(r300);
+
+    /* XXX Move this checking to a more approriate place. */
+    if (!r300_validate_aos(r300)) {
+        /* XXX We should fallback using Draw. */
+        assert(0);
+    }
+
     BEGIN_CS(2 + packet_size + aos_count * 2);
     OUT_CS_PKT3(R300_PACKET3_3D_LOAD_VBPNTR, packet_size);
     OUT_CS(aos_count);
+
     for (i = 0; i < aos_count - 1; i += 2) {
-        int buf_num1 = velem[i].vertex_buffer_index;
-        int buf_num2 = velem[i+1].vertex_buffer_index;
-        assert(vbuf[buf_num1].stride % 4 == 0 && pf_get_size(velem[i].src_format) % 4 == 0);
-        assert(vbuf[buf_num2].stride % 4 == 0 && pf_get_size(velem[i+1].src_format) % 4 == 0);
-        OUT_CS((pf_get_size(velem[i].src_format) >> 2) | (vbuf[buf_num1].stride << 6) |
-               (pf_get_size(velem[i+1].src_format) << 14) | (vbuf[buf_num2].stride << 22));
-        OUT_CS(vbuf[buf_num1].buffer_offset + velem[i].src_offset +
-               offset * vbuf[buf_num1].stride);
-        OUT_CS(vbuf[buf_num2].buffer_offset + velem[i+1].src_offset +
-               offset * vbuf[buf_num2].stride);
+        vb1 = &vbuf[velem[i].vertex_buffer_index];
+        vb2 = &vbuf[velem[i+1].vertex_buffer_index];
+        size1 = util_format_get_blocksize(velem[i].src_format);
+        size2 = util_format_get_blocksize(velem[i+1].src_format);
+
+        OUT_CS(R300_VBPNTR_SIZE0(size1) | R300_VBPNTR_STRIDE0(vb1->stride) |
+               R300_VBPNTR_SIZE1(size2) | R300_VBPNTR_STRIDE1(vb2->stride));
+        OUT_CS(vb1->buffer_offset + velem[i].src_offset   + offset * vb1->stride);
+        OUT_CS(vb2->buffer_offset + velem[i+1].src_offset + offset * vb2->stride);
     }
+
     if (aos_count & 1) {
-        int buf_num = velem[i].vertex_buffer_index;
-        assert(vbuf[buf_num].stride % 4 == 0 && pf_get_size(velem[i].src_format) % 4 == 0);
-        OUT_CS((pf_get_size(velem[i].src_format) >> 2) | (vbuf[buf_num].stride << 6));
-        OUT_CS(vbuf[buf_num].buffer_offset + velem[i].src_offset +
-               offset * vbuf[buf_num].stride);
+        vb1 = &vbuf[velem[i].vertex_buffer_index];
+        size1 = util_format_get_blocksize(velem[i].src_format);
+
+        OUT_CS(R300_VBPNTR_SIZE0(size1) | R300_VBPNTR_STRIDE0(vb1->stride));
+        OUT_CS(vb1->buffer_offset + velem[i].src_offset + offset * vb1->stride);
     }
 
-    /* XXX bare CS reloc */
     for (i = 0; i < aos_count; i++) {
-        cs_winsys->write_cs_reloc(cs_winsys,
-                                  vbuf[velem[i].vertex_buffer_index].buffer,
-                                  RADEON_GEM_DOMAIN_GTT,
-                                  0,
-                                  0);
-        cs_count -= 2;
+        OUT_CS_RELOC_NO_OFFSET(vbuf[velem[i].vertex_buffer_index].buffer,
+                               RADEON_GEM_DOMAIN_GTT, 0, 0);
     }
     END_CS;
 }
+
 #if 0
 void r300_emit_draw_packet(struct r300_context* r300)
 {
@@ -690,12 +779,22 @@ void r300_emit_vertex_format_state(struct r300_context* r300)
     END_CS;
 }
 
+
 void r300_emit_vertex_program_code(struct r300_context* r300,
                                    struct r300_vertex_program_code* code)
 {
     int i;
     struct r300_screen* r300screen = r300_screen(r300->context.screen);
     unsigned instruction_count = code->length / 4;
+
+    int vtx_mem_size = r300screen->caps->is_r500 ? 128 : 72;
+    int input_count = MAX2(util_bitcount(code->InputsRead), 1);
+    int output_count = MAX2(util_bitcount(code->OutputsWritten), 1);
+    int temp_count = MAX2(code->num_temporaries, 1);
+    int pvs_num_slots = MIN3(vtx_mem_size / input_count,
+                             vtx_mem_size / output_count, 10);
+    int pvs_num_controllers = MIN2(vtx_mem_size / temp_count, 6);
+
     CS_LOCALS(r300);
 
     if (!r300screen->caps->has_tcl) {
@@ -708,8 +807,7 @@ void r300_emit_vertex_program_code(struct r300_context* r300,
     /* R300_VAP_PVS_CODE_CNTL_0
      * R300_VAP_PVS_CONST_CNTL
      * R300_VAP_PVS_CODE_CNTL_1
-     * See the r5xx docs for instructions on how to use these.
-     * XXX these could be optimized to select better values... */
+     * See the r5xx docs for instructions on how to use these. */
     OUT_CS_REG_SEQ(R300_VAP_PVS_CODE_CNTL_0, 3);
     OUT_CS(R300_PVS_FIRST_INST(0) |
             R300_PVS_XYZW_VALID_INST(instruction_count - 1) |
@@ -722,10 +820,11 @@ void r300_emit_vertex_program_code(struct r300_context* r300,
     for (i = 0; i < code->length; i++)
         OUT_CS(code->body.d[i]);
 
-    OUT_CS_REG(R300_VAP_CNTL, R300_PVS_NUM_SLOTS(10) |
-            R300_PVS_NUM_CNTLRS(5) |
+    OUT_CS_REG(R300_VAP_CNTL, R300_PVS_NUM_SLOTS(pvs_num_slots) |
+            R300_PVS_NUM_CNTLRS(pvs_num_controllers) |
             R300_PVS_NUM_FPUS(r300screen->caps->num_vert_fpus) |
-            R300_PVS_VF_MAX_VTX_NUM(12));
+            R300_PVS_VF_MAX_VTX_NUM(12) |
+            (r300screen->caps->is_r500 ? R500_TCL_STATE_OPTIMIZATION : 0));
     END_CS;
 }
 
@@ -790,13 +889,33 @@ void r300_emit_viewport_state(struct r300_context* r300,
     END_CS;
 }
 
+void r300_emit_texture_count(struct r300_context* r300)
+{
+    uint32_t tx_enable = 0;
+    int i;
+    CS_LOCALS(r300);
+
+    /* Notice that texture_count and sampler_count are just sizes
+     * of the respective arrays. We still have to check for the individual
+     * elements. */
+    for (i = 0; i < MIN2(r300->sampler_count, r300->texture_count); i++) {
+        if (r300->textures[i]) {
+            tx_enable |= 1 << i;
+        }
+    }
+
+    BEGIN_CS(2);
+    OUT_CS_REG(R300_TX_ENABLE, tx_enable);
+    END_CS;
+
+}
+
 void r300_flush_textures(struct r300_context* r300)
 {
     CS_LOCALS(r300);
 
-    BEGIN_CS(4);
+    BEGIN_CS(2);
     OUT_CS_REG(R300_TX_INVALTAGS, 0);
-    OUT_CS_REG(R300_TX_ENABLE, (1 << r300->texture_count) - 1);
     END_CS;
 }
 
@@ -821,10 +940,17 @@ void r300_emit_dirty_state(struct r300_context* r300)
         return;
     }
 
+    /* Check size of CS. */
+    /* Make sure we have at least 8*1024 spare dwords. */
+    /* XXX It would be nice to know the number of dwords we really need to
+     * XXX emit. */
+    if (!r300->winsys->check_cs(r300->winsys, 8*1024)) {
+        r300->context.flush(&r300->context, 0, NULL);
+    }
+
     /* Clean out BOs. */
     r300->winsys->reset_bos(r300->winsys);
 
-    /* XXX check size */
 validate:
     /* Color buffers... */
     for (i = 0; i < r300->framebuffer_state.nr_cbufs; i++) {
@@ -911,18 +1037,20 @@ validate:
 
     if (r300->dirty_state & R300_NEW_FRAGMENT_SHADER) {
         if (r300screen->caps->is_r500) {
-            r500_emit_fragment_program_code(r300, &r300->fs->code);
+            r500_emit_fragment_program_code(r300, &r300->fs->shader->code);
         } else {
-            r300_emit_fragment_program_code(r300, &r300->fs->code);
+            r300_emit_fragment_program_code(r300, &r300->fs->shader->code);
         }
         r300->dirty_state &= ~R300_NEW_FRAGMENT_SHADER;
     }
 
     if (r300->dirty_state & R300_NEW_FRAGMENT_SHADER_CONSTANTS) {
         if (r300screen->caps->is_r500) {
-            r500_emit_fs_constant_buffer(r300, &r300->fs->code.constants);
+            r500_emit_fs_constant_buffer(r300,
+                                         &r300->fs->shader->code.constants);
         } else {
-            r300_emit_fs_constant_buffer(r300, &r300->fs->code.constants);
+            r300_emit_fs_constant_buffer(r300,
+                                         &r300->fs->shader->code.constants);
         }
         r300->dirty_state &= ~R300_NEW_FRAGMENT_SHADER_CONSTANTS;
     }
@@ -950,6 +1078,8 @@ validate:
     /* Samplers and textures are tracked separately but emitted together. */
     if (r300->dirty_state &
             (R300_ANY_NEW_SAMPLERS | R300_ANY_NEW_TEXTURES)) {
+        r300_emit_texture_count(r300);
+
         for (i = 0; i < MIN2(r300->sampler_count, r300->texture_count); i++) {
            if (r300->dirty_state &
                ((R300_NEW_SAMPLER << i) | (R300_NEW_TEXTURE << i))) {
index 7c83c5166de9d8d17e7bc2458a4bc4719fb2537c..3797d3d332addfdb133ce978854145b99932d71b 100644 (file)
@@ -92,6 +92,8 @@ void r300_emit_vertex_shader(struct r300_context* r300,
 void r300_emit_viewport_state(struct r300_context* r300,
                               struct r300_viewport_state* viewport);
 
+void r300_emit_texture_count(struct r300_context* r300);
+
 void r300_flush_textures(struct r300_context* r300);
 
 /* Emit all dirty state. */
index 29ddc84c411e13edf5117ddfd02b876b5e43ceb5..4e1b61ca409bced507670f1235e91908e35afbc4 100644 (file)
@@ -1,6 +1,7 @@
 /*
  * Copyright 2008 Corbin Simpson <MostAwesomeDude@gmail.com>
  *                Joakim Sindholt <opensource@zhasha.com>
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -21,6 +22,9 @@
  * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
  * USE OR OTHER DEALINGS IN THE SOFTWARE. */
 
+#include "util/u_math.h"
+#include "util/u_memory.h"
+
 #include "tgsi/tgsi_dump.h"
 
 #include "r300_context.h"
 #include "radeon_code.h"
 #include "radeon_compiler.h"
 
+/* Convert info about FS input semantics to r300_shader_semantics. */
+void r300_shader_read_fs_inputs(struct tgsi_shader_info* info,
+                                struct r300_shader_semantics* fs_inputs)
+{
+    int i;
+    unsigned index;
+
+    r300_shader_semantics_reset(fs_inputs);
+
+    for (i = 0; i < info->num_inputs; i++) {
+        index = info->input_semantic_index[i];
+
+        switch (info->input_semantic_name[i]) {
+            case TGSI_SEMANTIC_COLOR:
+                assert(index <= ATTR_COLOR_COUNT);
+                fs_inputs->color[index] = i;
+                break;
+
+            case TGSI_SEMANTIC_GENERIC:
+                assert(index <= ATTR_GENERIC_COUNT);
+                fs_inputs->generic[index] = i;
+                break;
+
+            case TGSI_SEMANTIC_FOG:
+                assert(index == 0);
+                fs_inputs->fog = i;
+                break;
+
+            default:
+                assert(0);
+        }
+    }
+}
+
 static void find_output_registers(struct r300_fragment_program_compiler * compiler,
                                   struct r300_fragment_shader * fs)
 {
@@ -58,61 +96,65 @@ static void allocate_hardware_inputs(
     void (*allocate)(void * data, unsigned input, unsigned hwreg),
     void * mydata)
 {
-    struct tgsi_shader_info* info = &((struct r300_fragment_shader*)c->UserData)->info;
-    int total_colors = 0;
-    int colors = 0;
-    int total_generic = 0;
-    int generic = 0;
-    int i;
-
-    for (i = 0; i < info->num_inputs; i++) {
-        switch (info->input_semantic_name[i]) {
-            case TGSI_SEMANTIC_COLOR:
-                total_colors++;
-                break;
-            case TGSI_SEMANTIC_FOG:
-            case TGSI_SEMANTIC_GENERIC:
-                total_generic++;
-                break;
+    struct r300_shader_semantics* inputs =
+        (struct r300_shader_semantics*)c->UserData;
+    int i, reg = 0;
+
+    /* Allocate input registers. */
+    for (i = 0; i < ATTR_COLOR_COUNT; i++) {
+        if (inputs->color[i] != ATTR_UNUSED) {
+            allocate(mydata, inputs->color[i], reg++);
         }
     }
+    for (i = 0; i < ATTR_GENERIC_COUNT; i++) {
+        if (inputs->generic[i] != ATTR_UNUSED) {
+            allocate(mydata, inputs->generic[i], reg++);
+        }
+    }
+    if (inputs->fog != ATTR_UNUSED) {
+        allocate(mydata, inputs->fog, reg++);
+    }
+}
 
-    for(i = 0; i < info->num_inputs; i++) {
-        switch (info->input_semantic_name[i]) {
-            case TGSI_SEMANTIC_COLOR:
-                allocate(mydata, i, colors);
-                colors++;
-                break;
-            case TGSI_SEMANTIC_FOG:
-            case TGSI_SEMANTIC_GENERIC:
-                allocate(mydata, i, total_colors + generic);
-                generic++;
-                break;
+static void get_compare_state(
+    struct r300_context* r300,
+    struct r300_fragment_program_external_state* state,
+    unsigned shadow_samplers)
+{
+    memset(state, 0, sizeof(*state));
+
+    for (int i = 0; i < r300->sampler_count; i++) {
+        struct r300_sampler_state* s = r300->sampler_states[i];
+
+        if (s && s->state.compare_mode == PIPE_TEX_COMPARE_R_TO_TEXTURE) {
+            /* XXX Gallium doesn't provide us with any information regarding
+             * this mode, so we are screwed. I'm setting 0 = LUMINANCE. */
+            state->unit[i].depth_texture_mode = 0;
+
+            /* Fortunately, no need to translate this. */
+            state->unit[i].texture_compare_func = s->state.compare_func;
         }
     }
 }
 
-void r300_translate_fragment_shader(struct r300_context* r300,
-                                    struct r300_fragment_shader* fs)
+static void r300_translate_fragment_shader(
+    struct r300_context* r300,
+    struct r300_fragment_shader_code* shader)
 {
+    struct r300_fragment_shader* fs = r300->fs;
     struct r300_fragment_program_compiler compiler;
     struct tgsi_to_rc ttr;
 
+    /* Setup the compiler. */
     memset(&compiler, 0, sizeof(compiler));
     rc_init(&compiler.Base);
     compiler.Base.Debug = DBG_ON(r300, DBG_FP);
 
-    compiler.code = &fs->code;
+    compiler.code = &shader->code;
+    compiler.state = shader->compare_state;
     compiler.is_r500 = r300_screen(r300->context.screen)->caps->is_r500;
     compiler.AllocateHwInputs = &allocate_hardware_inputs;
-    compiler.UserData = fs;
-
-    /* TODO: Program compilation depends on texture compare modes,
-     * which are sampler state. Therefore, programs need to be recompiled
-     * depending on this state as in the classic Mesa driver.
-     *
-     * This is not yet handled correctly.
-     */
+    compiler.UserData = &fs->inputs;
 
     find_output_registers(&compiler, fs);
 
@@ -127,15 +169,64 @@ void r300_translate_fragment_shader(struct r300_context* r300,
 
     r300_tgsi_to_rc(&ttr, fs->state.tokens);
 
+    fs->shadow_samplers = compiler.Base.Program.ShadowSamplers;
+
     /* Invoke the compiler */
     r3xx_compile_fragment_program(&compiler);
     if (compiler.Base.Error) {
         /* XXX failover maybe? */
         DBG(r300, DBG_FP, "r300: Error compiling fragment program: %s\n",
             compiler.Base.ErrorMsg);
+        assert(0);
     }
 
     /* And, finally... */
     rc_destroy(&compiler.Base);
-    fs->translated = TRUE;
+}
+
+boolean r300_pick_fragment_shader(struct r300_context* r300)
+{
+    struct r300_fragment_shader* fs = r300->fs;
+    struct r300_fragment_program_external_state state;
+    struct r300_fragment_shader_code* ptr;
+
+    if (!fs->first) {
+        /* Build the fragment shader for the first time. */
+        fs->first = fs->shader = CALLOC_STRUCT(r300_fragment_shader_code);
+
+        /* BTW shadow samplers will be known after the first translation,
+         * therefore we set ~0, which means it should look at all sampler
+         * states. This choice doesn't have any impact on the correctness. */
+        get_compare_state(r300, &fs->shader->compare_state, ~0);
+        r300_translate_fragment_shader(r300, fs->shader);
+        return TRUE;
+
+    } else if (fs->shadow_samplers) {
+        get_compare_state(r300, &state, fs->shadow_samplers);
+
+        /* Check if the currently-bound shader has been compiled
+         * with the texture-compare state we need. */
+        if (memcmp(&fs->shader->compare_state, &state, sizeof(state)) != 0) {
+            /* Search for the right shader. */
+            ptr = fs->first;
+            while (ptr) {
+                if (memcmp(&ptr->compare_state, &state, sizeof(state)) == 0) {
+                    fs->shader = ptr;
+                    return TRUE;
+                }
+                ptr = ptr->next;
+            }
+
+            /* Not found, gotta compile a new one. */
+            ptr = CALLOC_STRUCT(r300_fragment_shader_code);
+            ptr->next = fs->first;
+            fs->first = fs->shader = ptr;
+
+            ptr->compare_state = state;
+            r300_translate_fragment_shader(r300, ptr);
+            return TRUE;
+        }
+    }
+
+    return FALSE;
 }
index e831c30301b2fd6386d1024d2a7bf017bca2a873..40ce874353c08bf2cb429bc4c0acc816ad26e475 100644 (file)
@@ -1,6 +1,7 @@
 /*
  * Copyright 2008 Corbin Simpson <MostAwesomeDude@gmail.com>
  *                Joakim Sindholt <opensource@zhasha.com>
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
 #define R300_FS_H
 
 #include "pipe/p_state.h"
-
 #include "tgsi/tgsi_scan.h"
-
 #include "radeon_code.h"
+#include "r300_shader_semantics.h"
+
+struct r300_fragment_shader_code {
+    struct r300_fragment_program_external_state compare_state;
+    struct rX00_fragment_program_code code;
+
+    struct r300_fragment_shader_code* next;
+};
 
 struct r300_fragment_shader {
     /* Parent class */
     struct pipe_shader_state state;
+
     struct tgsi_shader_info info;
+    struct r300_shader_semantics inputs;
 
-    /* Has this shader been translated yet? */
-    boolean translated;
+    /* Bits 0-15: TRUE if it's a shadow sampler, FALSE otherwise. */
+    unsigned shadow_samplers;
 
-    /* Compiled code */
-    struct rX00_fragment_program_code code;
+    /* Currently-bound fragment shader. */
+    struct r300_fragment_shader_code* shader;
+
+    /* List of the same shaders compiled with different texture-compare
+     * states. */
+    struct r300_fragment_shader_code* first;
 };
 
+void r300_shader_read_fs_inputs(struct tgsi_shader_info* info,
+                                struct r300_shader_semantics* fs_inputs);
 
-void r300_translate_fragment_shader(struct r300_context* r300,
-                                    struct r300_fragment_shader* fs);
+/* Return TRUE if the shader was switched and should be re-emitted. */
+boolean r300_pick_fragment_shader(struct r300_context* r300);
 
-static inline boolean r300_fragment_shader_writes_depth(struct r300_fragment_shader *fs)
+static INLINE boolean r300_fragment_shader_writes_depth(struct r300_fragment_shader *fs)
 {
     if (!fs)
-       return FALSE;
-    return (fs->code.writes_depth) ? TRUE : FALSE;
+        return FALSE;
+    return (fs->shader->code.writes_depth) ? TRUE : FALSE;
 }
+
 #endif /* R300_FS_H */
index 8ca785cb587b4db2b802372898176d3e4c328d0b..d8d08fbe26417ff889ceab23295542b57a05cc88 100644 (file)
@@ -661,20 +661,20 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #      define R300_GB_SUPER_TILE_B             (1 << 15)
 #      define R300_GB_SUBPIXEL_1_12            (0 << 16)
 #      define R300_GB_SUBPIXEL_1_16            (1 << 16)
-#      define GB_TILE_CONFIG_QUADS_PER_RAS_4   (0 << 17)
-#      define GB_TILE_CONFIG_QUADS_PER_RAS_8   (1 << 17)
-#      define GB_TILE_CONFIG_QUADS_PER_RAS_16  (2 << 17)
-#      define GB_TILE_CONFIG_QUADS_PER_RAS_32  (3 << 17)
-#      define GB_TILE_CONFIG_BB_SCAN_INTERCEPT (0 << 19)
-#      define GB_TILE_CONFIG_BB_SCAN_BOUND_BOX (1 << 19)
-#      define GB_TILE_CONFIG_ALT_SCAN_EN_LR    (0 << 20)
-#      define GB_TILE_CONFIG_ALT_SCAN_EN_LRL   (1 << 20)
-#      define GB_TILE_CONFIG_ALT_OFFSET        (0 << 21)
-#      define GB_TILE_CONFIG_SUBPRECISION      (0 << 22)
-#      define GB_TILE_CONFIG_ALT_TILING_DEF    (0 << 23)
-#      define GB_TILE_CONFIG_ALT_TILING_3_2    (1 << 23)
-#      define GB_TILE_CONFIG_Z_EXTENDED_24_1   (0 << 24)
-#      define GB_TILE_CONFIG_Z_EXTENDED_S25_1  (1 << 24)
+#      define R300_GB_TILE_CONFIG_QUADS_PER_RAS_4   (0 << 17)
+#      define R300_GB_TILE_CONFIG_QUADS_PER_RAS_8   (1 << 17)
+#      define R300_GB_TILE_CONFIG_QUADS_PER_RAS_16  (2 << 17)
+#      define R300_GB_TILE_CONFIG_QUADS_PER_RAS_32  (3 << 17)
+#      define R300_GB_TILE_CONFIG_BB_SCAN_INTERCEPT (0 << 19)
+#      define R300_GB_TILE_CONFIG_BB_SCAN_BOUND_BOX (1 << 19)
+#      define R300_GB_TILE_CONFIG_ALT_SCAN_EN_LR    (0 << 20)
+#      define R300_GB_TILE_CONFIG_ALT_SCAN_EN_LRL   (1 << 20)
+#      define R300_GB_TILE_CONFIG_ALT_OFFSET        (0 << 21)
+#      define R300_GB_TILE_CONFIG_SUBPRECISION      (0 << 22)
+#      define R300_GB_TILE_CONFIG_ALT_TILING_DEF    (0 << 23)
+#      define R300_GB_TILE_CONFIG_ALT_TILING_3_2    (1 << 23)
+#      define R300_GB_TILE_CONFIG_Z_EXTENDED_24_1   (0 << 24)
+#      define R300_GB_TILE_CONFIG_Z_EXTENDED_S25_1  (1 << 24)
 
 /* Specifies the sizes of the various FIFO`s in the sc/rs/us. This register must be the first one written */
 #define R300_GB_FIFO_SIZE      0x4024
@@ -700,9 +700,9 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #      define R300_OFIFO_HIGHWATER_SHIFT       22      /* two bits only */
 #      define R300_CUBE_FIFO_HIGHWATER_COL_SHIFT       24
 
-#define GB_Z_PEQ_CONFIG                          0x4028
-#      define GB_Z_PEQ_CONFIG_Z_PEQ_SIZE_4_4    (0 << 0)
-#      define GB_Z_PEQ_CONFIG_Z_PEQ_SIZE_8_8    (1 << 0)
+#define R300_GB_Z_PEQ_CONFIG                          0x4028
+#      define R300_GB_Z_PEQ_CONFIG_Z_PEQ_SIZE_4_4    (0 << 0)
+#      define R300_GB_Z_PEQ_CONFIG_Z_PEQ_SIZE_8_8    (1 << 0)
 
 /* Specifies various polygon specific selects (fog, depth, perspective). */
 #define R300_GB_SELECT                           0x401c
@@ -725,39 +725,39 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 /* Specifies the graphics pipeline configuration for antialiasing. */
 #define R300_GB_AA_CONFIG                         0x4020
-#      define GB_AA_CONFIG_AA_DISABLE           (0 << 0)
-#      define GB_AA_CONFIG_AA_ENABLE            (1 << 0)
-#      define GB_AA_CONFIG_NUM_AA_SUBSAMPLES_2  (0 << 1)
-#      define GB_AA_CONFIG_NUM_AA_SUBSAMPLES_3  (1 << 1)
-#      define GB_AA_CONFIG_NUM_AA_SUBSAMPLES_4  (2 << 1)
-#      define GB_AA_CONFIG_NUM_AA_SUBSAMPLES_6  (3 << 1)
+#      define R300_GB_AA_CONFIG_AA_DISABLE           (0 << 0)
+#      define R300_GB_AA_CONFIG_AA_ENABLE            (1 << 0)
+#      define R300_GB_AA_CONFIG_NUM_AA_SUBSAMPLES_2  (0 << 1)
+#      define R300_GB_AA_CONFIG_NUM_AA_SUBSAMPLES_3  (1 << 1)
+#      define R300_GB_AA_CONFIG_NUM_AA_SUBSAMPLES_4  (2 << 1)
+#      define R300_GB_AA_CONFIG_NUM_AA_SUBSAMPLES_6  (3 << 1)
 
 /* Selects which of 4 pipes are active. */
-#define GB_PIPE_SELECT                           0x402c
-#      define GB_PIPE_SELECT_PIPE0_ID_SHIFT  0
-#      define GB_PIPE_SELECT_PIPE1_ID_SHIFT  2
-#      define GB_PIPE_SELECT_PIPE2_ID_SHIFT  4
-#      define GB_PIPE_SELECT_PIPE3_ID_SHIFT  6
-#      define GB_PIPE_SELECT_PIPE_MASK_SHIFT 8
-#      define GB_PIPE_SELECT_MAX_PIPE        12
-#      define GB_PIPE_SELECT_BAD_PIPES       14
-#      define GB_PIPE_SELECT_CONFIG_PIPES    18
+#define R300_GB_PIPE_SELECT                           0x402c
+#      define R300_GB_PIPE_SELECT_PIPE0_ID_SHIFT  0
+#      define R300_GB_PIPE_SELECT_PIPE1_ID_SHIFT  2
+#      define R300_GB_PIPE_SELECT_PIPE2_ID_SHIFT  4
+#      define R300_GB_PIPE_SELECT_PIPE3_ID_SHIFT  6
+#      define R300_GB_PIPE_SELECT_PIPE_MASK_SHIFT 8
+#      define R300_GB_PIPE_SELECT_MAX_PIPE        12
+#      define R300_GB_PIPE_SELECT_BAD_PIPES       14
+#      define R300_GB_PIPE_SELECT_CONFIG_PIPES    18
 
 
 /* Specifies the sizes of the various FIFO`s in the sc/rs. */
-#define GB_FIFO_SIZE1                            0x4070
+#define R300_GB_FIFO_SIZE1                            0x4070
 /* High water mark for SC input fifo */
-#      define GB_FIFO_SIZE1_SC_HIGHWATER_IFIFO_SHIFT 0
-#      define GB_FIFO_SIZE1_SC_HIGHWATER_IFIFO_MASK  0x0000003f
+#      define R300_GB_FIFO_SIZE1_SC_HIGHWATER_IFIFO_SHIFT 0
+#      define R300_GB_FIFO_SIZE1_SC_HIGHWATER_IFIFO_MASK  0x0000003f
 /* High water mark for SC input fifo (B) */
-#      define GB_FIFO_SIZE1_SC_HIGHWATER_BFIFO_SHIFT 6
-#      define GB_FIFO_SIZE1_SC_HIGHWATER_BFIFO_MASK  0x00000fc0
+#      define R300_GB_FIFO_SIZE1_SC_HIGHWATER_BFIFO_SHIFT 6
+#      define R300_GB_FIFO_SIZE1_SC_HIGHWATER_BFIFO_MASK  0x00000fc0
 /* High water mark for RS colors' fifo */
-#      define GB_FIFO_SIZE1_SC_HIGHWATER_COL_SHIFT   12
-#      define GB_FIFO_SIZE1_SC_HIGHWATER_COL_MASK    0x0003f000
+#      define R300_GB_FIFO_SIZE1_SC_HIGHWATER_COL_SHIFT   12
+#      define R300_GB_FIFO_SIZE1_SC_HIGHWATER_COL_MASK    0x0003f000
 /* High water mark for RS textures' fifo */
-#      define GB_FIFO_SIZE1_SC_HIGHWATER_TEX_SHIFT   18
-#      define GB_FIFO_SIZE1_SC_HIGHWATER_TEX_MASK    0x00fc0000
+#      define R300_GB_FIFO_SIZE1_SC_HIGHWATER_TEX_SHIFT   18
+#      define R300_GB_FIFO_SIZE1_SC_HIGHWATER_TEX_MASK    0x00fc0000
 
 /* This table specifies the source location and format for up to 16 texture
  * addresses (i[0]:i[15]) and four colors (c[0]:c[3])
@@ -1293,7 +1293,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #        define R500_RS_INST_TEX_ID(x)                  ((x) << 0)
 #define R500_RS_INST_TEX_CN_WRITE                      (1 << 4)
 #define R500_RS_INST_TEX_ADDR_SHIFT                    5
-#        define R500_RS_INST_TEX_ADDR(x)                ((x) << 0)
+#        define R500_RS_INST_TEX_ADDR(x)                ((x) << 5)
 #define R500_RS_INST_COL_ID_SHIFT                      12
 #        define R500_RS_INST_COL_ID(x)                  ((x) << 12)
 #define R500_RS_INST_COL_CN_NO_WRITE                   (0 << 16)
@@ -1463,6 +1463,8 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #      define R300_TX_MIN_FILTER_MIP_NEAREST    (1 << 13)
 #      define R300_TX_MIN_FILTER_MIP_LINEAR     (2 << 13)
 #      define R300_TX_MIN_FILTER_MIP_MASK       (3 << 13)
+#       define R300_TX_MAX_MIP_LEVEL_SHIFT       17
+#       define R300_TX_MAX_MIP_LEVEL_MASK        (0xf << 17)
 #      define R300_TX_MAX_ANISO_1_TO_1          (0 << 21)
 #      define R300_TX_MAX_ANISO_2_TO_1          (1 << 21)
 #      define R300_TX_MAX_ANISO_4_TO_1          (2 << 21)
@@ -1471,6 +1473,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #      define R300_TX_MAX_ANISO_MASK            (7 << 21)
 #       define R300_TX_WRAP_S(x)                 ((x) << 0)
 #       define R300_TX_WRAP_T(x)                 ((x) << 3)
+#       define R300_TX_MAX_MIP_LEVEL(x)          ((x) << 17)
 
 #define R300_TX_FILTER1_0                      0x4440
 #      define R300_CHROMA_KEY_MODE_DISABLE    0
@@ -1500,8 +1503,6 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #       define R300_TX_HEIGHTMASK_MASK           (2047 << 11)
 #      define R300_TX_DEPTHMASK_SHIFT           22
 #      define R300_TX_DEPTHMASK_MASK            (0xf << 22)
-#       define R300_TX_MAX_MIP_LEVEL_SHIFT       26
-#       define R300_TX_MAX_MIP_LEVEL_MASK        (0xf << 26)
 #       define R300_TX_SIZE_PROJECTED            (1 << 30)
 #       define R300_TX_PITCH_EN                  (1 << 31)
 #       define R300_TX_WIDTH(x)                  ((x) << 0)
@@ -2144,6 +2145,7 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 /* Unpipelined. */
 #define R300_RB3D_CCTL                      0x4e00
+#      define R300_RB3D_CCTL_NUM_MULTIWRITES(x)       (MAX2(((x)-1), 0) << 5)
 #      define R300_RB3D_CCTL_NUM_MULTIWRITES_1_BUFFER                (0 << 5)
 #      define R300_RB3D_CCTL_NUM_MULTIWRITES_2_BUFFERS               (1 << 5)
 #      define R300_RB3D_CCTL_NUM_MULTIWRITES_3_BUFFERS               (2 << 5)
@@ -3292,6 +3294,11 @@ enum {
  */
 #define R300_PACKET3_3D_LOAD_VBPNTR         0x00002F00
 
+#   define R300_VBPNTR_SIZE0(x)    ((x) >> 2)
+#   define R300_VBPNTR_STRIDE0(x)  (((x) >> 2) << 8)
+#   define R300_VBPNTR_SIZE1(x)    (((x) >> 2) << 16)
+#   define R300_VBPNTR_STRIDE1(x)  (((x) >> 2) << 24)
+
 #define R300_PACKET3_INDX_BUFFER            0x00003300
 #    define R300_INDX_BUFFER_DST_SHIFT          0
 #    define R300_INDX_BUFFER_SKIP_SHIFT         16
index 62e1456ed366406415c09fc9b8825de1e8869cc0..82089f91617408026ddb1d5bf9e300685b2f2ace 100644 (file)
@@ -37,7 +37,6 @@
 #include "r300_reg.h"
 #include "r300_render.h"
 #include "r300_state_derived.h"
-#include "r300_vbo.h"
 
 /* r300_render: Vertex and index buffer primitive emission. */
 #define R300_MAX_VBO_SIZE  (1024 * 1024)
@@ -70,14 +69,67 @@ uint32_t r300_translate_primitive(unsigned prim)
     }
 }
 
+static boolean r300_nothing_to_draw(struct r300_context *r300)
+{
+    return r300->rs_state->rs.scissor &&
+           r300->scissor_state->scissor.empty_area;
+}
+
+static uint32_t r300_provoking_vertex_fixes(struct r300_context *r300,
+                                            unsigned mode)
+{
+    uint32_t color_control = r300->rs_state->color_control;
+
+    /* By default (see r300_state.c:r300_create_rs_state) color_control is
+     * initialized to provoking the first vertex.
+     *
+     * Triangle fans must be reduced to the second vertex, not the first, in
+     * Gallium flatshade-first mode, as per the GL spec.
+     * (http://www.opengl.org/registry/specs/ARB/provoking_vertex.txt)
+     *
+     * Quads never provoke correctly in flatshade-first mode. The first
+     * vertex is never considered as provoking, so only the second, third,
+     * and fourth vertices can be selected, and both "third" and "last" modes
+     * select the fourth vertex. This is probably due to D3D lacking quads.
+     *
+     * Similarly, polygons reduce to the first, not the last, vertex, when in
+     * "last" mode, and all other modes start from the second vertex.
+     *
+     * ~ C.
+     */
+
+    if (r300->rs_state->rs.flatshade_first) {
+        switch (mode) {
+            case PIPE_PRIM_TRIANGLE_FAN:
+                color_control |= R300_GA_COLOR_CONTROL_PROVOKING_VERTEX_SECOND;
+                break;
+            case PIPE_PRIM_QUADS:
+            case PIPE_PRIM_QUAD_STRIP:
+            case PIPE_PRIM_POLYGON:
+                color_control |= R300_GA_COLOR_CONTROL_PROVOKING_VERTEX_LAST;
+                break;
+            default:
+                color_control |= R300_GA_COLOR_CONTROL_PROVOKING_VERTEX_FIRST;
+                break;
+        }
+    } else {
+        color_control |= R300_GA_COLOR_CONTROL_PROVOKING_VERTEX_LAST;
+    }
+
+    return color_control;
+}
+
 static void r300_emit_draw_arrays(struct r300_context *r300,
                                   unsigned mode,
                                   unsigned count)
 {
     CS_LOCALS(r300);
 
-    BEGIN_CS(4);
-    OUT_CS_REG(R300_VAP_VF_MAX_VTX_INDX, count);
+    BEGIN_CS(8);
+    OUT_CS_REG(R300_GA_COLOR_CONTROL,
+            r300_provoking_vertex_fixes(r300, mode));
+    OUT_CS_REG(R300_VAP_VF_MIN_VTX_INDX, 0);
+    OUT_CS_REG(R300_VAP_VF_MAX_VTX_INDX, count - 1);
     OUT_CS_PKT3(R300_PACKET3_3D_DRAW_VBUF_2, 0);
     OUT_CS(R300_VAP_VF_CNTL__PRIM_WALK_VERTEX_LIST | (count << 16) |
            r300_translate_primitive(mode));
@@ -102,7 +154,10 @@ static void r300_emit_draw_elements(struct r300_context *r300,
     assert((start * indexSize)  % 4 == 0);
     assert(offset_dwords == 0);
 
-    BEGIN_CS(10);
+    BEGIN_CS(14);
+    OUT_CS_REG(R300_GA_COLOR_CONTROL,
+            r300_provoking_vertex_fixes(r300, mode));
+    OUT_CS_REG(R300_VAP_VF_MIN_VTX_INDX, minIndex);
     OUT_CS_REG(R300_VAP_VF_MAX_VTX_INDX, maxIndex);
     OUT_CS_PKT3(R300_PACKET3_3D_DRAW_INDX_2, 0);
     if (indexSize == 4) {
@@ -177,15 +232,24 @@ boolean r300_draw_range_elements(struct pipe_context* pipe,
         return FALSE;
     }
 
+    if (r300_nothing_to_draw(r300)) {
+        return TRUE;
+    }
+
     r300_update_derived_state(r300);
 
     if (!r300_setup_vertex_buffers(r300)) {
         return FALSE;
     }
 
-    setup_vertex_attributes(r300);
+    if (!r300->winsys->add_buffer(r300->winsys, indexBuffer,
+                                  RADEON_GEM_DOMAIN_GTT, 0)) {
+        return FALSE;
+    }
 
-    setup_index_buffer(r300, indexBuffer, indexSize);
+    if (!r300->winsys->validate(r300->winsys)) {
+        return FALSE;
+    }
 
     r300_emit_dirty_state(r300);
 
@@ -220,14 +284,16 @@ boolean r300_draw_arrays(struct pipe_context* pipe, unsigned mode,
         return FALSE;
     }
 
+    if (r300_nothing_to_draw(r300)) {
+        return TRUE;
+    }
+
     r300_update_derived_state(r300);
 
     if (!r300_setup_vertex_buffers(r300)) {
         return FALSE;
     }
 
-    setup_vertex_attributes(r300);
-
     r300_emit_dirty_state(r300);
 
     r300_emit_aos(r300, start);
@@ -255,6 +321,10 @@ boolean r300_swtcl_draw_arrays(struct pipe_context* pipe,
         return FALSE;
     }
 
+    if (r300_nothing_to_draw(r300)) {
+        return TRUE;
+    }
+
     for (i = 0; i < r300->vertex_buffer_count; i++) {
         void* buf = pipe_buffer_map(pipe->screen,
                                     r300->vertex_buffer[i].buffer,
@@ -265,8 +335,9 @@ boolean r300_swtcl_draw_arrays(struct pipe_context* pipe,
     draw_set_mapped_element_buffer(r300->draw, 0, NULL);
 
     draw_set_mapped_constant_buffer(r300->draw,
-            r300->shader_constants[PIPE_SHADER_VERTEX].constants,
-            r300->shader_constants[PIPE_SHADER_VERTEX].count *
+                                   PIPE_SHADER_VERTEX,
+                                   r300->shader_constants[PIPE_SHADER_VERTEX].constants,
+                                   r300->shader_constants[PIPE_SHADER_VERTEX].count *
                 (sizeof(float) * 4));
 
     draw_arrays(r300->draw, mode, start, count);
@@ -296,6 +367,10 @@ boolean r300_swtcl_draw_range_elements(struct pipe_context* pipe,
         return FALSE;
     }
 
+    if (r300_nothing_to_draw(r300)) {
+        return TRUE;
+    }
+
     for (i = 0; i < r300->vertex_buffer_count; i++) {
         void* buf = pipe_buffer_map(pipe->screen,
                                     r300->vertex_buffer[i].buffer,
@@ -309,6 +384,7 @@ boolean r300_swtcl_draw_range_elements(struct pipe_context* pipe,
                                          minIndex, maxIndex, indices);
 
     draw_set_mapped_constant_buffer(r300->draw,
+                                   PIPE_SHADER_VERTEX,
             r300->shader_constants[PIPE_SHADER_VERTEX].constants,
             r300->shader_constants[PIPE_SHADER_VERTEX].count *
                 (sizeof(float) * 4));
index 390b63007e5f2ee50660de39c6dc165382a00d19..2a8667d483556cdb1915da3c6fd35f78f56eb651 100644 (file)
  * USE OR OTHER DEALINGS IN THE SOFTWARE. */
 
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_memory.h"
 #include "util/u_simple_screen.h"
 
 #include "r300_context.h"
 #include "r300_screen.h"
 #include "r300_texture.h"
-#include "r300_winsys.h"
+
+#include "radeon_winsys.h"
 
 /* Return the identifier behind whom the brave coders responsible for this
  * amalgamation of code, sweat, and duct tape, routinely obscure their names.
@@ -140,6 +142,10 @@ static int r300_get_param(struct pipe_screen* pscreen, int param)
             return 0;
         case PIPE_CAP_BLEND_EQUATION_SEPARATE:
             return 1;
+        case PIPE_CAP_SM3:
+            return 1;
+        case PIPE_CAP_MAX_COMBINED_SAMPLERS:
+            return 8;
         default:
             debug_printf("r300: Implementation error: Bad param %d\n",
                 param);
@@ -219,12 +225,18 @@ static boolean check_tex_format(enum pipe_format format, uint32_t usage,
 
         /* Z buffer or texture */
         case PIPE_FORMAT_Z16_UNORM:
+            retval = usage &
+                (PIPE_TEXTURE_USAGE_DEPTH_STENCIL |
+                 PIPE_TEXTURE_USAGE_SAMPLER);
+            break;
+
+        /* 24bit Z buffer can only be used as a texture on R500. */
         case PIPE_FORMAT_Z24X8_UNORM:
         /* Z buffer with stencil or texture */
         case PIPE_FORMAT_Z24S8_UNORM:
             retval = usage &
                 (PIPE_TEXTURE_USAGE_DEPTH_STENCIL |
-                 PIPE_TEXTURE_USAGE_SAMPLER);
+                 (is_r500 ? PIPE_TEXTURE_USAGE_SAMPLER : 0));
             break;
 
         /* Definitely unsupported formats. */
@@ -311,14 +323,10 @@ r300_get_tex_transfer(struct pipe_screen *screen,
     trans = CALLOC_STRUCT(r300_transfer);
     if (trans) {
         pipe_texture_reference(&trans->transfer.texture, texture);
-        trans->transfer.format = texture->format;
         trans->transfer.x = x;
         trans->transfer.y = y;
         trans->transfer.width = w;
         trans->transfer.height = h;
-        trans->transfer.block = texture->block;
-        trans->transfer.nblocksx = texture->nblocksx[level];
-        trans->transfer.nblocksy = texture->nblocksy[level];
         trans->transfer.stride = r300_texture_get_stride(tex, level);
         trans->transfer.usage = usage;
 
@@ -344,6 +352,7 @@ static void* r300_transfer_map(struct pipe_screen* screen,
 {
     struct r300_texture* tex = (struct r300_texture*)transfer->texture;
     char* map;
+    enum pipe_format format = tex->tex.format;
 
     map = pipe_buffer_map(screen, tex->buffer,
                           pipe_transfer_buffer_flags(transfer));
@@ -353,8 +362,8 @@ static void* r300_transfer_map(struct pipe_screen* screen,
     }
 
     return map + r300_transfer(transfer)->offset +
-        transfer->y / transfer->block.height * transfer->stride +
-        transfer->x / transfer->block.width * transfer->block.size;
+        transfer->y / util_format_get_blockheight(format) * transfer->stride +
+        transfer->x / util_format_get_blockwidth(format) * util_format_get_blocksize(format);
 }
 
 static void r300_transfer_unmap(struct pipe_screen* screen,
@@ -372,7 +381,7 @@ static void r300_destroy_screen(struct pipe_screen* pscreen)
     FREE(r300screen);
 }
 
-struct pipe_screen* r300_create_screen(struct r300_winsys* r300_winsys)
+struct pipe_screen* r300_create_screen(struct radeon_winsys* radeon_winsys)
 {
     struct r300_screen* r300screen = CALLOC_STRUCT(r300_screen);
     struct r300_capabilities* caps = CALLOC_STRUCT(r300_capabilities);
@@ -380,14 +389,14 @@ struct pipe_screen* r300_create_screen(struct r300_winsys* r300_winsys)
     if (!r300screen || !caps)
         return NULL;
 
-    caps->pci_id = r300_winsys->pci_id;
-    caps->num_frag_pipes = r300_winsys->gb_pipes;
-    caps->num_z_pipes = r300_winsys->z_pipes;
+    caps->pci_id = radeon_winsys->pci_id;
+    caps->num_frag_pipes = radeon_winsys->gb_pipes;
+    caps->num_z_pipes = radeon_winsys->z_pipes;
 
     r300_parse_chipset(caps);
 
     r300screen->caps = caps;
-    r300screen->screen.winsys = (struct pipe_winsys*)r300_winsys;
+    r300screen->screen.winsys = (struct pipe_winsys*)radeon_winsys;
     r300screen->screen.destroy = r300_destroy_screen;
     r300screen->screen.get_name = r300_get_name;
     r300screen->screen.get_vendor = r300_get_vendor;
index 41df31f670fdeff294c2a80c6d61c7c896a05c67..2217988adddb596758b2c83d1d1709d112c0dc8c 100644 (file)
@@ -27,6 +27,8 @@
 
 #include "r300_chipset.h"
 
+struct radeon_winsys;
+
 struct r300_screen {
     /* Parent class */
     struct pipe_screen screen;
@@ -56,6 +58,6 @@ r300_transfer(struct pipe_transfer* transfer)
 }
 
 /* Creates a new r300 screen. */
-struct pipe_screen* r300_create_screen(struct r300_winsys* r300_winsys);
+struct pipe_screen* r300_create_screen(struct radeon_winsys* radeon_winsys);
 
 #endif /* R300_SCREEN_H */
diff --git a/src/gallium/drivers/r300/r300_shader_semantics.h b/src/gallium/drivers/r300/r300_shader_semantics.h
new file mode 100644 (file)
index 0000000..85184e2
--- /dev/null
@@ -0,0 +1,64 @@
+/*
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the "Software"),
+ * to deal in the Software without restriction, including without limitation
+ * on the rights to use, copy, modify, merge, publish, distribute, sub
+ * license, and/or sell copies of the Software, and to permit persons to whom
+ * the Software is furnished to do so, subject to the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the next
+ * paragraph) shall be included in all copies or substantial portions of the
+ * Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE. */
+
+#ifndef R300_SHADER_SEMANTICS_H
+#define R300_SHADER_SEMANTICS_H
+
+#define ATTR_UNUSED             (-1)
+#define ATTR_COLOR_COUNT        2
+#define ATTR_GENERIC_COUNT      16
+
+/* This structure contains information about what attributes are written by VS
+ * or read by FS. (but not both) It's much easier to work with than
+ * tgsi_shader_info.
+ *
+ * The variables contain indices to tgsi_shader_info semantics and those
+ * indices are nothing else than input/output register numbers. */
+struct r300_shader_semantics {
+    int pos;
+    int psize;
+    int color[ATTR_COLOR_COUNT];
+    int bcolor[ATTR_COLOR_COUNT];
+    int generic[ATTR_GENERIC_COUNT];
+    int fog;
+};
+
+static INLINE void r300_shader_semantics_reset(
+    struct r300_shader_semantics* info)
+{
+    int i;
+
+    info->pos = ATTR_UNUSED;
+    info->psize = ATTR_UNUSED;
+    info->fog = ATTR_UNUSED;
+
+    for (i = 0; i < ATTR_COLOR_COUNT; i++) {
+        info->color[i] = ATTR_UNUSED;
+        info->bcolor[i] = ATTR_UNUSED;
+    }
+
+    for (i = 0; i < ATTR_GENERIC_COUNT; i++) {
+        info->generic[i] = ATTR_UNUSED;
+    }
+}
+
+#endif
index d1eced61db1a1ec45e71af72e7a69256010d355e..49072462ec3ff0e27ddd2d7a3fd80a9099ecdd4f 100644 (file)
@@ -151,9 +151,10 @@ static void r300_set_blend_color(struct pipe_context* pipe,
                                  const struct pipe_blend_color* color)
 {
     struct r300_context* r300 = r300_context(pipe);
+    union util_color uc;
 
-    util_pack_color(color->color, PIPE_FORMAT_A8R8G8B8_UNORM,
-            &r300->blend_color_state->blend_color);
+    util_pack_color(color->color, PIPE_FORMAT_A8R8G8B8_UNORM, &uc);
+    r300->blend_color_state->blend_color = uc.ui;
 
     /* XXX if FP16 blending is enabled, we should use the FP16 format */
     r300->blend_color_state->blend_color_red_alpha =
@@ -282,11 +283,29 @@ static void r300_delete_dsa_state(struct pipe_context* pipe,
     FREE(state);
 }
 
-static void r300_set_edgeflags(struct pipe_context* pipe,
-                               const unsigned* bitfield)
+static void r300_set_scissor_regs(const struct pipe_scissor_state* state,
+                                  struct r300_scissor_regs *scissor,
+                                  boolean is_r500)
 {
-    /* XXX you know it's bad when i915 has this blank too */
-    /* XXX and even worse, I have no idea WTF the bitfield is */
+    if (is_r500) {
+        scissor->top_left =
+            (state->minx << R300_SCISSORS_X_SHIFT) |
+            (state->miny << R300_SCISSORS_Y_SHIFT);
+        scissor->bottom_right =
+            ((state->maxx - 1) << R300_SCISSORS_X_SHIFT) |
+            ((state->maxy - 1) << R300_SCISSORS_Y_SHIFT);
+    } else {
+        /* Offset of 1440 in non-R500 chipsets. */
+        scissor->top_left =
+            ((state->minx + 1440) << R300_SCISSORS_X_SHIFT) |
+            ((state->miny + 1440) << R300_SCISSORS_Y_SHIFT);
+        scissor->bottom_right =
+            (((state->maxx - 1) + 1440) << R300_SCISSORS_X_SHIFT) |
+            (((state->maxy - 1) + 1440) << R300_SCISSORS_Y_SHIFT);
+    }
+
+    scissor->empty_area = state->minx >= state->maxx ||
+                          state->miny >= state->maxy;
 }
 
 static void
@@ -294,6 +313,7 @@ static void
                                const struct pipe_framebuffer_state* state)
 {
     struct r300_context* r300 = r300_context(pipe);
+    struct pipe_scissor_state scissor;
 
     if (r300->draw) {
         draw_flush(r300->draw);
@@ -301,7 +321,18 @@ static void
 
     r300->framebuffer_state = *state;
 
+    scissor.minx = scissor.miny = 0;
+    scissor.maxx = state->width;
+    scissor.maxy = state->height;
+    r300_set_scissor_regs(&scissor, &r300->scissor_state->framebuffer,
+                          r300_screen(r300->context.screen)->caps->is_r500);
+
+    /* Don't rely on the order of states being set for the first time. */
+    if (!r300->rs_state || !r300->rs_state->rs.scissor) {
+        r300->dirty_state |= R300_NEW_SCISSOR;
+    }
     r300->dirty_state |= R300_NEW_FRAMEBUFFERS;
+    r300->dirty_state |= R300_NEW_BLEND;
 }
 
 /* Create fragment shader state. */
@@ -317,6 +348,7 @@ static void* r300_create_fs_state(struct pipe_context* pipe,
     fs->state.tokens = tgsi_dup_tokens(shader->tokens);
 
     tgsi_scan_shader(shader->tokens, &fs->info);
+    r300_shader_read_fs_inputs(&fs->info, &fs->inputs);
 
     return (void*)fs;
 }
@@ -330,11 +362,10 @@ static void r300_bind_fs_state(struct pipe_context* pipe, void* shader)
     if (fs == NULL) {
         r300->fs = NULL;
         return;
-    } else if (!fs->translated) {
-        r300_translate_fragment_shader(r300, fs);
     }
 
     r300->fs = fs;
+    r300_pick_fragment_shader(r300);
 
     r300->dirty_state |= R300_NEW_FRAGMENT_SHADER | R300_NEW_FRAGMENT_SHADER_CONSTANTS;
 }
@@ -343,7 +374,14 @@ static void r300_bind_fs_state(struct pipe_context* pipe, void* shader)
 static void r300_delete_fs_state(struct pipe_context* pipe, void* shader)
 {
     struct r300_fragment_shader* fs = (struct r300_fragment_shader*)shader;
-    rc_constants_destroy(&fs->code.constants);
+    struct r300_fragment_shader_code *tmp, *ptr = fs->first;
+
+    while (ptr) {
+        tmp = ptr;
+        ptr = ptr->next;
+        rc_constants_destroy(&tmp->code.constants);
+        FREE(tmp);
+    }
     FREE((void*)fs->state.tokens);
     FREE(shader);
 }
@@ -382,8 +420,6 @@ static void* r300_create_rs_state(struct pipe_context* pipe,
     if (state->bypass_vs_clip_and_viewport ||
             !r300_screen(pipe->screen)->caps->has_tcl) {
         rs->vap_control_status |= R300_VAP_TCL_BYPASS;
-    } else {
-        rs->rs.bypass_vs_clip_and_viewport = TRUE;
     }
 
     rs->point_size = pack_float_16_6x(state->point_size) |
@@ -474,10 +510,6 @@ static void* r300_create_rs_state(struct pipe_context* pipe,
         rs->color_control = R300_SHADE_MODEL_SMOOTH;
     }
 
-    if (!state->flatshade_first) {
-        rs->color_control |= R300_GA_COLOR_CONTROL_PROVOKING_VERTEX_LAST;
-    }
-
     return (void*)rs;
 }
 
@@ -513,6 +545,9 @@ static void*
     struct r300_context* r300 = r300_context(pipe);
     struct r300_sampler_state* sampler = CALLOC_STRUCT(r300_sampler_state);
     int lod_bias;
+    union util_color uc;
+
+    sampler->state = *state;
 
     sampler->filter0 |=
         (r300_translate_wrap(state->wrap_s) << R300_TX_WRAP_S_SHIFT) |
@@ -523,14 +558,19 @@ static void*
                                                    state->mag_img_filter,
                                                    state->min_mip_filter);
 
+    /* Unfortunately, r300-r500 don't support floating-point mipmap lods. */
+    /* We must pass these to the emit function to clamp them properly. */
+    sampler->min_lod = MAX2((unsigned)state->min_lod, 0);
+    sampler->max_lod = MAX2((unsigned)ceilf(state->max_lod), 0);
+
     lod_bias = CLAMP((int)(state->lod_bias * 32), -(1 << 9), (1 << 9) - 1);
 
     sampler->filter1 |= lod_bias << R300_LOD_BIAS_SHIFT;
 
     sampler->filter1 |= r300_anisotropy(state->max_anisotropy);
 
-    util_pack_color(state->border_color, PIPE_FORMAT_A8R8G8B8_UNORM,
-                    &sampler->border_color);
+    util_pack_color(state->border_color, PIPE_FORMAT_A8R8G8B8_UNORM, &uc);
+    sampler->border_color = uc.ui;
 
     /* R500-specific fixups and optimizations */
     if (r300_screen(r300->context.screen)->caps->is_r500) {
@@ -559,6 +599,20 @@ static void r300_bind_sampler_states(struct pipe_context* pipe,
     }
 
     r300->sampler_count = count;
+
+    /* Pick a fragment shader based on the texture compare state. */
+    if (r300->fs && (r300->dirty_state & R300_ANY_NEW_SAMPLERS)) {
+        if (r300_pick_fragment_shader(r300)) {
+            r300->dirty_state |= R300_NEW_FRAGMENT_SHADER |
+                                 R300_NEW_FRAGMENT_SHADER_CONSTANTS;
+        }
+    }
+}
+
+static void r300_lacks_vertex_textures(struct pipe_context* pipe,
+                                       unsigned count,
+                                       void** states)
+{
 }
 
 static void r300_delete_sampler_state(struct pipe_context* pipe, void* state)
@@ -571,6 +625,7 @@ static void r300_set_sampler_textures(struct pipe_context* pipe,
                                       struct pipe_texture** texture)
 {
     struct r300_context* r300 = r300_context(pipe);
+    boolean is_r500 = r300_screen(r300->context.screen)->caps->is_r500;
     int i;
 
     /* XXX magic num */
@@ -578,13 +633,18 @@ static void r300_set_sampler_textures(struct pipe_context* pipe,
         return;
     }
     
-    r300->context.flush(&r300->context, 0, NULL);
-
     for (i = 0; i < count; i++) {
         if (r300->textures[i] != (struct r300_texture*)texture[i]) {
             pipe_texture_reference((struct pipe_texture**)&r300->textures[i],
                 texture[i]);
             r300->dirty_state |= (R300_NEW_TEXTURE << i);
+
+            /* R300-specific - set the texrect factor in a fragment shader */
+            if (!is_r500 && r300->textures[i]->is_npot) {
+                /* XXX It would be nice to re-emit just 1 constant,
+                 * XXX not all of them */
+                r300->dirty_state |= R300_NEW_FRAGMENT_SHADER_CONSTANTS;
+            }
         }
     }
 
@@ -604,24 +664,13 @@ static void r300_set_scissor_state(struct pipe_context* pipe,
 {
     struct r300_context* r300 = r300_context(pipe);
 
-    if (r300_screen(r300->context.screen)->caps->is_r500) {
-        r300->scissor_state->scissor_top_left =
-            (state->minx << R300_SCISSORS_X_SHIFT) |
-            (state->miny << R300_SCISSORS_Y_SHIFT);
-        r300->scissor_state->scissor_bottom_right =
-            ((state->maxx - 1) << R300_SCISSORS_X_SHIFT) |
-            ((state->maxy - 1) << R300_SCISSORS_Y_SHIFT);
-    } else {
-        /* Offset of 1440 in non-R500 chipsets. */
-        r300->scissor_state->scissor_top_left =
-            ((state->minx + 1440) << R300_SCISSORS_X_SHIFT) |
-            ((state->miny + 1440) << R300_SCISSORS_Y_SHIFT);
-        r300->scissor_state->scissor_bottom_right =
-            (((state->maxx - 1) + 1440) << R300_SCISSORS_X_SHIFT) |
-            (((state->maxy - 1) + 1440) << R300_SCISSORS_Y_SHIFT);
-    }
+    r300_set_scissor_regs(state, &r300->scissor_state->scissor,
+                          r300_screen(r300->context.screen)->caps->is_r500);
 
-    r300->dirty_state |= R300_NEW_SCISSOR;
+    /* Don't rely on the order of states being set for the first time. */
+    if (!r300->rs_state || r300->rs_state->rs.scissor) {
+        r300->dirty_state |= R300_NEW_SCISSOR;
+    }
 }
 
 static void r300_set_viewport_state(struct pipe_context* pipe,
@@ -674,6 +723,8 @@ static void r300_set_vertex_buffers(struct pipe_context* pipe,
         draw_flush(r300->draw);
         draw_set_vertex_buffers(r300->draw, count, buffers);
     }
+
+    r300->dirty_state |= R300_NEW_VERTEX_FORMAT;
 }
 
 static void r300_set_vertex_elements(struct pipe_context* pipe,
@@ -706,9 +757,6 @@ static void* r300_create_vs_state(struct pipe_context* pipe,
 
         tgsi_scan_shader(shader->tokens, &vs->info);
 
-        /* Appease Draw. */
-        vs->draw = draw_create_vertex_shader(r300->draw, shader);
-
         return (void*)vs;
     } else {
         return draw_create_vertex_shader(r300->draw, shader);
@@ -719,8 +767,6 @@ static void r300_bind_vs_state(struct pipe_context* pipe, void* shader)
 {
     struct r300_context* r300 = r300_context(pipe);
 
-    draw_flush(r300->draw);
-
     if (r300_screen(pipe->screen)->caps->has_tcl) {
         struct r300_vertex_shader* vs = (struct r300_vertex_shader*)shader;
 
@@ -731,10 +777,10 @@ static void r300_bind_vs_state(struct pipe_context* pipe, void* shader)
             r300_translate_vertex_shader(r300, vs);
         }
 
-        draw_bind_vertex_shader(r300->draw, vs->draw);
         r300->vs = vs;
         r300->dirty_state |= R300_NEW_VERTEX_SHADER | R300_NEW_VERTEX_SHADER_CONSTANTS;
     } else {
+        draw_flush(r300->draw);
         draw_bind_vertex_shader(r300->draw,
                 (struct draw_vertex_shader*)shader);
     }
@@ -748,7 +794,6 @@ static void r300_delete_vs_state(struct pipe_context* pipe, void* shader)
         struct r300_vertex_shader* vs = (struct r300_vertex_shader*)shader;
 
         rc_constants_destroy(&vs->code.constants);
-        draw_delete_vertex_shader(r300->draw, vs->draw);
         FREE((void*)vs->state.tokens);
         FREE(shader);
     } else {
@@ -798,8 +843,6 @@ void r300_init_state_functions(struct r300_context* r300)
     r300->context.bind_depth_stencil_alpha_state = r300_bind_dsa_state;
     r300->context.delete_depth_stencil_alpha_state = r300_delete_dsa_state;
 
-    r300->context.set_edgeflags = r300_set_edgeflags;
-
     r300->context.set_framebuffer_state = r300_set_framebuffer_state;
 
     r300->context.create_fs_state = r300_create_fs_state;
@@ -813,10 +856,11 @@ void r300_init_state_functions(struct r300_context* r300)
     r300->context.delete_rasterizer_state = r300_delete_rs_state;
 
     r300->context.create_sampler_state = r300_create_sampler_state;
-    r300->context.bind_sampler_states = r300_bind_sampler_states;
+    r300->context.bind_fragment_sampler_states = r300_bind_sampler_states;
+    r300->context.bind_vertex_sampler_states = r300_lacks_vertex_textures;
     r300->context.delete_sampler_state = r300_delete_sampler_state;
 
-    r300->context.set_sampler_textures = r300_set_sampler_textures;
+    r300->context.set_fragment_sampler_textures = r300_set_sampler_textures;
 
     r300->context.set_scissor_state = r300_set_scissor_state;
 
index 7166694edf46646662f9568c5cd9195c546ef89b..727ae7ade6d57e7e62681e0af55a7ebdf4d99b40 100644 (file)
@@ -1,5 +1,6 @@
 /*
  * Copyright 2008 Corbin Simpson <MostAwesomeDude@gmail.com>
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
@@ -28,6 +29,7 @@
 #include "r300_context.h"
 #include "r300_fs.h"
 #include "r300_screen.h"
+#include "r300_shader_semantics.h"
 #include "r300_state_derived.h"
 #include "r300_state_inlines.h"
 #include "r300_vs.h"
@@ -47,8 +49,8 @@ struct r300_shader_derived_value {
 
 unsigned r300_shader_key_hash(void* key) {
     struct r300_shader_key* shader_key = (struct r300_shader_key*)key;
-    unsigned vs = (unsigned)shader_key->vs;
-    unsigned fs = (unsigned)shader_key->fs;
+    unsigned vs = (intptr_t)shader_key->vs;
+    unsigned fs = (intptr_t)shader_key->fs;
 
     return (vs << 16) | (fs & 0xffff);
 }
@@ -61,209 +63,152 @@ int r300_shader_key_compare(void* key1, void* key2) {
         (shader_key1->fs == shader_key2->fs);
 }
 
-/* Set up the vs_tab and routes. */
-static void r300_vs_tab_routes(struct r300_context* r300,
-                               struct r300_vertex_info* vformat)
+static void r300_draw_emit_attrib(struct r300_context* r300,
+                                  enum attrib_emit emit,
+                                  enum interp_mode interp,
+                                  int index)
 {
-    struct r300_screen* r300screen = r300_screen(r300->context.screen);
-    struct vertex_info* vinfo = &vformat->vinfo;
-    int* tab = vformat->vs_tab;
-    boolean pos = FALSE, psize = FALSE, fog = FALSE;
-    int i, texs = 0, cols = 0;
-    struct tgsi_shader_info* info;
+    struct tgsi_shader_info* info = &r300->vs->info;
+    int output;
 
-    if (r300screen->caps->has_tcl) {
-        /* Use vertex shader to determine required routes. */
-        info = &r300->vs->info;
+    output = draw_find_shader_output(r300->draw,
+                                     info->output_semantic_name[index],
+                                     info->output_semantic_index[index]);
+    draw_emit_vertex_attr(&r300->vertex_info->vinfo, emit, interp, output);
+}
+
+static void r300_draw_emit_all_attribs(struct r300_context* r300)
+{
+    struct r300_shader_semantics* vs_outputs = &r300->vs->outputs;
+    int i, gen_count;
+
+    /* Position. */
+    if (vs_outputs->pos != ATTR_UNUSED) {
+        r300_draw_emit_attrib(r300, EMIT_4F, INTERP_PERSPECTIVE,
+                              vs_outputs->pos);
     } else {
-        /* Use fragment shader to determine required routes. */
-        info = &r300->fs->info;
+        assert(0);
     }
 
-    assert(info->num_inputs <= 16);
-
-    if (!r300screen->caps->has_tcl || !r300->rs_state->enable_vte)
-    {
-        for (i = 0; i < info->num_inputs; i++) {
-            switch (r300->vs->code.inputs[i]) {
-                case TGSI_SEMANTIC_POSITION:
-                    pos = TRUE;
-                    tab[i] = 0;
-                    break;
-                case TGSI_SEMANTIC_COLOR:
-                    tab[i] = 2 + cols;
-                    cols++;
-                    break;
-                case TGSI_SEMANTIC_PSIZE:
-                    assert(psize == FALSE);
-                    psize = TRUE;
-                    tab[i] = 15;
-                    break;
-                case TGSI_SEMANTIC_FOG:
-                    assert(fog == FALSE);
-                    fog = TRUE;
-                    /* Fall through */
-                case TGSI_SEMANTIC_GENERIC:
-                    tab[i] = 6 + texs;
-                    texs++;
-                    break;
-                default:
-                    debug_printf("r300: Unknown vertex input %d\n",
-                        info->input_semantic_name[i]);
-                    break;
-            }
-        }
+    /* Point size. */
+    if (vs_outputs->psize != ATTR_UNUSED) {
+        r300_draw_emit_attrib(r300, EMIT_1F_PSIZE, INTERP_POS,
+                              vs_outputs->psize);
     }
-    else
-    {
-        /* Just copy vert attribs over as-is. */
-        for (i = 0; i < info->num_inputs; i++) {
-            tab[i] = i;
-        }
 
-        for (i = 0; i < info->num_outputs; i++) {
-            switch (info->output_semantic_name[i]) {
-                case TGSI_SEMANTIC_POSITION:
-                    pos = TRUE;
-                    break;
-                case TGSI_SEMANTIC_COLOR:
-                    cols++;
-                    break;
-                case TGSI_SEMANTIC_PSIZE:
-                    psize = TRUE;
-                    break;
-                case TGSI_SEMANTIC_FOG:
-                    fog = TRUE;
-                    /* Fall through */
-                case TGSI_SEMANTIC_GENERIC:
-                    texs++;
-                    break;
-                default:
-                    debug_printf("r300: Unknown vertex output %d\n",
-                        info->output_semantic_name[i]);
-                    break;
-            }
+    /* Colors. */
+    for (i = 0; i < ATTR_COLOR_COUNT; i++) {
+        if (vs_outputs->color[i] != ATTR_UNUSED) {
+            r300_draw_emit_attrib(r300, EMIT_4F, INTERP_LINEAR,
+                                  vs_outputs->color[i]);
         }
     }
 
-    /* XXX magic */
-    assert(texs <= 8);
+    /* XXX Back-face colors. */
 
-    /* Do the actual vertex_info setup.
-     *
-     * vertex_info has four uints of hardware-specific data in it.
-     * vinfo.hwfmt[0] is R300_VAP_VTX_STATE_CNTL
-     * vinfo.hwfmt[1] is R300_VAP_VSM_VTX_ASSM
-     * vinfo.hwfmt[2] is R300_VAP_OUTPUT_VTX_FMT_0
-     * vinfo.hwfmt[3] is R300_VAP_OUTPUT_VTX_FMT_1 */
-
-    vinfo->hwfmt[0] = 0x5555; /* XXX this is classic Mesa bonghits */
-
-    /* We need to add vertex position attribute only for SW TCL case,
-     * for HW TCL case it could be generated by vertex shader */
-    if (!pos && !r300screen->caps->has_tcl) {
-        debug_printf("r300: Forcing vertex position attribute emit...\n");
-        /* Make room for the position attribute
-         * at the beginning of the tab. */
-        for (i = 15; i > 0; i--) {
-            tab[i] = tab[i-1];
+    /* Texture coordinates. */
+    gen_count = 0;
+    for (i = 0; i < ATTR_GENERIC_COUNT; i++) {
+        if (vs_outputs->generic[i] != ATTR_UNUSED) {
+            r300_draw_emit_attrib(r300, EMIT_4F, INTERP_PERSPECTIVE,
+                                  vs_outputs->generic[i]);
+            gen_count++;
         }
-        tab[0] = 0;
     }
 
-    /* Position. */
-    if (r300->draw) {
-        draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_PERSPECTIVE,
-            draw_find_vs_output(r300->draw, TGSI_SEMANTIC_POSITION, 0));
+    /* Fog coordinates. */
+    if (vs_outputs->fog != ATTR_UNUSED) {
+        r300_draw_emit_attrib(r300, EMIT_4F, INTERP_PERSPECTIVE,
+                              vs_outputs->fog);
+        gen_count++;
     }
-    vinfo->hwfmt[1] |= R300_INPUT_CNTL_POS;
-    vinfo->hwfmt[2] |= R300_VAP_OUTPUT_VTX_FMT_0__POS_PRESENT;
 
-    /* Point size. */
-    if (psize) {
-        if (r300->draw) {
-            draw_emit_vertex_attr(vinfo, EMIT_1F_PSIZE, INTERP_POS,
-                draw_find_vs_output(r300->draw, TGSI_SEMANTIC_PSIZE, 0));
-        }
-        vinfo->hwfmt[2] |= R300_VAP_OUTPUT_VTX_FMT_0__PT_SIZE_PRESENT;
-    }
+    /* XXX magic */
+    assert(gen_count <= 8);
+}
 
-    /* Colors. */
-    for (i = 0; i < cols; i++) {
-        if (r300->draw) {
-            draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_LINEAR,
-                draw_find_vs_output(r300->draw, TGSI_SEMANTIC_COLOR, i));
-        }
-        vinfo->hwfmt[1] |= R300_INPUT_CNTL_COLOR;
-        vinfo->hwfmt[2] |= (R300_VAP_OUTPUT_VTX_FMT_0__COLOR_0_PRESENT << i);
+/* Update the PSC tables. */
+static void r300_vertex_psc(struct r300_context* r300)
+{
+    struct r300_vertex_info *vformat = r300->vertex_info;
+    uint16_t type, swizzle;
+    enum pipe_format format;
+    unsigned i;
+    int identity[16] = {0, 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15};
+    int* stream_tab;
+
+    /* If TCL is bypassed, map vertex streams to equivalent VS output
+     * locations. */
+    if (r300->rs_state->enable_vte) {
+        stream_tab = identity;
+    } else {
+        stream_tab = r300->vs->stream_loc_notcl;
     }
 
-    /* Init i right here, increment it if fog is enabled.
-     * This gets around a double-increment problem. */
-    i = 0;
+    /* Vertex shaders have no semantics on their inputs,
+     * so PSC should just route stuff based on the vertex elements,
+     * and not on attrib information. */
+    DBG(r300, DBG_DRAW, "r300: vs expects %d attribs, routing %d elements"
+            " in psc\n",
+            r300->vs->info.num_inputs,
+            r300->vertex_element_count);
 
-    /* Fog. This is a special-cased texcoord. */
-    if (fog) {
-        i++;
-        if (r300->draw) {
-            draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_PERSPECTIVE,
-                draw_find_vs_output(r300->draw, TGSI_SEMANTIC_FOG, 0));
-        }
-        vinfo->hwfmt[1] |= (R300_INPUT_CNTL_TC0 << i);
-        vinfo->hwfmt[3] |= (4 << (3 * i));
-    }
+    for (i = 0; i < r300->vertex_element_count; i++) {
+        format = r300->vertex_element[i].src_format;
+
+        type = r300_translate_vertex_data_type(format) |
+            (stream_tab[i] << R300_DST_VEC_LOC_SHIFT);
+        swizzle = r300_translate_vertex_data_swizzle(format);
 
-    /* Texcoords. */
-    for (; i < texs; i++) {
-        if (r300->draw) {
-            draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_PERSPECTIVE,
-                draw_find_vs_output(r300->draw, TGSI_SEMANTIC_GENERIC, i));
+        if (i & 1) {
+            vformat->vap_prog_stream_cntl[i >> 1] |= type << 16;
+            vformat->vap_prog_stream_cntl_ext[i >> 1] |= swizzle << 16;
+        } else {
+            vformat->vap_prog_stream_cntl[i >> 1] |= type;
+            vformat->vap_prog_stream_cntl_ext[i >> 1] |= swizzle;
         }
-        vinfo->hwfmt[1] |= (R300_INPUT_CNTL_TC0 << i);
-        vinfo->hwfmt[3] |= (4 << (3 * i));
     }
 
-    draw_compute_vertex_size(vinfo);
+    assert(i <= 15);
+
+    /* Set the last vector in the PSC. */
+    if (i) {
+        i -= 1;
+    }
+    vformat->vap_prog_stream_cntl[i >> 1] |=
+        (R300_LAST_VEC << (i & 1 ? 16 : 0));
 }
 
-/* Update the PSC tables. */
-static void r300_vertex_psc(struct r300_context* r300,
-                            struct r300_vertex_info* vformat)
+/* Update the PSC tables for SW TCL, using Draw. */
+static void r300_swtcl_vertex_psc(struct r300_context* r300)
 {
-    struct r300_screen* r300screen = r300_screen(r300->context.screen);
+    struct r300_vertex_info *vformat = r300->vertex_info;
     struct vertex_info* vinfo = &vformat->vinfo;
-    int* tab = vformat->vs_tab;
     uint16_t type, swizzle;
     enum pipe_format format;
     unsigned i, attrib_count;
+    int* vs_output_tab = r300->vs->stream_loc_notcl;
 
-    /* Vertex shaders have no semantics on their inputs,
-     * so PSC should just route stuff based on their info,
-     * and not on attrib information. */
-    if (r300screen->caps->has_tcl) {
-        attrib_count = r300->vs->info.num_inputs;
-        DBG(r300, DBG_DRAW, "r300: routing %d attribs in psc for vs\n",
-                attrib_count);
-    } else {
-        attrib_count = vinfo->num_attribs;
-        DBG(r300, DBG_DRAW, "r300: attrib count: %d\n", attrib_count);
-        for (i = 0; i < attrib_count; i++) {
-            DBG(r300, DBG_DRAW, "r300: attrib: offset %d, interp %d, size %d,"
-                   " tab %d\n", vinfo->attrib[i].src_index,
-                   vinfo->attrib[i].interp_mode, vinfo->attrib[i].emit,
-                   tab[i]);
-        }
+    /* For each Draw attribute, route it to the fragment shader according
+     * to the vs_output_tab. */
+    attrib_count = vinfo->num_attribs;
+    DBG(r300, DBG_DRAW, "r300: attrib count: %d\n", attrib_count);
+    for (i = 0; i < attrib_count; i++) {
+        DBG(r300, DBG_DRAW, "r300: attrib: offset %d, interp %d, size %d,"
+               " vs_output_tab %d\n", vinfo->attrib[i].src_index,
+               vinfo->attrib[i].interp_mode, vinfo->attrib[i].emit,
+               vs_output_tab[i]);
     }
 
     for (i = 0; i < attrib_count; i++) {
         /* Make sure we have a proper destination for our attribute. */
-        assert(tab[i] != -1);
+        assert(vs_output_tab[i] != -1);
 
         format = draw_translate_vinfo_format(vinfo->attrib[i].emit);
 
         /* Obtain the type of data in this attribute. */
         type = r300_translate_vertex_data_type(format) |
-            tab[i] << R300_DST_VEC_LOC_SHIFT;
+            vs_output_tab[i] << R300_DST_VEC_LOC_SHIFT;
 
         /* Obtain the swizzle for this attribute. Note that the default
          * swizzle in the hardware is not XYZW! */
@@ -272,12 +217,10 @@ static void r300_vertex_psc(struct r300_context* r300,
         /* Add the attribute to the PSC table. */
         if (i & 1) {
             vformat->vap_prog_stream_cntl[i >> 1] |= type << 16;
-
             vformat->vap_prog_stream_cntl_ext[i >> 1] |= swizzle << 16;
         } else {
-            vformat->vap_prog_stream_cntl[i >> 1] |= type <<  0;
-
-            vformat->vap_prog_stream_cntl_ext[i >> 1] |= swizzle << 0;
+            vformat->vap_prog_stream_cntl[i >> 1] |= type;
+            vformat->vap_prog_stream_cntl_ext[i >> 1] |= swizzle;
         }
     }
 
@@ -289,185 +232,200 @@ static void r300_vertex_psc(struct r300_context* r300,
         (R300_LAST_VEC << (i & 1 ? 16 : 0));
 }
 
-/* Set up the mappings from GB to US, for RS block. */
-static void r300_update_fs_tab(struct r300_context* r300,
-                               struct r300_vertex_info* vformat)
+static void r300_rs_col(struct r300_rs_block* rs, int id, int ptr,
+                        boolean swizzle_0001)
+{
+    rs->ip[id] |= R300_RS_COL_PTR(ptr);
+    if (swizzle_0001) {
+        rs->ip[id] |= R300_RS_COL_FMT(R300_RS_COL_FMT_0001);
+    } else {
+        rs->ip[id] |= R300_RS_COL_FMT(R300_RS_COL_FMT_RGBA);
+    }
+    rs->inst[id] |= R300_RS_INST_COL_ID(id);
+}
+
+static void r300_rs_col_write(struct r300_rs_block* rs, int id, int fp_offset)
 {
-    struct tgsi_shader_info* info = &r300->fs->info;
-    int i, cols = 0, texs = 0, cols_emitted = 0;
-    int* tab = vformat->fs_tab;
+    rs->inst[id] |= R300_RS_INST_COL_CN_WRITE |
+                    R300_RS_INST_COL_ADDR(fp_offset);
+}
 
-    for (i = 0; i < 16; i++) {
-        tab[i] = -1;
+static void r300_rs_tex(struct r300_rs_block* rs, int id, int ptr,
+                        boolean swizzle_X001)
+{
+    if (swizzle_X001) {
+        rs->ip[id] |= R300_RS_TEX_PTR(ptr*4) |
+                      R300_RS_SEL_S(R300_RS_SEL_C0) |
+                      R300_RS_SEL_T(R300_RS_SEL_K0) |
+                      R300_RS_SEL_R(R300_RS_SEL_K0) |
+                      R300_RS_SEL_Q(R300_RS_SEL_K1);
+    } else {
+        rs->ip[id] |= R300_RS_TEX_PTR(ptr*4) |
+                      R300_RS_SEL_S(R300_RS_SEL_C0) |
+                      R300_RS_SEL_T(R300_RS_SEL_C1) |
+                      R300_RS_SEL_R(R300_RS_SEL_C2) |
+                      R300_RS_SEL_Q(R300_RS_SEL_C3);
     }
+    rs->inst[id] |= R300_RS_INST_TEX_ID(id);
+}
 
-    assert(info->num_inputs <= 16);
-    for (i = 0; i < info->num_inputs; i++) {
-        switch (info->input_semantic_name[i]) {
-            case TGSI_SEMANTIC_COLOR:
-                tab[i] = INTERP_LINEAR;
-                cols++;
-                break;
-            case TGSI_SEMANTIC_POSITION:
-            case TGSI_SEMANTIC_PSIZE:
-                debug_printf("r300: Implementation error: Can't use "
-                        "pos attribs in fragshader yet!\n");
-                /* Pass through for now */
-            case TGSI_SEMANTIC_FOG:
-            case TGSI_SEMANTIC_GENERIC:
-                tab[i] = INTERP_PERSPECTIVE;
-                break;
-            default:
-                debug_printf("r300: Unknown vertex input %d\n",
-                    info->input_semantic_name[i]);
-                break;
-        }
+static void r300_rs_tex_write(struct r300_rs_block* rs, int id, int fp_offset)
+{
+    rs->inst[id] |= R300_RS_INST_TEX_CN_WRITE |
+                    R300_RS_INST_TEX_ADDR(fp_offset);
+}
+
+static void r500_rs_col(struct r300_rs_block* rs, int id, int ptr,
+                        boolean swizzle_0001)
+{
+    rs->ip[id] |= R500_RS_COL_PTR(ptr);
+    if (swizzle_0001) {
+        rs->ip[id] |= R500_RS_COL_FMT(R300_RS_COL_FMT_0001);
+    } else {
+        rs->ip[id] |= R500_RS_COL_FMT(R300_RS_COL_FMT_RGBA);
     }
+    rs->inst[id] |= R500_RS_INST_COL_ID(id);
+}
 
-    /* Now that we know where everything is... */
-    DBG(r300, DBG_DRAW, "r300: fp input count: %d\n", info->num_inputs);
-    for (i = 0; i < info->num_inputs; i++) {
-        switch (tab[i]) {
-            case INTERP_LINEAR:
-                DBG(r300, DBG_DRAW, "r300: attrib: "
-                        "stack offset %d, color,    tab %d\n",
-                        i, cols_emitted);
-                tab[i] = cols_emitted;
-                cols_emitted++;
-                break;
-            case INTERP_PERSPECTIVE:
-                DBG(r300, DBG_DRAW, "r300: attrib: "
-                        "stack offset %d, texcoord, tab %d\n",
-                        i, cols + texs);
-                tab[i] = cols + texs;
-                texs++;
-                break;
-            case -1:
-                debug_printf("r300: Implementation error: Bad fp interp!\n");
-            default:
-                break;
-        }
+static void r500_rs_col_write(struct r300_rs_block* rs, int id, int fp_offset)
+{
+    rs->inst[id] |= R500_RS_INST_COL_CN_WRITE |
+                    R500_RS_INST_COL_ADDR(fp_offset);
+}
+
+static void r500_rs_tex(struct r300_rs_block* rs, int id, int ptr,
+                        boolean swizzle_X001)
+{
+    int rs_tex_comp = ptr*4;
+
+    if (swizzle_X001) {
+        rs->ip[id] |= R500_RS_SEL_S(rs_tex_comp) |
+                      R500_RS_SEL_T(R500_RS_IP_PTR_K0) |
+                      R500_RS_SEL_R(R500_RS_IP_PTR_K0) |
+                      R500_RS_SEL_Q(R500_RS_IP_PTR_K1);
+    } else {
+        rs->ip[id] |= R500_RS_SEL_S(rs_tex_comp) |
+                      R500_RS_SEL_T(rs_tex_comp + 1) |
+                      R500_RS_SEL_R(rs_tex_comp + 2) |
+                      R500_RS_SEL_Q(rs_tex_comp + 3);
     }
+    rs->inst[id] |= R500_RS_INST_TEX_ID(id);
+}
 
+static void r500_rs_tex_write(struct r300_rs_block* rs, int id, int fp_offset)
+{
+    rs->inst[id] |= R500_RS_INST_TEX_CN_WRITE |
+                    R500_RS_INST_TEX_ADDR(fp_offset);
 }
 
-/* Set up the RS block. This is the part of the chipset that actually does
- * the rasterization of vertices into fragments. This is also the part of the
- * chipset that locks up if any part of it is even slightly wrong. */
+/* Set up the RS block.
+ *
+ * This is the part of the chipset that actually does the rasterization
+ * of vertices into fragments. This is also the part of the chipset that
+ * locks up if any part of it is even slightly wrong. */
 static void r300_update_rs_block(struct r300_context* r300,
-                                 struct r300_rs_block* rs)
+                                 struct r300_shader_semantics* vs_outputs,
+                                 struct r300_shader_semantics* fs_inputs)
 {
-    struct tgsi_shader_info* info = &r300->fs->info;
-    int col_count = 0, fp_offset = 0, i, tex_count = 0;
-    int rs_tex_comp = 0;
+    struct r300_rs_block* rs = r300->rs_block;
+    int i, col_count = 0, tex_count = 0, fp_offset = 0;
+    void (*rX00_rs_col)(struct r300_rs_block*, int, int, boolean);
+    void (*rX00_rs_col_write)(struct r300_rs_block*, int, int);
+    void (*rX00_rs_tex)(struct r300_rs_block*, int, int, boolean);
+    void (*rX00_rs_tex_write)(struct r300_rs_block*, int, int);
 
     if (r300_screen(r300->context.screen)->caps->is_r500) {
-        for (i = 0; i < info->num_inputs; i++) {
-            switch (info->input_semantic_name[i]) {
-                case TGSI_SEMANTIC_COLOR:
-                    rs->ip[col_count] |=
-                        R500_RS_COL_PTR(col_count) |
-                        R500_RS_COL_FMT(R300_RS_COL_FMT_RGBA);
-                    col_count++;
-                    break;
-                case TGSI_SEMANTIC_GENERIC:
-                    rs->ip[tex_count] |=
-                        R500_RS_SEL_S(rs_tex_comp) |
-                        R500_RS_SEL_T(rs_tex_comp + 1) |
-                        R500_RS_SEL_R(rs_tex_comp + 2) |
-                        R500_RS_SEL_Q(rs_tex_comp + 3);
-                    tex_count++;
-                    rs_tex_comp += 4;
-                    break;
-                default:
-                    break;
-            }
-        }
+        rX00_rs_col       = r500_rs_col;
+        rX00_rs_col_write = r500_rs_col_write;
+        rX00_rs_tex       = r500_rs_tex;
+        rX00_rs_tex_write = r500_rs_tex_write;
+    } else {
+        rX00_rs_col       = r300_rs_col;
+        rX00_rs_col_write = r300_rs_col_write;
+        rX00_rs_tex       = r300_rs_tex;
+        rX00_rs_tex_write = r300_rs_tex_write;
+    }
 
-        /* Rasterize at least one color, or bad things happen. */
-        if ((col_count == 0) && (tex_count == 0)) {
-            rs->ip[0] |= R500_RS_COL_FMT(R300_RS_COL_FMT_0001);
+    /* Rasterize colors. */
+    for (i = 0; i < ATTR_COLOR_COUNT; i++) {
+        if (vs_outputs->color[i] != ATTR_UNUSED) {
+            /* Always rasterize if it's written by the VS,
+             * otherwise it locks up. */
+            rX00_rs_col(rs, col_count, i, FALSE);
+
+            /* Write it to the FS input register if it's used by the FS. */
+            if (fs_inputs->color[i] != ATTR_UNUSED) {
+                rX00_rs_col_write(rs, col_count, fp_offset);
+                fp_offset++;
+            }
             col_count++;
-        }
-
-        for (i = 0; i < tex_count; i++) {
-            rs->inst[i] |= R500_RS_INST_TEX_ID(i) |
-                R500_RS_INST_TEX_CN_WRITE | R500_RS_INST_TEX_ADDR(fp_offset);
-            fp_offset++;
-        }
-
-        for (i = 0; i < col_count; i++) {
-            rs->inst[i] |= R500_RS_INST_COL_ID(i) |
-                R500_RS_INST_COL_CN_WRITE | R500_RS_INST_COL_ADDR(fp_offset);
-            fp_offset++;
-        }
-    } else {
-        for (i = 0; i < info->num_inputs; i++) {
-            switch (info->input_semantic_name[i]) {
-                case TGSI_SEMANTIC_COLOR:
-                    rs->ip[col_count] |=
-                        R300_RS_COL_PTR(col_count) |
-                        R300_RS_COL_FMT(R300_RS_COL_FMT_RGBA);
-                    col_count++;
-                    break;
-                case TGSI_SEMANTIC_GENERIC:
-                    rs->ip[tex_count] |=
-                        R300_RS_TEX_PTR(rs_tex_comp) |
-                        R300_RS_SEL_S(R300_RS_SEL_C0) |
-                        R300_RS_SEL_T(R300_RS_SEL_C1) |
-                        R300_RS_SEL_R(R300_RS_SEL_C2) |
-                        R300_RS_SEL_Q(R300_RS_SEL_C3);
-                    tex_count++;
-                    rs_tex_comp+=4;
-                    break;
-                default:
-                    break;
+        } else {
+            /* Skip the FS input register, leave it uninitialized. */
+            /* If we try to set it to (0,0,0,1), it will lock up. */
+            if (fs_inputs->color[i] != ATTR_UNUSED) {
+                fp_offset++;
             }
         }
+    }
 
-        if (col_count == 0) {
-            rs->ip[0] |= R300_RS_COL_FMT(R300_RS_COL_FMT_0001);
-        }
-
-        if (tex_count == 0) {
-            rs->ip[0] |=
-                R300_RS_SEL_S(R300_RS_SEL_K0) |
-                R300_RS_SEL_T(R300_RS_SEL_K0) |
-                R300_RS_SEL_R(R300_RS_SEL_K0) |
-                R300_RS_SEL_Q(R300_RS_SEL_K1);
+    /* Rasterize texture coordinates. */
+    for (i = 0; i < ATTR_GENERIC_COUNT; i++) {
+        if (vs_outputs->generic[i] != ATTR_UNUSED) {
+            /* Always rasterize if it's written by the VS,
+             * otherwise it locks up. */
+            rX00_rs_tex(rs, tex_count, tex_count, FALSE);
+
+            /* Write it to the FS input register if it's used by the FS. */
+            if (fs_inputs->generic[i] != ATTR_UNUSED) {
+                rX00_rs_tex_write(rs, tex_count, fp_offset);
+                fp_offset++;
+            }
+            tex_count++;
+        } else {
+            /* Skip the FS input register, leave it uninitialized. */
+            /* If we try to set it to (0,0,0,1), it will lock up. */
+            if (fs_inputs->generic[i] != ATTR_UNUSED) {
+                fp_offset++;
+            }
         }
+    }
 
-        /* Rasterize at least one color, or bad things happen. */
-        if ((col_count == 0) && (tex_count == 0)) {
-            col_count++;
-        }
+    /* Rasterize fog coordinates. */
+    if (vs_outputs->fog != ATTR_UNUSED) {
+        /* Always rasterize if it's written by the VS,
+         * otherwise it locks up. */
+        rX00_rs_tex(rs, tex_count, tex_count, TRUE);
 
-        for (i = 0; i < tex_count; i++) {
-            rs->inst[i] |= R300_RS_INST_TEX_ID(i) |
-                R300_RS_INST_TEX_CN_WRITE | R300_RS_INST_TEX_ADDR(fp_offset);
+        /* Write it to the FS input register if it's used by the FS. */
+        if (fs_inputs->fog != ATTR_UNUSED) {
+            rX00_rs_tex_write(rs, tex_count, fp_offset);
             fp_offset++;
         }
-
-        for (i = 0; i < col_count; i++) {
-            rs->inst[i] |= R300_RS_INST_COL_ID(i) |
-                R300_RS_INST_COL_CN_WRITE | R300_RS_INST_COL_ADDR(fp_offset);
+        tex_count++;
+    } else {
+        /* Skip the FS input register, leave it uninitialized. */
+        /* If we try to set it to (0,0,0,1), it will lock up. */
+        if (fs_inputs->fog != ATTR_UNUSED) {
             fp_offset++;
         }
     }
 
-    rs->count = (rs_tex_comp) | (col_count << R300_IC_COUNT_SHIFT) |
+    /* Rasterize at least one color, or bad things happen. */
+    if (col_count == 0 && tex_count == 0) {
+        rX00_rs_col(rs, 0, 0, TRUE);
+        col_count++;
+    }
+
+    rs->count = (tex_count*4) | (col_count << R300_IC_COUNT_SHIFT) |
         R300_HIRES_EN;
 
-    rs->inst_count = MAX2(MAX2(col_count - 1, tex_count - 1), 0);
+    rs->inst_count = MAX3(col_count - 1, tex_count - 1, 0);
 }
 
 /* Update the vertex format. */
 static void r300_update_derived_shader_state(struct r300_context* r300)
 {
     struct r300_screen* r300screen = r300_screen(r300->context.screen);
-    struct r300_vertex_info* vformat;
-    struct r300_rs_block* rs_block;
-    int i;
 
     /*
     struct r300_shader_key* key;
@@ -495,27 +453,45 @@ static void r300_update_derived_shader_state(struct r300_context* r300)
             (void*)key, (void*)value);
     } */
 
-    /* XXX This will be refactored ASAP. */
-    vformat = CALLOC_STRUCT(r300_vertex_info);
-    rs_block = CALLOC_STRUCT(r300_rs_block);
+    /* Reset structures */
+    memset(r300->rs_block, 0, sizeof(struct r300_rs_block));
+    memset(r300->vertex_info, 0, sizeof(struct r300_vertex_info));
+    memcpy(r300->vertex_info->vinfo.hwfmt, r300->vs->hwfmt, sizeof(uint)*4);
 
-    for (i = 0; i < 16; i++) {
-        vformat->vs_tab[i] = -1;
-        vformat->fs_tab[i] = -1;
+    r300_update_rs_block(r300, &r300->vs->outputs, &r300->fs->inputs);
+
+    if (r300screen->caps->has_tcl) {
+        r300_vertex_psc(r300);
+    } else {
+        r300_draw_emit_all_attribs(r300);
+        draw_compute_vertex_size(&r300->vertex_info->vinfo);
+        r300_swtcl_vertex_psc(r300);
     }
 
-    r300_vs_tab_routes(r300, vformat);
-    r300_vertex_psc(r300, vformat);
-    r300_update_fs_tab(r300, vformat);
+    r300->dirty_state |= R300_NEW_RS_BLOCK;
+}
 
-    r300_update_rs_block(r300, rs_block);
+static boolean r300_dsa_writes_depth_stencil(struct r300_dsa_state* dsa)
+{
+    /* We are interested only in the cases when a new depth or stencil value
+     * can be written and changed. */
+
+    /* We might optionally check for [Z func: never] and inspect the stencil
+     * state in a similar fashion, but it's not terribly important. */
+    return (dsa->z_buffer_control & R300_Z_WRITE_ENABLE) ||
+           (dsa->stencil_ref_mask & R300_STENCILWRITEMASK_MASK) ||
+           ((dsa->z_buffer_control & R500_STENCIL_REFMASK_FRONT_BACK) &&
+            (dsa->stencil_ref_bf & R300_STENCILWRITEMASK_MASK));
+}
 
-    FREE(r300->vertex_info);
-    FREE(r300->rs_block);
+static boolean r300_dsa_alpha_test_enabled(struct r300_dsa_state* dsa)
+{
+    /* We are interested only in the cases when alpha testing can kill
+     * a fragment. */
+    uint32_t af = dsa->alpha_function;
 
-    r300->vertex_info = vformat;
-    r300->rs_block = rs_block;
-    r300->dirty_state |= (R300_NEW_VERTEX_FORMAT | R300_NEW_RS_BLOCK);
+    return (af & R300_FG_ALPHA_FUNC_ENABLE) &&
+           (af & R300_FG_ALPHA_FUNC_ALWAYS) != R300_FG_ALPHA_FUNC_ALWAYS;
 }
 
 static void r300_update_ztop(struct r300_context* r300)
@@ -534,28 +510,34 @@ static void r300_update_ztop(struct r300_context* r300)
      * The docs claim that for the first three cases, if no ZS writes happen,
      * then ZTOP can be used.
      *
+     * (3) will never apply since we do not support chroma-keyed operations.
+     * (4) will need to be re-examined (and this comment updated) if/when
+     * Hyper-Z becomes supported.
+     *
      * Additionally, the following conditions require disabled ZTOP:
-     * ~) Depth writes in fragment shader
-     * ~) Outstanding occlusion queries
+     * 5) Depth writes in fragment shader
+     * 6) Outstanding occlusion queries
      *
      * ~C.
      */
-    if (r300->dsa_state->alpha_function) {
-        r300->ztop_state.z_buffer_top = R300_ZTOP_DISABLE;
-    } else if (r300->fs->info.uses_kill) {
+
+    /* ZS writes */
+    if (r300_dsa_writes_depth_stencil(r300->dsa_state) &&
+           (r300_dsa_alpha_test_enabled(r300->dsa_state) ||   /* (1) */
+            r300->fs->info.uses_kill)) {                      /* (2) */
         r300->ztop_state.z_buffer_top = R300_ZTOP_DISABLE;
-    } else if (r300_fragment_shader_writes_depth(r300->fs)) {
+    } else if (r300_fragment_shader_writes_depth(r300->fs)) { /* (5) */
         r300->ztop_state.z_buffer_top = R300_ZTOP_DISABLE;
-    } else if (r300->query_current) {
+    } else if (r300->query_current) {                         /* (6) */
         r300->ztop_state.z_buffer_top = R300_ZTOP_DISABLE;
     }
 }
 
 void r300_update_derived_state(struct r300_context* r300)
 {
-    /* XXX */
-    if (TRUE || r300->dirty_state &
-        (R300_NEW_FRAGMENT_SHADER | R300_NEW_VERTEX_SHADER)) {
+    if (r300->dirty_state &
+        (R300_NEW_FRAGMENT_SHADER | R300_NEW_VERTEX_SHADER |
+         R300_NEW_VERTEX_FORMAT)) {
         r300_update_derived_shader_state(r300);
     }
 
index e6c1cb54dac3b29881ddfc2e3fb55242c8512f1f..dbe42edd91068917c5797fa8c3ca3934114ab601 100644 (file)
@@ -28,6 +28,8 @@
 
 #include "pipe/p_format.h"
 
+#include "util/u_format.h"
+
 #include "r300_reg.h"
 
 /* Some maths. These should probably find their way to u_math, if needed. */
@@ -443,20 +445,22 @@ static INLINE uint32_t r300_translate_gb_pipes(int pipe_count)
 static INLINE unsigned pf_component_count(enum pipe_format format) {
     unsigned count = 0;
 
-    if (pf_layout(format) != PIPE_FORMAT_LAYOUT_RGBAZS) {
-        return count;
+    if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0)) {
+        count++;
     }
-
-    if (pf_size_x(format)) {
+    if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 1)) {
+        count++;
+    }
+    if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 2)) {
         count++;
     }
-    if (pf_size_y(format)) {
+    if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 3)) {
         count++;
     }
-    if (pf_size_z(format)) {
+    if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_ZS, 0)) {
         count++;
     }
-    if (pf_size_w(format)) {
+    if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_ZS, 1)) {
         count++;
     }
 
@@ -467,19 +471,23 @@ static INLINE unsigned pf_component_count(enum pipe_format format) {
 static INLINE uint16_t
 r300_translate_vertex_data_type(enum pipe_format format) {
     uint32_t result = 0;
+    const struct util_format_description *desc;
     unsigned components = pf_component_count(format);
 
-    if (pf_layout(format) != PIPE_FORMAT_LAYOUT_RGBAZS) {
+    desc = util_format_description(format);
+
+    if (desc->layout != UTIL_FORMAT_LAYOUT_ARITH &&
+        desc->layout != UTIL_FORMAT_LAYOUT_ARRAY) {
         debug_printf("r300: Bad format %s in %s:%d\n", pf_name(format),
             __FUNCTION__, __LINE__);
         assert(0);
     }
 
-    switch (pf_type(format)) {
+    switch (desc->channel[0].type) {
         /* Half-floats, floats, doubles */
-        case PIPE_FORMAT_TYPE_FLOAT:
-            switch (pf_size_x(format)) {
-                case 4:
+        case UTIL_FORMAT_TYPE_FLOAT:
+            switch (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0)) {
+                case 32:
                     result = R300_DATA_TYPE_FLOAT_1 + (components - 1);
                     break;
                 default:
@@ -488,19 +496,15 @@ r300_translate_vertex_data_type(enum pipe_format format) {
                     assert(0);
             }
             break;
-        /* Normalized unsigned ints */
-        case PIPE_FORMAT_TYPE_UNORM:
-        /* Normalized signed ints */
-        case PIPE_FORMAT_TYPE_SNORM:
-        /* Non-normalized unsigned ints */
-        case PIPE_FORMAT_TYPE_USCALED:
-        /* Non-normalized signed ints */
-        case PIPE_FORMAT_TYPE_SSCALED:
-            switch (pf_size_x(format)) {
-                case 1:
+        /* Unsigned ints */
+        case UTIL_FORMAT_TYPE_UNSIGNED:
+        /* Signed ints */
+        case UTIL_FORMAT_TYPE_SIGNED:
+            switch (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0)) {
+                case 8:
                     result = R300_DATA_TYPE_BYTE;
                     break;
-                case 2:
+                case 16:
                     if (components > 2) {
                         result = R300_DATA_TYPE_SHORT_4;
                     } else {
@@ -510,8 +514,8 @@ r300_translate_vertex_data_type(enum pipe_format format) {
                 default:
                     debug_printf("r300: Bad format %s in %s:%d\n",
                         pf_name(format), __FUNCTION__, __LINE__);
-                    debug_printf("r300: pf_size_x(format) == %d\n",
-                        pf_size_x(format));
+                    debug_printf("r300: util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0) == %d\n",
+                        util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0));
                     assert(0);
             }
             break;
@@ -521,12 +525,11 @@ r300_translate_vertex_data_type(enum pipe_format format) {
             assert(0);
     }
 
-    if (pf_type(format) == PIPE_FORMAT_TYPE_SSCALED) {
+    if (desc->channel[0].type == UTIL_FORMAT_TYPE_SIGNED) {
         result |= R300_SIGNED;
-    } else if (pf_type(format) == PIPE_FORMAT_TYPE_UNORM) {
+    }
+    if (desc->channel[0].normalized) {
         result |= R300_NORMALIZE;
-    } else if (pf_type(format) == PIPE_FORMAT_TYPE_SNORM) {
-        result |= (R300_SIGNED | R300_NORMALIZE);
     }
 
     return result;
@@ -534,17 +537,21 @@ r300_translate_vertex_data_type(enum pipe_format format) {
 
 static INLINE uint16_t
 r300_translate_vertex_data_swizzle(enum pipe_format format) {
+    const struct util_format_description *desc = util_format_description(format);
+
+    assert(format);
 
-    if (pf_layout(format) != PIPE_FORMAT_LAYOUT_RGBAZS) {
+    if (desc->layout != UTIL_FORMAT_LAYOUT_ARITH &&
+        desc->layout != UTIL_FORMAT_LAYOUT_ARRAY) {
         debug_printf("r300: Bad format %s in %s:%d\n",
             pf_name(format), __FUNCTION__, __LINE__);
         return 0;
     }
 
-    return ((pf_swizzle_x(format) << R300_SWIZZLE_SELECT_X_SHIFT) |
-        (pf_swizzle_y(format) << R300_SWIZZLE_SELECT_Y_SHIFT) |
-        (pf_swizzle_z(format) << R300_SWIZZLE_SELECT_Z_SHIFT) |
-        (pf_swizzle_w(format) << R300_SWIZZLE_SELECT_W_SHIFT) |
+    return ((desc->swizzle[0] << R300_SWIZZLE_SELECT_X_SHIFT) |
+        (desc->swizzle[1] << R300_SWIZZLE_SELECT_Y_SHIFT) |
+        (desc->swizzle[2] << R300_SWIZZLE_SELECT_Z_SHIFT) |
+        (desc->swizzle[3] << R300_SWIZZLE_SELECT_W_SHIFT) |
         (0xf << R300_WRITE_ENA_SHIFT));
 }
 
index c07e6ae676d962c61039975ecf5eb0ab847b8fc9..bcd4c030f9c0a5c25135d03f4300b4779359f6ce 100644 (file)
@@ -43,7 +43,7 @@ void r300_emit_invariant_state(struct r300_context* r300)
     struct r300_capabilities* caps = r300_screen(r300->context.screen)->caps;
     CS_LOCALS(r300);
 
-    BEGIN_CS(24 + (caps->has_tcl ? 2: 0));
+    BEGIN_CS(20 + (caps->has_tcl ? 2: 0));
 
     /*** Graphics Backend (GB) ***/
     /* Various GB enables */
@@ -70,9 +70,6 @@ void r300_emit_invariant_state(struct r300_context* r300)
     OUT_CS_REG(R300_US_W_FMT, 0x0);
 
     /*** VAP ***/
-    /* Max and min vertex index clamp. */
-    OUT_CS_REG(R300_VAP_VF_MIN_VTX_INDX, 0x0);
-    OUT_CS_REG(R300_VAP_VF_MAX_VTX_INDX, 0xffffff);
     /* Sign/normalize control */
     OUT_CS_REG(R300_VAP_PSC_SGN_NORM_CNTL, R300_SGN_NORM_NO_ZERO);
     /* TCL-only stuff */
@@ -84,15 +81,11 @@ void r300_emit_invariant_state(struct r300_context* r300)
     END_CS;
 
     /* XXX unsorted stuff from surface_fill */
-    BEGIN_CS(60 + (caps->has_tcl ? 5 : 0) + (caps->is_r500 ? 4 : 0));
-    /* Flush PVS. */
-    OUT_CS_REG(R300_VAP_PVS_STATE_FLUSH_REG, 0x0);
+    BEGIN_CS(44 + (caps->has_tcl ? 7 : 0) + (caps->is_r500 ? 4 : 0));
 
-    OUT_CS_REG(R300_SE_VTE_CNTL, R300_VPORT_X_SCALE_ENA |
-        R300_VPORT_X_OFFSET_ENA | R300_VPORT_Y_SCALE_ENA |
-        R300_VPORT_Y_OFFSET_ENA | R300_VPORT_Z_SCALE_ENA |
-        R300_VPORT_Z_OFFSET_ENA | R300_VTX_W0_FMT);
     if (caps->has_tcl) {
+        /*Flushing PVS is required before the VAP_GB registers can be changed*/
+        OUT_CS_REG(R300_VAP_PVS_STATE_FLUSH_REG, 0);
         OUT_CS_REG_SEQ(R300_VAP_GB_VERT_CLIP_ADJ, 4);
         OUT_CS_32F(1.0);
         OUT_CS_32F(1.0);
@@ -123,21 +116,15 @@ void r300_emit_invariant_state(struct r300_context* r300)
     OUT_CS_REG(R300_SU_DEPTH_OFFSET, 0x00000000);
     OUT_CS_REG(R300_SC_HYPERZ, 0x0000001C);
     OUT_CS_REG(R300_SC_EDGERULE, 0x2DA49525);
-    OUT_CS_REG(R300_RB3D_CCTL, 0x00000000);
     OUT_CS_REG(R300_RB3D_AARESOLVE_CTL, 0x00000000);
     if (caps->is_r500) {
         OUT_CS_REG(R500_RB3D_DISCARD_SRC_PIXEL_LTE_THRESHOLD, 0x00000000);
         OUT_CS_REG(R500_RB3D_DISCARD_SRC_PIXEL_GTE_THRESHOLD, 0xFFFFFFFF);
     }
-    OUT_CS_REG(R300_ZB_FORMAT, 0x00000002);
-    OUT_CS_REG(R300_ZB_ZCACHE_CTLSTAT, 0x00000003);
     OUT_CS_REG(R300_ZB_BW_CNTL, 0x00000000);
     OUT_CS_REG(R300_ZB_DEPTHCLEARVALUE, 0x00000000);
     OUT_CS_REG(R300_ZB_HIZ_OFFSET, 0x00000000);
     OUT_CS_REG(R300_ZB_HIZ_PITCH, 0x00000000);
-    OUT_CS_REG(R300_VAP_VTX_STATE_CNTL, 0x1);
-    OUT_CS_REG(R300_VAP_VSM_VTX_ASSM, 0x405);
-    OUT_CS_REG(R300_SE_VTE_CNTL, 0x0000043F);
 
     /* XXX */
     OUT_CS_REG(R300_SC_CLIP_RULE, 0xaaaa);
index aea25cf71dd48a7c16563bc254d2d3594e5dd3d3..9a96206a4dc186f72828191488466321cfe9e205 100644 (file)
@@ -22,6 +22,7 @@
 
 #include "pipe/p_screen.h"
 
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -34,8 +35,8 @@ static void r300_setup_texture_state(struct r300_texture* tex, boolean is_r500)
     struct r300_texture_state* state = &tex->state;
     struct pipe_texture *pt = &tex->tex;
 
-    state->format0 = R300_TX_WIDTH((pt->width[0] - 1) & 0x7ff) |
-                     R300_TX_HEIGHT((pt->height[0] - 1) & 0x7ff);
+    state->format0 = R300_TX_WIDTH((pt->width0 - 1) & 0x7ff) |
+                     R300_TX_HEIGHT((pt->height0 - 1) & 0x7ff);
 
     if (tex->is_npot) {
         /* rectangles love this */
@@ -43,8 +44,7 @@ static void r300_setup_texture_state(struct r300_texture* tex, boolean is_r500)
         state->format2 = (tex->pitch[0] - 1) & 0x1fff;
     } else {
         /* power of two textures (3D, mipmaps, and no pitch) */
-        state->format0 |= R300_TX_DEPTH(util_logbase2(pt->depth[0]) & 0xf) |
-                          R300_TX_NUM_LEVELS(pt->last_level & 0xf);
+        state->format0 |= R300_TX_DEPTH(util_logbase2(pt->depth0) & 0xf);
     }
 
     state->format1 = r300_translate_texformat(pt->format);
@@ -58,17 +58,17 @@ static void r300_setup_texture_state(struct r300_texture* tex, boolean is_r500)
     /* large textures on r500 */
     if (is_r500)
     {
-        if (pt->width[0] > 2048) {
+        if (pt->width0 > 2048) {
             state->format2 |= R500_TXWIDTH_BIT11;
         }
-        if (pt->height[0] > 2048) {
+        if (pt->height0 > 2048) {
             state->format2 |= R500_TXHEIGHT_BIT11;
         }
     }
-    assert(is_r500 || (pt->width[0] <= 2048 && pt->height[0] <= 2048));
+    assert(is_r500 || (pt->width0 <= 2048 && pt->height0 <= 2048));
 
     debug_printf("r300: Set texture state (%dx%d, %d levels)\n",
-                pt->width[0], pt->height[0], pt->last_level);
+                pt->width0, pt->height0, pt->last_level);
 }
 
 unsigned r300_texture_get_offset(struct r300_texture* tex, unsigned level,
@@ -106,7 +106,7 @@ unsigned r300_texture_get_stride(struct r300_texture* tex, unsigned level)
         return 0;
     }
 
-    return align(pf_get_stride(&tex->tex.block, tex->tex.width[level]), 32);
+    return align(util_format_get_stride(tex->tex.format, u_minify(tex->tex.width0, level)), 32);
 }
 
 static void r300_setup_miptree(struct r300_texture* tex)
@@ -116,39 +116,32 @@ static void r300_setup_miptree(struct r300_texture* tex)
     int i;
 
     for (i = 0; i <= base->last_level; i++) {
-        if (i > 0) {
-            base->width[i] = minify(base->width[i-1]);
-            base->height[i] = minify(base->height[i-1]);
-            base->depth[i] = minify(base->depth[i-1]);
-        }
-
-        base->nblocksx[i] = pf_get_nblocksx(&base->block, base->width[i]);
-        base->nblocksy[i] = pf_get_nblocksy(&base->block, base->height[i]);
+        unsigned nblocksy = util_format_get_nblocksy(base->format, u_minify(base->height0, i));
 
         stride = r300_texture_get_stride(tex, i);
-        layer_size = stride * base->nblocksy[i];
+        layer_size = stride * nblocksy;
 
         if (base->target == PIPE_TEXTURE_CUBE)
             size = layer_size * 6;
         else
-            size = layer_size * base->depth[i];
+            size = layer_size * u_minify(base->depth0, i);
 
         tex->offset[i] = align(tex->size, 32);
         tex->size = tex->offset[i] + size;
         tex->layer_size[i] = layer_size;
-        tex->pitch[i] = stride / base->block.size;
+        tex->pitch[i] = stride / util_format_get_blocksize(base->format);
 
         debug_printf("r300: Texture miptree: Level %d "
                 "(%dx%dx%d px, pitch %d bytes)\n",
-                i, base->width[i], base->height[i], base->depth[i],
-                stride);
+                i, u_minify(base->width0, i), u_minify(base->height0, i),
+                u_minify(base->depth0, i), stride);
     }
 }
 
 static void r300_setup_flags(struct r300_texture* tex)
 {
-    tex->is_npot = !util_is_power_of_two(tex->tex.width[0]) ||
-                   !util_is_power_of_two(tex->tex.height[0]);
+    tex->is_npot = !util_is_power_of_two(tex->tex.width0) ||
+                   !util_is_power_of_two(tex->tex.height0);
 }
 
 /* Create a new texture. */
@@ -208,8 +201,8 @@ static struct pipe_surface* r300_get_tex_surface(struct pipe_screen* screen,
         pipe_reference_init(&surface->reference, 1);
         pipe_texture_reference(&surface->texture, texture);
         surface->format = texture->format;
-        surface->width = texture->width[level];
-        surface->height = texture->height[level];
+        surface->width = u_minify(texture->width0, level);
+        surface->height = u_minify(texture->height0, level);
         surface->offset = offset;
         surface->usage = flags;
         surface->zslice = zslice;
@@ -237,7 +230,7 @@ static struct pipe_texture*
 
     /* Support only 2D textures without mipmaps */
     if (base->target != PIPE_TEXTURE_2D ||
-        base->depth[0] != 1 ||
+        base->depth0 != 1 ||
         base->last_level != 0) {
         return NULL;
     }
@@ -252,7 +245,7 @@ static struct pipe_texture*
     tex->tex.screen = screen;
 
     tex->stride_override = *stride;
-    tex->pitch[0] = *stride / base->block.size;
+    tex->pitch[0] = *stride / util_format_get_blocksize(base->format);
 
     r300_setup_flags(tex);
     r300_setup_texture_state(tex, r300_screen(screen)->caps->is_r500);
@@ -287,10 +280,9 @@ r300_video_surface_create(struct pipe_screen *screen,
     template.target = PIPE_TEXTURE_2D;
     template.format = PIPE_FORMAT_X8R8G8B8_UNORM;
     template.last_level = 0;
-    template.width[0] = util_next_power_of_two(width);
-    template.height[0] = util_next_power_of_two(height);
-    template.depth[0] = 1;
-    pf_get_block(template.format, &template.block);
+    template.width0 = util_next_power_of_two(width);
+    template.height0 = util_next_power_of_two(height);
+    template.depth0 = 1;
     template.tex_usage = PIPE_TEXTURE_USAGE_SAMPLER |
                          PIPE_TEXTURE_USAGE_RENDER_TARGET;
 
index 589f1984ee32f7a033761e5d50751ed3231705de..096cdb20bbe5a496fa0e6bb33332537dc8f54fc3 100644 (file)
@@ -190,10 +190,10 @@ static void transform_dstreg(
     struct rc_dst_register * dst,
     struct tgsi_full_dst_register * src)
 {
-    dst->File = translate_register_file(src->DstRegister.File);
-    dst->Index = translate_register_index(ttr, src->DstRegister.File, src->DstRegister.Index);
-    dst->WriteMask = src->DstRegister.WriteMask;
-    dst->RelAddr = src->DstRegister.Indirect;
+    dst->File = translate_register_file(src->Register.File);
+    dst->Index = translate_register_index(ttr, src->Register.File, src->Register.Index);
+    dst->WriteMask = src->Register.WriteMask;
+    dst->RelAddr = src->Register.Indirect;
 }
 
 static void transform_srcreg(
@@ -201,18 +201,19 @@ static void transform_srcreg(
     struct rc_src_register * dst,
     struct tgsi_full_src_register * src)
 {
-    dst->File = translate_register_file(src->SrcRegister.File);
-    dst->Index = translate_register_index(ttr, src->SrcRegister.File, src->SrcRegister.Index);
-    dst->RelAddr = src->SrcRegister.Indirect;
+    dst->File = translate_register_file(src->Register.File);
+    dst->Index = translate_register_index(ttr, src->Register.File, src->Register.Index);
+    dst->RelAddr = src->Register.Indirect;
     dst->Swizzle = tgsi_util_get_full_src_register_swizzle(src, 0);
     dst->Swizzle |= tgsi_util_get_full_src_register_swizzle(src, 1) << 3;
     dst->Swizzle |= tgsi_util_get_full_src_register_swizzle(src, 2) << 6;
     dst->Swizzle |= tgsi_util_get_full_src_register_swizzle(src, 3) << 9;
-    dst->Abs = src->SrcRegisterExtMod.Absolute;
-    dst->Negate = src->SrcRegister.Negate ? RC_MASK_XYZW : 0;
+    dst->Abs = src->Register.Absolute;
+    dst->Negate = src->Register.Negate ? RC_MASK_XYZW : 0;
 }
 
-static void transform_texture(struct rc_instruction * dst, struct tgsi_instruction_ext_texture src)
+static void transform_texture(struct rc_instruction * dst, struct tgsi_instruction_texture src,
+                              uint32_t *shadowSamplers)
 {
     switch(src.Texture) {
         case TGSI_TEXTURE_1D:
@@ -233,14 +234,17 @@ static void transform_texture(struct rc_instruction * dst, struct tgsi_instructi
         case TGSI_TEXTURE_SHADOW1D:
             dst->U.I.TexSrcTarget = RC_TEXTURE_1D;
             dst->U.I.TexShadow = 1;
+            *shadowSamplers |= 1 << dst->U.I.TexSrcUnit;
             break;
         case TGSI_TEXTURE_SHADOW2D:
             dst->U.I.TexSrcTarget = RC_TEXTURE_2D;
             dst->U.I.TexShadow = 1;
+            *shadowSamplers |= 1 << dst->U.I.TexSrcUnit;
             break;
         case TGSI_TEXTURE_SHADOWRECT:
             dst->U.I.TexSrcTarget = RC_TEXTURE_RECT;
             dst->U.I.TexShadow = 1;
+            *shadowSamplers |= 1 << dst->U.I.TexSrcUnit;
             break;
     }
 }
@@ -258,17 +262,19 @@ static void transform_instruction(struct tgsi_to_rc * ttr, struct tgsi_full_inst
     dst->U.I.SaturateMode = translate_saturate(src->Instruction.Saturate);
 
     if (src->Instruction.NumDstRegs)
-        transform_dstreg(ttr, &dst->U.I.DstReg, &src->FullDstRegisters[0]);
+        transform_dstreg(ttr, &dst->U.I.DstReg, &src->Dst[0]);
 
     for(i = 0; i < src->Instruction.NumSrcRegs; ++i) {
-        if (src->FullSrcRegisters[i].SrcRegister.File == TGSI_FILE_SAMPLER)
-            dst->U.I.TexSrcUnit = src->FullSrcRegisters[i].SrcRegister.Index;
+        if (src->Src[i].Register.File == TGSI_FILE_SAMPLER)
+            dst->U.I.TexSrcUnit = src->Src[i].Register.Index;
         else
-            transform_srcreg(ttr, &dst->U.I.SrcReg[i], &src->FullSrcRegisters[i]);
+            transform_srcreg(ttr, &dst->U.I.SrcReg[i], &src->Src[i]);
     }
 
     /* Texturing. */
-    transform_texture(dst, src->InstructionExtTexture);
+    if (src->Instruction.Texture)
+        transform_texture(dst, src->Texture,
+                          &ttr->compiler->Program.ShadowSamplers);
 }
 
 static void handle_immediate(struct tgsi_to_rc * ttr, struct tgsi_full_immediate * imm)
diff --git a/src/gallium/drivers/r300/r300_vbo.c b/src/gallium/drivers/r300/r300_vbo.c
deleted file mode 100644 (file)
index a6a1596..0000000
+++ /dev/null
@@ -1,132 +0,0 @@
-/*
- * Copyright 2009 Maciej Cencora <m.cencora@gmail.com>
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * on the rights to use, copy, modify, merge, publish, distribute, sub
- * license, and/or sell copies of the Software, and to permit persons to whom
- * the Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/* r300_vbo: Various helpers for emitting vertex buffers. Needs cleanup,
- * refactoring, etc. */
-
-#include "r300_vbo.h"
-
-#include "pipe/p_format.h"
-
-#include "r300_cs.h"
-#include "r300_context.h"
-#include "r300_state_inlines.h"
-#include "r300_reg.h"
-#include "r300_winsys.h"
-
-static INLINE void setup_vertex_attribute(struct r300_vertex_info *vinfo,
-                                          struct pipe_vertex_element *vert_elem,
-                                          unsigned attr_num)
-{
-    uint16_t hw_fmt1, hw_fmt2;
-
-    hw_fmt1 = r300_translate_vertex_data_type(vert_elem->src_format) |
-        (attr_num << R300_DST_VEC_LOC_SHIFT);
-    hw_fmt2 = r300_translate_vertex_data_swizzle(vert_elem->src_format);
-
-    if (attr_num % 2 == 0)
-    {
-        vinfo->vap_prog_stream_cntl[attr_num >> 1] = hw_fmt1;
-        vinfo->vap_prog_stream_cntl_ext[attr_num >> 1] = hw_fmt2;
-    }
-    else
-    {
-        vinfo->vap_prog_stream_cntl[attr_num >> 1] |= hw_fmt1 << 16;
-        vinfo->vap_prog_stream_cntl_ext[attr_num >> 1] |= hw_fmt2 << 16;
-    }
-}
-
-static void finish_vertex_attribs_setup(struct r300_vertex_info *vinfo,
-                                        unsigned attribs_num)
-{
-    uint32_t last_vec_bit = (attribs_num % 2 == 0) ?
-        (R300_LAST_VEC << 16) : R300_LAST_VEC;
-
-    assert(attribs_num > 0 && attribs_num <= 16);
-    vinfo->vap_prog_stream_cntl[(attribs_num - 1) >> 1] |= last_vec_bit;
-}
-
-void setup_vertex_attributes(struct r300_context *r300)
-{
-    struct pipe_vertex_element *vert_elem;
-    int i;
-
-    for (i = 0; i < r300->vertex_element_count; i++) {
-        vert_elem = &r300->vertex_element[i];
-        setup_vertex_attribute(r300->vertex_info, vert_elem, i);
-    }
-
-    finish_vertex_attribs_setup(r300->vertex_info,
-        r300->vertex_element_count);
-}
-
-static INLINE int get_buffer_offset(struct r300_context *r300,
-                                    unsigned int buf_nr,
-                                    unsigned int elem_offset)
-{
-    return r300->vertex_buffer[buf_nr].buffer_offset + elem_offset;
-}
-#if 0
-/* XXX not called at all */
-static void setup_vertex_buffers(struct r300_context *r300)
-{
-    struct pipe_vertex_element *vert_elem;
-    int i;
-
-    for (i = 0; i < r300->aos_count; i++)
-    {
-        vert_elem = &r300->vertex_element[i];
-            /* XXX use translate module to convert the data */
-        if (!format_is_supported(vert_elem->src_format,
-                                 vert_elem->nr_components)) {
-            assert(0);
-            /*
-            struct pipe_buffer *buf;
-            const unsigned int max_index = r300->vertex_buffers[vert_elem->vertex_buffer_index].max_index;
-            buf = pipe_buffer_create(r300->context.screen, 4, usage, vert_elem->nr_components * max_index * sizeof(float));
-            */
-        }
-
-        if (get_buffer_offset(r300,
-                              vert_elem->vertex_buffer_index,
-                              vert_elem->src_offset) % 4) {
-            /* XXX need to align buffer */
-            assert(0);
-        }
-    }
-}
-#endif
-/* XXX these shouldn't be asserts since we can work around bad indexbufs */
-void setup_index_buffer(struct r300_context *r300,
-                        struct pipe_buffer* indexBuffer,
-                        unsigned indexSize)
-{
-    if (!r300->winsys->add_buffer(r300->winsys, indexBuffer,
-                                  RADEON_GEM_DOMAIN_GTT, 0)) {
-        assert(0);
-    }
-
-    if (!r300->winsys->validate(r300->winsys)) {
-        assert(0);
-    }
-}
diff --git a/src/gallium/drivers/r300/r300_vbo.h b/src/gallium/drivers/r300/r300_vbo.h
deleted file mode 100644 (file)
index 7afa758..0000000
+++ /dev/null
@@ -1,36 +0,0 @@
-/*
- * Copyright 2009 Maciej Cencora <m.cencora@gmail.com>
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * on the rights to use, copy, modify, merge, publish, distribute, sub
- * license, and/or sell copies of the Software, and to permit persons to whom
- * the Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice (including the next
- * paragraph) shall be included in all copies or substantial portions of the
- * Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE AUTHOR(S) AND/OR THEIR SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-#ifndef R300_VBO_H
-#define R300_VBO_H
-
-struct r300_context;
-struct pipe_buffer;
-
-void setup_vertex_attributes(struct r300_context *r300);
-
-void setup_index_buffer(struct r300_context *r300,
-                        struct pipe_buffer* indexBuffer,
-                        unsigned indexSize);
-
-#endif
index 74ef416dc140149cb9272d9cf0313565aa6e3815..c4ed0d712f9cb84c6e3d15b5ed33f1207f27ceee 100644 (file)
@@ -1,5 +1,6 @@
 /*
  * Copyright 2009 Corbin Simpson <MostAwesomeDude@gmail.com>
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
 #include "r300_vs.h"
 
 #include "r300_context.h"
+#include "r300_screen.h"
 #include "r300_tgsi_to_rc.h"
+#include "r300_reg.h"
 
 #include "tgsi/tgsi_dump.h"
 #include "tgsi/tgsi_parse.h"
 
 #include "radeon_compiler.h"
 
-
-static void set_vertex_inputs_outputs(struct r300_vertex_program_compiler * c)
+/* Convert info about VS output semantics into r300_shader_semantics. */
+static void r300_shader_read_vs_outputs(
+    struct tgsi_shader_info* info,
+    struct r300_shader_semantics* vs_outputs)
 {
-    struct r300_vertex_shader * vs = c->UserData;
-    struct tgsi_shader_info* info = &vs->info;
-    struct tgsi_parse_context parser;
-    struct tgsi_full_declaration * decl;
-    boolean pointsize = FALSE;
-    int out_colors = 0;
-    int colors = 0;
-    int out_generic = 0;
-    int generic = 0;
     int i;
+    unsigned index;
 
-    /* Fill in the input mapping */
-    for (i = 0; i < info->num_inputs; i++)
-        c->code->inputs[i] = i;
+    r300_shader_semantics_reset(vs_outputs);
 
-    /* Fill in the output mapping */
     for (i = 0; i < info->num_outputs; i++) {
+        index = info->output_semantic_index[i];
+
         switch (info->output_semantic_name[i]) {
+            case TGSI_SEMANTIC_POSITION:
+                assert(index == 0);
+                vs_outputs->pos = i;
+                break;
+
             case TGSI_SEMANTIC_PSIZE:
-                pointsize = TRUE;
+                assert(index == 0);
+                vs_outputs->psize = i;
                 break;
+
             case TGSI_SEMANTIC_COLOR:
-                out_colors++;
+                assert(index <= ATTR_COLOR_COUNT);
+                vs_outputs->color[index] = i;
                 break;
-            case TGSI_SEMANTIC_FOG:
+
+            case TGSI_SEMANTIC_BCOLOR:
+                assert(index <= ATTR_COLOR_COUNT);
+                vs_outputs->bcolor[index] = i;
+                break;
+
             case TGSI_SEMANTIC_GENERIC:
-                out_generic++;
+                assert(index <= ATTR_GENERIC_COUNT);
+                vs_outputs->generic[index] = i;
                 break;
+
+            case TGSI_SEMANTIC_FOG:
+                assert(index == 0);
+                vs_outputs->fog = i;
+                break;
+
+            case TGSI_SEMANTIC_EDGEFLAG:
+                assert(index == 0);
+                fprintf(stderr, "r300 VP: cannot handle edgeflag output\n");
+                assert(0);
+                break;
+            default:
+                assert(0);
         }
     }
+}
 
-    tgsi_parse_init(&parser, vs->state.tokens);
+static void r300_shader_vap_output_fmt(
+    struct r300_shader_semantics* vs_outputs,
+    uint* hwfmt)
+{
+    int i, gen_count;
 
-    while (!tgsi_parse_end_of_tokens(&parser)) {
-        tgsi_parse_token(&parser);
+    /* Do the actual vertex_info setup.
+     *
+     * vertex_info has four uints of hardware-specific data in it.
+     * vinfo.hwfmt[0] is R300_VAP_VTX_STATE_CNTL
+     * vinfo.hwfmt[1] is R300_VAP_VSM_VTX_ASSM
+     * vinfo.hwfmt[2] is R300_VAP_OUTPUT_VTX_FMT_0
+     * vinfo.hwfmt[3] is R300_VAP_OUTPUT_VTX_FMT_1 */
 
-        if (parser.FullToken.Token.Type != TGSI_TOKEN_TYPE_DECLARATION)
-            continue;
+    hwfmt[0] = 0x5555; /* XXX this is classic Mesa bonghits */
 
-        decl = &parser.FullToken.FullDeclaration;
+    /* Position. */
+    if (vs_outputs->pos != ATTR_UNUSED) {
+        hwfmt[1] |= R300_INPUT_CNTL_POS;
+        hwfmt[2] |= R300_VAP_OUTPUT_VTX_FMT_0__POS_PRESENT;
+    } else {
+        assert(0);
+    }
 
-        if (decl->Declaration.File != TGSI_FILE_OUTPUT)
-            continue;
+    /* Point size. */
+    if (vs_outputs->psize != ATTR_UNUSED) {
+        hwfmt[2] |= R300_VAP_OUTPUT_VTX_FMT_0__PT_SIZE_PRESENT;
+    }
 
-        switch (decl->Semantic.SemanticName) {
-            case TGSI_SEMANTIC_POSITION:
-                c->code->outputs[decl->DeclarationRange.First] = 0;
-                break;
-            case TGSI_SEMANTIC_PSIZE:
-                c->code->outputs[decl->DeclarationRange.First] = 1;
-                break;
-            case TGSI_SEMANTIC_COLOR:
-                c->code->outputs[decl->DeclarationRange.First] = 1 +
-                    (pointsize ? 1 : 0) +
-                    colors++;
-                break;
-            case TGSI_SEMANTIC_FOG:
-            case TGSI_SEMANTIC_GENERIC:
-                c->code->outputs[decl->DeclarationRange.First] = 1 +
-                    (pointsize ? 1 : 0) +
-                    out_colors +
-                    generic++;
-                break;
-            default:
-                debug_printf("r300: vs: Bad semantic declaration %d\n",
-                    decl->Semantic.SemanticName);
-                break;
+    /* Colors. */
+    for (i = 0; i < ATTR_COLOR_COUNT; i++) {
+        if (vs_outputs->color[i] != ATTR_UNUSED) {
+            hwfmt[1] |= R300_INPUT_CNTL_COLOR;
+            hwfmt[2] |= R300_VAP_OUTPUT_VTX_FMT_0__COLOR_0_PRESENT << i;
+        }
+    }
+
+    /* XXX Back-face colors. */
+
+    /* Texture coordinates. */
+    gen_count = 0;
+    for (i = 0; i < ATTR_GENERIC_COUNT; i++) {
+        if (vs_outputs->generic[i] != ATTR_UNUSED) {
+            hwfmt[1] |= (R300_INPUT_CNTL_TC0 << gen_count);
+            hwfmt[3] |= (4 << (3 * gen_count));
+            gen_count++;
+        }
+    }
+
+    /* Fog coordinates. */
+    if (vs_outputs->fog != ATTR_UNUSED) {
+        hwfmt[1] |= (R300_INPUT_CNTL_TC0 << gen_count);
+        hwfmt[3] |= (4 << (3 * gen_count));
+        gen_count++;
+    }
+
+    /* XXX magic */
+    assert(gen_count <= 8);
+}
+
+/* Sets up stream mapping to equivalent VS outputs if TCL is bypassed
+ * or isn't present. */
+static void r300_stream_locations_notcl(
+    struct r300_shader_semantics* vs_outputs,
+    int* stream_loc)
+{
+    int i, tabi = 0, gen_count;
+
+    /* Position. */
+    stream_loc[tabi++] = 0;
+
+    /* Point size. */
+    if (vs_outputs->psize != ATTR_UNUSED) {
+        stream_loc[tabi++] = 1;
+    }
+
+    /* Colors. */
+    for (i = 0; i < ATTR_COLOR_COUNT; i++) {
+        if (vs_outputs->color[i] != ATTR_UNUSED) {
+            stream_loc[tabi++] = 2 + i;
+        }
+    }
+
+    /* Back-face colors. */
+    for (i = 0; i < ATTR_COLOR_COUNT; i++) {
+        if (vs_outputs->bcolor[i] != ATTR_UNUSED) {
+            stream_loc[tabi++] = 4 + i;
+        }
+    }
+
+    /* Texture coordinates. */
+    gen_count = 0;
+    for (i = 0; i < ATTR_GENERIC_COUNT; i++) {
+        if (vs_outputs->bcolor[i] != ATTR_UNUSED) {
+            assert(tabi < 16);
+            stream_loc[tabi++] = 6 + gen_count;
+            gen_count++;
         }
     }
 
-    tgsi_parse_free(&parser);
+    /* Fog coordinates. */
+    if (vs_outputs->fog != ATTR_UNUSED) {
+        assert(tabi < 16);
+        stream_loc[tabi++] = 6 + gen_count;
+        gen_count++;
+    }
+
+    /* XXX magic */
+    assert(gen_count <= 8);
+
+    for (; tabi < 16;) {
+        stream_loc[tabi++] = -1;
+    }
 }
 
+static void set_vertex_inputs_outputs(struct r300_vertex_program_compiler * c)
+{
+    struct r300_vertex_shader * vs = c->UserData;
+    struct r300_shader_semantics* outputs = &vs->outputs;
+    struct tgsi_shader_info* info = &vs->info;
+    int i, reg = 0;
+
+    /* Fill in the input mapping */
+    for (i = 0; i < info->num_inputs; i++)
+        c->code->inputs[i] = i;
+
+    /* Position. */
+    if (outputs->pos != ATTR_UNUSED) {
+        c->code->outputs[outputs->pos] = reg++;
+    } else {
+        assert(0);
+    }
+
+    /* Point size. */
+    if (outputs->psize != ATTR_UNUSED) {
+        c->code->outputs[outputs->psize] = reg++;
+    }
+
+    /* Colors. */
+    for (i = 0; i < ATTR_COLOR_COUNT; i++) {
+        if (outputs->color[i] != ATTR_UNUSED) {
+            c->code->outputs[outputs->color[i]] = reg++;
+        }
+    }
+
+    /* XXX Back-face colors. */
+
+    /* Texture coordinates. */
+    for (i = 0; i < ATTR_GENERIC_COUNT; i++) {
+        if (outputs->generic[i] != ATTR_UNUSED) {
+            c->code->outputs[outputs->generic[i]] = reg++;
+        }
+    }
+
+    /* Fog coordinates. */
+    if (outputs->fog != ATTR_UNUSED) {
+        c->code->outputs[outputs->fog] = reg++;
+    }
+}
 
 void r300_translate_vertex_shader(struct r300_context* r300,
                                   struct r300_vertex_shader* vs)
@@ -113,6 +254,11 @@ void r300_translate_vertex_shader(struct r300_context* r300,
     struct r300_vertex_program_compiler compiler;
     struct tgsi_to_rc ttr;
 
+    /* Initialize. */
+    r300_shader_read_vs_outputs(&vs->info, &vs->outputs);
+    r300_shader_vap_output_fmt(&vs->outputs, vs->hwfmt);
+    r300_stream_locations_notcl(&vs->outputs, vs->stream_loc_notcl);
+
     /* Setup the compiler */
     rc_init(&compiler.Base);
 
@@ -137,7 +283,7 @@ void r300_translate_vertex_shader(struct r300_context* r300,
     /* Invoke the compiler */
     r3xx_compile_vertex_program(&compiler);
     if (compiler.Base.Error) {
-        /* Todo: Fail gracefully */
+        /* XXX We should fallback using Draw. */
         fprintf(stderr, "r300 VP: Compiler error\n");
         abort();
     }
index 2a4ce315e3213c06f85244d9ddffd8e381e95538..67e9db5366fd9ad6a85eeea5c859d1091f606773 100644 (file)
@@ -1,5 +1,6 @@
 /*
  * Copyright 2009 Corbin Simpson <MostAwesomeDude@gmail.com>
+ * Copyright 2009 Marek Olšák <maraeo@gmail.com>
  *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the "Software"),
 
 #include "pipe/p_state.h"
 #include "tgsi/tgsi_scan.h"
-
 #include "radeon_code.h"
 
+#include "r300_shader_semantics.h"
+
 struct r300_context;
 
 struct r300_vertex_shader {
     /* Parent class */
     struct pipe_shader_state state;
+
     struct tgsi_shader_info info;
+    struct r300_shader_semantics outputs;
+    uint hwfmt[4];
 
-    /* Fallback shader, because Draw has issues */
-    struct draw_vertex_shader* draw;
+    /* Stream locations for SWTCL or if TCL is bypassed. */
+    int stream_loc_notcl[16];
 
     /* Has this shader been translated yet? */
     boolean translated;
@@ -45,9 +50,6 @@ struct r300_vertex_shader {
     struct r300_vertex_program_code code;
 };
 
-
-extern struct r300_vertex_program_code r300_passthrough_vertex_shader;
-
 void r300_translate_vertex_shader(struct r300_context* r300,
                                   struct r300_vertex_shader* vs);
 
index 864a6146b22644a05ae521de29c95d954f0c54af..1ae6de70feeb8d460ab5c85c0a0fadf15dd37673 100644 (file)
@@ -35,76 +35,10 @@ extern "C" {
 #include "pipe/p_state.h"
 #include "pipe/internal/p_winsys_screen.h"
 
-struct r300_winsys {
-    /* Parent class */
-    struct pipe_winsys base;
-
-    /* Opaque Radeon-specific winsys object. */
-    void* radeon_winsys;
-
-    /* PCI ID */
-    uint32_t pci_id;
-
-    /* GB pipe count */
-    uint32_t gb_pipes;
-
-    /* Z pipe count (rv530 only) */
-    uint32_t z_pipes;
-
-    /* GART size. */
-    uint32_t gart_size;
-
-    /* VRAM size. */
-    uint32_t vram_size;
-
-    /* Add a pipe_buffer to the list of buffer objects to validate. */
-    boolean (*add_buffer)(struct r300_winsys* winsys,
-                          struct pipe_buffer* pbuffer,
-                          uint32_t rd,
-                          uint32_t wd);
-
-    /* Revalidate all currently setup pipe_buffers.
-     * Returns TRUE if a flush is required. */
-    boolean (*validate)(struct r300_winsys* winsys);
-
-    /* Check to see if there's room for commands. */
-    boolean (*check_cs)(struct r300_winsys* winsys, int size);
-
-    /* Start a command emit. */
-    void (*begin_cs)(struct r300_winsys* winsys,
-                     int size,
-                     const char* file,
-                     const char* function,
-                     int line);
-
-    /* Write a dword to the command buffer. */
-    void (*write_cs_dword)(struct r300_winsys* winsys, uint32_t dword);
-
-    /* Write a relocated dword to the command buffer. */
-    void (*write_cs_reloc)(struct r300_winsys* winsys,
-                           struct pipe_buffer* bo,
-                           uint32_t rd,
-                           uint32_t wd,
-                           uint32_t flags);
-
-    /* Finish a command emit. */
-    void (*end_cs)(struct r300_winsys* winsys,
-                   const char* file,
-                   const char* function,
-                   int line);
-
-    /* Flush the CS. */
-    void (*flush_cs)(struct r300_winsys* winsys);
-
-    /* winsys flush - callback from winsys when flush required */
-    void (*set_flush_cb)(struct r300_winsys *winsys,
-                        void (*flush_cb)(void *), void *data);
-
-    void (*reset_bos)(struct r300_winsys *winsys);
-};
+#include "radeon_winsys.h"
 
 struct pipe_context* r300_create_context(struct pipe_screen* screen,
-                                         struct r300_winsys* r300_winsys);
+                                         struct radeon_winsys* radeon_winsys);
 
 boolean r300_get_texture_buffer(struct pipe_texture* texture,
                                 struct pipe_buffer** buffer,
index 8fac8e6e05ffd528510f699f6f4b6a5c50be9854..f98087deb8c8c3094ca04bac354af36d1aa2ac64 100644 (file)
@@ -48,6 +48,7 @@ softpipe_clear(struct pipe_context *pipe, unsigned buffers, const float *rgba,
                double depth, unsigned stencil)
 {
    struct softpipe_context *softpipe = softpipe_context(pipe);
+   union util_color uc;
    unsigned cv;
    uint i;
 
@@ -62,12 +63,12 @@ softpipe_clear(struct pipe_context *pipe, unsigned buffers, const float *rgba,
       for (i = 0; i < softpipe->framebuffer.nr_cbufs; i++) {
          struct pipe_surface *ps = softpipe->framebuffer.cbufs[i];
 
-         util_pack_color(rgba, ps->format, &cv);
-         sp_tile_cache_clear(softpipe->cbuf_cache[i], rgba, cv);
+         util_pack_color(rgba, ps->format, &uc);
+         sp_tile_cache_clear(softpipe->cbuf_cache[i], rgba, uc.ui);
 
 #if !TILE_CLEAR_OPTIMIZATION
          /* non-cached surface */
-         pipe->surface_fill(pipe, ps, 0, 0, ps->width, ps->height, cv);
+         pipe->surface_fill(pipe, ps, 0, 0, ps->width, ps->height, uc.ui);
 #endif
       }
    }
index 5f60139968a5e72c4e109add2d92de940ff36719..82173a3c2a3ab7010d849e189f0cf97dc3ea221a 100644 (file)
@@ -90,14 +90,15 @@ softpipe_destroy( struct pipe_context *pipe )
    if (softpipe->draw)
       draw_destroy( softpipe->draw );
 
-      softpipe->quad.shade->destroy( softpipe->quad.shade );
-      softpipe->quad.depth_test->destroy( softpipe->quad.depth_test );
-      softpipe->quad.blend->destroy( softpipe->quad.blend );
+   softpipe->quad.shade->destroy( softpipe->quad.shade );
+   softpipe->quad.depth_test->destroy( softpipe->quad.depth_test );
+   softpipe->quad.blend->destroy( softpipe->quad.blend );
 
    for (i = 0; i < PIPE_MAX_COLOR_BUFS; i++) {
       sp_destroy_tile_cache(softpipe->cbuf_cache[i]);
       pipe_surface_reference(&softpipe->framebuffer.cbufs[i], NULL);
    }
+
    sp_destroy_tile_cache(softpipe->zsbuf_cache);
    pipe_surface_reference(&softpipe->framebuffer.zsbuf, NULL);
 
@@ -106,6 +107,11 @@ softpipe_destroy( struct pipe_context *pipe )
       pipe_texture_reference(&softpipe->texture[i], NULL);
    }
 
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      sp_destroy_tex_tile_cache(softpipe->vertex_tex_cache[i]);
+      pipe_texture_reference(&softpipe->vertex_textures[i], NULL);
+   }
+
    for (i = 0; i < Elements(softpipe->constants); i++) {
       if (softpipe->constants[i].buffer) {
          pipe_buffer_reference(&softpipe->constants[i].buffer, NULL);
@@ -152,6 +158,11 @@ softpipe_is_texture_referenced( struct pipe_context *pipe,
           softpipe->tex_cache[i]->texture == texture)
          return PIPE_REFERENCED_FOR_READ;
    }
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      if (softpipe->vertex_tex_cache[i] &&
+          softpipe->vertex_tex_cache[i]->texture == texture)
+         return PIPE_REFERENCED_FOR_READ;
+   }
    
    return PIPE_UNREFERENCED;
 }
@@ -180,6 +191,7 @@ softpipe_create( struct pipe_screen *screen )
 #endif
 
    softpipe->dump_fs = debug_get_bool_option( "GALLIUM_DUMP_FS", FALSE );
+   softpipe->dump_gs = debug_get_bool_option( "SOFTPIPE_DUMP_GS", FALSE );
 
    softpipe->pipe.winsys = screen->winsys;
    softpipe->pipe.screen = screen;
@@ -191,7 +203,8 @@ softpipe_create( struct pipe_screen *screen )
    softpipe->pipe.delete_blend_state = softpipe_delete_blend_state;
 
    softpipe->pipe.create_sampler_state = softpipe_create_sampler_state;
-   softpipe->pipe.bind_sampler_states  = softpipe_bind_sampler_states;
+   softpipe->pipe.bind_fragment_sampler_states  = softpipe_bind_sampler_states;
+   softpipe->pipe.bind_vertex_sampler_states = softpipe_bind_vertex_sampler_states;
    softpipe->pipe.delete_sampler_state = softpipe_delete_sampler_state;
 
    softpipe->pipe.create_depth_stencil_alpha_state = softpipe_create_depth_stencil_state;
@@ -210,13 +223,18 @@ softpipe_create( struct pipe_screen *screen )
    softpipe->pipe.bind_vs_state   = softpipe_bind_vs_state;
    softpipe->pipe.delete_vs_state = softpipe_delete_vs_state;
 
+   softpipe->pipe.create_gs_state = softpipe_create_gs_state;
+   softpipe->pipe.bind_gs_state   = softpipe_bind_gs_state;
+   softpipe->pipe.delete_gs_state = softpipe_delete_gs_state;
+
    softpipe->pipe.set_blend_color = softpipe_set_blend_color;
    softpipe->pipe.set_clip_state = softpipe_set_clip_state;
    softpipe->pipe.set_constant_buffer = softpipe_set_constant_buffer;
    softpipe->pipe.set_framebuffer_state = softpipe_set_framebuffer_state;
    softpipe->pipe.set_polygon_stipple = softpipe_set_polygon_stipple;
    softpipe->pipe.set_scissor_state = softpipe_set_scissor_state;
-   softpipe->pipe.set_sampler_textures = softpipe_set_sampler_textures;
+   softpipe->pipe.set_fragment_sampler_textures = softpipe_set_sampler_textures;
+   softpipe->pipe.set_vertex_sampler_textures = softpipe_set_vertex_sampler_textures;
    softpipe->pipe.set_viewport_state = softpipe_set_viewport_state;
 
    softpipe->pipe.set_vertex_buffers = softpipe_set_vertex_buffers;
@@ -225,8 +243,6 @@ softpipe_create( struct pipe_screen *screen )
    softpipe->pipe.draw_arrays = softpipe_draw_arrays;
    softpipe->pipe.draw_elements = softpipe_draw_elements;
    softpipe->pipe.draw_range_elements = softpipe_draw_range_elements;
-   softpipe->pipe.set_edgeflags = softpipe_set_edgeflags;
-
 
    softpipe->pipe.clear = softpipe_clear;
    softpipe->pipe.flush = softpipe_flush;
@@ -246,7 +262,9 @@ softpipe_create( struct pipe_screen *screen )
 
    for (i = 0; i < PIPE_MAX_SAMPLERS; i++)
       softpipe->tex_cache[i] = sp_create_tex_tile_cache( screen );
-
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      softpipe->vertex_tex_cache[i] = sp_create_tex_tile_cache(screen);
+   }
 
    /* setup quad rendering stages */
    softpipe->quad.shade = sp_quad_shade_stage(softpipe);
@@ -262,7 +280,7 @@ softpipe_create( struct pipe_screen *screen )
       goto fail;
 
    draw_texture_samplers(softpipe->draw,
-                         PIPE_MAX_SAMPLERS,
+                         PIPE_MAX_VERTEX_SAMPLERS,
                          (struct tgsi_sampler **)
                             softpipe->tgsi.vert_samplers_list);
 
index a735573d6fbe432012c6637ecf9cb10efc4764f0..6a89bd4b06213d80cd614f1b83aacd0606f5d698 100644 (file)
@@ -53,10 +53,12 @@ struct softpipe_context {
    /** Constant state objects */
    struct pipe_blend_state *blend;
    struct pipe_sampler_state *sampler[PIPE_MAX_SAMPLERS];
+   struct pipe_sampler_state *vertex_samplers[PIPE_MAX_VERTEX_SAMPLERS];
    struct pipe_depth_stencil_alpha_state *depth_stencil;
    struct pipe_rasterizer_state *rasterizer;
    struct sp_fragment_shader *fs;
    struct sp_vertex_shader *vs;
+   struct sp_geometry_shader *gs;
 
    /** Other rendering state */
    struct pipe_blend_color blend_color;
@@ -66,12 +68,15 @@ struct softpipe_context {
    struct pipe_poly_stipple poly_stipple;
    struct pipe_scissor_state scissor;
    struct pipe_texture *texture[PIPE_MAX_SAMPLERS];
+   struct pipe_texture *vertex_textures[PIPE_MAX_VERTEX_SAMPLERS];
    struct pipe_viewport_state viewport;
    struct pipe_vertex_buffer vertex_buffer[PIPE_MAX_ATTRIBS];
    struct pipe_vertex_element vertex_element[PIPE_MAX_ATTRIBS];
 
    unsigned num_samplers;
    unsigned num_textures;
+   unsigned num_vertex_samplers;
+   unsigned num_vertex_textures;
    unsigned num_vertex_elements;
    unsigned num_vertex_buffers;
 
@@ -121,7 +126,7 @@ struct softpipe_context {
 
    /** TGSI exec things */
    struct {
-      struct sp_sampler_varient *vert_samplers_list[PIPE_MAX_SAMPLERS];
+      struct sp_sampler_varient *vert_samplers_list[PIPE_MAX_VERTEX_SAMPLERS];
       struct sp_sampler_varient *frag_samplers_list[PIPE_MAX_SAMPLERS];
    } tgsi;
 
@@ -139,9 +144,11 @@ struct softpipe_context {
 
    unsigned tex_timestamp;
    struct softpipe_tex_tile_cache *tex_cache[PIPE_MAX_SAMPLERS];
+   struct softpipe_tex_tile_cache *vertex_tex_cache[PIPE_MAX_VERTEX_SAMPLERS];
 
    unsigned use_sse : 1;
    unsigned dump_fs : 1;
+   unsigned dump_gs : 1;
    unsigned no_rast : 1;
 };
 
index d4045816d03ed7a428d8ea6849d057ad84c3b956..9ea5d6fb9fbd3c16bce74668b6640f5065f7699c 100644 (file)
@@ -48,7 +48,7 @@ static void
 softpipe_map_constant_buffers(struct softpipe_context *sp)
 {
    struct pipe_winsys *ws = sp->pipe.winsys;
-   uint i, size;
+   uint i, vssize, gssize;
 
    for (i = 0; i < PIPE_SHADER_TYPES; i++) {
       if (sp->constants[i].buffer && sp->constants[i].buffer->size)
@@ -57,13 +57,21 @@ softpipe_map_constant_buffers(struct softpipe_context *sp)
    }
 
    if (sp->constants[PIPE_SHADER_VERTEX].buffer)
-      size = sp->constants[PIPE_SHADER_VERTEX].buffer->size;
+      vssize = sp->constants[PIPE_SHADER_VERTEX].buffer->size;
    else
-      size = 0;
+      vssize = 0;
 
-   draw_set_mapped_constant_buffer(sp->draw,
+   if (sp->constants[PIPE_SHADER_GEOMETRY].buffer)
+      gssize = sp->constants[PIPE_SHADER_GEOMETRY].buffer->size;
+   else
+      gssize = 0;
+
+   draw_set_mapped_constant_buffer(sp->draw, PIPE_SHADER_VERTEX,
                                    sp->mapped_constants[PIPE_SHADER_VERTEX],
-                                   size);
+                                   vssize);
+   draw_set_mapped_constant_buffer(sp->draw, PIPE_SHADER_GEOMETRY,
+                                   sp->mapped_constants[PIPE_SHADER_GEOMETRY],
+                                   gssize);
 }
 
 
@@ -78,9 +86,10 @@ softpipe_unmap_constant_buffers(struct softpipe_context *sp)
     */
    draw_flush(sp->draw);
 
-   draw_set_mapped_constant_buffer(sp->draw, NULL, 0);
+   draw_set_mapped_constant_buffer(sp->draw, PIPE_SHADER_VERTEX, NULL, 0);
+   draw_set_mapped_constant_buffer(sp->draw, PIPE_SHADER_GEOMETRY, NULL, 0);
 
-   for (i = 0; i < 2; i++) {
+   for (i = 0; i < PIPE_SHADER_TYPES; i++) {
       if (sp->constants[i].buffer && sp->constants[i].buffer->size)
          ws->buffer_unmap(ws, sp->constants[i].buffer);
       sp->mapped_constants[i] = NULL;
@@ -184,11 +193,3 @@ softpipe_draw_elements(struct pipe_context *pipe,
                                         0, 0xffffffff,
                                         mode, start, count );
 }
-
-
-void
-softpipe_set_edgeflags(struct pipe_context *pipe, const unsigned *edgeflags)
-{
-   struct softpipe_context *sp = softpipe_context(pipe);
-   draw_set_edgeflags(sp->draw, edgeflags);
-}
index e38b767cf2c03fed2b9177f1dbb5f69e803f9042..75dac810a12afe00d5b33abaed4974abe6cf5680 100644 (file)
@@ -55,6 +55,9 @@ softpipe_flush( struct pipe_context *pipe,
       for (i = 0; i < softpipe->num_textures; i++) {
          sp_flush_tex_tile_cache(softpipe->tex_cache[i]);
       }
+      for (i = 0; i < softpipe->num_vertex_textures; i++) {
+         sp_flush_tex_tile_cache(softpipe->vertex_tex_cache[i]);
+      }
    }
 
    if (flags & PIPE_FLUSH_SWAPBUFFERS) {
index 4076114d392a894165e5e6c3419eb63fc4d6a0e5..27fa126b7c391e5e8e77e34b74af414b804c9206 100644 (file)
@@ -126,7 +126,10 @@ exec_run( const struct sp_fragment_shader *base,
    setup_pos_vector(quad->posCoef, 
                     (float)quad->input.x0, (float)quad->input.y0, 
                     &machine->QuadPos);
-   
+
+   /* convert 0 to 1.0 and 1 to -1.0 */
+   machine->Face = (float) (quad->input.facing * -2 + 1);
+
    quad->inout.mask &= tgsi_exec_machine_run( machine );
    if (quad->inout.mask == 0)
       return FALSE;
index 81fb7aa20c67e2ec4c1fdeffc0800ab071ab63c2..bd3532de4f445f1b90ad4581c5a3c11619e5714a 100644 (file)
@@ -58,7 +58,9 @@ softpipe_get_param(struct pipe_screen *screen, int param)
    case PIPE_CAP_MAX_TEXTURE_IMAGE_UNITS:
       return PIPE_MAX_SAMPLERS;
    case PIPE_CAP_MAX_VERTEX_TEXTURE_UNITS:
-      return PIPE_MAX_SAMPLERS;
+      return PIPE_MAX_VERTEX_SAMPLERS;
+   case PIPE_CAP_MAX_COMBINED_SAMPLERS:
+      return PIPE_MAX_SAMPLERS + PIPE_MAX_VERTEX_SAMPLERS;
    case PIPE_CAP_NPOT_TEXTURES:
       return 1;
    case PIPE_CAP_TWO_SIDED_STENCIL:
@@ -143,6 +145,11 @@ softpipe_is_format_supported( struct pipe_screen *screen,
    case PIPE_FORMAT_DXT3_RGBA:
    case PIPE_FORMAT_DXT5_RGBA:
    case PIPE_FORMAT_Z32_FLOAT:
+   case PIPE_FORMAT_R8G8_SNORM:
+   case PIPE_FORMAT_B6UG5SR5S_NORM:
+   case PIPE_FORMAT_X8UB8UG8SR8S_NORM:
+   case PIPE_FORMAT_A8B8G8R8_SNORM:
+   case PIPE_FORMAT_NONE:
       return FALSE;
    default:
       return TRUE;
index 615581b95f92b554a2342ca1d1fadfae72381af2..3da75364c5d7090c274244e47768408ce5d96dd7 100644 (file)
@@ -1268,7 +1268,7 @@ void sp_setup_prepare( struct setup_context *setup )
    }
 
    /* Note: nr_attrs is only used for debugging (vertex printing) */
-   setup->nr_vertex_attrs = draw_num_vs_outputs(sp->draw);
+   setup->nr_vertex_attrs = draw_num_shader_outputs(sp->draw);
 
    sp->quad.first->begin( sp->quad.first );
 
index 77ee3c1136bba64015eec2e157d1a70e2ef2aae8..5a32d211d6da9851859488175adf7f8b282bec1d 100644 (file)
@@ -50,6 +50,7 @@
 #define SP_NEW_VERTEX        0x1000
 #define SP_NEW_VS            0x2000
 #define SP_NEW_QUERY         0x4000
+#define SP_NEW_GS            0x8000
 
 
 struct tgsi_sampler;
@@ -90,6 +91,11 @@ struct sp_vertex_shader {
    int max_sampler;             /* -1 if no samplers */
 };
 
+/** Subclass of pipe_shader_state */
+struct sp_geometry_shader {
+   struct pipe_shader_state shader;
+   struct draw_geometry_shader *draw_data;
+};
 
 
 void *
@@ -104,6 +110,10 @@ void *
 softpipe_create_sampler_state(struct pipe_context *,
                               const struct pipe_sampler_state *);
 void softpipe_bind_sampler_states(struct pipe_context *, unsigned, void **);
+void
+softpipe_bind_vertex_sampler_states(struct pipe_context *,
+                                    unsigned num_samplers,
+                                    void **samplers);
 void softpipe_delete_sampler_state(struct pipe_context *, void *);
 
 void *
@@ -139,6 +149,10 @@ void *softpipe_create_vs_state(struct pipe_context *,
                                const struct pipe_shader_state *);
 void softpipe_bind_vs_state(struct pipe_context *, void *);
 void softpipe_delete_vs_state(struct pipe_context *, void *);
+void *softpipe_create_gs_state(struct pipe_context *,
+                               const struct pipe_shader_state *);
+void softpipe_bind_gs_state(struct pipe_context *, void *);
+void softpipe_delete_gs_state(struct pipe_context *, void *);
 
 void softpipe_set_polygon_stipple( struct pipe_context *,
                                  const struct pipe_poly_stipple * );
@@ -150,6 +164,11 @@ void softpipe_set_sampler_textures( struct pipe_context *,
                                     unsigned num,
                                     struct pipe_texture ** );
 
+void
+softpipe_set_vertex_sampler_textures(struct pipe_context *,
+                                     unsigned num_textures,
+                                     struct pipe_texture **);
+
 void softpipe_set_viewport_state( struct pipe_context *,
                                   const struct pipe_viewport_state * );
 
@@ -180,10 +199,6 @@ softpipe_draw_range_elements(struct pipe_context *pipe,
                              unsigned max_index,
                              unsigned mode, unsigned start, unsigned count);
 
-void
-softpipe_set_edgeflags(struct pipe_context *pipe, const unsigned *edgeflags);
-
-
 void
 softpipe_map_transfers(struct softpipe_context *sp);
 
index efed082f823f05d8c383a933d642b0a517af7b84..95ab32343375aaf76c592a12e478b68fa3604d1f 100644 (file)
@@ -29,6 +29,7 @@
  */
 
 #include "util/u_memory.h"
+#include "draw/draw_context.h"
 #include "sp_context.h"
 #include "sp_state.h"
 
@@ -45,6 +46,8 @@ void softpipe_bind_blend_state( struct pipe_context *pipe,
 {
    struct softpipe_context *softpipe = softpipe_context(pipe);
 
+   draw_flush(softpipe->draw);
+
    softpipe->blend = (struct pipe_blend_state *)blend;
 
    softpipe->dirty |= SP_NEW_BLEND;
@@ -62,6 +65,8 @@ void softpipe_set_blend_color( struct pipe_context *pipe,
 {
    struct softpipe_context *softpipe = softpipe_context(pipe);
 
+   draw_flush(softpipe->draw);
+
    softpipe->blend_color = *blend_color;
 
    softpipe->dirty |= SP_NEW_BLEND;
index 3bc96b95385302216a296f9cb074bd9e6afe9396..f6856a5f69102dd22bf1d5649840f5057a1bbdaf 100644 (file)
@@ -67,7 +67,7 @@ softpipe_get_vertex_info(struct softpipe_context *softpipe)
       /* compute vertex layout now */
       const struct sp_fragment_shader *spfs = softpipe->fs;
       struct vertex_info *vinfo_vbuf = &softpipe->vertex_info_vbuf;
-      const uint num = draw_num_vs_outputs(softpipe->draw);
+      const uint num = draw_current_shader_outputs(softpipe->draw);
       uint i;
 
       /* Tell draw_vbuf to simply emit the whole post-xform vertex
@@ -117,13 +117,13 @@ softpipe_get_vertex_info(struct softpipe_context *softpipe)
          }
 
          /* this includes texcoords and varying vars */
-         src = draw_find_vs_output(softpipe->draw,
-                                   spfs->info.input_semantic_name[i],
-                                   spfs->info.input_semantic_index[i]);
+         src = draw_find_shader_output(softpipe->draw,
+                                       spfs->info.input_semantic_name[i],
+                                       spfs->info.input_semantic_index[i]);
          draw_emit_vertex_attr(vinfo, EMIT_4F, interp, src);
       }
 
-      softpipe->psize_slot = draw_find_vs_output(softpipe->draw,
+      softpipe->psize_slot = draw_find_shader_output(softpipe->draw,
                                                  TGSI_SEMANTIC_PSIZE, 0);
       if (softpipe->psize_slot > 0) {
          draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_CONSTANT,
@@ -213,6 +213,19 @@ update_tgsi_samplers( struct softpipe_context *softpipe )
          }
       }
    }
+
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      struct softpipe_tex_tile_cache *tc = softpipe->vertex_tex_cache[i];
+
+      if (tc->texture) {
+         struct softpipe_texture *spt = softpipe_texture(tc->texture);
+
+         if (spt->timestamp != tc->timestamp) {
+           sp_tex_tile_cache_validate_texture(tc);
+            tc->timestamp = spt->timestamp;
+         }
+      }
+   }
 }
 
 
index b41f7e8ab726c5feacffb388479bcd13c97ac684..aa12bb215a8e4db13f842fe73099a83824ebb134 100644 (file)
@@ -69,7 +69,14 @@ softpipe_bind_fs_state(struct pipe_context *pipe, void *fs)
 {
    struct softpipe_context *softpipe = softpipe_context(pipe);
 
-   softpipe->fs = (struct sp_fragment_shader *) fs;
+   draw_flush(softpipe->draw);
+
+   if (softpipe->fs == fs)
+      return;
+
+   draw_flush(softpipe->draw);
+
+   softpipe->fs = fs;
 
    softpipe->dirty |= SP_NEW_FS;
 }
@@ -159,9 +166,75 @@ softpipe_set_constant_buffer(struct pipe_context *pipe,
    assert(shader < PIPE_SHADER_TYPES);
    assert(index == 0);
 
+   draw_flush(softpipe->draw);
+
    /* note: reference counting */
    pipe_buffer_reference(&softpipe->constants[shader].buffer,
                         buf ? buf->buffer : NULL);
 
    softpipe->dirty |= SP_NEW_CONSTANTS;
 }
+
+void *
+softpipe_create_gs_state(struct pipe_context *pipe,
+                         const struct pipe_shader_state *templ)
+{
+   struct softpipe_context *softpipe = softpipe_context(pipe);
+   struct sp_geometry_shader *state;
+
+   state = CALLOC_STRUCT(sp_geometry_shader);
+   if (state == NULL )
+      goto fail;
+
+   /* debug */
+   if (softpipe->dump_gs)
+      tgsi_dump(templ->tokens, 0);
+
+   /* copy shader tokens, the ones passed in will go away.
+    */
+   state->shader.tokens = tgsi_dup_tokens(templ->tokens);
+   if (state->shader.tokens == NULL)
+      goto fail;
+
+   state->draw_data = draw_create_geometry_shader(softpipe->draw, templ);
+   if (state->draw_data == NULL)
+      goto fail;
+
+   return state;
+
+fail:
+   if (state) {
+      FREE( (void *)state->shader.tokens );
+      FREE( state->draw_data );
+      FREE( state );
+   }
+   return NULL;
+}
+
+
+void
+softpipe_bind_gs_state(struct pipe_context *pipe, void *gs)
+{
+   struct softpipe_context *softpipe = softpipe_context(pipe);
+
+   softpipe->gs = (struct sp_geometry_shader *)gs;
+
+   draw_bind_geometry_shader(softpipe->draw,
+                             (softpipe->gs ? softpipe->gs->draw_data : NULL));
+
+   softpipe->dirty |= SP_NEW_GS;
+}
+
+
+void
+softpipe_delete_gs_state(struct pipe_context *pipe, void *gs)
+{
+   struct softpipe_context *softpipe = softpipe_context(pipe);
+
+   struct sp_geometry_shader *state =
+      (struct sp_geometry_shader *)gs;
+
+   draw_delete_geometry_shader(softpipe->draw,
+                               (state) ? state->draw_data : 0);
+   FREE(state);
+}
index 87b72196838f6865df06890f0ca3dd6639160b2d..a5b00336d44d9d6660244beab73110ea9d1ddcdf 100644 (file)
@@ -41,14 +41,17 @@ softpipe_create_rasterizer_state(struct pipe_context *pipe,
 }
 
 void softpipe_bind_rasterizer_state(struct pipe_context *pipe,
-                                    void *setup)
+                                    void *rasterizer)
 {
    struct softpipe_context *softpipe = softpipe_context(pipe);
 
+   if (softpipe->rasterizer == rasterizer)
+      return;
+
    /* pass-through to draw module */
-   draw_set_rasterizer_state(softpipe->draw, setup);
+   draw_set_rasterizer_state(softpipe->draw, rasterizer);
 
-   softpipe->rasterizer = (struct pipe_rasterizer_state *)setup;
+   softpipe->rasterizer = rasterizer;
 
    softpipe->dirty |= SP_NEW_RASTERIZER;
 }
index db0b8ab76b1722a039b99cb779cf956f927c4e78..ceb4e338f1a5b4c893387975d5c8cf21ed14ecd0 100644 (file)
@@ -93,6 +93,34 @@ softpipe_bind_sampler_states(struct pipe_context *pipe,
 }
 
 
+void
+softpipe_bind_vertex_sampler_states(struct pipe_context *pipe,
+                                    unsigned num_samplers,
+                                    void **samplers)
+{
+   struct softpipe_context *softpipe = softpipe_context(pipe);
+   unsigned i;
+
+   assert(num_samplers <= PIPE_MAX_VERTEX_SAMPLERS);
+
+   /* Check for no-op */
+   if (num_samplers == softpipe->num_vertex_samplers &&
+       !memcmp(softpipe->vertex_samplers, samplers, num_samplers * sizeof(void *)))
+      return;
+
+   draw_flush(softpipe->draw);
+
+   for (i = 0; i < num_samplers; ++i)
+      softpipe->vertex_samplers[i] = samplers[i];
+   for (i = num_samplers; i < PIPE_MAX_VERTEX_SAMPLERS; ++i)
+      softpipe->vertex_samplers[i] = NULL;
+
+   softpipe->num_vertex_samplers = num_samplers;
+
+   softpipe->dirty |= SP_NEW_SAMPLER;
+}
+
+
 void
 softpipe_set_sampler_textures(struct pipe_context *pipe,
                               unsigned num, struct pipe_texture **texture)
@@ -122,6 +150,37 @@ softpipe_set_sampler_textures(struct pipe_context *pipe,
 }
 
 
+void
+softpipe_set_vertex_sampler_textures(struct pipe_context *pipe,
+                                     unsigned num_textures,
+                                     struct pipe_texture **textures)
+{
+   struct softpipe_context *softpipe = softpipe_context(pipe);
+   uint i;
+
+   assert(num_textures <= PIPE_MAX_VERTEX_SAMPLERS);
+
+   /* Check for no-op */
+   if (num_textures == softpipe->num_vertex_textures &&
+       !memcmp(softpipe->vertex_textures, textures, num_textures * sizeof(struct pipe_texture *))) {
+      return;
+   }
+
+   draw_flush(softpipe->draw);
+
+   for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++) {
+      struct pipe_texture *tex = i < num_textures ? textures[i] : NULL;
+
+      pipe_texture_reference(&softpipe->vertex_textures[i], tex);
+      sp_tex_tile_cache_set_texture(softpipe->vertex_tex_cache[i], tex);
+   }
+
+   softpipe->num_vertex_textures = num_textures;
+
+   softpipe->dirty |= SP_NEW_TEXTURE;
+}
+
+
 /**
  * Find/create an sp_sampler_varient object for sampling the given texture,
  * sampler and tex unit.
@@ -185,16 +244,16 @@ softpipe_reset_sampler_varients(struct softpipe_context *softpipe)
     * fragment programs.
     */
    for (i = 0; i <= softpipe->vs->max_sampler; i++) {
-      if (softpipe->sampler[i]) {
+      if (softpipe->vertex_samplers[i]) {
          softpipe->tgsi.vert_samplers_list[i] = 
             get_sampler_varient( i,
-                                 sp_sampler(softpipe->sampler[i]),
-                                 softpipe->texture[i],
+                                sp_sampler(softpipe->vertex_samplers[i]),
+                                softpipe->vertex_textures[i],
                                  TGSI_PROCESSOR_VERTEX );
 
          sp_sampler_varient_bind_texture( softpipe->tgsi.vert_samplers_list[i], 
-                                          softpipe->tex_cache[i],
-                                          softpipe->texture[i] );
+                                         softpipe->vertex_tex_cache[i],
+                                         softpipe->vertex_textures[i] );
       }
    }
 
index bc0e2011300910c1af2a434562d0361f29c595de..f6154109ea831dae4d555249261bd46de91f96f9 100644 (file)
@@ -35,6 +35,8 @@
 
 #include "draw/draw_context.h"
 
+#include "util/u_format.h"
+
 
 /**
  * XXX this might get moved someday
@@ -49,6 +51,8 @@ softpipe_set_framebuffer_state(struct pipe_context *pipe,
    struct softpipe_context *sp = softpipe_context(pipe);
    uint i;
 
+   draw_flush(sp->draw);
+
    for (i = 0; i < PIPE_MAX_COLOR_BUFS; i++) {
       /* check if changing cbuf */
       if (sp->framebuffer.cbufs[i] != fb->cbufs[i]) {
@@ -80,8 +84,9 @@ softpipe_set_framebuffer_state(struct pipe_context *pipe,
       if (sp->framebuffer.zsbuf) {
          int depth_bits;
          double mrd;
-         depth_bits = pf_get_component_bits(sp->framebuffer.zsbuf->format,
-                                            PIPE_FORMAT_COMP_Z);
+         depth_bits = util_format_get_component_bits(sp->framebuffer.zsbuf->format,
+                                                     UTIL_FORMAT_COLORSPACE_ZS,
+                                                     0);
          if (depth_bits > 16) {
             mrd = 0.0000001;
          }
index c22ee86b66c4f670375ffcb796b24c6e3498b5b3..e26153b1d9096012ebf1a6a357efb867f82691eb 100644 (file)
@@ -521,7 +521,7 @@ compute_lambda_1d(const struct sp_sampler_varient *samp,
    const struct pipe_sampler_state *sampler = samp->sampler;
    float dsdx = fabsf(s[QUAD_BOTTOM_RIGHT] - s[QUAD_BOTTOM_LEFT]);
    float dsdy = fabsf(s[QUAD_TOP_LEFT]     - s[QUAD_BOTTOM_LEFT]);
-   float rho = MAX2(dsdx, dsdy) * texture->width[0];
+   float rho = MAX2(dsdx, dsdy) * texture->width0;
    float lambda;
 
    lambda = util_fast_log2(rho);
@@ -545,8 +545,8 @@ compute_lambda_2d(const struct sp_sampler_varient *samp,
    float dsdy = fabsf(s[QUAD_TOP_LEFT]     - s[QUAD_BOTTOM_LEFT]);
    float dtdx = fabsf(t[QUAD_BOTTOM_RIGHT] - t[QUAD_BOTTOM_LEFT]);
    float dtdy = fabsf(t[QUAD_TOP_LEFT]     - t[QUAD_BOTTOM_LEFT]);
-   float maxx = MAX2(dsdx, dsdy) * texture->width[0];
-   float maxy = MAX2(dtdx, dtdy) * texture->height[0];
+   float maxx = MAX2(dsdx, dsdy) * texture->width0;
+   float maxy = MAX2(dtdx, dtdy) * texture->height0;
    float rho  = MAX2(maxx, maxy);
    float lambda;
 
@@ -573,9 +573,9 @@ compute_lambda_3d(const struct sp_sampler_varient *samp,
    float dtdy = fabsf(t[QUAD_TOP_LEFT]     - t[QUAD_BOTTOM_LEFT]);
    float dpdx = fabsf(p[QUAD_BOTTOM_RIGHT] - p[QUAD_BOTTOM_LEFT]);
    float dpdy = fabsf(p[QUAD_TOP_LEFT]     - p[QUAD_BOTTOM_LEFT]);
-   float maxx = MAX2(dsdx, dsdy) * texture->width[0];
-   float maxy = MAX2(dtdx, dtdy) * texture->height[0];
-   float maxz = MAX2(dpdx, dpdy) * texture->depth[0];
+   float maxx = MAX2(dsdx, dsdy) * texture->width0;
+   float maxy = MAX2(dtdx, dtdy) * texture->height0;
+   float maxz = MAX2(dpdx, dpdy) * texture->depth0;
    float rho, lambda;
 
    rho = MAX2(maxx, maxy);
@@ -644,8 +644,8 @@ get_texel_2d(const struct sp_sampler_varient *samp,
    const struct pipe_texture *texture = samp->texture;
    unsigned level = addr.bits.level;
 
-   if (x < 0 || x >= (int) texture->width[level] ||
-       y < 0 || y >= (int) texture->height[level]) {
+   if (x < 0 || x >= (int) u_minify(texture->width0, level) ||
+       y < 0 || y >= (int) u_minify(texture->height0, level)) {
       return samp->sampler->border_color;
    }
    else {
@@ -737,9 +737,9 @@ get_texel_3d(const struct sp_sampler_varient *samp,
    const struct pipe_texture *texture = samp->texture;
    unsigned level = addr.bits.level;
 
-   if (x < 0 || x >= (int) texture->width[level] ||
-       y < 0 || y >= (int) texture->height[level] ||
-       z < 0 || z >= (int) texture->depth[level]) {
+   if (x < 0 || x >= (int) u_minify(texture->width0, level) ||
+       y < 0 || y >= (int) u_minify(texture->height0, level) ||
+       z < 0 || z >= (int) u_minify(texture->depth0, level)) {
       return samp->sampler->border_color;
    }
    else {
@@ -925,7 +925,7 @@ img_filter_1d_nearest(struct tgsi_sampler *tgsi_sampler,
    union tex_tile_address addr;
 
    level0 = samp->level;
-   width = texture->width[level0];
+   width = u_minify(texture->width0, level0);
 
    assert(width > 0);
 
@@ -961,8 +961,8 @@ img_filter_2d_nearest(struct tgsi_sampler *tgsi_sampler,
 
 
    level0 = samp->level;
-   width = texture->width[level0];
-   height = texture->height[level0];
+   width = u_minify(texture->width0, level0);
+   height = u_minify(texture->height0, level0);
 
    assert(width > 0);
    assert(height > 0);
@@ -1008,8 +1008,8 @@ img_filter_cube_nearest(struct tgsi_sampler *tgsi_sampler,
    union tex_tile_address addr;
 
    level0 = samp->level;
-   width = texture->width[level0];
-   height = texture->height[level0];
+   width = u_minify(texture->width0, level0);
+   height = u_minify(texture->height0, level0);
 
    assert(width > 0);
    assert(height > 0);
@@ -1046,9 +1046,9 @@ img_filter_3d_nearest(struct tgsi_sampler *tgsi_sampler,
    union tex_tile_address addr;
 
    level0 = samp->level;
-   width = texture->width[level0];
-   height = texture->height[level0];
-   depth = texture->depth[level0];
+   width = u_minify(texture->width0, level0);
+   height = u_minify(texture->height0, level0);
+   depth = u_minify(texture->depth0, level0);
 
    assert(width > 0);
    assert(height > 0);
@@ -1088,7 +1088,7 @@ img_filter_1d_linear(struct tgsi_sampler *tgsi_sampler,
    union tex_tile_address addr;
 
    level0 = samp->level;
-   width = texture->width[level0];
+   width = u_minify(texture->width0, level0);
 
    assert(width > 0);
 
@@ -1127,8 +1127,8 @@ img_filter_2d_linear(struct tgsi_sampler *tgsi_sampler,
    union tex_tile_address addr;
 
    level0 = samp->level;
-   width = texture->width[level0];
-   height = texture->height[level0];
+   width = u_minify(texture->width0, level0);
+   height = u_minify(texture->height0, level0);
 
    assert(width > 0);
    assert(height > 0);
@@ -1174,8 +1174,8 @@ img_filter_cube_linear(struct tgsi_sampler *tgsi_sampler,
    union tex_tile_address addr;
 
    level0 = samp->level;
-   width = texture->width[level0];
-   height = texture->height[level0];
+   width = u_minify(texture->width0, level0);
+   height = u_minify(texture->height0, level0);
 
    assert(width > 0);
    assert(height > 0);
@@ -1221,9 +1221,9 @@ img_filter_3d_linear(struct tgsi_sampler *tgsi_sampler,
    union tex_tile_address addr;
 
    level0 = samp->level;
-   width = texture->width[level0];
-   height = texture->height[level0];
-   depth = texture->depth[level0];
+   width = u_minify(texture->width0, level0);
+   height = u_minify(texture->height0, level0);
+   depth = u_minify(texture->depth0, level0);
 
    addr.value = 0;
    addr.bits.level = level0;
@@ -1778,8 +1778,8 @@ sp_sampler_varient_bind_texture( struct sp_sampler_varient *samp,
 
    samp->texture = texture;
    samp->cache = tex_cache;
-   samp->xpot = util_unsigned_logbase2( texture->width[0] );
-   samp->ypot = util_unsigned_logbase2( texture->height[0] );
+   samp->xpot = util_unsigned_logbase2( texture->width0 );
+   samp->ypot = util_unsigned_logbase2( texture->height0 );
    samp->level = CLAMP((int) sampler->min_lod, 0, (int) texture->last_level);
 }
 
index 407a22a9f4bd02c3c6bb857c137dae05d6ba67d7..e50a76a73bc9b9809d525a50d66dfb374a4c0ea6 100644 (file)
@@ -35,6 +35,7 @@
 #include "pipe/p_inlines.h"
 #include "util/u_memory.h"
 #include "util/u_tile.h"
+#include "util/u_math.h"
 #include "sp_context.h"
 #include "sp_surface.h"
 #include "sp_texture.h"
@@ -246,9 +247,9 @@ sp_find_cached_tile_tex(struct softpipe_tex_tile_cache *tc,
                                      addr.bits.level, 
                                      addr.bits.z, 
                                      PIPE_TRANSFER_READ, 0, 0,
-                                     tc->texture->width[addr.bits.level],
-                                     tc->texture->height[addr.bits.level]);
-
+                                     u_minify(tc->texture->width0, addr.bits.level),
+                                     u_minify(tc->texture->height0, addr.bits.level));
+         
          tc->tex_trans_map = screen->transfer_map(screen, tc->tex_trans);
 
          tc->tex_face = addr.bits.face;
index 7caf2928b4ba11ac8c2ca59d6d7749476c39ea15..a9436a33942b299f0ada7ee10b190ed2e8b190ff 100644 (file)
@@ -32,6 +32,8 @@
 
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
+
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -52,29 +54,28 @@ softpipe_texture_layout(struct pipe_screen *screen,
 {
    struct pipe_texture *pt = &spt->base;
    unsigned level;
-   unsigned width = pt->width[0];
-   unsigned height = pt->height[0];
-   unsigned depth = pt->depth[0];
+   unsigned width = pt->width0;
+   unsigned height = pt->height0;
+   unsigned depth = pt->depth0;
 
    unsigned buffer_size = 0;
 
+   pt->width0 = width;
+   pt->height0 = height;
+   pt->depth0 = depth;
+
    for (level = 0; level <= pt->last_level; level++) {
-      pt->width[level] = width;
-      pt->height[level] = height;
-      pt->depth[level] = depth;
-      pt->nblocksx[level] = pf_get_nblocksx(&pt->block, width);  
-      pt->nblocksy[level] = pf_get_nblocksy(&pt->block, height);  
-      spt->stride[level] = pt->nblocksx[level]*pt->block.size;
+      spt->stride[level] = util_format_get_stride(pt->format, width);
 
       spt->level_offset[level] = buffer_size;
 
-      buffer_size += (pt->nblocksy[level] *
+      buffer_size += (util_format_get_nblocksy(pt->format, height) *
                       ((pt->target == PIPE_TEXTURE_CUBE) ? 6 : depth) *
                       spt->stride[level]);
 
-      width  = minify(width);
-      height = minify(height);
-      depth = minify(depth);
+      width  = u_minify(width, 1);
+      height = u_minify(height, 1);
+      depth = u_minify(depth, 1);
    }
 
    spt->buffer = screen->buffer_create(screen, 32,
@@ -96,12 +97,9 @@ softpipe_displaytarget_layout(struct pipe_screen *screen,
                      PIPE_BUFFER_USAGE_GPU_READ_WRITE);
    unsigned tex_usage = spt->base.tex_usage;
 
-   spt->base.nblocksx[0] = pf_get_nblocksx(&spt->base.block, spt->base.width[0]);  
-   spt->base.nblocksy[0] = pf_get_nblocksy(&spt->base.block, spt->base.height[0]);  
-
    spt->buffer = screen->surface_buffer_create( screen, 
-                                                spt->base.width[0]
-                                                spt->base.height[0],
+                                                spt->base.width0
+                                                spt->base.height0,
                                                 spt->base.format,
                                                 usage,
                                                 tex_usage,
@@ -126,9 +124,9 @@ softpipe_texture_create(struct pipe_screen *screen,
    pipe_reference_init(&spt->base.reference, 1);
    spt->base.screen = screen;
 
-   spt->pot = (util_is_power_of_two(template->width[0]) &&
-               util_is_power_of_two(template->height[0]) &&
-               util_is_power_of_two(template->depth[0]));
+   spt->pot = (util_is_power_of_two(template->width0) &&
+               util_is_power_of_two(template->height0) &&
+               util_is_power_of_two(template->depth0));
 
    if (spt->base.tex_usage & (PIPE_TEXTURE_USAGE_DISPLAY_TARGET |
                               PIPE_TEXTURE_USAGE_PRIMARY)) {
@@ -163,7 +161,7 @@ softpipe_texture_blanket(struct pipe_screen * screen,
    /* Only supports one type */
    if (base->target != PIPE_TEXTURE_2D ||
        base->last_level != 0 ||
-       base->depth[0] != 1) {
+       base->depth0 != 1) {
       return NULL;
    }
 
@@ -174,8 +172,6 @@ softpipe_texture_blanket(struct pipe_screen * screen,
    spt->base = *base;
    pipe_reference_init(&spt->base.reference, 1);
    spt->base.screen = screen;
-   spt->base.nblocksx[0] = pf_get_nblocksx(&spt->base.block, spt->base.width[0]);  
-   spt->base.nblocksy[0] = pf_get_nblocksy(&spt->base.block, spt->base.height[0]);  
    spt->stride[0] = stride[0];
 
    pipe_buffer_reference(&spt->buffer, buffer);
@@ -213,8 +209,8 @@ softpipe_get_tex_surface(struct pipe_screen *screen,
       pipe_reference_init(&ps->reference, 1);
       pipe_texture_reference(&ps->texture, pt);
       ps->format = pt->format;
-      ps->width = pt->width[level];
-      ps->height = pt->height[level];
+      ps->width = u_minify(pt->width0, level);
+      ps->height = u_minify(pt->height0, level);
       ps->offset = spt->level_offset[level];
       ps->usage = usage;
 
@@ -243,10 +239,12 @@ softpipe_get_tex_surface(struct pipe_screen *screen,
       ps->zslice = zslice;
 
       if (pt->target == PIPE_TEXTURE_CUBE) {
-         ps->offset += face * pt->nblocksy[level] * spt->stride[level];
+         ps->offset += face * util_format_get_nblocksy(pt->format, u_minify(pt->height0, level)) *
+                       spt->stride[level];
       }
       else if (pt->target == PIPE_TEXTURE_3D) {
-         ps->offset += zslice * pt->nblocksy[level] * spt->stride[level];
+         ps->offset += zslice * util_format_get_nblocksy(pt->format, u_minify(pt->height0, level)) *
+                       spt->stride[level];
       }
       else {
          assert(face == 0);
@@ -301,15 +299,12 @@ softpipe_get_tex_transfer(struct pipe_screen *screen,
    spt = CALLOC_STRUCT(softpipe_transfer);
    if (spt) {
       struct pipe_transfer *pt = &spt->base;
+      int nblocksy = util_format_get_nblocksy(texture->format, u_minify(texture->height0, level));
       pipe_texture_reference(&pt->texture, texture);
-      pt->format = texture->format;
-      pt->block = texture->block;
       pt->x = x;
       pt->y = y;
       pt->width = w;
       pt->height = h;
-      pt->nblocksx = texture->nblocksx[level];
-      pt->nblocksy = texture->nblocksy[level];
       pt->stride = sptex->stride[level];
       pt->usage = usage;
       pt->face = face;
@@ -319,10 +314,10 @@ softpipe_get_tex_transfer(struct pipe_screen *screen,
       spt->offset = sptex->level_offset[level];
 
       if (texture->target == PIPE_TEXTURE_CUBE) {
-         spt->offset += face * pt->nblocksy * pt->stride;
+         spt->offset += face * nblocksy * pt->stride;
       }
       else if (texture->target == PIPE_TEXTURE_3D) {
-         spt->offset += zslice * pt->nblocksy * pt->stride;
+         spt->offset += zslice * nblocksy * pt->stride;
       }
       else {
          assert(face == 0);
@@ -360,9 +355,11 @@ softpipe_transfer_map( struct pipe_screen *screen,
 {
    ubyte *map, *xfer_map;
    struct softpipe_texture *spt;
+   enum pipe_format format;
 
    assert(transfer->texture);
    spt = softpipe_texture(transfer->texture);
+   format = transfer->texture->format;
 
    map = pipe_buffer_map(screen, spt->buffer, pipe_transfer_buffer_flags(transfer));
    if (map == NULL)
@@ -379,8 +376,8 @@ softpipe_transfer_map( struct pipe_screen *screen,
    }
 
    xfer_map = map + softpipe_transfer(transfer)->offset +
-      transfer->y / transfer->block.height * transfer->stride +
-      transfer->x / transfer->block.width * transfer->block.size;
+      transfer->y / util_format_get_blockheight(format) * transfer->stride +
+      transfer->x / util_format_get_blockwidth(format) * util_format_get_blocksize(format);
    /*printf("map = %p  xfer map = %p\n", map, xfer_map);*/
    return xfer_map;
 }
@@ -434,10 +431,9 @@ softpipe_video_surface_create(struct pipe_screen *screen,
    template.format = PIPE_FORMAT_X8R8G8B8_UNORM;
    template.last_level = 0;
    /* vl_mpeg12_mc_renderer expects this when it's initialized with pot_buffers=true */
-   template.width[0] = util_next_power_of_two(width);
-   template.height[0] = util_next_power_of_two(height);
-   template.depth[0] = 1;
-   pf_get_block(template.format, &template.block);
+   template.width0 = util_next_power_of_two(width);
+   template.height0 = util_next_power_of_two(height);
+   template.depth0 = 1;
    template.tex_usage = PIPE_TEXTURE_USAGE_SAMPLER | PIPE_TEXTURE_USAGE_RENDER_TARGET;
 
    sp_vsfc->tex = screen->texture_create(screen, &template);
index 65872cecc4fdfd0fc4ada4af11de0e1b1453808b..112a6fe0cf36cf09425a953fbf6538b3a04e8a05 100644 (file)
@@ -33,6 +33,7 @@
  */
 
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_memory.h"
 #include "util/u_tile.h"
 #include "sp_tile_cache.h"
@@ -238,7 +239,7 @@ clear_tile(struct softpipe_cached_tile *tile,
 {
    uint i, j;
 
-   switch (pf_get_size(format)) {
+   switch (util_format_get_blocksize(format)) {
    case 1:
       memset(tile->data.any, clear_value, TILE_SIZE * TILE_SIZE);
       break;
@@ -284,8 +285,9 @@ sp_tile_cache_flush_clear(struct softpipe_tile_cache *tc)
    uint x, y;
    uint numCleared = 0;
 
+   assert(pt->texture);
    /* clear the scratch tile to the clear value */
-   clear_tile(&tc->tile, pt->format, tc->clear_val);
+   clear_tile(&tc->tile, pt->texture->format, tc->clear_val);
 
    /* push the tile to all positions marked as clear */
    for (y = 0; y < h; y += TILE_SIZE) {
@@ -372,6 +374,7 @@ sp_find_cached_tile(struct softpipe_tile_cache *tc,
 
    if (addr.value != tile->addr.value) {
 
+      assert(pt->texture);
       if (tile->addr.bits.invalid == 0) {
          /* put dirty tile back in framebuffer */
          if (tc->depth_stencil) {
@@ -395,10 +398,10 @@ sp_find_cached_tile(struct softpipe_tile_cache *tc,
       if (is_clear_flag_set(tc->clear_flags, addr)) {
          /* don't get tile from framebuffer, just clear it */
          if (tc->depth_stencil) {
-            clear_tile(tile, pt->format, tc->clear_val);
+            clear_tile(tile, pt->texture->format, tc->clear_val);
          }
          else {
-            clear_tile_rgba(tile, pt->format, tc->clear_color);
+            clear_tile_rgba(tile, pt->texture->format, tc->clear_color);
          }
          clear_clear_flag(tc->clear_flags, addr);
       }
index 9e571862b7516f7a0fbd33119c8fb78532be9ff0..f203ded29eee72a6bdf674117faa2804d6ad1e9a 100644 (file)
 #ifndef SP_WINSYS_H
 #define SP_WINSYS_H
 
-
 #ifdef __cplusplus
 extern "C" {
 #endif
 
+#include "pipe/p_defines.h"
 
 struct pipe_screen;
 struct pipe_winsys;
 struct pipe_context;
+struct pipe_texture;
+struct pipe_buffer;
 
 
 struct pipe_context *softpipe_create( struct pipe_screen * );
index e650a251d1975a134754702c5302f58fd56c1603..0885d9ca74101d3f1872b49772c0ecd22608276f 100644 (file)
@@ -202,8 +202,6 @@ struct svga_state
    struct pipe_clip_state clip;
    struct pipe_viewport_state viewport;
 
-   const unsigned *edgeflags;
-
    unsigned num_samplers;
    unsigned num_textures;
    unsigned num_vertex_elements;
@@ -369,7 +367,7 @@ struct svga_context
 #define SVGA_NEW_FRAME_BUFFER        0x800
 #define SVGA_NEW_STIPPLE             0x1000
 #define SVGA_NEW_SCISSOR             0x2000
-#define SVGA_NEW_BLEND_COLOR         0x5000
+#define SVGA_NEW_BLEND_COLOR         0x4000
 #define SVGA_NEW_CLIP                0x8000
 #define SVGA_NEW_VIEWPORT            0x10000
 #define SVGA_NEW_PRESCALE            0x20000
@@ -380,9 +378,8 @@ struct svga_context
 #define SVGA_NEW_NEED_SWTNL          0x400000
 #define SVGA_NEW_FS_RESULT           0x800000
 #define SVGA_NEW_VS_RESULT           0x1000000
-#define SVGA_NEW_EDGEFLAGS           0x2000000
-#define SVGA_NEW_ZERO_STRIDE         0x4000000
-#define SVGA_NEW_TEXTURE_FLAGS       0x8000000
+#define SVGA_NEW_ZERO_STRIDE         0x2000000
+#define SVGA_NEW_TEXTURE_FLAGS       0x4000000
 
 
 
index 6195c3897ed2bb764fddc3d282064afd3b05e372..409b3b41cbc9d142ef976d468ea196599cee58d3 100644 (file)
@@ -46,7 +46,7 @@ try_clear(struct svga_context *svga,
    boolean restore_viewport = FALSE;
    SVGA3dClearFlag flags = 0;
    struct pipe_framebuffer_state *fb = &svga->curr.framebuffer;
-   unsigned color = 0;
+   union util_color uc;
 
    ret = svga_update_state(svga, SVGA_STATE_HW_CLEAR);
    if (ret)
@@ -54,7 +54,7 @@ try_clear(struct svga_context *svga,
 
    if ((buffers & PIPE_CLEAR_COLOR) && fb->cbufs[0]) {
       flags |= SVGA3D_CLEAR_COLOR;
-      util_pack_color(rgba, PIPE_FORMAT_A8R8G8B8_UNORM, &color);
+      util_pack_color(rgba, PIPE_FORMAT_A8R8G8B8_UNORM, &uc);
 
       rect.w = fb->cbufs[0]->width;
       rect.h = fb->cbufs[0]->height;
@@ -77,7 +77,7 @@ try_clear(struct svga_context *svga,
          return ret;
    }
 
-   ret = SVGA3D_ClearRect(svga->swc, flags, color, depth, stencil,
+   ret = SVGA3D_ClearRect(svga->swc, flags, uc.ui, depth, stencil,
                           rect.x, rect.y, rect.w, rect.h);
    if (ret != PIPE_OK)
       return ret;
index 3eeca6b784b1a517efd39b7a4a4c2dfc785fe6ef..78053e755e288d644cbcb1674b66ef055a2acffa 100644 (file)
@@ -101,6 +101,7 @@ svga_create_sampler_state(struct pipe_context *pipe,
 {
    struct svga_context *svga = svga_context(pipe);
    struct svga_sampler_state *cso = CALLOC_STRUCT( svga_sampler_state );
+   union util_color uc;
    
    cso->mipfilter = translate_mip_filter(sampler->min_mip_filter);
    cso->magfilter = translate_img_filter( sampler->mag_img_filter );
@@ -121,8 +122,8 @@ svga_create_sampler_state(struct pipe_context *pipe,
       ubyte a = float_to_ubyte(sampler->border_color[3]);
 
       util_pack_color_ub( r, g, b, a,
-                          PIPE_FORMAT_B8G8R8A8_UNORM,
-                          &cso->bordercolor );
+                          PIPE_FORMAT_B8G8R8A8_UNORM, &uc);
+      cso->bordercolor = uc.ui;
    }
 
    /* No SVGA3D support for:
@@ -234,9 +235,9 @@ static void svga_set_sampler_textures(struct pipe_context *pipe,
 void svga_init_sampler_functions( struct svga_context *svga )
 {
    svga->pipe.create_sampler_state = svga_create_sampler_state;
-   svga->pipe.bind_sampler_states = svga_bind_sampler_states;
+   svga->pipe.bind_fragment_sampler_states = svga_bind_sampler_states;
    svga->pipe.delete_sampler_state = svga_delete_sampler_state;
-   svga->pipe.set_sampler_textures = svga_set_sampler_textures;
+   svga->pipe.set_fragment_sampler_textures = svga_set_sampler_textures;
 }
 
 
index 28e2787e0d3d77c2f20c57f6ff0c223789524466..42f290d162af4def6327f2fbcb1badcd887cb59b 100644 (file)
@@ -84,18 +84,6 @@ static void svga_set_vertex_elements(struct pipe_context *pipe,
 }
 
 
-static void svga_set_edgeflags(struct pipe_context *pipe,
-                               const unsigned *bitfield)
-{
-   struct svga_context *svga = svga_context(pipe);
-
-   if (bitfield != NULL || svga->curr.edgeflags != NULL) {
-      svga->curr.edgeflags = bitfield;
-      svga->dirty |= SVGA_NEW_EDGEFLAGS;
-   }
-}
-
-
 void svga_cleanup_vertex_state( struct svga_context *svga )
 {
    unsigned i;
@@ -109,7 +97,6 @@ void svga_init_vertex_functions( struct svga_context *svga )
 {
    svga->pipe.set_vertex_buffers = svga_set_vertex_buffers;
    svga->pipe.set_vertex_elements = svga_set_vertex_elements;
-   svga->pipe.set_edgeflags = svga_set_edgeflags;
 }
 
 
index e5ffe668c355ae07d9425a959e2fdfe9b04705a8..c104c41f5f87343b97553dee471bdf37b51a49b9 100644 (file)
@@ -48,7 +48,7 @@ static const struct tgsi_token *substitute_vs(
    static struct tgsi_token tokens[300];
 
    const char *text = 
-      "VERT1.1\n"
+      "VERT\n"
       "DCL IN[0]\n"
       "DCL IN[1]\n"
       "DCL IN[2]\n"
index 1f8a88967238e9f3430a7fa15c7ca1f98a3ce6a7..58a1aba464b8fb4be3e29a0294414eb272082f9c 100644 (file)
@@ -356,7 +356,8 @@ svga_buffer_upload_flush(struct svga_context *svga,
    sbuf->hw.boxes = NULL;
 
    /* Decrement reference count */
-   pipe_buffer_reference((struct pipe_buffer **)&sbuf, NULL);
+   pipe_reference(&(sbuf->base.reference), NULL);
+   sbuf = NULL;
 }
 
 
index 1eb03db2806b182ce1c8f3e4b2a9819f931f3546..2224c2d3945d96d102dea8a032a9be3f311be2bb 100644 (file)
@@ -29,6 +29,7 @@
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
 #include "pipe/p_thread.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -158,7 +159,8 @@ svga_transfer_dma_band(struct svga_transfer *st,
                 st->base.x + st->base.width,
                 y + h,
                 st->base.zslice + 1,
-                texture->base.block.size*8/(texture->base.block.width*texture->base.block.height));
+                util_format_get_blocksize(texture->base.format)*8/
+                (util_format_get_blockwidth(texture->base.format)*util_format_get_blockheight(texture->base.format)));
    
    box.x = st->base.x;
    box.y = y;
@@ -208,7 +210,8 @@ svga_transfer_dma(struct svga_transfer *st,
    }
    else {
       unsigned y, h, srcy;
-      h = st->hw_nblocksy * st->base.block.height;
+      unsigned blockheight = util_format_get_blockheight(st->base.texture->format);
+      h = st->hw_nblocksy * blockheight;
       srcy = 0;
       for(y = 0; y < st->base.height; y += h) {
          unsigned offset, length;
@@ -218,11 +221,11 @@ svga_transfer_dma(struct svga_transfer *st,
             h = st->base.height - y;
 
          /* Transfer band must be aligned to pixel block boundaries */
-         assert(y % st->base.block.height == 0);
-         assert(h % st->base.block.height == 0);
+         assert(y % blockheight == 0);
+         assert(h % blockheight == 0);
          
-         offset = y * st->base.stride / st->base.block.height;
-         length = h * st->base.stride / st->base.block.height;
+         offset = y * st->base.stride / blockheight;
+         length = h * st->base.stride / blockheight;
 
          sw = (uint8_t *)st->swbuf + offset;
          
@@ -281,24 +284,19 @@ svga_texture_create(struct pipe_screen *screen,
    if(templat->last_level >= SVGA_MAX_TEXTURE_LEVELS)
       goto error2;
    
-   width = templat->width[0];
-   height = templat->height[0];
-   depth = templat->depth[0];
+   width = templat->width0;
+   height = templat->height0;
+   depth = templat->depth0;
    for(level = 0; level <= templat->last_level; ++level) {
-      tex->base.width[level] = width;
-      tex->base.height[level] = height;
-      tex->base.depth[level] = depth;
-      tex->base.nblocksx[level] = pf_get_nblocksx(&tex->base.block, width);  
-      tex->base.nblocksy[level] = pf_get_nblocksy(&tex->base.block, height);  
-      width  = minify(width);
-      height = minify(height);
-      depth = minify(depth);
+      width = u_minify(width, 1);
+      height = u_minify(height, 1);
+      depth = u_minify(depth, 1);
    }
    
    tex->key.flags = 0;
-   tex->key.size.width = templat->width[0];
-   tex->key.size.height = templat->height[0];
-   tex->key.size.depth = templat->depth[0];
+   tex->key.size.width = templat->width0;
+   tex->key.size.height = templat->height0;
+   tex->key.size.depth = templat->depth0;
    
    if(templat->target == PIPE_TEXTURE_CUBE) {
       tex->key.flags |= SVGA3D_SURFACE_CUBEMAP;
@@ -322,7 +320,7 @@ svga_texture_create(struct pipe_screen *screen,
     */
 #if 0
    if((templat->tex_usage & PIPE_TEXTURE_USAGE_RENDER_TARGET) &&
-      !pf_is_compressed(templat->format))
+      !util_format_is_compressed(templat->format))
       tex->key.flags |= SVGA3D_SURFACE_HINT_RENDERTARGET;
 #endif
    
@@ -365,7 +363,7 @@ svga_texture_blanket(struct pipe_screen * screen,
    /* Only supports one type */
    if (base->target != PIPE_TEXTURE_2D ||
        base->last_level != 0 ||
-       base->depth[0] != 1) {
+       base->depth0 != 1) {
       return NULL;
    }
 
@@ -538,9 +536,9 @@ svga_texture_view_surface(struct pipe_context *pipe,
    key->flags = 0;
    key->format = format;
    key->numMipLevels = num_mip;
-   key->size.width = tex->base.width[start_mip];
-   key->size.height = tex->base.height[start_mip];
-   key->size.depth = zslice_pick < 0 ? tex->base.depth[start_mip] : 1;
+   key->size.width = u_minify(tex->base.width0, start_mip);
+   key->size.height = u_minify(tex->base.height0, start_mip);
+   key->size.depth = zslice_pick < 0 ? u_minify(tex->base.depth0, start_mip) : 1;
    key->cachable = 1;
    assert(key->size.depth == 1);
    
@@ -574,7 +572,10 @@ svga_texture_view_surface(struct pipe_context *pipe,
    for (i = 0; i < key->numMipLevels; i++) {
       for (j = 0; j < key->numFaces; j++) {
          if(tex->defined[j + face_pick][i + start_mip]) {
-            unsigned depth = zslice_pick < 0 ? tex->base.depth[i + start_mip] : 1;
+            unsigned depth = (zslice_pick < 0 ?
+                              u_minify(tex->base.depth0, i + start_mip) :
+                              1);
+
             svga_texture_copy_handle(svga_context(pipe),
                                      ss,
                                      tex->handle, 
@@ -582,8 +583,8 @@ svga_texture_view_surface(struct pipe_context *pipe,
                                      i + start_mip, 
                                      j + face_pick,
                                      handle, 0, 0, 0, i, j,
-                                     tex->base.width[i + start_mip],
-                                     tex->base.height[i + start_mip],
+                                     u_minify(tex->base.width0, i + start_mip),
+                                     u_minify(tex->base.height0, i + start_mip),
                                      depth);
          }
       }
@@ -612,8 +613,8 @@ svga_get_tex_surface(struct pipe_screen *screen,
    pipe_reference_init(&s->base.reference, 1);
    pipe_texture_reference(&s->base.texture, pt);
    s->base.format = pt->format;
-   s->base.width = pt->width[level];
-   s->base.height = pt->height[level];
+   s->base.width = u_minify(pt->width0, level);
+   s->base.height = u_minify(pt->height0, level);
    s->base.usage = flags;
    s->base.level = level;
    s->base.face = face;
@@ -742,7 +743,8 @@ svga_propagate_surface(struct pipe_context *pipe, struct pipe_surface *surf)
       svga_texture_copy_handle(svga_context(pipe), ss,
                                s->handle, 0, 0, 0, s->real_level, s->real_face,
                                tex->handle, 0, 0, surf->zslice, surf->level, surf->face,
-                               tex->base.width[surf->level], tex->base.height[surf->level], 1);
+                               u_minify(tex->base.width0, surf->level),
+                               u_minify(tex->base.height0, surf->level), 1);
       tex->defined[surf->face][surf->level] = TRUE;
    }
 }
@@ -770,6 +772,8 @@ svga_get_tex_transfer(struct pipe_screen *screen,
    struct svga_screen *ss = svga_screen(screen);
    struct svga_winsys_screen *sws = ss->sws;
    struct svga_transfer *st;
+   unsigned nblocksx = util_format_get_nblocksx(texture->format, w);
+   unsigned nblocksy = util_format_get_nblocksy(texture->format, h);
 
    /* We can't map texture storage directly */
    if (usage & PIPE_TRANSFER_MAP_DIRECTLY)
@@ -779,21 +783,17 @@ svga_get_tex_transfer(struct pipe_screen *screen,
    if (!st)
       return NULL;
    
-   st->base.format = texture->format;
-   st->base.block = texture->block;
    st->base.x = x;
    st->base.y = y;
    st->base.width = w;
    st->base.height = h;
-   st->base.nblocksx = pf_get_nblocksx(&texture->block, w);
-   st->base.nblocksy = pf_get_nblocksy(&texture->block, h);
-   st->base.stride = st->base.nblocksx*st->base.block.size;
+   st->base.stride = nblocksx*util_format_get_blocksize(texture->format);
    st->base.usage = usage;
    st->base.face = face;
    st->base.level = level;
    st->base.zslice = zslice;
 
-   st->hw_nblocksy = st->base.nblocksy;
+   st->hw_nblocksy = nblocksy;
    
    st->hwbuf = svga_winsys_buffer_create(ss, 
                                          1, 
@@ -809,15 +809,15 @@ svga_get_tex_transfer(struct pipe_screen *screen,
    if(!st->hwbuf)
       goto no_hwbuf;
 
-   if(st->hw_nblocksy < st->base.nblocksy) {
+   if(st->hw_nblocksy < nblocksy) {
       /* We couldn't allocate a hardware buffer big enough for the transfer, 
        * so allocate regular malloc memory instead */
       debug_printf("%s: failed to allocate %u KB of DMA, splitting into %u x %u KB DMA transfers\n",
                    __FUNCTION__,
-                   (st->base.nblocksy*st->base.stride + 1023)/1024,
-                   (st->base.nblocksy + st->hw_nblocksy - 1)/st->hw_nblocksy,
+                   (nblocksy*st->base.stride + 1023)/1024,
+                   (nblocksy + st->hw_nblocksy - 1)/st->hw_nblocksy,
                    (st->hw_nblocksy*st->base.stride + 1023)/1024);
-      st->swbuf = MALLOC(st->base.nblocksy*st->base.stride);
+      st->swbuf = MALLOC(nblocksy*st->base.stride);
       if(!st->swbuf)
          goto no_swbuf;
    }
@@ -932,7 +932,7 @@ svga_get_tex_sampler_view(struct pipe_context *pipe, struct pipe_texture *pt,
       if (min_lod == 0 && max_lod >= pt->last_level)
          view = FALSE;
 
-      if (pf_is_compressed(pt->format) && view) {
+      if (util_format_is_compressed(pt->format) && view) {
          format = svga_translate_format_render(pt->format);
       }
 
@@ -971,9 +971,9 @@ svga_get_tex_sampler_view(struct pipe_context *pipe, struct pipe_texture *pt,
                "svga: Sampler view: no %p, mips %u..%u, nr %u, size (%ux%ux%u), last %u\n",
                pt, min_lod, max_lod,
                max_lod - min_lod + 1,
-               pt->width[0],
-               pt->height[0],
-               pt->depth[0],
+               pt->width0,
+               pt->height0,
+               pt->depth0,
                pt->last_level);
       sv->key.cachable = 0;
       sv->handle = tex->handle;
@@ -984,9 +984,9 @@ svga_get_tex_sampler_view(struct pipe_context *pipe, struct pipe_texture *pt,
             "svga: Sampler view: yes %p, mips %u..%u, nr %u, size (%ux%ux%u), last %u\n",
             pt, min_lod, max_lod,
             max_lod - min_lod + 1,
-            pt->width[0],
-            pt->height[0],
-            pt->depth[0],
+            pt->width0,
+            pt->height0,
+            pt->depth0,
             pt->last_level);
 
    sv->age = tex->age;
@@ -1036,9 +1036,9 @@ svga_validate_sampler_view(struct svga_context *svga, struct svga_sampler_view *
             svga_texture_copy_handle(svga, NULL,
                                      tex->handle, 0, 0, 0, i, k,
                                      v->handle, 0, 0, 0, i - v->min_lod, k,
-                                     tex->base.width[i],
-                                     tex->base.height[i],
-                                     tex->base.depth[i]);
+                                     u_minify(tex->base.width0, i),
+                                     u_minify(tex->base.height0, i),
+                                     u_minify(tex->base.depth0, i));
       }
    }
 
@@ -1071,8 +1071,7 @@ svga_screen_buffer_from_texture(struct pipe_texture *texture,
        svga_translate_format(texture->format),
        stex->handle);
 
-   *stride = pf_get_nblocksx(&texture->block, texture->width[0]) *
-      texture->block.size;
+   *stride = util_format_get_stride(texture->format, texture->width0);
 
    return *buffer != NULL;
 }
index 8cfdfea6937da2ce0db2d7d1c9aa5cf205fe471b..89ae24219fd7f23fdd09da7035712b41f1a60b2c 100644 (file)
@@ -171,8 +171,9 @@ svga_sampler_view_reference(struct svga_sampler_view **ptr, struct svga_sampler_
 {
    struct svga_sampler_view *old = *ptr;
 
-   if (pipe_reference((struct pipe_reference **)ptr, &v->reference))
+   if (pipe_reference(&(*ptr)->reference, &v->reference))
       svga_destroy_sampler_view_priv(old);
+   *ptr = v;
 }
 
 extern void
index a5777d4fbdcf4f3510795fb352cbc43873450025..6b0e511cec1d51554c35a7445f02a2bf55cbe21c 100644 (file)
@@ -140,8 +140,8 @@ static int emit_fs_consts( struct svga_context *svga,
             struct pipe_texture *tex = svga->curr.texture[i];
             float data[4];
 
-            data[0] = 1.0 / (float)tex->width[0];
-            data[1] = 1.0 / (float)tex->height[0];
+            data[0] = 1.0 / (float)tex->width0;
+            data[1] = 1.0 / (float)tex->height0;
             data[2] = 1.0;
             data[3] = 1.0;
 
index 00201b8091d9bba23b79489e1c24e79d6fb96ccf..3c35a8579f714f0826f33b45472a56ad482b00c0 100644 (file)
@@ -108,6 +108,7 @@ static int update_need_pipeline( struct svga_context *svga,
 {
    
    boolean need_pipeline = FALSE;
+   struct svga_vertex_shader *vs = svga->curr.vs;
 
    /* SVGA_NEW_RAST, SVGA_NEW_REDUCED_PRIMITIVE
     */
@@ -119,11 +120,9 @@ static int update_need_pipeline( struct svga_context *svga,
       need_pipeline = TRUE;
    }
 
-   /* SVGA_NEW_EDGEFLAGS
+   /* EDGEFLAGS
     */
-   if (svga->curr.rast->hw_unfilled != PIPE_POLYGON_MODE_FILL &&
-       svga->curr.reduced_prim == PIPE_PRIM_TRIANGLES && 
-       svga->curr.edgeflags != NULL) {
+    if (vs->base.info.writes_edgeflag) {
       SVGA_DBG(DEBUG_SWTNL, "%s: edgeflags\n", __FUNCTION__);
       need_pipeline = TRUE;
    }
@@ -150,6 +149,7 @@ struct svga_tracked_state svga_update_need_pipeline =
    "need pipeline",
    (SVGA_NEW_RAST |
     SVGA_NEW_CLIP |
+    SVGA_NEW_VS |
     SVGA_NEW_REDUCED_PRIMITIVE),
    update_need_pipeline
 };
index a947745732c78c700d1992b12ee8123e67feef56..44b7ceb4fa407ae5657b1c9281336ce9b6792642 100644 (file)
@@ -25,6 +25,7 @@
 
 #include "pipe/p_inlines.h"
 #include "pipe/p_defines.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "translate/translate.h"
 
@@ -210,7 +211,7 @@ static int update_zero_stride( struct svga_context *svga,
          mapped_buffer = pipe_buffer_map_range(svga->pipe.screen, 
                                                vbuffer->buffer,
                                                vel->src_offset,
-                                               pf_get_size(vel->src_format),
+                                               util_format_get_blocksize(vel->src_format),
                                                PIPE_BUFFER_USAGE_CPU_READ);
          translate->set_buffer(translate, vel->vertex_buffer_index,
                                mapped_buffer,
index 8b14c913f726b465b326e0dc5089a9c3f94dc4e1..7655121bec17fb6ea448cc93f8e4ae88896a4c4c 100644 (file)
@@ -90,7 +90,7 @@ svga_swtnl_draw_range_elements(struct svga_context *svga,
                             PIPE_BUFFER_USAGE_CPU_READ);
       assert(map);
       draw_set_mapped_constant_buffer(
-         draw, 
+         draw, PIPE_SHADER_VERTEX,
          map,
          svga->curr.cb[PIPE_SHADER_VERTEX]->size);
    }
index 161631211310298304bfd7ffd13e91f29166062d..94b6ccc62dd7fef172858c44b86152aed5dc10e4 100644 (file)
@@ -120,10 +120,6 @@ static int update_swtnl_draw( struct svga_context *svga,
       draw_set_mrd(svga->swtnl.draw, 
                    svga->curr.depthscale);
 
-   if (dirty & SVGA_NEW_EDGEFLAGS)
-      draw_set_edgeflags( svga->swtnl.draw, 
-                          svga->curr.edgeflags );
-
    return 0;
 }
 
@@ -138,8 +134,7 @@ struct svga_tracked_state svga_update_swtnl_draw =
     SVGA_NEW_VIEWPORT |
     SVGA_NEW_RAST |
     SVGA_NEW_FRAME_BUFFER |
-    SVGA_NEW_REDUCED_PRIMITIVE |
-    SVGA_NEW_EDGEFLAGS),
+    SVGA_NEW_REDUCED_PRIMITIVE),
    update_swtnl_draw
 };
 
@@ -161,7 +156,7 @@ int svga_swtnl_update_vdecl( struct svga_context *svga )
    memset(vdecl, 0, sizeof(vdecl));
 
    /* always add position */
-   src = draw_find_vs_output(draw, TGSI_SEMANTIC_POSITION, 0);
+   src = draw_find_shader_output(draw, TGSI_SEMANTIC_POSITION, 0);
    draw_emit_vertex_attr(vinfo, EMIT_4F, INTERP_LINEAR, src);
    vinfo->attrib[0].emit = EMIT_4F;
    vdecl[0].array.offset = offset;
@@ -174,7 +169,7 @@ int svga_swtnl_update_vdecl( struct svga_context *svga )
    for (i = 0; i < fs->base.info.num_inputs; i++) {
       unsigned name = fs->base.info.input_semantic_name[i];
       unsigned index = fs->base.info.input_semantic_index[i];
-      src = draw_find_vs_output(draw, name, index);
+      src = draw_find_shader_output(draw, name, index);
       vdecl[nr_decls].array.offset = offset;
       vdecl[nr_decls].identity.usageIndex = fs->base.info.input_semantic_index[i];
 
index 54457082a06da77c56e276c4f8c7d4dbffc01811..23b3ace7f30dd2ff571e83f0d96be1ae224fb8fe 100644 (file)
@@ -46,7 +46,7 @@ static boolean ps20_input( struct svga_shader_emitter *emit,
    dcl.values[0] = 0;
    dcl.values[1] = 0;
 
-   switch (semantic.SemanticName) {
+   switch (semantic.Name) {
    case TGSI_SEMANTIC_POSITION:
       /* Special case:
        */
@@ -55,15 +55,15 @@ static boolean ps20_input( struct svga_shader_emitter *emit,
       break;
    case TGSI_SEMANTIC_COLOR:
       reg = src_register( SVGA3DREG_INPUT, 
-                          semantic.SemanticIndex );
+                          semantic.Index );
       break;
    case TGSI_SEMANTIC_FOG:
-      assert(semantic.SemanticIndex == 0);
+      assert(semantic.Index == 0);
       reg = src_register( SVGA3DREG_TEXTURE, 0 );
       break;
    case TGSI_SEMANTIC_GENERIC:
       reg = src_register( SVGA3DREG_TEXTURE,
-                          semantic.SemanticIndex + 1 );
+                          semantic.Index + 1 );
       break;
    default:
       assert(0);
@@ -90,16 +90,16 @@ static boolean ps20_output( struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken reg;
 
-   switch (semantic.SemanticName) {
+   switch (semantic.Name) {
    case TGSI_SEMANTIC_COLOR:
-      if (semantic.SemanticIndex < PIPE_MAX_COLOR_BUFS) {
-         unsigned cbuf = semantic.SemanticIndex;
+      if (semantic.Index < PIPE_MAX_COLOR_BUFS) {
+         unsigned cbuf = semantic.Index;
 
          emit->output_map[idx] = dst_register( SVGA3DREG_TEMP,
                                                emit->nr_hw_temp++ );
          emit->temp_col[cbuf] = emit->output_map[idx];
          emit->true_col[cbuf] = dst_register( SVGA3DREG_COLOROUT, 
-                                              semantic.SemanticIndex );
+                                              semantic.Index );
       }
       else {
          assert(0);
@@ -111,7 +111,7 @@ static boolean ps20_output( struct svga_shader_emitter *emit,
                                             emit->nr_hw_temp++ );
       emit->temp_pos = emit->output_map[idx];
       emit->true_pos = dst_register( SVGA3DREG_DEPTHOUT, 
-                                     semantic.SemanticIndex );
+                                     semantic.Index );
       break;
    default:
       assert(0);
@@ -169,9 +169,9 @@ static boolean vs20_output( struct svga_shader_emitter *emit,
 
    /* Just build the register map table: 
     */
-   switch (semantic.SemanticName) {
+   switch (semantic.Name) {
    case TGSI_SEMANTIC_POSITION:
-      assert(semantic.SemanticIndex == 0);
+      assert(semantic.Index == 0);
       emit->output_map[idx] = dst_register( SVGA3DREG_TEMP,
                                             emit->nr_hw_temp++ );
       emit->temp_pos = emit->output_map[idx];
@@ -179,7 +179,7 @@ static boolean vs20_output( struct svga_shader_emitter *emit,
                                      SVGA3DRASTOUT_POSITION);
       break;
    case TGSI_SEMANTIC_PSIZE:
-      assert(semantic.SemanticIndex == 0);
+      assert(semantic.Index == 0);
       emit->output_map[idx] = dst_register( SVGA3DREG_TEMP,
                                             emit->nr_hw_temp++ );
       emit->temp_psiz = emit->output_map[idx];
@@ -187,17 +187,17 @@ static boolean vs20_output( struct svga_shader_emitter *emit,
                                       SVGA3DRASTOUT_PSIZE );
       break;
    case TGSI_SEMANTIC_FOG:
-      assert(semantic.SemanticIndex == 0);
+      assert(semantic.Index == 0);
       emit->output_map[idx] = dst_register( SVGA3DREG_TEXCRDOUT, 0 );
       break;
    case TGSI_SEMANTIC_COLOR:
       /* oD0 */
       emit->output_map[idx] = dst_register( SVGA3DREG_ATTROUT,
-                                            semantic.SemanticIndex );
+                                            semantic.Index );
       break;
    case TGSI_SEMANTIC_GENERIC:
       emit->output_map[idx] = dst_register( SVGA3DREG_TEXCRDOUT,
-                                            semantic.SemanticIndex + 1 );
+                                            semantic.Index + 1 );
       break;
    default:
       assert(0);
@@ -230,15 +230,15 @@ static boolean ps20_sampler( struct svga_shader_emitter *emit,
 boolean svga_translate_decl_sm20( struct svga_shader_emitter *emit,
                              const struct tgsi_full_declaration *decl )
 {
-   unsigned first = decl->DeclarationRange.First;
-   unsigned last = decl->DeclarationRange.Last;
+   unsigned first = decl->Range.First;
+   unsigned last = decl->Range.Last;
    unsigned semantic = 0;
    unsigned semantic_idx = 0;
    unsigned idx;
    
    if (decl->Declaration.Semantic) {
-      semantic = decl->Semantic.SemanticName;
-      semantic_idx = decl->Semantic.SemanticIndex;
+      semantic = decl->Semantic.Name;
+      semantic_idx = decl->Semantic.Index;
    }
 
    for( idx = first; idx <= last; idx++ ) {
index 08e7dfb117cdb13be85f6b6a02492c39bdcf172d..d1c7336dec497b0861b32cc74fb31d27cfc864fc 100644 (file)
@@ -35,35 +35,35 @@ static boolean translate_vs_ps_semantic( struct tgsi_declaration_semantic semant
                                          unsigned *usage,
                                          unsigned *idx )
 {
-   switch (semantic.SemanticName) {
+   switch (semantic.Name) {
    case TGSI_SEMANTIC_POSITION:  
-      *idx = semantic.SemanticIndex;
+      *idx = semantic.Index;
       *usage = SVGA3D_DECLUSAGE_POSITION;
       break;
    case TGSI_SEMANTIC_COLOR:     
 
-      *idx = semantic.SemanticIndex;
+      *idx = semantic.Index;
       *usage = SVGA3D_DECLUSAGE_COLOR;
       break;
    case TGSI_SEMANTIC_BCOLOR:
-      *idx = semantic.SemanticIndex + 2; /* sharing with COLOR */
+      *idx = semantic.Index + 2; /* sharing with COLOR */
       *usage = SVGA3D_DECLUSAGE_COLOR;
       break;
    case TGSI_SEMANTIC_FOG:       
       *idx = 0;
-      assert(semantic.SemanticIndex == 0);
+      assert(semantic.Index == 0);
       *usage = SVGA3D_DECLUSAGE_TEXCOORD;
       break;
    case TGSI_SEMANTIC_PSIZE:     
-      *idx = semantic.SemanticIndex;
+      *idx = semantic.Index;
       *usage = SVGA3D_DECLUSAGE_PSIZE;
       break;
    case TGSI_SEMANTIC_GENERIC:   
-      *idx = semantic.SemanticIndex + 1; /* texcoord[0] is reserved for fog */
+      *idx = semantic.Index + 1; /* texcoord[0] is reserved for fog */
       *usage = SVGA3D_DECLUSAGE_TEXCOORD;
       break;
    case TGSI_SEMANTIC_NORMAL:    
-      *idx = semantic.SemanticIndex;
+      *idx = semantic.Index;
       *usage = SVGA3D_DECLUSAGE_NORMAL;
       break;
    default:
@@ -120,7 +120,7 @@ static boolean ps30_input( struct svga_shader_emitter *emit,
    unsigned usage, index;
    SVGA3dShaderDestToken reg;
 
-   if (semantic.SemanticName == TGSI_SEMANTIC_POSITION) {
+   if (semantic.Name == TGSI_SEMANTIC_POSITION) {
       emit->input_map[idx] = src_register( SVGA3DREG_MISCTYPE,
                                            SVGA3DMISCREG_POSITION );
 
@@ -135,7 +135,7 @@ static boolean ps30_input( struct svga_shader_emitter *emit,
       return emit_decl( emit, reg, 0, 0 );
    }
    else if (emit->key.fkey.light_twoside &&
-            (semantic.SemanticName == TGSI_SEMANTIC_COLOR)) {
+            (semantic.Name == TGSI_SEMANTIC_COLOR)) {
 
       if (!translate_vs_ps_semantic( semantic, &usage, &index ))
          return FALSE;
@@ -150,7 +150,7 @@ static boolean ps30_input( struct svga_shader_emitter *emit,
       if (!emit_decl( emit, reg, usage, index ))
          return FALSE;
 
-      semantic.SemanticName = TGSI_SEMANTIC_BCOLOR;
+      semantic.Name = TGSI_SEMANTIC_BCOLOR;
       if (!translate_vs_ps_semantic( semantic, &usage, &index ))
          return FALSE;
 
@@ -164,7 +164,7 @@ static boolean ps30_input( struct svga_shader_emitter *emit,
 
       return TRUE;
    }
-   else if (semantic.SemanticName == TGSI_SEMANTIC_FACE) {
+   else if (semantic.Name == TGSI_SEMANTIC_FACE) {
       if (!emit_vface_decl( emit ))
          return FALSE;
       emit->emit_frontface = TRUE;
@@ -193,17 +193,17 @@ static boolean ps30_output( struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken reg;
 
-   switch (semantic.SemanticName) {
+   switch (semantic.Name) {
    case TGSI_SEMANTIC_COLOR:
       emit->output_map[idx] = dst_register( SVGA3DREG_COLOROUT, 
-                                            semantic.SemanticIndex );
+                                            semantic.Index );
       break;
    case TGSI_SEMANTIC_POSITION:
       emit->output_map[idx] = dst_register( SVGA3DREG_TEMP,
                                             emit->nr_hw_temp++ );
       emit->temp_pos = emit->output_map[idx];
       emit->true_pos = dst_register( SVGA3DREG_DEPTHOUT, 
-                                     semantic.SemanticIndex );
+                                     semantic.Index );
       break;
    default:
       assert(0);
@@ -283,14 +283,14 @@ static boolean vs30_output( struct svga_shader_emitter *emit,
    dcl.index = index;
    dcl.values[0] |= 1<<31;
 
-   if (semantic.SemanticName == TGSI_SEMANTIC_POSITION) {
+   if (semantic.Name == TGSI_SEMANTIC_POSITION) {
       assert(idx == 0);
       emit->output_map[idx] = dst_register( SVGA3DREG_TEMP,
                                             emit->nr_hw_temp++ );
       emit->temp_pos = emit->output_map[idx];
       emit->true_pos = dcl.dst;
    }
-   else if (semantic.SemanticName == TGSI_SEMANTIC_PSIZE) {
+   else if (semantic.Name == TGSI_SEMANTIC_PSIZE) {
       emit->output_map[idx] = dst_register( SVGA3DREG_TEMP,
                                             emit->nr_hw_temp++ );
       emit->temp_psiz = emit->output_map[idx];
@@ -335,15 +335,15 @@ static boolean ps30_sampler( struct svga_shader_emitter *emit,
 boolean svga_translate_decl_sm30( struct svga_shader_emitter *emit,
                              const struct tgsi_full_declaration *decl )
 {
-   unsigned first = decl->DeclarationRange.First;
-   unsigned last = decl->DeclarationRange.Last;
+   unsigned first = decl->Range.First;
+   unsigned last = decl->Range.Last;
    unsigned semantic = 0;
    unsigned semantic_idx = 0;
    unsigned idx;
 
    if (decl->Declaration.Semantic) {
-      semantic = decl->Semantic.SemanticName;
-      semantic_idx = decl->Semantic.SemanticIndex;
+      semantic = decl->Semantic.Name;
+      semantic_idx = decl->Semantic.Index;
    }
 
    for( idx = first; idx <= last; idx++ ) {
index ea409b7e165a5d548a5026476b0fbe8b22c9b00d..1670da8bfa995436a48c0b0a66c353dcf60fbdea 100644 (file)
@@ -96,24 +96,24 @@ translate_dst_register( struct svga_shader_emitter *emit,
                         const struct tgsi_full_instruction *insn,
                         unsigned idx )
 {
-   const struct tgsi_full_dst_register *reg = &insn->FullDstRegisters[idx];
+   const struct tgsi_full_dst_register *reg = &insn->Dst[idx];
    SVGA3dShaderDestToken dest;
 
-   switch (reg->DstRegister.File) {
+   switch (reg->Register.File) {
    case TGSI_FILE_OUTPUT:
       /* Output registers encode semantic information in their name.
        * Need to lookup a table built at decl time:
        */
-      dest = emit->output_map[reg->DstRegister.Index];
+      dest = emit->output_map[reg->Register.Index];
       break;
 
    default:
-      dest = dst_register( translate_file( reg->DstRegister.File ),
-                           reg->DstRegister.Index );
+      dest = dst_register( translate_file( reg->Register.File ),
+                           reg->Register.Index );
       break;
    }
 
-   dest.mask = reg->DstRegister.WriteMask;
+   dest.mask = reg->Register.WriteMask;
 
    if (insn->Instruction.Saturate) 
       dest.dstMod = SVGA3DDSTMOD_SATURATE;
@@ -176,33 +176,33 @@ translate_src_register( const struct svga_shader_emitter *emit,
 {
    struct src_register src;
 
-   switch (reg->SrcRegister.File) {
+   switch (reg->Register.File) {
    case TGSI_FILE_INPUT:
       /* Input registers are referred to by their semantic name rather
        * than by index.  Use the mapping build up from the decls:
        */
-      src = emit->input_map[reg->SrcRegister.Index];
+      src = emit->input_map[reg->Register.Index];
       break;
        
    case TGSI_FILE_IMMEDIATE:
       /* Immediates are appended after TGSI constants in the D3D
        * constant buffer.
        */
-      src = src_register( translate_file( reg->SrcRegister.File ),
-                          reg->SrcRegister.Index + 
+      src = src_register( translate_file( reg->Register.File ),
+                          reg->Register.Index + 
                           emit->imm_start );
       break;
 
    default:
-      src = src_register( translate_file( reg->SrcRegister.File ),
-                          reg->SrcRegister.Index );
+      src = src_register( translate_file( reg->Register.File ),
+                          reg->Register.Index );
 
       break;
    }
 
    /* Indirect addressing (for coninstant buffer lookups only)
     */
-   if (reg->SrcRegister.Indirect)
+   if (reg->Register.Indirect)
    {
       /* we shift the offset towards the minimum */
       if (svga_arl_needs_adjustment( emit )) {
@@ -213,28 +213,28 @@ translate_src_register( const struct svga_shader_emitter *emit,
       /* Not really sure what should go in the second token:
        */
       src.indirect = src_token( SVGA3DREG_ADDR,
-                                reg->SrcRegisterInd.Index );
+                                reg->Indirect.Index );
 
       src.indirect.swizzle = SWIZZLE_XXXX;
    }
 
    src = swizzle( src,
-                  reg->SrcRegister.SwizzleX,
-                  reg->SrcRegister.SwizzleY,
-                  reg->SrcRegister.SwizzleZ,
-                  reg->SrcRegister.SwizzleW );
+                  reg->Register.SwizzleX,
+                  reg->Register.SwizzleY,
+                  reg->Register.SwizzleZ,
+                  reg->Register.SwizzleW );
 
    /* src.mod isn't a bitfield, unfortunately:
     * See tgsi_util_get_full_src_register_sign_mode for implementation details.
     */
-   if (reg->SrcRegisterExtMod.Absolute) {
-      if (reg->SrcRegisterExtMod.Negate)
+   if (reg->Register.Absolute) {
+      if (reg->Register.Negate)
          src.base.srcMod = SVGA3DSRCMOD_ABSNEG;
       else
          src.base.srcMod = SVGA3DSRCMOD_ABS;
    }
    else {
-      if (reg->SrcRegister.Negate != reg->SrcRegisterExtMod.Negate)
+      if (reg->Register.Negate)
          src.base.srcMod = SVGA3DSRCMOD_NEG;
       else
          src.base.srcMod = SVGA3DSRCMOD_NONE;
@@ -629,7 +629,7 @@ static boolean emit_fake_arl(struct svga_shader_emitter *emit,
                              const struct tgsi_full_instruction *insn)
 {
    const struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    struct src_register src1 = get_fake_arl_const( emit );
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    SVGA3dShaderDestToken tmp = get_temp( emit );
@@ -653,7 +653,7 @@ static boolean emit_if(struct svga_shader_emitter *emit,
                        const struct tgsi_full_instruction *insn)
 {
    const struct src_register src = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    struct src_register zero = get_zero_immediate( emit );
    SVGA3dShaderInstToken if_token = inst_token( SVGA3DOP_IFC );
 
@@ -690,7 +690,7 @@ static boolean emit_floor(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    const struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    SVGA3dShaderDestToken temp = get_temp( emit );
 
    /* FRC  TMP, SRC */
@@ -716,11 +716,11 @@ static boolean emit_cmp(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    const struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    const struct src_register src1 = translate_src_register(
-      emit, &insn->FullSrcRegisters[1] );
+      emit, &insn->Src[1] );
    const struct src_register src2 = translate_src_register(
-      emit, &insn->FullSrcRegisters[2] );
+      emit, &insn->Src[2] );
 
    /* CMP  DST, SRC0, SRC2, SRC1 */
    return submit_op3( emit, inst_token( SVGA3DOP_CMP ), dst, src0, src2, src1);
@@ -740,9 +740,9 @@ static boolean emit_div(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    const struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    const struct src_register src1 = translate_src_register(
-      emit, &insn->FullSrcRegisters[1] );
+      emit, &insn->Src[1] );
    SVGA3dShaderDestToken temp = get_temp( emit );
    int i;
 
@@ -782,9 +782,9 @@ static boolean emit_dp2(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    const struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    const struct src_register src1 = translate_src_register(
-      emit, &insn->FullSrcRegisters[1] );
+      emit, &insn->Src[1] );
    SVGA3dShaderDestToken temp = get_temp( emit );
    struct src_register temp_src0, temp_src1;
 
@@ -815,9 +815,9 @@ static boolean emit_dph(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    const struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    struct src_register src1 = translate_src_register(
-      emit, &insn->FullSrcRegisters[1] );
+      emit, &insn->Src[1] );
    SVGA3dShaderDestToken temp = get_temp( emit );
 
    /* DP3  TMP, SRC1, SRC2 */
@@ -846,7 +846,7 @@ static boolean emit_nrm(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    const struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    SVGA3dShaderDestToken temp = get_temp( emit );
 
    /* DP3  TMP, SRC, SRC */
@@ -889,7 +889,7 @@ static boolean emit_sincos(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    SVGA3dShaderDestToken temp = get_temp( emit );
 
    /* SCS TMP SRC */
@@ -912,7 +912,7 @@ static boolean emit_sin(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    SVGA3dShaderDestToken temp = get_temp( emit );
 
    /* SCS TMP SRC */
@@ -937,7 +937,7 @@ static boolean emit_cos(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    SVGA3dShaderDestToken temp = get_temp( emit );
 
    /* SCS TMP SRC */
@@ -962,9 +962,9 @@ static boolean emit_sub(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    struct src_register src1 = translate_src_register(
-      emit, &insn->FullSrcRegisters[1] );
+      emit, &insn->Src[1] );
 
    src1 = negate(src1);
 
@@ -980,19 +980,19 @@ static boolean emit_kil(struct svga_shader_emitter *emit,
                         const struct tgsi_full_instruction *insn )
 {
    SVGA3dShaderInstToken inst;
-   const struct tgsi_full_src_register *reg = &insn->FullSrcRegisters[0];
+   const struct tgsi_full_src_register *reg = &insn->Src[0];
    struct src_register src0;
 
    inst = inst_token( SVGA3DOP_TEXKILL );
    src0 = translate_src_register( emit, reg );
 
-   if (reg->SrcRegisterExtMod.Absolute ||
-       reg->SrcRegister.Negate != reg->SrcRegisterExtMod.Negate ||
-       reg->SrcRegister.Indirect ||
-       reg->SrcRegister.SwizzleX != 0 ||
-       reg->SrcRegister.SwizzleY != 1 ||
-       reg->SrcRegister.SwizzleZ != 2 ||
-       reg->SrcRegister.File != TGSI_FILE_TEMPORARY)
+   if (reg->Register.Absolute ||
+       reg->Register.Negate ||
+       reg->Register.Indirect ||
+       reg->Register.SwizzleX != 0 ||
+       reg->Register.SwizzleY != 1 ||
+       reg->Register.SwizzleZ != 2 ||
+       reg->Register.File != TGSI_FILE_TEMPORARY)
    {
       SVGA3dShaderDestToken temp = get_temp( emit );
 
@@ -1154,9 +1154,9 @@ static boolean emit_select_op(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    struct src_register src1 = translate_src_register(
-      emit, &insn->FullSrcRegisters[1] );
+      emit, &insn->Src[1] );
       
    return emit_select( emit, compare, dst, src0, src1 );
 }
@@ -1189,8 +1189,8 @@ static boolean emit_tex2(struct svga_shader_emitter *emit,
       return FALSE;
    }
 
-   src0 = translate_src_register( emit, &insn->FullSrcRegisters[0] );
-   src1 = translate_src_register( emit, &insn->FullSrcRegisters[1] );
+   src0 = translate_src_register( emit, &insn->Src[0] );
+   src1 = translate_src_register( emit, &insn->Src[1] );
 
    if (emit->key.fkey.tex[src1.base.num].unnormalized) {
       struct src_register wh = get_tex_dimensions( emit, src1.base.num );
@@ -1231,9 +1231,9 @@ static boolean emit_tex3(struct svga_shader_emitter *emit,
       break;
    }
 
-   src0 = translate_src_register( emit, &insn->FullSrcRegisters[0] );
-   src1 = translate_src_register( emit, &insn->FullSrcRegisters[1] );
-   src2 = translate_src_register( emit, &insn->FullSrcRegisters[2] );
+   src0 = translate_src_register( emit, &insn->Src[0] );
+   src1 = translate_src_register( emit, &insn->Src[1] );
+   src2 = translate_src_register( emit, &insn->Src[2] );
 
    return submit_op3( emit, inst, dst, src0, src1, src2 );
 }
@@ -1245,9 +1245,9 @@ static boolean emit_tex(struct svga_shader_emitter *emit,
    SVGA3dShaderDestToken dst = 
       translate_dst_register( emit, insn, 0 );
    struct src_register src0 =
-      translate_src_register( emit, &insn->FullSrcRegisters[0] );
+      translate_src_register( emit, &insn->Src[0] );
    struct src_register src1 =
-      translate_src_register( emit, &insn->FullSrcRegisters[1] );
+      translate_src_register( emit, &insn->Src[1] );
 
    SVGA3dShaderDestToken tex_result;
 
@@ -1359,7 +1359,7 @@ static boolean emit_scalar_op1( struct svga_shader_emitter *emit,
 
    inst = inst_token( opcode );
    dst = translate_dst_register( emit, insn, 0 );
-   src = translate_src_register( emit, &insn->FullSrcRegisters[0] );
+   src = translate_src_register( emit, &insn->Src[0] );
    src = scalar( src, TGSI_SWIZZLE_X );
 
    return submit_op1( emit, inst, dst, src );
@@ -1370,7 +1370,7 @@ static boolean emit_simple_instruction(struct svga_shader_emitter *emit,
                                        unsigned opcode,
                                        const struct tgsi_full_instruction *insn )
 {
-   const struct tgsi_full_src_register *src = insn->FullSrcRegisters;
+   const struct tgsi_full_src_register *src = insn->Src;
    SVGA3dShaderInstToken inst;
    SVGA3dShaderDestToken dst;
 
@@ -1428,13 +1428,13 @@ static boolean emit_pow(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    struct src_register src1 = translate_src_register(
-      emit, &insn->FullSrcRegisters[1] );
+      emit, &insn->Src[1] );
    boolean need_tmp = FALSE;
    
    /* POW can only output to a temporary */
-   if (insn->FullDstRegisters[0].DstRegister.File != TGSI_FILE_TEMPORARY)
+   if (insn->Dst[0].Register.File != TGSI_FILE_TEMPORARY)
       need_tmp = TRUE;
    
    /* POW src1 must not be the same register as dst */
@@ -1463,9 +1463,9 @@ static boolean emit_xpd(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    const struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    const struct src_register src1 = translate_src_register(
-      emit, &insn->FullSrcRegisters[1] );
+      emit, &insn->Src[1] );
    boolean need_dst_tmp = FALSE;
 
    /* XPD can only output to a temporary */
@@ -1517,11 +1517,11 @@ static boolean emit_lrp(struct svga_shader_emitter *emit,
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    SVGA3dShaderDestToken tmp;
    const struct src_register src0 = translate_src_register(
-      emit, &insn->FullSrcRegisters[0] );
+      emit, &insn->Src[0] );
    const struct src_register src1 = translate_src_register(
-      emit, &insn->FullSrcRegisters[1] );
+      emit, &insn->Src[1] );
    const struct src_register src2 = translate_src_register(
-      emit, &insn->FullSrcRegisters[2] );
+      emit, &insn->Src[2] );
    boolean need_dst_tmp = FALSE;
 
    /* The dst reg must not be the same as src0 or src2 */
@@ -1568,9 +1568,9 @@ static boolean emit_dst_insn(struct svga_shader_emitter *emit,
       SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
       SVGA3dShaderDestToken tmp;
       const struct src_register src0 = translate_src_register(
-         emit, &insn->FullSrcRegisters[0] );
+         emit, &insn->Src[0] );
       const struct src_register src1 = translate_src_register(
-         emit, &insn->FullSrcRegisters[1] );
+         emit, &insn->Src[1] );
       struct src_register zero = get_zero_immediate( emit );
       boolean need_tmp = FALSE;
 
@@ -1633,7 +1633,7 @@ static boolean emit_exp(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    struct src_register src0 =
-      translate_src_register( emit, &insn->FullSrcRegisters[0] );
+      translate_src_register( emit, &insn->Src[0] );
    struct src_register zero = get_zero_immediate( emit );
    SVGA3dShaderDestToken fraction;
 
@@ -1723,7 +1723,7 @@ static boolean emit_lit(struct svga_shader_emitter *emit,
       SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
       SVGA3dShaderDestToken tmp = get_temp( emit );
       const struct src_register src0 = translate_src_register(
-         emit, &insn->FullSrcRegisters[0] );
+         emit, &insn->Src[0] );
       struct src_register zero = get_zero_immediate( emit );
 
       /* tmp = pow(src.y, src.w)
@@ -1806,7 +1806,7 @@ static boolean emit_ex2( struct svga_shader_emitter *emit,
 
    inst = inst_token( SVGA3DOP_EXP );
    dst = translate_dst_register( emit, insn, 0 );
-   src0 = translate_src_register( emit, &insn->FullSrcRegisters[0] );
+   src0 = translate_src_register( emit, &insn->Src[0] );
    src0 = scalar( src0, TGSI_SWIZZLE_X );
 
    if (dst.mask != TGSI_WRITEMASK_XYZW) {
@@ -1829,7 +1829,7 @@ static boolean emit_log(struct svga_shader_emitter *emit,
 {
    SVGA3dShaderDestToken dst = translate_dst_register( emit, insn, 0 );
    struct src_register src0 =
-      translate_src_register( emit, &insn->FullSrcRegisters[0] );
+      translate_src_register( emit, &insn->Src[0] );
    struct src_register zero = get_zero_immediate( emit );
    SVGA3dShaderDestToken abs_tmp;
    struct src_register abs_src0;
@@ -1953,7 +1953,7 @@ static boolean emit_bgnsub( struct svga_shader_emitter *emit,
 static boolean emit_call( struct svga_shader_emitter *emit,
                            const struct tgsi_full_instruction *insn )
 {
-   unsigned position = insn->InstructionExtLabel.Label;
+   unsigned position = insn->Label.Label;
    unsigned i;
    
    for (i = 0; i < emit->nr_labels; i++) {
@@ -2543,27 +2543,27 @@ pre_parse_instruction( struct svga_shader_emitter *emit,
                        const struct tgsi_full_instruction *insn,
                        int current_arl)
 {
-   if (insn->FullSrcRegisters[0].SrcRegister.Indirect &&
-       insn->FullSrcRegisters[0].SrcRegisterInd.File == TGSI_FILE_ADDRESS) {
-      const struct tgsi_full_src_register *reg = &insn->FullSrcRegisters[0];
-      if (reg->SrcRegister.Index < 0) {
-         pre_parse_add_indirect(emit, reg->SrcRegister.Index, current_arl);
+   if (insn->Src[0].Register.Indirect &&
+       insn->Src[0].Indirect.File == TGSI_FILE_ADDRESS) {
+      const struct tgsi_full_src_register *reg = &insn->Src[0];
+      if (reg->Register.Index < 0) {
+         pre_parse_add_indirect(emit, reg->Register.Index, current_arl);
       }
    }
 
-   if (insn->FullSrcRegisters[1].SrcRegister.Indirect &&
-       insn->FullSrcRegisters[1].SrcRegisterInd.File == TGSI_FILE_ADDRESS) {
-      const struct tgsi_full_src_register *reg = &insn->FullSrcRegisters[1];
-      if (reg->SrcRegister.Index < 0) {
-         pre_parse_add_indirect(emit, reg->SrcRegister.Index, current_arl);
+   if (insn->Src[1].Register.Indirect &&
+       insn->Src[1].Indirect.File == TGSI_FILE_ADDRESS) {
+      const struct tgsi_full_src_register *reg = &insn->Src[1];
+      if (reg->Register.Index < 0) {
+         pre_parse_add_indirect(emit, reg->Register.Index, current_arl);
       }
    }
 
-   if (insn->FullSrcRegisters[2].SrcRegister.Indirect &&
-       insn->FullSrcRegisters[2].SrcRegisterInd.File == TGSI_FILE_ADDRESS) {
-      const struct tgsi_full_src_register *reg = &insn->FullSrcRegisters[2];
-      if (reg->SrcRegister.Index < 0) {
-         pre_parse_add_indirect(emit, reg->SrcRegister.Index, current_arl);
+   if (insn->Src[2].Register.Indirect &&
+       insn->Src[2].Indirect.File == TGSI_FILE_ADDRESS) {
+      const struct tgsi_full_src_register *reg = &insn->Src[2];
+      if (reg->Register.Index < 0) {
+         pre_parse_add_indirect(emit, reg->Register.Index, current_arl);
       }
    }
 
index 910afa25287d494a2346399f6cf7bb9f924b4d2d..d59fb89a58cb2bfc1c9076948ee2779f7207c127 100644 (file)
@@ -42,554 +42,554 @@ dump_SVGA3dVertexDecl(const SVGA3dVertexDecl *cmd)
 {
    switch((*cmd).identity.type) {
    case SVGA3D_DECLTYPE_FLOAT1:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT1\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT1\n");
       break;
    case SVGA3D_DECLTYPE_FLOAT2:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT2\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT2\n");
       break;
    case SVGA3D_DECLTYPE_FLOAT3:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT3\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT3\n");
       break;
    case SVGA3D_DECLTYPE_FLOAT4:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT4\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT4\n");
       break;
    case SVGA3D_DECLTYPE_D3DCOLOR:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_D3DCOLOR\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_D3DCOLOR\n");
       break;
    case SVGA3D_DECLTYPE_UBYTE4:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_UBYTE4\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_UBYTE4\n");
       break;
    case SVGA3D_DECLTYPE_SHORT2:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_SHORT2\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_SHORT2\n");
       break;
    case SVGA3D_DECLTYPE_SHORT4:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_SHORT4\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_SHORT4\n");
       break;
    case SVGA3D_DECLTYPE_UBYTE4N:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_UBYTE4N\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_UBYTE4N\n");
       break;
    case SVGA3D_DECLTYPE_SHORT2N:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_SHORT2N\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_SHORT2N\n");
       break;
    case SVGA3D_DECLTYPE_SHORT4N:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_SHORT4N\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_SHORT4N\n");
       break;
    case SVGA3D_DECLTYPE_USHORT2N:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_USHORT2N\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_USHORT2N\n");
       break;
    case SVGA3D_DECLTYPE_USHORT4N:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_USHORT4N\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_USHORT4N\n");
       break;
    case SVGA3D_DECLTYPE_UDEC3:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_UDEC3\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_UDEC3\n");
       break;
    case SVGA3D_DECLTYPE_DEC3N:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_DEC3N\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_DEC3N\n");
       break;
    case SVGA3D_DECLTYPE_FLOAT16_2:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT16_2\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT16_2\n");
       break;
    case SVGA3D_DECLTYPE_FLOAT16_4:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT16_4\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_FLOAT16_4\n");
       break;
    case SVGA3D_DECLTYPE_MAX:
-      debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_MAX\n");
+      _debug_printf("\t\t.identity.type = SVGA3D_DECLTYPE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.identity.type = %i\n", (*cmd).identity.type);
+      _debug_printf("\t\t.identity.type = %i\n", (*cmd).identity.type);
       break;
    }
    switch((*cmd).identity.method) {
    case SVGA3D_DECLMETHOD_DEFAULT:
-      debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_DEFAULT\n");
+      _debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_DEFAULT\n");
       break;
    case SVGA3D_DECLMETHOD_PARTIALU:
-      debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_PARTIALU\n");
+      _debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_PARTIALU\n");
       break;
    case SVGA3D_DECLMETHOD_PARTIALV:
-      debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_PARTIALV\n");
+      _debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_PARTIALV\n");
       break;
    case SVGA3D_DECLMETHOD_CROSSUV:
-      debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_CROSSUV\n");
+      _debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_CROSSUV\n");
       break;
    case SVGA3D_DECLMETHOD_UV:
-      debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_UV\n");
+      _debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_UV\n");
       break;
    case SVGA3D_DECLMETHOD_LOOKUP:
-      debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_LOOKUP\n");
+      _debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_LOOKUP\n");
       break;
    case SVGA3D_DECLMETHOD_LOOKUPPRESAMPLED:
-      debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_LOOKUPPRESAMPLED\n");
+      _debug_printf("\t\t.identity.method = SVGA3D_DECLMETHOD_LOOKUPPRESAMPLED\n");
       break;
    default:
-      debug_printf("\t\t.identity.method = %i\n", (*cmd).identity.method);
+      _debug_printf("\t\t.identity.method = %i\n", (*cmd).identity.method);
       break;
    }
    switch((*cmd).identity.usage) {
    case SVGA3D_DECLUSAGE_POSITION:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_POSITION\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_POSITION\n");
       break;
    case SVGA3D_DECLUSAGE_BLENDWEIGHT:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_BLENDWEIGHT\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_BLENDWEIGHT\n");
       break;
    case SVGA3D_DECLUSAGE_BLENDINDICES:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_BLENDINDICES\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_BLENDINDICES\n");
       break;
    case SVGA3D_DECLUSAGE_NORMAL:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_NORMAL\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_NORMAL\n");
       break;
    case SVGA3D_DECLUSAGE_PSIZE:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_PSIZE\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_PSIZE\n");
       break;
    case SVGA3D_DECLUSAGE_TEXCOORD:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_TEXCOORD\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_TEXCOORD\n");
       break;
    case SVGA3D_DECLUSAGE_TANGENT:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_TANGENT\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_TANGENT\n");
       break;
    case SVGA3D_DECLUSAGE_BINORMAL:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_BINORMAL\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_BINORMAL\n");
       break;
    case SVGA3D_DECLUSAGE_TESSFACTOR:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_TESSFACTOR\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_TESSFACTOR\n");
       break;
    case SVGA3D_DECLUSAGE_POSITIONT:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_POSITIONT\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_POSITIONT\n");
       break;
    case SVGA3D_DECLUSAGE_COLOR:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_COLOR\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_COLOR\n");
       break;
    case SVGA3D_DECLUSAGE_FOG:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_FOG\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_FOG\n");
       break;
    case SVGA3D_DECLUSAGE_DEPTH:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_DEPTH\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_DEPTH\n");
       break;
    case SVGA3D_DECLUSAGE_SAMPLE:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_SAMPLE\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_SAMPLE\n");
       break;
    case SVGA3D_DECLUSAGE_MAX:
-      debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_MAX\n");
+      _debug_printf("\t\t.identity.usage = SVGA3D_DECLUSAGE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.identity.usage = %i\n", (*cmd).identity.usage);
+      _debug_printf("\t\t.identity.usage = %i\n", (*cmd).identity.usage);
       break;
    }
-   debug_printf("\t\t.identity.usageIndex = %u\n", (*cmd).identity.usageIndex);
-   debug_printf("\t\t.array.surfaceId = %u\n", (*cmd).array.surfaceId);
-   debug_printf("\t\t.array.offset = %u\n", (*cmd).array.offset);
-   debug_printf("\t\t.array.stride = %u\n", (*cmd).array.stride);
-   debug_printf("\t\t.rangeHint.first = %u\n", (*cmd).rangeHint.first);
-   debug_printf("\t\t.rangeHint.last = %u\n", (*cmd).rangeHint.last);
+   _debug_printf("\t\t.identity.usageIndex = %u\n", (*cmd).identity.usageIndex);
+   _debug_printf("\t\t.array.surfaceId = %u\n", (*cmd).array.surfaceId);
+   _debug_printf("\t\t.array.offset = %u\n", (*cmd).array.offset);
+   _debug_printf("\t\t.array.stride = %u\n", (*cmd).array.stride);
+   _debug_printf("\t\t.rangeHint.first = %u\n", (*cmd).rangeHint.first);
+   _debug_printf("\t\t.rangeHint.last = %u\n", (*cmd).rangeHint.last);
 }
 
 static void
 dump_SVGA3dTextureState(const SVGA3dTextureState *cmd)
 {
-   debug_printf("\t\t.stage = %u\n", (*cmd).stage);
+   _debug_printf("\t\t.stage = %u\n", (*cmd).stage);
    switch((*cmd).name) {
    case SVGA3D_TS_INVALID:
-      debug_printf("\t\t.name = SVGA3D_TS_INVALID\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_INVALID\n");
       break;
    case SVGA3D_TS_BIND_TEXTURE:
-      debug_printf("\t\t.name = SVGA3D_TS_BIND_TEXTURE\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_BIND_TEXTURE\n");
       break;
    case SVGA3D_TS_COLOROP:
-      debug_printf("\t\t.name = SVGA3D_TS_COLOROP\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_COLOROP\n");
       break;
    case SVGA3D_TS_COLORARG1:
-      debug_printf("\t\t.name = SVGA3D_TS_COLORARG1\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_COLORARG1\n");
       break;
    case SVGA3D_TS_COLORARG2:
-      debug_printf("\t\t.name = SVGA3D_TS_COLORARG2\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_COLORARG2\n");
       break;
    case SVGA3D_TS_ALPHAOP:
-      debug_printf("\t\t.name = SVGA3D_TS_ALPHAOP\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_ALPHAOP\n");
       break;
    case SVGA3D_TS_ALPHAARG1:
-      debug_printf("\t\t.name = SVGA3D_TS_ALPHAARG1\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_ALPHAARG1\n");
       break;
    case SVGA3D_TS_ALPHAARG2:
-      debug_printf("\t\t.name = SVGA3D_TS_ALPHAARG2\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_ALPHAARG2\n");
       break;
    case SVGA3D_TS_ADDRESSU:
-      debug_printf("\t\t.name = SVGA3D_TS_ADDRESSU\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_ADDRESSU\n");
       break;
    case SVGA3D_TS_ADDRESSV:
-      debug_printf("\t\t.name = SVGA3D_TS_ADDRESSV\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_ADDRESSV\n");
       break;
    case SVGA3D_TS_MIPFILTER:
-      debug_printf("\t\t.name = SVGA3D_TS_MIPFILTER\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_MIPFILTER\n");
       break;
    case SVGA3D_TS_MAGFILTER:
-      debug_printf("\t\t.name = SVGA3D_TS_MAGFILTER\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_MAGFILTER\n");
       break;
    case SVGA3D_TS_MINFILTER:
-      debug_printf("\t\t.name = SVGA3D_TS_MINFILTER\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_MINFILTER\n");
       break;
    case SVGA3D_TS_BORDERCOLOR:
-      debug_printf("\t\t.name = SVGA3D_TS_BORDERCOLOR\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_BORDERCOLOR\n");
       break;
    case SVGA3D_TS_TEXCOORDINDEX:
-      debug_printf("\t\t.name = SVGA3D_TS_TEXCOORDINDEX\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_TEXCOORDINDEX\n");
       break;
    case SVGA3D_TS_TEXTURETRANSFORMFLAGS:
-      debug_printf("\t\t.name = SVGA3D_TS_TEXTURETRANSFORMFLAGS\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_TEXTURETRANSFORMFLAGS\n");
       break;
    case SVGA3D_TS_TEXCOORDGEN:
-      debug_printf("\t\t.name = SVGA3D_TS_TEXCOORDGEN\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_TEXCOORDGEN\n");
       break;
    case SVGA3D_TS_BUMPENVMAT00:
-      debug_printf("\t\t.name = SVGA3D_TS_BUMPENVMAT00\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_BUMPENVMAT00\n");
       break;
    case SVGA3D_TS_BUMPENVMAT01:
-      debug_printf("\t\t.name = SVGA3D_TS_BUMPENVMAT01\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_BUMPENVMAT01\n");
       break;
    case SVGA3D_TS_BUMPENVMAT10:
-      debug_printf("\t\t.name = SVGA3D_TS_BUMPENVMAT10\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_BUMPENVMAT10\n");
       break;
    case SVGA3D_TS_BUMPENVMAT11:
-      debug_printf("\t\t.name = SVGA3D_TS_BUMPENVMAT11\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_BUMPENVMAT11\n");
       break;
    case SVGA3D_TS_TEXTURE_MIPMAP_LEVEL:
-      debug_printf("\t\t.name = SVGA3D_TS_TEXTURE_MIPMAP_LEVEL\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_TEXTURE_MIPMAP_LEVEL\n");
       break;
    case SVGA3D_TS_TEXTURE_LOD_BIAS:
-      debug_printf("\t\t.name = SVGA3D_TS_TEXTURE_LOD_BIAS\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_TEXTURE_LOD_BIAS\n");
       break;
    case SVGA3D_TS_TEXTURE_ANISOTROPIC_LEVEL:
-      debug_printf("\t\t.name = SVGA3D_TS_TEXTURE_ANISOTROPIC_LEVEL\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_TEXTURE_ANISOTROPIC_LEVEL\n");
       break;
    case SVGA3D_TS_ADDRESSW:
-      debug_printf("\t\t.name = SVGA3D_TS_ADDRESSW\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_ADDRESSW\n");
       break;
    case SVGA3D_TS_GAMMA:
-      debug_printf("\t\t.name = SVGA3D_TS_GAMMA\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_GAMMA\n");
       break;
    case SVGA3D_TS_BUMPENVLSCALE:
-      debug_printf("\t\t.name = SVGA3D_TS_BUMPENVLSCALE\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_BUMPENVLSCALE\n");
       break;
    case SVGA3D_TS_BUMPENVLOFFSET:
-      debug_printf("\t\t.name = SVGA3D_TS_BUMPENVLOFFSET\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_BUMPENVLOFFSET\n");
       break;
    case SVGA3D_TS_COLORARG0:
-      debug_printf("\t\t.name = SVGA3D_TS_COLORARG0\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_COLORARG0\n");
       break;
    case SVGA3D_TS_ALPHAARG0:
-      debug_printf("\t\t.name = SVGA3D_TS_ALPHAARG0\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_ALPHAARG0\n");
       break;
    case SVGA3D_TS_MAX:
-      debug_printf("\t\t.name = SVGA3D_TS_MAX\n");
+      _debug_printf("\t\t.name = SVGA3D_TS_MAX\n");
       break;
    default:
-      debug_printf("\t\t.name = %i\n", (*cmd).name);
+      _debug_printf("\t\t.name = %i\n", (*cmd).name);
       break;
    }
-   debug_printf("\t\t.value = %u\n", (*cmd).value);
-   debug_printf("\t\t.floatValue = %f\n", (*cmd).floatValue);
+   _debug_printf("\t\t.value = %u\n", (*cmd).value);
+   _debug_printf("\t\t.floatValue = %f\n", (*cmd).floatValue);
 }
 
 static void
 dump_SVGA3dCopyBox(const SVGA3dCopyBox *cmd)
 {
-   debug_printf("\t\t.x = %u\n", (*cmd).x);
-   debug_printf("\t\t.y = %u\n", (*cmd).y);
-   debug_printf("\t\t.z = %u\n", (*cmd).z);
-   debug_printf("\t\t.w = %u\n", (*cmd).w);
-   debug_printf("\t\t.h = %u\n", (*cmd).h);
-   debug_printf("\t\t.d = %u\n", (*cmd).d);
-   debug_printf("\t\t.srcx = %u\n", (*cmd).srcx);
-   debug_printf("\t\t.srcy = %u\n", (*cmd).srcy);
-   debug_printf("\t\t.srcz = %u\n", (*cmd).srcz);
+   _debug_printf("\t\t.x = %u\n", (*cmd).x);
+   _debug_printf("\t\t.y = %u\n", (*cmd).y);
+   _debug_printf("\t\t.z = %u\n", (*cmd).z);
+   _debug_printf("\t\t.w = %u\n", (*cmd).w);
+   _debug_printf("\t\t.h = %u\n", (*cmd).h);
+   _debug_printf("\t\t.d = %u\n", (*cmd).d);
+   _debug_printf("\t\t.srcx = %u\n", (*cmd).srcx);
+   _debug_printf("\t\t.srcy = %u\n", (*cmd).srcy);
+   _debug_printf("\t\t.srcz = %u\n", (*cmd).srcz);
 }
 
 static void
 dump_SVGA3dCmdSetClipPlane(const SVGA3dCmdSetClipPlane *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
-   debug_printf("\t\t.index = %u\n", (*cmd).index);
-   debug_printf("\t\t.plane[0] = %f\n", (*cmd).plane[0]);
-   debug_printf("\t\t.plane[1] = %f\n", (*cmd).plane[1]);
-   debug_printf("\t\t.plane[2] = %f\n", (*cmd).plane[2]);
-   debug_printf("\t\t.plane[3] = %f\n", (*cmd).plane[3]);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.index = %u\n", (*cmd).index);
+   _debug_printf("\t\t.plane[0] = %f\n", (*cmd).plane[0]);
+   _debug_printf("\t\t.plane[1] = %f\n", (*cmd).plane[1]);
+   _debug_printf("\t\t.plane[2] = %f\n", (*cmd).plane[2]);
+   _debug_printf("\t\t.plane[3] = %f\n", (*cmd).plane[3]);
 }
 
 static void
 dump_SVGA3dCmdWaitForQuery(const SVGA3dCmdWaitForQuery *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
    switch((*cmd).type) {
    case SVGA3D_QUERYTYPE_OCCLUSION:
-      debug_printf("\t\t.type = SVGA3D_QUERYTYPE_OCCLUSION\n");
+      _debug_printf("\t\t.type = SVGA3D_QUERYTYPE_OCCLUSION\n");
       break;
    case SVGA3D_QUERYTYPE_MAX:
-      debug_printf("\t\t.type = SVGA3D_QUERYTYPE_MAX\n");
+      _debug_printf("\t\t.type = SVGA3D_QUERYTYPE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.type = %i\n", (*cmd).type);
+      _debug_printf("\t\t.type = %i\n", (*cmd).type);
       break;
    }
-   debug_printf("\t\t.guestResult.gmrId = %u\n", (*cmd).guestResult.gmrId);
-   debug_printf("\t\t.guestResult.offset = %u\n", (*cmd).guestResult.offset);
+   _debug_printf("\t\t.guestResult.gmrId = %u\n", (*cmd).guestResult.gmrId);
+   _debug_printf("\t\t.guestResult.offset = %u\n", (*cmd).guestResult.offset);
 }
 
 static void
 dump_SVGA3dCmdSetRenderTarget(const SVGA3dCmdSetRenderTarget *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
    switch((*cmd).type) {
    case SVGA3D_RT_DEPTH:
-      debug_printf("\t\t.type = SVGA3D_RT_DEPTH\n");
+      _debug_printf("\t\t.type = SVGA3D_RT_DEPTH\n");
       break;
    case SVGA3D_RT_STENCIL:
-      debug_printf("\t\t.type = SVGA3D_RT_STENCIL\n");
+      _debug_printf("\t\t.type = SVGA3D_RT_STENCIL\n");
       break;
    default:
-      debug_printf("\t\t.type = SVGA3D_RT_COLOR%u\n", (*cmd).type - SVGA3D_RT_COLOR0);
+      _debug_printf("\t\t.type = SVGA3D_RT_COLOR%u\n", (*cmd).type - SVGA3D_RT_COLOR0);
       break;
    }
-   debug_printf("\t\t.target.sid = %u\n", (*cmd).target.sid);
-   debug_printf("\t\t.target.face = %u\n", (*cmd).target.face);
-   debug_printf("\t\t.target.mipmap = %u\n", (*cmd).target.mipmap);
+   _debug_printf("\t\t.target.sid = %u\n", (*cmd).target.sid);
+   _debug_printf("\t\t.target.face = %u\n", (*cmd).target.face);
+   _debug_printf("\t\t.target.mipmap = %u\n", (*cmd).target.mipmap);
 }
 
 static void
 dump_SVGA3dCmdSetTextureState(const SVGA3dCmdSetTextureState *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
 }
 
 static void
 dump_SVGA3dCmdSurfaceCopy(const SVGA3dCmdSurfaceCopy *cmd)
 {
-   debug_printf("\t\t.src.sid = %u\n", (*cmd).src.sid);
-   debug_printf("\t\t.src.face = %u\n", (*cmd).src.face);
-   debug_printf("\t\t.src.mipmap = %u\n", (*cmd).src.mipmap);
-   debug_printf("\t\t.dest.sid = %u\n", (*cmd).dest.sid);
-   debug_printf("\t\t.dest.face = %u\n", (*cmd).dest.face);
-   debug_printf("\t\t.dest.mipmap = %u\n", (*cmd).dest.mipmap);
+   _debug_printf("\t\t.src.sid = %u\n", (*cmd).src.sid);
+   _debug_printf("\t\t.src.face = %u\n", (*cmd).src.face);
+   _debug_printf("\t\t.src.mipmap = %u\n", (*cmd).src.mipmap);
+   _debug_printf("\t\t.dest.sid = %u\n", (*cmd).dest.sid);
+   _debug_printf("\t\t.dest.face = %u\n", (*cmd).dest.face);
+   _debug_printf("\t\t.dest.mipmap = %u\n", (*cmd).dest.mipmap);
 }
 
 static void
 dump_SVGA3dCmdSetMaterial(const SVGA3dCmdSetMaterial *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
    switch((*cmd).face) {
    case SVGA3D_FACE_INVALID:
-      debug_printf("\t\t.face = SVGA3D_FACE_INVALID\n");
+      _debug_printf("\t\t.face = SVGA3D_FACE_INVALID\n");
       break;
    case SVGA3D_FACE_NONE:
-      debug_printf("\t\t.face = SVGA3D_FACE_NONE\n");
+      _debug_printf("\t\t.face = SVGA3D_FACE_NONE\n");
       break;
    case SVGA3D_FACE_FRONT:
-      debug_printf("\t\t.face = SVGA3D_FACE_FRONT\n");
+      _debug_printf("\t\t.face = SVGA3D_FACE_FRONT\n");
       break;
    case SVGA3D_FACE_BACK:
-      debug_printf("\t\t.face = SVGA3D_FACE_BACK\n");
+      _debug_printf("\t\t.face = SVGA3D_FACE_BACK\n");
       break;
    case SVGA3D_FACE_FRONT_BACK:
-      debug_printf("\t\t.face = SVGA3D_FACE_FRONT_BACK\n");
+      _debug_printf("\t\t.face = SVGA3D_FACE_FRONT_BACK\n");
       break;
    case SVGA3D_FACE_MAX:
-      debug_printf("\t\t.face = SVGA3D_FACE_MAX\n");
+      _debug_printf("\t\t.face = SVGA3D_FACE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.face = %i\n", (*cmd).face);
+      _debug_printf("\t\t.face = %i\n", (*cmd).face);
       break;
    }
-   debug_printf("\t\t.material.diffuse[0] = %f\n", (*cmd).material.diffuse[0]);
-   debug_printf("\t\t.material.diffuse[1] = %f\n", (*cmd).material.diffuse[1]);
-   debug_printf("\t\t.material.diffuse[2] = %f\n", (*cmd).material.diffuse[2]);
-   debug_printf("\t\t.material.diffuse[3] = %f\n", (*cmd).material.diffuse[3]);
-   debug_printf("\t\t.material.ambient[0] = %f\n", (*cmd).material.ambient[0]);
-   debug_printf("\t\t.material.ambient[1] = %f\n", (*cmd).material.ambient[1]);
-   debug_printf("\t\t.material.ambient[2] = %f\n", (*cmd).material.ambient[2]);
-   debug_printf("\t\t.material.ambient[3] = %f\n", (*cmd).material.ambient[3]);
-   debug_printf("\t\t.material.specular[0] = %f\n", (*cmd).material.specular[0]);
-   debug_printf("\t\t.material.specular[1] = %f\n", (*cmd).material.specular[1]);
-   debug_printf("\t\t.material.specular[2] = %f\n", (*cmd).material.specular[2]);
-   debug_printf("\t\t.material.specular[3] = %f\n", (*cmd).material.specular[3]);
-   debug_printf("\t\t.material.emissive[0] = %f\n", (*cmd).material.emissive[0]);
-   debug_printf("\t\t.material.emissive[1] = %f\n", (*cmd).material.emissive[1]);
-   debug_printf("\t\t.material.emissive[2] = %f\n", (*cmd).material.emissive[2]);
-   debug_printf("\t\t.material.emissive[3] = %f\n", (*cmd).material.emissive[3]);
-   debug_printf("\t\t.material.shininess = %f\n", (*cmd).material.shininess);
+   _debug_printf("\t\t.material.diffuse[0] = %f\n", (*cmd).material.diffuse[0]);
+   _debug_printf("\t\t.material.diffuse[1] = %f\n", (*cmd).material.diffuse[1]);
+   _debug_printf("\t\t.material.diffuse[2] = %f\n", (*cmd).material.diffuse[2]);
+   _debug_printf("\t\t.material.diffuse[3] = %f\n", (*cmd).material.diffuse[3]);
+   _debug_printf("\t\t.material.ambient[0] = %f\n", (*cmd).material.ambient[0]);
+   _debug_printf("\t\t.material.ambient[1] = %f\n", (*cmd).material.ambient[1]);
+   _debug_printf("\t\t.material.ambient[2] = %f\n", (*cmd).material.ambient[2]);
+   _debug_printf("\t\t.material.ambient[3] = %f\n", (*cmd).material.ambient[3]);
+   _debug_printf("\t\t.material.specular[0] = %f\n", (*cmd).material.specular[0]);
+   _debug_printf("\t\t.material.specular[1] = %f\n", (*cmd).material.specular[1]);
+   _debug_printf("\t\t.material.specular[2] = %f\n", (*cmd).material.specular[2]);
+   _debug_printf("\t\t.material.specular[3] = %f\n", (*cmd).material.specular[3]);
+   _debug_printf("\t\t.material.emissive[0] = %f\n", (*cmd).material.emissive[0]);
+   _debug_printf("\t\t.material.emissive[1] = %f\n", (*cmd).material.emissive[1]);
+   _debug_printf("\t\t.material.emissive[2] = %f\n", (*cmd).material.emissive[2]);
+   _debug_printf("\t\t.material.emissive[3] = %f\n", (*cmd).material.emissive[3]);
+   _debug_printf("\t\t.material.shininess = %f\n", (*cmd).material.shininess);
 }
 
 static void
 dump_SVGA3dCmdSetLightData(const SVGA3dCmdSetLightData *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
-   debug_printf("\t\t.index = %u\n", (*cmd).index);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.index = %u\n", (*cmd).index);
    switch((*cmd).data.type) {
    case SVGA3D_LIGHTTYPE_INVALID:
-      debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_INVALID\n");
+      _debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_INVALID\n");
       break;
    case SVGA3D_LIGHTTYPE_POINT:
-      debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_POINT\n");
+      _debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_POINT\n");
       break;
    case SVGA3D_LIGHTTYPE_SPOT1:
-      debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_SPOT1\n");
+      _debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_SPOT1\n");
       break;
    case SVGA3D_LIGHTTYPE_SPOT2:
-      debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_SPOT2\n");
+      _debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_SPOT2\n");
       break;
    case SVGA3D_LIGHTTYPE_DIRECTIONAL:
-      debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_DIRECTIONAL\n");
+      _debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_DIRECTIONAL\n");
       break;
    case SVGA3D_LIGHTTYPE_MAX:
-      debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_MAX\n");
+      _debug_printf("\t\t.data.type = SVGA3D_LIGHTTYPE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.data.type = %i\n", (*cmd).data.type);
+      _debug_printf("\t\t.data.type = %i\n", (*cmd).data.type);
       break;
    }
-   debug_printf("\t\t.data.inWorldSpace = %u\n", (*cmd).data.inWorldSpace);
-   debug_printf("\t\t.data.diffuse[0] = %f\n", (*cmd).data.diffuse[0]);
-   debug_printf("\t\t.data.diffuse[1] = %f\n", (*cmd).data.diffuse[1]);
-   debug_printf("\t\t.data.diffuse[2] = %f\n", (*cmd).data.diffuse[2]);
-   debug_printf("\t\t.data.diffuse[3] = %f\n", (*cmd).data.diffuse[3]);
-   debug_printf("\t\t.data.specular[0] = %f\n", (*cmd).data.specular[0]);
-   debug_printf("\t\t.data.specular[1] = %f\n", (*cmd).data.specular[1]);
-   debug_printf("\t\t.data.specular[2] = %f\n", (*cmd).data.specular[2]);
-   debug_printf("\t\t.data.specular[3] = %f\n", (*cmd).data.specular[3]);
-   debug_printf("\t\t.data.ambient[0] = %f\n", (*cmd).data.ambient[0]);
-   debug_printf("\t\t.data.ambient[1] = %f\n", (*cmd).data.ambient[1]);
-   debug_printf("\t\t.data.ambient[2] = %f\n", (*cmd).data.ambient[2]);
-   debug_printf("\t\t.data.ambient[3] = %f\n", (*cmd).data.ambient[3]);
-   debug_printf("\t\t.data.position[0] = %f\n", (*cmd).data.position[0]);
-   debug_printf("\t\t.data.position[1] = %f\n", (*cmd).data.position[1]);
-   debug_printf("\t\t.data.position[2] = %f\n", (*cmd).data.position[2]);
-   debug_printf("\t\t.data.position[3] = %f\n", (*cmd).data.position[3]);
-   debug_printf("\t\t.data.direction[0] = %f\n", (*cmd).data.direction[0]);
-   debug_printf("\t\t.data.direction[1] = %f\n", (*cmd).data.direction[1]);
-   debug_printf("\t\t.data.direction[2] = %f\n", (*cmd).data.direction[2]);
-   debug_printf("\t\t.data.direction[3] = %f\n", (*cmd).data.direction[3]);
-   debug_printf("\t\t.data.range = %f\n", (*cmd).data.range);
-   debug_printf("\t\t.data.falloff = %f\n", (*cmd).data.falloff);
-   debug_printf("\t\t.data.attenuation0 = %f\n", (*cmd).data.attenuation0);
-   debug_printf("\t\t.data.attenuation1 = %f\n", (*cmd).data.attenuation1);
-   debug_printf("\t\t.data.attenuation2 = %f\n", (*cmd).data.attenuation2);
-   debug_printf("\t\t.data.theta = %f\n", (*cmd).data.theta);
-   debug_printf("\t\t.data.phi = %f\n", (*cmd).data.phi);
+   _debug_printf("\t\t.data.inWorldSpace = %u\n", (*cmd).data.inWorldSpace);
+   _debug_printf("\t\t.data.diffuse[0] = %f\n", (*cmd).data.diffuse[0]);
+   _debug_printf("\t\t.data.diffuse[1] = %f\n", (*cmd).data.diffuse[1]);
+   _debug_printf("\t\t.data.diffuse[2] = %f\n", (*cmd).data.diffuse[2]);
+   _debug_printf("\t\t.data.diffuse[3] = %f\n", (*cmd).data.diffuse[3]);
+   _debug_printf("\t\t.data.specular[0] = %f\n", (*cmd).data.specular[0]);
+   _debug_printf("\t\t.data.specular[1] = %f\n", (*cmd).data.specular[1]);
+   _debug_printf("\t\t.data.specular[2] = %f\n", (*cmd).data.specular[2]);
+   _debug_printf("\t\t.data.specular[3] = %f\n", (*cmd).data.specular[3]);
+   _debug_printf("\t\t.data.ambient[0] = %f\n", (*cmd).data.ambient[0]);
+   _debug_printf("\t\t.data.ambient[1] = %f\n", (*cmd).data.ambient[1]);
+   _debug_printf("\t\t.data.ambient[2] = %f\n", (*cmd).data.ambient[2]);
+   _debug_printf("\t\t.data.ambient[3] = %f\n", (*cmd).data.ambient[3]);
+   _debug_printf("\t\t.data.position[0] = %f\n", (*cmd).data.position[0]);
+   _debug_printf("\t\t.data.position[1] = %f\n", (*cmd).data.position[1]);
+   _debug_printf("\t\t.data.position[2] = %f\n", (*cmd).data.position[2]);
+   _debug_printf("\t\t.data.position[3] = %f\n", (*cmd).data.position[3]);
+   _debug_printf("\t\t.data.direction[0] = %f\n", (*cmd).data.direction[0]);
+   _debug_printf("\t\t.data.direction[1] = %f\n", (*cmd).data.direction[1]);
+   _debug_printf("\t\t.data.direction[2] = %f\n", (*cmd).data.direction[2]);
+   _debug_printf("\t\t.data.direction[3] = %f\n", (*cmd).data.direction[3]);
+   _debug_printf("\t\t.data.range = %f\n", (*cmd).data.range);
+   _debug_printf("\t\t.data.falloff = %f\n", (*cmd).data.falloff);
+   _debug_printf("\t\t.data.attenuation0 = %f\n", (*cmd).data.attenuation0);
+   _debug_printf("\t\t.data.attenuation1 = %f\n", (*cmd).data.attenuation1);
+   _debug_printf("\t\t.data.attenuation2 = %f\n", (*cmd).data.attenuation2);
+   _debug_printf("\t\t.data.theta = %f\n", (*cmd).data.theta);
+   _debug_printf("\t\t.data.phi = %f\n", (*cmd).data.phi);
 }
 
 static void
 dump_SVGA3dCmdSetViewport(const SVGA3dCmdSetViewport *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
-   debug_printf("\t\t.rect.x = %u\n", (*cmd).rect.x);
-   debug_printf("\t\t.rect.y = %u\n", (*cmd).rect.y);
-   debug_printf("\t\t.rect.w = %u\n", (*cmd).rect.w);
-   debug_printf("\t\t.rect.h = %u\n", (*cmd).rect.h);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.rect.x = %u\n", (*cmd).rect.x);
+   _debug_printf("\t\t.rect.y = %u\n", (*cmd).rect.y);
+   _debug_printf("\t\t.rect.w = %u\n", (*cmd).rect.w);
+   _debug_printf("\t\t.rect.h = %u\n", (*cmd).rect.h);
 }
 
 static void
 dump_SVGA3dCmdSetScissorRect(const SVGA3dCmdSetScissorRect *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
-   debug_printf("\t\t.rect.x = %u\n", (*cmd).rect.x);
-   debug_printf("\t\t.rect.y = %u\n", (*cmd).rect.y);
-   debug_printf("\t\t.rect.w = %u\n", (*cmd).rect.w);
-   debug_printf("\t\t.rect.h = %u\n", (*cmd).rect.h);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.rect.x = %u\n", (*cmd).rect.x);
+   _debug_printf("\t\t.rect.y = %u\n", (*cmd).rect.y);
+   _debug_printf("\t\t.rect.w = %u\n", (*cmd).rect.w);
+   _debug_printf("\t\t.rect.h = %u\n", (*cmd).rect.h);
 }
 
 static void
 dump_SVGA3dCopyRect(const SVGA3dCopyRect *cmd)
 {
-   debug_printf("\t\t.x = %u\n", (*cmd).x);
-   debug_printf("\t\t.y = %u\n", (*cmd).y);
-   debug_printf("\t\t.w = %u\n", (*cmd).w);
-   debug_printf("\t\t.h = %u\n", (*cmd).h);
-   debug_printf("\t\t.srcx = %u\n", (*cmd).srcx);
-   debug_printf("\t\t.srcy = %u\n", (*cmd).srcy);
+   _debug_printf("\t\t.x = %u\n", (*cmd).x);
+   _debug_printf("\t\t.y = %u\n", (*cmd).y);
+   _debug_printf("\t\t.w = %u\n", (*cmd).w);
+   _debug_printf("\t\t.h = %u\n", (*cmd).h);
+   _debug_printf("\t\t.srcx = %u\n", (*cmd).srcx);
+   _debug_printf("\t\t.srcy = %u\n", (*cmd).srcy);
 }
 
 static void
 dump_SVGA3dCmdSetShader(const SVGA3dCmdSetShader *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
    switch((*cmd).type) {
    case SVGA3D_SHADERTYPE_COMPILED_DX8:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_COMPILED_DX8\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_COMPILED_DX8\n");
       break;
    case SVGA3D_SHADERTYPE_VS:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_VS\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_VS\n");
       break;
    case SVGA3D_SHADERTYPE_PS:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_PS\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_PS\n");
       break;
    case SVGA3D_SHADERTYPE_MAX:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_MAX\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.type = %i\n", (*cmd).type);
+      _debug_printf("\t\t.type = %i\n", (*cmd).type);
       break;
    }
-   debug_printf("\t\t.shid = %u\n", (*cmd).shid);
+   _debug_printf("\t\t.shid = %u\n", (*cmd).shid);
 }
 
 static void
 dump_SVGA3dCmdEndQuery(const SVGA3dCmdEndQuery *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
    switch((*cmd).type) {
    case SVGA3D_QUERYTYPE_OCCLUSION:
-      debug_printf("\t\t.type = SVGA3D_QUERYTYPE_OCCLUSION\n");
+      _debug_printf("\t\t.type = SVGA3D_QUERYTYPE_OCCLUSION\n");
       break;
    case SVGA3D_QUERYTYPE_MAX:
-      debug_printf("\t\t.type = SVGA3D_QUERYTYPE_MAX\n");
+      _debug_printf("\t\t.type = SVGA3D_QUERYTYPE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.type = %i\n", (*cmd).type);
+      _debug_printf("\t\t.type = %i\n", (*cmd).type);
       break;
    }
-   debug_printf("\t\t.guestResult.gmrId = %u\n", (*cmd).guestResult.gmrId);
-   debug_printf("\t\t.guestResult.offset = %u\n", (*cmd).guestResult.offset);
+   _debug_printf("\t\t.guestResult.gmrId = %u\n", (*cmd).guestResult.gmrId);
+   _debug_printf("\t\t.guestResult.offset = %u\n", (*cmd).guestResult.offset);
 }
 
 static void
 dump_SVGA3dSize(const SVGA3dSize *cmd)
 {
-   debug_printf("\t\t.width = %u\n", (*cmd).width);
-   debug_printf("\t\t.height = %u\n", (*cmd).height);
-   debug_printf("\t\t.depth = %u\n", (*cmd).depth);
+   _debug_printf("\t\t.width = %u\n", (*cmd).width);
+   _debug_printf("\t\t.height = %u\n", (*cmd).height);
+   _debug_printf("\t\t.depth = %u\n", (*cmd).depth);
 }
 
 static void
 dump_SVGA3dCmdDestroySurface(const SVGA3dCmdDestroySurface *cmd)
 {
-   debug_printf("\t\t.sid = %u\n", (*cmd).sid);
+   _debug_printf("\t\t.sid = %u\n", (*cmd).sid);
 }
 
 static void
 dump_SVGA3dCmdDefineContext(const SVGA3dCmdDefineContext *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
 }
 
 static void
 dump_SVGA3dRect(const SVGA3dRect *cmd)
 {
-   debug_printf("\t\t.x = %u\n", (*cmd).x);
-   debug_printf("\t\t.y = %u\n", (*cmd).y);
-   debug_printf("\t\t.w = %u\n", (*cmd).w);
-   debug_printf("\t\t.h = %u\n", (*cmd).h);
+   _debug_printf("\t\t.x = %u\n", (*cmd).x);
+   _debug_printf("\t\t.y = %u\n", (*cmd).y);
+   _debug_printf("\t\t.w = %u\n", (*cmd).w);
+   _debug_printf("\t\t.h = %u\n", (*cmd).h);
 }
 
 static void
 dump_SVGA3dCmdBeginQuery(const SVGA3dCmdBeginQuery *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
    switch((*cmd).type) {
    case SVGA3D_QUERYTYPE_OCCLUSION:
-      debug_printf("\t\t.type = SVGA3D_QUERYTYPE_OCCLUSION\n");
+      _debug_printf("\t\t.type = SVGA3D_QUERYTYPE_OCCLUSION\n");
       break;
    case SVGA3D_QUERYTYPE_MAX:
-      debug_printf("\t\t.type = SVGA3D_QUERYTYPE_MAX\n");
+      _debug_printf("\t\t.type = SVGA3D_QUERYTYPE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.type = %i\n", (*cmd).type);
+      _debug_printf("\t\t.type = %i\n", (*cmd).type);
       break;
    }
 }
@@ -599,336 +599,336 @@ dump_SVGA3dRenderState(const SVGA3dRenderState *cmd)
 {
    switch((*cmd).state) {
    case SVGA3D_RS_INVALID:
-      debug_printf("\t\t.state = SVGA3D_RS_INVALID\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_INVALID\n");
       break;
    case SVGA3D_RS_ZENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_ZENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_ZENABLE\n");
       break;
    case SVGA3D_RS_ZWRITEENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_ZWRITEENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_ZWRITEENABLE\n");
       break;
    case SVGA3D_RS_ALPHATESTENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_ALPHATESTENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_ALPHATESTENABLE\n");
       break;
    case SVGA3D_RS_DITHERENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_DITHERENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_DITHERENABLE\n");
       break;
    case SVGA3D_RS_BLENDENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_BLENDENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_BLENDENABLE\n");
       break;
    case SVGA3D_RS_FOGENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_FOGENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_FOGENABLE\n");
       break;
    case SVGA3D_RS_SPECULARENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_SPECULARENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_SPECULARENABLE\n");
       break;
    case SVGA3D_RS_STENCILENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_STENCILENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_STENCILENABLE\n");
       break;
    case SVGA3D_RS_LIGHTINGENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_LIGHTINGENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_LIGHTINGENABLE\n");
       break;
    case SVGA3D_RS_NORMALIZENORMALS:
-      debug_printf("\t\t.state = SVGA3D_RS_NORMALIZENORMALS\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_NORMALIZENORMALS\n");
       break;
    case SVGA3D_RS_POINTSPRITEENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_POINTSPRITEENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_POINTSPRITEENABLE\n");
       break;
    case SVGA3D_RS_POINTSCALEENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_POINTSCALEENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_POINTSCALEENABLE\n");
       break;
    case SVGA3D_RS_STENCILREF:
-      debug_printf("\t\t.state = SVGA3D_RS_STENCILREF\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_STENCILREF\n");
       break;
    case SVGA3D_RS_STENCILMASK:
-      debug_printf("\t\t.state = SVGA3D_RS_STENCILMASK\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_STENCILMASK\n");
       break;
    case SVGA3D_RS_STENCILWRITEMASK:
-      debug_printf("\t\t.state = SVGA3D_RS_STENCILWRITEMASK\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_STENCILWRITEMASK\n");
       break;
    case SVGA3D_RS_FOGSTART:
-      debug_printf("\t\t.state = SVGA3D_RS_FOGSTART\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_FOGSTART\n");
       break;
    case SVGA3D_RS_FOGEND:
-      debug_printf("\t\t.state = SVGA3D_RS_FOGEND\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_FOGEND\n");
       break;
    case SVGA3D_RS_FOGDENSITY:
-      debug_printf("\t\t.state = SVGA3D_RS_FOGDENSITY\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_FOGDENSITY\n");
       break;
    case SVGA3D_RS_POINTSIZE:
-      debug_printf("\t\t.state = SVGA3D_RS_POINTSIZE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_POINTSIZE\n");
       break;
    case SVGA3D_RS_POINTSIZEMIN:
-      debug_printf("\t\t.state = SVGA3D_RS_POINTSIZEMIN\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_POINTSIZEMIN\n");
       break;
    case SVGA3D_RS_POINTSIZEMAX:
-      debug_printf("\t\t.state = SVGA3D_RS_POINTSIZEMAX\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_POINTSIZEMAX\n");
       break;
    case SVGA3D_RS_POINTSCALE_A:
-      debug_printf("\t\t.state = SVGA3D_RS_POINTSCALE_A\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_POINTSCALE_A\n");
       break;
    case SVGA3D_RS_POINTSCALE_B:
-      debug_printf("\t\t.state = SVGA3D_RS_POINTSCALE_B\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_POINTSCALE_B\n");
       break;
    case SVGA3D_RS_POINTSCALE_C:
-      debug_printf("\t\t.state = SVGA3D_RS_POINTSCALE_C\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_POINTSCALE_C\n");
       break;
    case SVGA3D_RS_FOGCOLOR:
-      debug_printf("\t\t.state = SVGA3D_RS_FOGCOLOR\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_FOGCOLOR\n");
       break;
    case SVGA3D_RS_AMBIENT:
-      debug_printf("\t\t.state = SVGA3D_RS_AMBIENT\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_AMBIENT\n");
       break;
    case SVGA3D_RS_CLIPPLANEENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_CLIPPLANEENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_CLIPPLANEENABLE\n");
       break;
    case SVGA3D_RS_FOGMODE:
-      debug_printf("\t\t.state = SVGA3D_RS_FOGMODE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_FOGMODE\n");
       break;
    case SVGA3D_RS_FILLMODE:
-      debug_printf("\t\t.state = SVGA3D_RS_FILLMODE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_FILLMODE\n");
       break;
    case SVGA3D_RS_SHADEMODE:
-      debug_printf("\t\t.state = SVGA3D_RS_SHADEMODE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_SHADEMODE\n");
       break;
    case SVGA3D_RS_LINEPATTERN:
-      debug_printf("\t\t.state = SVGA3D_RS_LINEPATTERN\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_LINEPATTERN\n");
       break;
    case SVGA3D_RS_SRCBLEND:
-      debug_printf("\t\t.state = SVGA3D_RS_SRCBLEND\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_SRCBLEND\n");
       break;
    case SVGA3D_RS_DSTBLEND:
-      debug_printf("\t\t.state = SVGA3D_RS_DSTBLEND\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_DSTBLEND\n");
       break;
    case SVGA3D_RS_BLENDEQUATION:
-      debug_printf("\t\t.state = SVGA3D_RS_BLENDEQUATION\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_BLENDEQUATION\n");
       break;
    case SVGA3D_RS_CULLMODE:
-      debug_printf("\t\t.state = SVGA3D_RS_CULLMODE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_CULLMODE\n");
       break;
    case SVGA3D_RS_ZFUNC:
-      debug_printf("\t\t.state = SVGA3D_RS_ZFUNC\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_ZFUNC\n");
       break;
    case SVGA3D_RS_ALPHAFUNC:
-      debug_printf("\t\t.state = SVGA3D_RS_ALPHAFUNC\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_ALPHAFUNC\n");
       break;
    case SVGA3D_RS_STENCILFUNC:
-      debug_printf("\t\t.state = SVGA3D_RS_STENCILFUNC\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_STENCILFUNC\n");
       break;
    case SVGA3D_RS_STENCILFAIL:
-      debug_printf("\t\t.state = SVGA3D_RS_STENCILFAIL\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_STENCILFAIL\n");
       break;
    case SVGA3D_RS_STENCILZFAIL:
-      debug_printf("\t\t.state = SVGA3D_RS_STENCILZFAIL\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_STENCILZFAIL\n");
       break;
    case SVGA3D_RS_STENCILPASS:
-      debug_printf("\t\t.state = SVGA3D_RS_STENCILPASS\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_STENCILPASS\n");
       break;
    case SVGA3D_RS_ALPHAREF:
-      debug_printf("\t\t.state = SVGA3D_RS_ALPHAREF\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_ALPHAREF\n");
       break;
    case SVGA3D_RS_FRONTWINDING:
-      debug_printf("\t\t.state = SVGA3D_RS_FRONTWINDING\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_FRONTWINDING\n");
       break;
    case SVGA3D_RS_COORDINATETYPE:
-      debug_printf("\t\t.state = SVGA3D_RS_COORDINATETYPE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_COORDINATETYPE\n");
       break;
    case SVGA3D_RS_ZBIAS:
-      debug_printf("\t\t.state = SVGA3D_RS_ZBIAS\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_ZBIAS\n");
       break;
    case SVGA3D_RS_RANGEFOGENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_RANGEFOGENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_RANGEFOGENABLE\n");
       break;
    case SVGA3D_RS_COLORWRITEENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_COLORWRITEENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_COLORWRITEENABLE\n");
       break;
    case SVGA3D_RS_VERTEXMATERIALENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_VERTEXMATERIALENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_VERTEXMATERIALENABLE\n");
       break;
    case SVGA3D_RS_DIFFUSEMATERIALSOURCE:
-      debug_printf("\t\t.state = SVGA3D_RS_DIFFUSEMATERIALSOURCE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_DIFFUSEMATERIALSOURCE\n");
       break;
    case SVGA3D_RS_SPECULARMATERIALSOURCE:
-      debug_printf("\t\t.state = SVGA3D_RS_SPECULARMATERIALSOURCE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_SPECULARMATERIALSOURCE\n");
       break;
    case SVGA3D_RS_AMBIENTMATERIALSOURCE:
-      debug_printf("\t\t.state = SVGA3D_RS_AMBIENTMATERIALSOURCE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_AMBIENTMATERIALSOURCE\n");
       break;
    case SVGA3D_RS_EMISSIVEMATERIALSOURCE:
-      debug_printf("\t\t.state = SVGA3D_RS_EMISSIVEMATERIALSOURCE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_EMISSIVEMATERIALSOURCE\n");
       break;
    case SVGA3D_RS_TEXTUREFACTOR:
-      debug_printf("\t\t.state = SVGA3D_RS_TEXTUREFACTOR\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_TEXTUREFACTOR\n");
       break;
    case SVGA3D_RS_LOCALVIEWER:
-      debug_printf("\t\t.state = SVGA3D_RS_LOCALVIEWER\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_LOCALVIEWER\n");
       break;
    case SVGA3D_RS_SCISSORTESTENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_SCISSORTESTENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_SCISSORTESTENABLE\n");
       break;
    case SVGA3D_RS_BLENDCOLOR:
-      debug_printf("\t\t.state = SVGA3D_RS_BLENDCOLOR\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_BLENDCOLOR\n");
       break;
    case SVGA3D_RS_STENCILENABLE2SIDED:
-      debug_printf("\t\t.state = SVGA3D_RS_STENCILENABLE2SIDED\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_STENCILENABLE2SIDED\n");
       break;
    case SVGA3D_RS_CCWSTENCILFUNC:
-      debug_printf("\t\t.state = SVGA3D_RS_CCWSTENCILFUNC\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_CCWSTENCILFUNC\n");
       break;
    case SVGA3D_RS_CCWSTENCILFAIL:
-      debug_printf("\t\t.state = SVGA3D_RS_CCWSTENCILFAIL\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_CCWSTENCILFAIL\n");
       break;
    case SVGA3D_RS_CCWSTENCILZFAIL:
-      debug_printf("\t\t.state = SVGA3D_RS_CCWSTENCILZFAIL\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_CCWSTENCILZFAIL\n");
       break;
    case SVGA3D_RS_CCWSTENCILPASS:
-      debug_printf("\t\t.state = SVGA3D_RS_CCWSTENCILPASS\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_CCWSTENCILPASS\n");
       break;
    case SVGA3D_RS_VERTEXBLEND:
-      debug_printf("\t\t.state = SVGA3D_RS_VERTEXBLEND\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_VERTEXBLEND\n");
       break;
    case SVGA3D_RS_SLOPESCALEDEPTHBIAS:
-      debug_printf("\t\t.state = SVGA3D_RS_SLOPESCALEDEPTHBIAS\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_SLOPESCALEDEPTHBIAS\n");
       break;
    case SVGA3D_RS_DEPTHBIAS:
-      debug_printf("\t\t.state = SVGA3D_RS_DEPTHBIAS\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_DEPTHBIAS\n");
       break;
    case SVGA3D_RS_OUTPUTGAMMA:
-      debug_printf("\t\t.state = SVGA3D_RS_OUTPUTGAMMA\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_OUTPUTGAMMA\n");
       break;
    case SVGA3D_RS_ZVISIBLE:
-      debug_printf("\t\t.state = SVGA3D_RS_ZVISIBLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_ZVISIBLE\n");
       break;
    case SVGA3D_RS_LASTPIXEL:
-      debug_printf("\t\t.state = SVGA3D_RS_LASTPIXEL\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_LASTPIXEL\n");
       break;
    case SVGA3D_RS_CLIPPING:
-      debug_printf("\t\t.state = SVGA3D_RS_CLIPPING\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_CLIPPING\n");
       break;
    case SVGA3D_RS_WRAP0:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP0\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP0\n");
       break;
    case SVGA3D_RS_WRAP1:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP1\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP1\n");
       break;
    case SVGA3D_RS_WRAP2:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP2\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP2\n");
       break;
    case SVGA3D_RS_WRAP3:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP3\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP3\n");
       break;
    case SVGA3D_RS_WRAP4:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP4\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP4\n");
       break;
    case SVGA3D_RS_WRAP5:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP5\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP5\n");
       break;
    case SVGA3D_RS_WRAP6:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP6\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP6\n");
       break;
    case SVGA3D_RS_WRAP7:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP7\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP7\n");
       break;
    case SVGA3D_RS_WRAP8:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP8\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP8\n");
       break;
    case SVGA3D_RS_WRAP9:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP9\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP9\n");
       break;
    case SVGA3D_RS_WRAP10:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP10\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP10\n");
       break;
    case SVGA3D_RS_WRAP11:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP11\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP11\n");
       break;
    case SVGA3D_RS_WRAP12:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP12\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP12\n");
       break;
    case SVGA3D_RS_WRAP13:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP13\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP13\n");
       break;
    case SVGA3D_RS_WRAP14:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP14\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP14\n");
       break;
    case SVGA3D_RS_WRAP15:
-      debug_printf("\t\t.state = SVGA3D_RS_WRAP15\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_WRAP15\n");
       break;
    case SVGA3D_RS_MULTISAMPLEANTIALIAS:
-      debug_printf("\t\t.state = SVGA3D_RS_MULTISAMPLEANTIALIAS\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_MULTISAMPLEANTIALIAS\n");
       break;
    case SVGA3D_RS_MULTISAMPLEMASK:
-      debug_printf("\t\t.state = SVGA3D_RS_MULTISAMPLEMASK\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_MULTISAMPLEMASK\n");
       break;
    case SVGA3D_RS_INDEXEDVERTEXBLENDENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_INDEXEDVERTEXBLENDENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_INDEXEDVERTEXBLENDENABLE\n");
       break;
    case SVGA3D_RS_TWEENFACTOR:
-      debug_printf("\t\t.state = SVGA3D_RS_TWEENFACTOR\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_TWEENFACTOR\n");
       break;
    case SVGA3D_RS_ANTIALIASEDLINEENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_ANTIALIASEDLINEENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_ANTIALIASEDLINEENABLE\n");
       break;
    case SVGA3D_RS_COLORWRITEENABLE1:
-      debug_printf("\t\t.state = SVGA3D_RS_COLORWRITEENABLE1\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_COLORWRITEENABLE1\n");
       break;
    case SVGA3D_RS_COLORWRITEENABLE2:
-      debug_printf("\t\t.state = SVGA3D_RS_COLORWRITEENABLE2\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_COLORWRITEENABLE2\n");
       break;
    case SVGA3D_RS_COLORWRITEENABLE3:
-      debug_printf("\t\t.state = SVGA3D_RS_COLORWRITEENABLE3\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_COLORWRITEENABLE3\n");
       break;
    case SVGA3D_RS_SEPARATEALPHABLENDENABLE:
-      debug_printf("\t\t.state = SVGA3D_RS_SEPARATEALPHABLENDENABLE\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_SEPARATEALPHABLENDENABLE\n");
       break;
    case SVGA3D_RS_SRCBLENDALPHA:
-      debug_printf("\t\t.state = SVGA3D_RS_SRCBLENDALPHA\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_SRCBLENDALPHA\n");
       break;
    case SVGA3D_RS_DSTBLENDALPHA:
-      debug_printf("\t\t.state = SVGA3D_RS_DSTBLENDALPHA\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_DSTBLENDALPHA\n");
       break;
    case SVGA3D_RS_BLENDEQUATIONALPHA:
-      debug_printf("\t\t.state = SVGA3D_RS_BLENDEQUATIONALPHA\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_BLENDEQUATIONALPHA\n");
       break;
    case SVGA3D_RS_MAX:
-      debug_printf("\t\t.state = SVGA3D_RS_MAX\n");
+      _debug_printf("\t\t.state = SVGA3D_RS_MAX\n");
       break;
    default:
-      debug_printf("\t\t.state = %i\n", (*cmd).state);
+      _debug_printf("\t\t.state = %i\n", (*cmd).state);
       break;
    }
-   debug_printf("\t\t.uintValue = %u\n", (*cmd).uintValue);
-   debug_printf("\t\t.floatValue = %f\n", (*cmd).floatValue);
+   _debug_printf("\t\t.uintValue = %u\n", (*cmd).uintValue);
+   _debug_printf("\t\t.floatValue = %f\n", (*cmd).floatValue);
 }
 
 static void
 dump_SVGA3dVertexDivisor(const SVGA3dVertexDivisor *cmd)
 {
-   debug_printf("\t\t.value = %u\n", (*cmd).value);
-   debug_printf("\t\t.count = %u\n", (*cmd).count);
-   debug_printf("\t\t.indexedData = %u\n", (*cmd).indexedData);
-   debug_printf("\t\t.instanceData = %u\n", (*cmd).instanceData);
+   _debug_printf("\t\t.value = %u\n", (*cmd).value);
+   _debug_printf("\t\t.count = %u\n", (*cmd).count);
+   _debug_printf("\t\t.indexedData = %u\n", (*cmd).indexedData);
+   _debug_printf("\t\t.instanceData = %u\n", (*cmd).instanceData);
 }
 
 static void
 dump_SVGA3dCmdDefineShader(const SVGA3dCmdDefineShader *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
-   debug_printf("\t\t.shid = %u\n", (*cmd).shid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.shid = %u\n", (*cmd).shid);
    switch((*cmd).type) {
    case SVGA3D_SHADERTYPE_COMPILED_DX8:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_COMPILED_DX8\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_COMPILED_DX8\n");
       break;
    case SVGA3D_SHADERTYPE_VS:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_VS\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_VS\n");
       break;
    case SVGA3D_SHADERTYPE_PS:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_PS\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_PS\n");
       break;
    case SVGA3D_SHADERTYPE_MAX:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_MAX\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.type = %i\n", (*cmd).type);
+      _debug_printf("\t\t.type = %i\n", (*cmd).type);
       break;
    }
 }
@@ -936,53 +936,53 @@ dump_SVGA3dCmdDefineShader(const SVGA3dCmdDefineShader *cmd)
 static void
 dump_SVGA3dCmdSetShaderConst(const SVGA3dCmdSetShaderConst *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
-   debug_printf("\t\t.reg = %u\n", (*cmd).reg);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.reg = %u\n", (*cmd).reg);
    switch((*cmd).type) {
    case SVGA3D_SHADERTYPE_COMPILED_DX8:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_COMPILED_DX8\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_COMPILED_DX8\n");
       break;
    case SVGA3D_SHADERTYPE_VS:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_VS\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_VS\n");
       break;
    case SVGA3D_SHADERTYPE_PS:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_PS\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_PS\n");
       break;
    case SVGA3D_SHADERTYPE_MAX:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_MAX\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.type = %i\n", (*cmd).type);
+      _debug_printf("\t\t.type = %i\n", (*cmd).type);
       break;
    }
    switch((*cmd).ctype) {
    case SVGA3D_CONST_TYPE_FLOAT:
-      debug_printf("\t\t.ctype = SVGA3D_CONST_TYPE_FLOAT\n");
-      debug_printf("\t\t.values[0] = %f\n", *(const float *)&(*cmd).values[0]);
-      debug_printf("\t\t.values[1] = %f\n", *(const float *)&(*cmd).values[1]);
-      debug_printf("\t\t.values[2] = %f\n", *(const float *)&(*cmd).values[2]);
-      debug_printf("\t\t.values[3] = %f\n", *(const float *)&(*cmd).values[3]);
+      _debug_printf("\t\t.ctype = SVGA3D_CONST_TYPE_FLOAT\n");
+      _debug_printf("\t\t.values[0] = %f\n", *(const float *)&(*cmd).values[0]);
+      _debug_printf("\t\t.values[1] = %f\n", *(const float *)&(*cmd).values[1]);
+      _debug_printf("\t\t.values[2] = %f\n", *(const float *)&(*cmd).values[2]);
+      _debug_printf("\t\t.values[3] = %f\n", *(const float *)&(*cmd).values[3]);
       break;
    case SVGA3D_CONST_TYPE_INT:
-      debug_printf("\t\t.ctype = SVGA3D_CONST_TYPE_INT\n");
-      debug_printf("\t\t.values[0] = %u\n", (*cmd).values[0]);
-      debug_printf("\t\t.values[1] = %u\n", (*cmd).values[1]);
-      debug_printf("\t\t.values[2] = %u\n", (*cmd).values[2]);
-      debug_printf("\t\t.values[3] = %u\n", (*cmd).values[3]);
+      _debug_printf("\t\t.ctype = SVGA3D_CONST_TYPE_INT\n");
+      _debug_printf("\t\t.values[0] = %u\n", (*cmd).values[0]);
+      _debug_printf("\t\t.values[1] = %u\n", (*cmd).values[1]);
+      _debug_printf("\t\t.values[2] = %u\n", (*cmd).values[2]);
+      _debug_printf("\t\t.values[3] = %u\n", (*cmd).values[3]);
       break;
    case SVGA3D_CONST_TYPE_BOOL:
-      debug_printf("\t\t.ctype = SVGA3D_CONST_TYPE_BOOL\n");
-      debug_printf("\t\t.values[0] = %u\n", (*cmd).values[0]);
-      debug_printf("\t\t.values[1] = %u\n", (*cmd).values[1]);
-      debug_printf("\t\t.values[2] = %u\n", (*cmd).values[2]);
-      debug_printf("\t\t.values[3] = %u\n", (*cmd).values[3]);
+      _debug_printf("\t\t.ctype = SVGA3D_CONST_TYPE_BOOL\n");
+      _debug_printf("\t\t.values[0] = %u\n", (*cmd).values[0]);
+      _debug_printf("\t\t.values[1] = %u\n", (*cmd).values[1]);
+      _debug_printf("\t\t.values[2] = %u\n", (*cmd).values[2]);
+      _debug_printf("\t\t.values[3] = %u\n", (*cmd).values[3]);
       break;
    default:
-      debug_printf("\t\t.ctype = %i\n", (*cmd).ctype);
-      debug_printf("\t\t.values[0] = %u\n", (*cmd).values[0]);
-      debug_printf("\t\t.values[1] = %u\n", (*cmd).values[1]);
-      debug_printf("\t\t.values[2] = %u\n", (*cmd).values[2]);
-      debug_printf("\t\t.values[3] = %u\n", (*cmd).values[3]);
+      _debug_printf("\t\t.ctype = %i\n", (*cmd).ctype);
+      _debug_printf("\t\t.values[0] = %u\n", (*cmd).values[0]);
+      _debug_printf("\t\t.values[1] = %u\n", (*cmd).values[1]);
+      _debug_printf("\t\t.values[2] = %u\n", (*cmd).values[2]);
+      _debug_printf("\t\t.values[3] = %u\n", (*cmd).values[3]);
       break;
    }
 }
@@ -990,25 +990,25 @@ dump_SVGA3dCmdSetShaderConst(const SVGA3dCmdSetShaderConst *cmd)
 static void
 dump_SVGA3dCmdSetZRange(const SVGA3dCmdSetZRange *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
-   debug_printf("\t\t.zRange.min = %f\n", (*cmd).zRange.min);
-   debug_printf("\t\t.zRange.max = %f\n", (*cmd).zRange.max);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.zRange.min = %f\n", (*cmd).zRange.min);
+   _debug_printf("\t\t.zRange.max = %f\n", (*cmd).zRange.max);
 }
 
 static void
 dump_SVGA3dCmdDrawPrimitives(const SVGA3dCmdDrawPrimitives *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
-   debug_printf("\t\t.numVertexDecls = %u\n", (*cmd).numVertexDecls);
-   debug_printf("\t\t.numRanges = %u\n", (*cmd).numRanges);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.numVertexDecls = %u\n", (*cmd).numVertexDecls);
+   _debug_printf("\t\t.numRanges = %u\n", (*cmd).numRanges);
 }
 
 static void
 dump_SVGA3dCmdSetLightEnabled(const SVGA3dCmdSetLightEnabled *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
-   debug_printf("\t\t.index = %u\n", (*cmd).index);
-   debug_printf("\t\t.enabled = %u\n", (*cmd).enabled);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.index = %u\n", (*cmd).index);
+   _debug_printf("\t\t.enabled = %u\n", (*cmd).enabled);
 }
 
 static void
@@ -1016,86 +1016,86 @@ dump_SVGA3dPrimitiveRange(const SVGA3dPrimitiveRange *cmd)
 {
    switch((*cmd).primType) {
    case SVGA3D_PRIMITIVE_INVALID:
-      debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_INVALID\n");
+      _debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_INVALID\n");
       break;
    case SVGA3D_PRIMITIVE_TRIANGLELIST:
-      debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_TRIANGLELIST\n");
+      _debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_TRIANGLELIST\n");
       break;
    case SVGA3D_PRIMITIVE_POINTLIST:
-      debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_POINTLIST\n");
+      _debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_POINTLIST\n");
       break;
    case SVGA3D_PRIMITIVE_LINELIST:
-      debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_LINELIST\n");
+      _debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_LINELIST\n");
       break;
    case SVGA3D_PRIMITIVE_LINESTRIP:
-      debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_LINESTRIP\n");
+      _debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_LINESTRIP\n");
       break;
    case SVGA3D_PRIMITIVE_TRIANGLESTRIP:
-      debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_TRIANGLESTRIP\n");
+      _debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_TRIANGLESTRIP\n");
       break;
    case SVGA3D_PRIMITIVE_TRIANGLEFAN:
-      debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_TRIANGLEFAN\n");
+      _debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_TRIANGLEFAN\n");
       break;
    case SVGA3D_PRIMITIVE_MAX:
-      debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_MAX\n");
+      _debug_printf("\t\t.primType = SVGA3D_PRIMITIVE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.primType = %i\n", (*cmd).primType);
+      _debug_printf("\t\t.primType = %i\n", (*cmd).primType);
       break;
    }
-   debug_printf("\t\t.primitiveCount = %u\n", (*cmd).primitiveCount);
-   debug_printf("\t\t.indexArray.surfaceId = %u\n", (*cmd).indexArray.surfaceId);
-   debug_printf("\t\t.indexArray.offset = %u\n", (*cmd).indexArray.offset);
-   debug_printf("\t\t.indexArray.stride = %u\n", (*cmd).indexArray.stride);
-   debug_printf("\t\t.indexWidth = %u\n", (*cmd).indexWidth);
-   debug_printf("\t\t.indexBias = %i\n", (*cmd).indexBias);
+   _debug_printf("\t\t.primitiveCount = %u\n", (*cmd).primitiveCount);
+   _debug_printf("\t\t.indexArray.surfaceId = %u\n", (*cmd).indexArray.surfaceId);
+   _debug_printf("\t\t.indexArray.offset = %u\n", (*cmd).indexArray.offset);
+   _debug_printf("\t\t.indexArray.stride = %u\n", (*cmd).indexArray.stride);
+   _debug_printf("\t\t.indexWidth = %u\n", (*cmd).indexWidth);
+   _debug_printf("\t\t.indexBias = %i\n", (*cmd).indexBias);
 }
 
 static void
 dump_SVGA3dCmdPresent(const SVGA3dCmdPresent *cmd)
 {
-   debug_printf("\t\t.sid = %u\n", (*cmd).sid);
+   _debug_printf("\t\t.sid = %u\n", (*cmd).sid);
 }
 
 static void
 dump_SVGA3dCmdSetRenderState(const SVGA3dCmdSetRenderState *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
 }
 
 static void
 dump_SVGA3dCmdSurfaceStretchBlt(const SVGA3dCmdSurfaceStretchBlt *cmd)
 {
-   debug_printf("\t\t.src.sid = %u\n", (*cmd).src.sid);
-   debug_printf("\t\t.src.face = %u\n", (*cmd).src.face);
-   debug_printf("\t\t.src.mipmap = %u\n", (*cmd).src.mipmap);
-   debug_printf("\t\t.dest.sid = %u\n", (*cmd).dest.sid);
-   debug_printf("\t\t.dest.face = %u\n", (*cmd).dest.face);
-   debug_printf("\t\t.dest.mipmap = %u\n", (*cmd).dest.mipmap);
-   debug_printf("\t\t.boxSrc.x = %u\n", (*cmd).boxSrc.x);
-   debug_printf("\t\t.boxSrc.y = %u\n", (*cmd).boxSrc.y);
-   debug_printf("\t\t.boxSrc.z = %u\n", (*cmd).boxSrc.z);
-   debug_printf("\t\t.boxSrc.w = %u\n", (*cmd).boxSrc.w);
-   debug_printf("\t\t.boxSrc.h = %u\n", (*cmd).boxSrc.h);
-   debug_printf("\t\t.boxSrc.d = %u\n", (*cmd).boxSrc.d);
-   debug_printf("\t\t.boxDest.x = %u\n", (*cmd).boxDest.x);
-   debug_printf("\t\t.boxDest.y = %u\n", (*cmd).boxDest.y);
-   debug_printf("\t\t.boxDest.z = %u\n", (*cmd).boxDest.z);
-   debug_printf("\t\t.boxDest.w = %u\n", (*cmd).boxDest.w);
-   debug_printf("\t\t.boxDest.h = %u\n", (*cmd).boxDest.h);
-   debug_printf("\t\t.boxDest.d = %u\n", (*cmd).boxDest.d);
+   _debug_printf("\t\t.src.sid = %u\n", (*cmd).src.sid);
+   _debug_printf("\t\t.src.face = %u\n", (*cmd).src.face);
+   _debug_printf("\t\t.src.mipmap = %u\n", (*cmd).src.mipmap);
+   _debug_printf("\t\t.dest.sid = %u\n", (*cmd).dest.sid);
+   _debug_printf("\t\t.dest.face = %u\n", (*cmd).dest.face);
+   _debug_printf("\t\t.dest.mipmap = %u\n", (*cmd).dest.mipmap);
+   _debug_printf("\t\t.boxSrc.x = %u\n", (*cmd).boxSrc.x);
+   _debug_printf("\t\t.boxSrc.y = %u\n", (*cmd).boxSrc.y);
+   _debug_printf("\t\t.boxSrc.z = %u\n", (*cmd).boxSrc.z);
+   _debug_printf("\t\t.boxSrc.w = %u\n", (*cmd).boxSrc.w);
+   _debug_printf("\t\t.boxSrc.h = %u\n", (*cmd).boxSrc.h);
+   _debug_printf("\t\t.boxSrc.d = %u\n", (*cmd).boxSrc.d);
+   _debug_printf("\t\t.boxDest.x = %u\n", (*cmd).boxDest.x);
+   _debug_printf("\t\t.boxDest.y = %u\n", (*cmd).boxDest.y);
+   _debug_printf("\t\t.boxDest.z = %u\n", (*cmd).boxDest.z);
+   _debug_printf("\t\t.boxDest.w = %u\n", (*cmd).boxDest.w);
+   _debug_printf("\t\t.boxDest.h = %u\n", (*cmd).boxDest.h);
+   _debug_printf("\t\t.boxDest.d = %u\n", (*cmd).boxDest.d);
    switch((*cmd).mode) {
    case SVGA3D_STRETCH_BLT_POINT:
-      debug_printf("\t\t.mode = SVGA3D_STRETCH_BLT_POINT\n");
+      _debug_printf("\t\t.mode = SVGA3D_STRETCH_BLT_POINT\n");
       break;
    case SVGA3D_STRETCH_BLT_LINEAR:
-      debug_printf("\t\t.mode = SVGA3D_STRETCH_BLT_LINEAR\n");
+      _debug_printf("\t\t.mode = SVGA3D_STRETCH_BLT_LINEAR\n");
       break;
    case SVGA3D_STRETCH_BLT_MAX:
-      debug_printf("\t\t.mode = SVGA3D_STRETCH_BLT_MAX\n");
+      _debug_printf("\t\t.mode = SVGA3D_STRETCH_BLT_MAX\n");
       break;
    default:
-      debug_printf("\t\t.mode = %i\n", (*cmd).mode);
+      _debug_printf("\t\t.mode = %i\n", (*cmd).mode);
       break;
    }
 }
@@ -1103,21 +1103,21 @@ dump_SVGA3dCmdSurfaceStretchBlt(const SVGA3dCmdSurfaceStretchBlt *cmd)
 static void
 dump_SVGA3dCmdSurfaceDMA(const SVGA3dCmdSurfaceDMA *cmd)
 {
-   debug_printf("\t\t.guest.ptr.gmrId = %u\n", (*cmd).guest.ptr.gmrId);
-   debug_printf("\t\t.guest.ptr.offset = %u\n", (*cmd).guest.ptr.offset);
-   debug_printf("\t\t.guest.pitch = %u\n", (*cmd).guest.pitch);
-   debug_printf("\t\t.host.sid = %u\n", (*cmd).host.sid);
-   debug_printf("\t\t.host.face = %u\n", (*cmd).host.face);
-   debug_printf("\t\t.host.mipmap = %u\n", (*cmd).host.mipmap);
+   _debug_printf("\t\t.guest.ptr.gmrId = %u\n", (*cmd).guest.ptr.gmrId);
+   _debug_printf("\t\t.guest.ptr.offset = %u\n", (*cmd).guest.ptr.offset);
+   _debug_printf("\t\t.guest.pitch = %u\n", (*cmd).guest.pitch);
+   _debug_printf("\t\t.host.sid = %u\n", (*cmd).host.sid);
+   _debug_printf("\t\t.host.face = %u\n", (*cmd).host.face);
+   _debug_printf("\t\t.host.mipmap = %u\n", (*cmd).host.mipmap);
    switch((*cmd).transfer) {
    case SVGA3D_WRITE_HOST_VRAM:
-      debug_printf("\t\t.transfer = SVGA3D_WRITE_HOST_VRAM\n");
+      _debug_printf("\t\t.transfer = SVGA3D_WRITE_HOST_VRAM\n");
       break;
    case SVGA3D_READ_HOST_VRAM:
-      debug_printf("\t\t.transfer = SVGA3D_READ_HOST_VRAM\n");
+      _debug_printf("\t\t.transfer = SVGA3D_READ_HOST_VRAM\n");
       break;
    default:
-      debug_printf("\t\t.transfer = %i\n", (*cmd).transfer);
+      _debug_printf("\t\t.transfer = %i\n", (*cmd).transfer);
       break;
    }
 }
@@ -1125,107 +1125,107 @@ dump_SVGA3dCmdSurfaceDMA(const SVGA3dCmdSurfaceDMA *cmd)
 static void
 dump_SVGA3dCmdSurfaceDMASuffix(const SVGA3dCmdSurfaceDMASuffix *cmd)
 {
-   debug_printf("\t\t.suffixSize = %u\n", (*cmd).suffixSize);
-   debug_printf("\t\t.maximumOffset = %u\n", (*cmd).maximumOffset);
-   debug_printf("\t\t.flags.discard = %u\n", (*cmd).flags.discard);
-   debug_printf("\t\t.flags.unsynchronized = %u\n", (*cmd).flags.unsynchronized);
+   _debug_printf("\t\t.suffixSize = %u\n", (*cmd).suffixSize);
+   _debug_printf("\t\t.maximumOffset = %u\n", (*cmd).maximumOffset);
+   _debug_printf("\t\t.flags.discard = %u\n", (*cmd).flags.discard);
+   _debug_printf("\t\t.flags.unsynchronized = %u\n", (*cmd).flags.unsynchronized);
 }
 
 static void
 dump_SVGA3dCmdSetTransform(const SVGA3dCmdSetTransform *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
    switch((*cmd).type) {
    case SVGA3D_TRANSFORM_INVALID:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_INVALID\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_INVALID\n");
       break;
    case SVGA3D_TRANSFORM_WORLD:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_WORLD\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_WORLD\n");
       break;
    case SVGA3D_TRANSFORM_VIEW:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_VIEW\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_VIEW\n");
       break;
    case SVGA3D_TRANSFORM_PROJECTION:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_PROJECTION\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_PROJECTION\n");
       break;
    case SVGA3D_TRANSFORM_TEXTURE0:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE0\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE0\n");
       break;
    case SVGA3D_TRANSFORM_TEXTURE1:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE1\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE1\n");
       break;
    case SVGA3D_TRANSFORM_TEXTURE2:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE2\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE2\n");
       break;
    case SVGA3D_TRANSFORM_TEXTURE3:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE3\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE3\n");
       break;
    case SVGA3D_TRANSFORM_TEXTURE4:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE4\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE4\n");
       break;
    case SVGA3D_TRANSFORM_TEXTURE5:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE5\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE5\n");
       break;
    case SVGA3D_TRANSFORM_TEXTURE6:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE6\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE6\n");
       break;
    case SVGA3D_TRANSFORM_TEXTURE7:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE7\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_TEXTURE7\n");
       break;
    case SVGA3D_TRANSFORM_WORLD1:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_WORLD1\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_WORLD1\n");
       break;
    case SVGA3D_TRANSFORM_WORLD2:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_WORLD2\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_WORLD2\n");
       break;
    case SVGA3D_TRANSFORM_WORLD3:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_WORLD3\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_WORLD3\n");
       break;
    case SVGA3D_TRANSFORM_MAX:
-      debug_printf("\t\t.type = SVGA3D_TRANSFORM_MAX\n");
+      _debug_printf("\t\t.type = SVGA3D_TRANSFORM_MAX\n");
       break;
    default:
-      debug_printf("\t\t.type = %i\n", (*cmd).type);
+      _debug_printf("\t\t.type = %i\n", (*cmd).type);
       break;
    }
-   debug_printf("\t\t.matrix[0] = %f\n", (*cmd).matrix[0]);
-   debug_printf("\t\t.matrix[1] = %f\n", (*cmd).matrix[1]);
-   debug_printf("\t\t.matrix[2] = %f\n", (*cmd).matrix[2]);
-   debug_printf("\t\t.matrix[3] = %f\n", (*cmd).matrix[3]);
-   debug_printf("\t\t.matrix[4] = %f\n", (*cmd).matrix[4]);
-   debug_printf("\t\t.matrix[5] = %f\n", (*cmd).matrix[5]);
-   debug_printf("\t\t.matrix[6] = %f\n", (*cmd).matrix[6]);
-   debug_printf("\t\t.matrix[7] = %f\n", (*cmd).matrix[7]);
-   debug_printf("\t\t.matrix[8] = %f\n", (*cmd).matrix[8]);
-   debug_printf("\t\t.matrix[9] = %f\n", (*cmd).matrix[9]);
-   debug_printf("\t\t.matrix[10] = %f\n", (*cmd).matrix[10]);
-   debug_printf("\t\t.matrix[11] = %f\n", (*cmd).matrix[11]);
-   debug_printf("\t\t.matrix[12] = %f\n", (*cmd).matrix[12]);
-   debug_printf("\t\t.matrix[13] = %f\n", (*cmd).matrix[13]);
-   debug_printf("\t\t.matrix[14] = %f\n", (*cmd).matrix[14]);
-   debug_printf("\t\t.matrix[15] = %f\n", (*cmd).matrix[15]);
+   _debug_printf("\t\t.matrix[0] = %f\n", (*cmd).matrix[0]);
+   _debug_printf("\t\t.matrix[1] = %f\n", (*cmd).matrix[1]);
+   _debug_printf("\t\t.matrix[2] = %f\n", (*cmd).matrix[2]);
+   _debug_printf("\t\t.matrix[3] = %f\n", (*cmd).matrix[3]);
+   _debug_printf("\t\t.matrix[4] = %f\n", (*cmd).matrix[4]);
+   _debug_printf("\t\t.matrix[5] = %f\n", (*cmd).matrix[5]);
+   _debug_printf("\t\t.matrix[6] = %f\n", (*cmd).matrix[6]);
+   _debug_printf("\t\t.matrix[7] = %f\n", (*cmd).matrix[7]);
+   _debug_printf("\t\t.matrix[8] = %f\n", (*cmd).matrix[8]);
+   _debug_printf("\t\t.matrix[9] = %f\n", (*cmd).matrix[9]);
+   _debug_printf("\t\t.matrix[10] = %f\n", (*cmd).matrix[10]);
+   _debug_printf("\t\t.matrix[11] = %f\n", (*cmd).matrix[11]);
+   _debug_printf("\t\t.matrix[12] = %f\n", (*cmd).matrix[12]);
+   _debug_printf("\t\t.matrix[13] = %f\n", (*cmd).matrix[13]);
+   _debug_printf("\t\t.matrix[14] = %f\n", (*cmd).matrix[14]);
+   _debug_printf("\t\t.matrix[15] = %f\n", (*cmd).matrix[15]);
 }
 
 static void
 dump_SVGA3dCmdDestroyShader(const SVGA3dCmdDestroyShader *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
-   debug_printf("\t\t.shid = %u\n", (*cmd).shid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.shid = %u\n", (*cmd).shid);
    switch((*cmd).type) {
    case SVGA3D_SHADERTYPE_COMPILED_DX8:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_COMPILED_DX8\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_COMPILED_DX8\n");
       break;
    case SVGA3D_SHADERTYPE_VS:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_VS\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_VS\n");
       break;
    case SVGA3D_SHADERTYPE_PS:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_PS\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_PS\n");
       break;
    case SVGA3D_SHADERTYPE_MAX:
-      debug_printf("\t\t.type = SVGA3D_SHADERTYPE_MAX\n");
+      _debug_printf("\t\t.type = SVGA3D_SHADERTYPE_MAX\n");
       break;
    default:
-      debug_printf("\t\t.type = %i\n", (*cmd).type);
+      _debug_printf("\t\t.type = %i\n", (*cmd).type);
       break;
    }
 }
@@ -1233,187 +1233,519 @@ dump_SVGA3dCmdDestroyShader(const SVGA3dCmdDestroyShader *cmd)
 static void
 dump_SVGA3dCmdDestroyContext(const SVGA3dCmdDestroyContext *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
 }
 
 static void
 dump_SVGA3dCmdClear(const SVGA3dCmdClear *cmd)
 {
-   debug_printf("\t\t.cid = %u\n", (*cmd).cid);
+   _debug_printf("\t\t.cid = %u\n", (*cmd).cid);
    switch((*cmd).clearFlag) {
    case SVGA3D_CLEAR_COLOR:
-      debug_printf("\t\t.clearFlag = SVGA3D_CLEAR_COLOR\n");
+      _debug_printf("\t\t.clearFlag = SVGA3D_CLEAR_COLOR\n");
       break;
    case SVGA3D_CLEAR_DEPTH:
-      debug_printf("\t\t.clearFlag = SVGA3D_CLEAR_DEPTH\n");
+      _debug_printf("\t\t.clearFlag = SVGA3D_CLEAR_DEPTH\n");
       break;
    case SVGA3D_CLEAR_STENCIL:
-      debug_printf("\t\t.clearFlag = SVGA3D_CLEAR_STENCIL\n");
+      _debug_printf("\t\t.clearFlag = SVGA3D_CLEAR_STENCIL\n");
       break;
    default:
-      debug_printf("\t\t.clearFlag = %i\n", (*cmd).clearFlag);
+      _debug_printf("\t\t.clearFlag = %i\n", (*cmd).clearFlag);
       break;
    }
-   debug_printf("\t\t.color = %u\n", (*cmd).color);
-   debug_printf("\t\t.depth = %f\n", (*cmd).depth);
-   debug_printf("\t\t.stencil = %u\n", (*cmd).stencil);
+   _debug_printf("\t\t.color = %u\n", (*cmd).color);
+   _debug_printf("\t\t.depth = %f\n", (*cmd).depth);
+   _debug_printf("\t\t.stencil = %u\n", (*cmd).stencil);
 }
 
 static void
 dump_SVGA3dCmdDefineSurface(const SVGA3dCmdDefineSurface *cmd)
 {
-   debug_printf("\t\t.sid = %u\n", (*cmd).sid);
+   _debug_printf("\t\t.sid = %u\n", (*cmd).sid);
    switch((*cmd).surfaceFlags) {
    case SVGA3D_SURFACE_CUBEMAP:
-      debug_printf("\t\t.surfaceFlags = SVGA3D_SURFACE_CUBEMAP\n");
+      _debug_printf("\t\t.surfaceFlags = SVGA3D_SURFACE_CUBEMAP\n");
       break;
    case SVGA3D_SURFACE_HINT_STATIC:
-      debug_printf("\t\t.surfaceFlags = SVGA3D_SURFACE_HINT_STATIC\n");
+      _debug_printf("\t\t.surfaceFlags = SVGA3D_SURFACE_HINT_STATIC\n");
       break;
    case SVGA3D_SURFACE_HINT_DYNAMIC:
-      debug_printf("\t\t.surfaceFlags = SVGA3D_SURFACE_HINT_DYNAMIC\n");
+      _debug_printf("\t\t.surfaceFlags = SVGA3D_SURFACE_HINT_DYNAMIC\n");
       break;
    case SVGA3D_SURFACE_HINT_INDEXBUFFER:
-      debug_printf("\t\t.surfaceFlags = SVGA3D_SURFACE_HINT_INDEXBUFFER\n");
+      _debug_printf("\t\t.surfaceFlags = SVGA3D_SURFACE_HINT_INDEXBUFFER\n");
       break;
    case SVGA3D_SURFACE_HINT_VERTEXBUFFER:
-      debug_printf("\t\t.surfaceFlags = SVGA3D_SURFACE_HINT_VERTEXBUFFER\n");
+      _debug_printf("\t\t.surfaceFlags = SVGA3D_SURFACE_HINT_VERTEXBUFFER\n");
       break;
    default:
-      debug_printf("\t\t.surfaceFlags = %i\n", (*cmd).surfaceFlags);
+      _debug_printf("\t\t.surfaceFlags = %i\n", (*cmd).surfaceFlags);
       break;
    }
    switch((*cmd).format) {
    case SVGA3D_FORMAT_INVALID:
-      debug_printf("\t\t.format = SVGA3D_FORMAT_INVALID\n");
+      _debug_printf("\t\t.format = SVGA3D_FORMAT_INVALID\n");
       break;
    case SVGA3D_X8R8G8B8:
-      debug_printf("\t\t.format = SVGA3D_X8R8G8B8\n");
+      _debug_printf("\t\t.format = SVGA3D_X8R8G8B8\n");
       break;
    case SVGA3D_A8R8G8B8:
-      debug_printf("\t\t.format = SVGA3D_A8R8G8B8\n");
+      _debug_printf("\t\t.format = SVGA3D_A8R8G8B8\n");
       break;
    case SVGA3D_R5G6B5:
-      debug_printf("\t\t.format = SVGA3D_R5G6B5\n");
+      _debug_printf("\t\t.format = SVGA3D_R5G6B5\n");
       break;
    case SVGA3D_X1R5G5B5:
-      debug_printf("\t\t.format = SVGA3D_X1R5G5B5\n");
+      _debug_printf("\t\t.format = SVGA3D_X1R5G5B5\n");
       break;
    case SVGA3D_A1R5G5B5:
-      debug_printf("\t\t.format = SVGA3D_A1R5G5B5\n");
+      _debug_printf("\t\t.format = SVGA3D_A1R5G5B5\n");
       break;
    case SVGA3D_A4R4G4B4:
-      debug_printf("\t\t.format = SVGA3D_A4R4G4B4\n");
+      _debug_printf("\t\t.format = SVGA3D_A4R4G4B4\n");
       break;
    case SVGA3D_Z_D32:
-      debug_printf("\t\t.format = SVGA3D_Z_D32\n");
+      _debug_printf("\t\t.format = SVGA3D_Z_D32\n");
       break;
    case SVGA3D_Z_D16:
-      debug_printf("\t\t.format = SVGA3D_Z_D16\n");
+      _debug_printf("\t\t.format = SVGA3D_Z_D16\n");
       break;
    case SVGA3D_Z_D24S8:
-      debug_printf("\t\t.format = SVGA3D_Z_D24S8\n");
+      _debug_printf("\t\t.format = SVGA3D_Z_D24S8\n");
       break;
    case SVGA3D_Z_D15S1:
-      debug_printf("\t\t.format = SVGA3D_Z_D15S1\n");
+      _debug_printf("\t\t.format = SVGA3D_Z_D15S1\n");
       break;
    case SVGA3D_LUMINANCE8:
-      debug_printf("\t\t.format = SVGA3D_LUMINANCE8\n");
+      _debug_printf("\t\t.format = SVGA3D_LUMINANCE8\n");
       break;
    case SVGA3D_LUMINANCE4_ALPHA4:
-      debug_printf("\t\t.format = SVGA3D_LUMINANCE4_ALPHA4\n");
+      _debug_printf("\t\t.format = SVGA3D_LUMINANCE4_ALPHA4\n");
       break;
    case SVGA3D_LUMINANCE16:
-      debug_printf("\t\t.format = SVGA3D_LUMINANCE16\n");
+      _debug_printf("\t\t.format = SVGA3D_LUMINANCE16\n");
       break;
    case SVGA3D_LUMINANCE8_ALPHA8:
-      debug_printf("\t\t.format = SVGA3D_LUMINANCE8_ALPHA8\n");
+      _debug_printf("\t\t.format = SVGA3D_LUMINANCE8_ALPHA8\n");
       break;
    case SVGA3D_DXT1:
-      debug_printf("\t\t.format = SVGA3D_DXT1\n");
+      _debug_printf("\t\t.format = SVGA3D_DXT1\n");
       break;
    case SVGA3D_DXT2:
-      debug_printf("\t\t.format = SVGA3D_DXT2\n");
+      _debug_printf("\t\t.format = SVGA3D_DXT2\n");
       break;
    case SVGA3D_DXT3:
-      debug_printf("\t\t.format = SVGA3D_DXT3\n");
+      _debug_printf("\t\t.format = SVGA3D_DXT3\n");
       break;
    case SVGA3D_DXT4:
-      debug_printf("\t\t.format = SVGA3D_DXT4\n");
+      _debug_printf("\t\t.format = SVGA3D_DXT4\n");
       break;
    case SVGA3D_DXT5:
-      debug_printf("\t\t.format = SVGA3D_DXT5\n");
+      _debug_printf("\t\t.format = SVGA3D_DXT5\n");
       break;
    case SVGA3D_BUMPU8V8:
-      debug_printf("\t\t.format = SVGA3D_BUMPU8V8\n");
+      _debug_printf("\t\t.format = SVGA3D_BUMPU8V8\n");
       break;
    case SVGA3D_BUMPL6V5U5:
-      debug_printf("\t\t.format = SVGA3D_BUMPL6V5U5\n");
+      _debug_printf("\t\t.format = SVGA3D_BUMPL6V5U5\n");
       break;
    case SVGA3D_BUMPX8L8V8U8:
-      debug_printf("\t\t.format = SVGA3D_BUMPX8L8V8U8\n");
+      _debug_printf("\t\t.format = SVGA3D_BUMPX8L8V8U8\n");
       break;
    case SVGA3D_BUMPL8V8U8:
-      debug_printf("\t\t.format = SVGA3D_BUMPL8V8U8\n");
+      _debug_printf("\t\t.format = SVGA3D_BUMPL8V8U8\n");
       break;
    case SVGA3D_ARGB_S10E5:
-      debug_printf("\t\t.format = SVGA3D_ARGB_S10E5\n");
+      _debug_printf("\t\t.format = SVGA3D_ARGB_S10E5\n");
       break;
    case SVGA3D_ARGB_S23E8:
-      debug_printf("\t\t.format = SVGA3D_ARGB_S23E8\n");
+      _debug_printf("\t\t.format = SVGA3D_ARGB_S23E8\n");
       break;
    case SVGA3D_A2R10G10B10:
-      debug_printf("\t\t.format = SVGA3D_A2R10G10B10\n");
+      _debug_printf("\t\t.format = SVGA3D_A2R10G10B10\n");
       break;
    case SVGA3D_V8U8:
-      debug_printf("\t\t.format = SVGA3D_V8U8\n");
+      _debug_printf("\t\t.format = SVGA3D_V8U8\n");
       break;
    case SVGA3D_Q8W8V8U8:
-      debug_printf("\t\t.format = SVGA3D_Q8W8V8U8\n");
+      _debug_printf("\t\t.format = SVGA3D_Q8W8V8U8\n");
       break;
    case SVGA3D_CxV8U8:
-      debug_printf("\t\t.format = SVGA3D_CxV8U8\n");
+      _debug_printf("\t\t.format = SVGA3D_CxV8U8\n");
       break;
    case SVGA3D_X8L8V8U8:
-      debug_printf("\t\t.format = SVGA3D_X8L8V8U8\n");
+      _debug_printf("\t\t.format = SVGA3D_X8L8V8U8\n");
       break;
    case SVGA3D_A2W10V10U10:
-      debug_printf("\t\t.format = SVGA3D_A2W10V10U10\n");
+      _debug_printf("\t\t.format = SVGA3D_A2W10V10U10\n");
       break;
    case SVGA3D_ALPHA8:
-      debug_printf("\t\t.format = SVGA3D_ALPHA8\n");
+      _debug_printf("\t\t.format = SVGA3D_ALPHA8\n");
       break;
    case SVGA3D_R_S10E5:
-      debug_printf("\t\t.format = SVGA3D_R_S10E5\n");
+      _debug_printf("\t\t.format = SVGA3D_R_S10E5\n");
       break;
    case SVGA3D_R_S23E8:
-      debug_printf("\t\t.format = SVGA3D_R_S23E8\n");
+      _debug_printf("\t\t.format = SVGA3D_R_S23E8\n");
       break;
    case SVGA3D_RG_S10E5:
-      debug_printf("\t\t.format = SVGA3D_RG_S10E5\n");
+      _debug_printf("\t\t.format = SVGA3D_RG_S10E5\n");
       break;
    case SVGA3D_RG_S23E8:
-      debug_printf("\t\t.format = SVGA3D_RG_S23E8\n");
+      _debug_printf("\t\t.format = SVGA3D_RG_S23E8\n");
       break;
    case SVGA3D_BUFFER:
-      debug_printf("\t\t.format = SVGA3D_BUFFER\n");
+      _debug_printf("\t\t.format = SVGA3D_BUFFER\n");
       break;
    case SVGA3D_Z_D24X8:
-      debug_printf("\t\t.format = SVGA3D_Z_D24X8\n");
+      _debug_printf("\t\t.format = SVGA3D_Z_D24X8\n");
       break;
    case SVGA3D_FORMAT_MAX:
-      debug_printf("\t\t.format = SVGA3D_FORMAT_MAX\n");
+      _debug_printf("\t\t.format = SVGA3D_FORMAT_MAX\n");
       break;
    default:
-      debug_printf("\t\t.format = %i\n", (*cmd).format);
+      _debug_printf("\t\t.format = %i\n", (*cmd).format);
       break;
    }
-   debug_printf("\t\t.face[0].numMipLevels = %u\n", (*cmd).face[0].numMipLevels);
-   debug_printf("\t\t.face[1].numMipLevels = %u\n", (*cmd).face[1].numMipLevels);
-   debug_printf("\t\t.face[2].numMipLevels = %u\n", (*cmd).face[2].numMipLevels);
-   debug_printf("\t\t.face[3].numMipLevels = %u\n", (*cmd).face[3].numMipLevels);
-   debug_printf("\t\t.face[4].numMipLevels = %u\n", (*cmd).face[4].numMipLevels);
-   debug_printf("\t\t.face[5].numMipLevels = %u\n", (*cmd).face[5].numMipLevels);
+   _debug_printf("\t\t.face[0].numMipLevels = %u\n", (*cmd).face[0].numMipLevels);
+   _debug_printf("\t\t.face[1].numMipLevels = %u\n", (*cmd).face[1].numMipLevels);
+   _debug_printf("\t\t.face[2].numMipLevels = %u\n", (*cmd).face[2].numMipLevels);
+   _debug_printf("\t\t.face[3].numMipLevels = %u\n", (*cmd).face[3].numMipLevels);
+   _debug_printf("\t\t.face[4].numMipLevels = %u\n", (*cmd).face[4].numMipLevels);
+   _debug_printf("\t\t.face[5].numMipLevels = %u\n", (*cmd).face[5].numMipLevels);
+}
+
+static void
+dump_SVGASignedRect(const SVGASignedRect *cmd)
+{
+   _debug_printf("\t\t.left = %i\n", (*cmd).left);
+   _debug_printf("\t\t.top = %i\n", (*cmd).top);
+   _debug_printf("\t\t.right = %i\n", (*cmd).right);
+   _debug_printf("\t\t.bottom = %i\n", (*cmd).bottom);
+}
+
+static void
+dump_SVGA3dCmdBlitSurfaceToScreen(const SVGA3dCmdBlitSurfaceToScreen *cmd)
+{
+   _debug_printf("\t\t.srcImage.sid = %u\n", (*cmd).srcImage.sid);
+   _debug_printf("\t\t.srcImage.face = %u\n", (*cmd).srcImage.face);
+   _debug_printf("\t\t.srcImage.mipmap = %u\n", (*cmd).srcImage.mipmap);
+   _debug_printf("\t\t.srcRect.left = %i\n", (*cmd).srcRect.left);
+   _debug_printf("\t\t.srcRect.top = %i\n", (*cmd).srcRect.top);
+   _debug_printf("\t\t.srcRect.right = %i\n", (*cmd).srcRect.right);
+   _debug_printf("\t\t.srcRect.bottom = %i\n", (*cmd).srcRect.bottom);
+   _debug_printf("\t\t.destScreenId = %u\n", (*cmd).destScreenId);
+   _debug_printf("\t\t.destRect.left = %i\n", (*cmd).destRect.left);
+   _debug_printf("\t\t.destRect.top = %i\n", (*cmd).destRect.top);
+   _debug_printf("\t\t.destRect.right = %i\n", (*cmd).destRect.right);
+   _debug_printf("\t\t.destRect.bottom = %i\n", (*cmd).destRect.bottom);
+}
+
+
+void            
+svga_dump_command(uint32_t cmd_id, const void *data, uint32_t size)
+{
+   const uint8_t *body = (const uint8_t *)data;
+   const uint8_t *next = body + size;
+  
+   switch(cmd_id) {
+   case SVGA_3D_CMD_SURFACE_DEFINE:
+      _debug_printf("\tSVGA_3D_CMD_SURFACE_DEFINE\n");
+      {
+         const SVGA3dCmdDefineSurface *cmd = (const SVGA3dCmdDefineSurface *)body;
+         dump_SVGA3dCmdDefineSurface(cmd);
+         body = (const uint8_t *)&cmd[1];
+         while(body + sizeof(SVGA3dSize) <= next) {
+            dump_SVGA3dSize((const SVGA3dSize *)body);
+            body += sizeof(SVGA3dSize);
+         }
+      }
+      break;
+   case SVGA_3D_CMD_SURFACE_DESTROY:
+      _debug_printf("\tSVGA_3D_CMD_SURFACE_DESTROY\n");
+      {
+         const SVGA3dCmdDestroySurface *cmd = (const SVGA3dCmdDestroySurface *)body;
+         dump_SVGA3dCmdDestroySurface(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SURFACE_COPY:
+      _debug_printf("\tSVGA_3D_CMD_SURFACE_COPY\n");
+      {
+         const SVGA3dCmdSurfaceCopy *cmd = (const SVGA3dCmdSurfaceCopy *)body;
+         dump_SVGA3dCmdSurfaceCopy(cmd);
+         body = (const uint8_t *)&cmd[1];
+         while(body + sizeof(SVGA3dCopyBox) <= next) {
+            dump_SVGA3dCopyBox((const SVGA3dCopyBox *)body);
+            body += sizeof(SVGA3dCopyBox);
+         }
+      }
+      break;
+   case SVGA_3D_CMD_SURFACE_STRETCHBLT:
+      _debug_printf("\tSVGA_3D_CMD_SURFACE_STRETCHBLT\n");
+      {
+         const SVGA3dCmdSurfaceStretchBlt *cmd = (const SVGA3dCmdSurfaceStretchBlt *)body;
+         dump_SVGA3dCmdSurfaceStretchBlt(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SURFACE_DMA:
+      _debug_printf("\tSVGA_3D_CMD_SURFACE_DMA\n");
+      {
+         const SVGA3dCmdSurfaceDMA *cmd = (const SVGA3dCmdSurfaceDMA *)body;
+         dump_SVGA3dCmdSurfaceDMA(cmd);
+         body = (const uint8_t *)&cmd[1];
+         while(body + sizeof(SVGA3dCopyBox) <= next) {
+            dump_SVGA3dCopyBox((const SVGA3dCopyBox *)body);
+            body += sizeof(SVGA3dCopyBox);
+         }
+         while(body + sizeof(SVGA3dCmdSurfaceDMASuffix) <= next) {
+            dump_SVGA3dCmdSurfaceDMASuffix((const SVGA3dCmdSurfaceDMASuffix *)body);
+            body += sizeof(SVGA3dCmdSurfaceDMASuffix);
+         }
+      }
+      break;
+   case SVGA_3D_CMD_CONTEXT_DEFINE:
+      _debug_printf("\tSVGA_3D_CMD_CONTEXT_DEFINE\n");
+      {
+         const SVGA3dCmdDefineContext *cmd = (const SVGA3dCmdDefineContext *)body;
+         dump_SVGA3dCmdDefineContext(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_CONTEXT_DESTROY:
+      _debug_printf("\tSVGA_3D_CMD_CONTEXT_DESTROY\n");
+      {
+         const SVGA3dCmdDestroyContext *cmd = (const SVGA3dCmdDestroyContext *)body;
+         dump_SVGA3dCmdDestroyContext(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SETTRANSFORM:
+      _debug_printf("\tSVGA_3D_CMD_SETTRANSFORM\n");
+      {
+         const SVGA3dCmdSetTransform *cmd = (const SVGA3dCmdSetTransform *)body;
+         dump_SVGA3dCmdSetTransform(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SETZRANGE:
+      _debug_printf("\tSVGA_3D_CMD_SETZRANGE\n");
+      {
+         const SVGA3dCmdSetZRange *cmd = (const SVGA3dCmdSetZRange *)body;
+         dump_SVGA3dCmdSetZRange(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SETRENDERSTATE:
+      _debug_printf("\tSVGA_3D_CMD_SETRENDERSTATE\n");
+      {
+         const SVGA3dCmdSetRenderState *cmd = (const SVGA3dCmdSetRenderState *)body;
+         dump_SVGA3dCmdSetRenderState(cmd);
+         body = (const uint8_t *)&cmd[1];
+         while(body + sizeof(SVGA3dRenderState) <= next) {
+            dump_SVGA3dRenderState((const SVGA3dRenderState *)body);
+            body += sizeof(SVGA3dRenderState);
+         }
+      }
+      break;
+   case SVGA_3D_CMD_SETRENDERTARGET:
+      _debug_printf("\tSVGA_3D_CMD_SETRENDERTARGET\n");
+      {
+         const SVGA3dCmdSetRenderTarget *cmd = (const SVGA3dCmdSetRenderTarget *)body;
+         dump_SVGA3dCmdSetRenderTarget(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SETTEXTURESTATE:
+      _debug_printf("\tSVGA_3D_CMD_SETTEXTURESTATE\n");
+      {
+         const SVGA3dCmdSetTextureState *cmd = (const SVGA3dCmdSetTextureState *)body;
+         dump_SVGA3dCmdSetTextureState(cmd);
+         body = (const uint8_t *)&cmd[1];
+         while(body + sizeof(SVGA3dTextureState) <= next) {
+            dump_SVGA3dTextureState((const SVGA3dTextureState *)body);
+            body += sizeof(SVGA3dTextureState);
+         }
+      }
+      break;
+   case SVGA_3D_CMD_SETMATERIAL:
+      _debug_printf("\tSVGA_3D_CMD_SETMATERIAL\n");
+      {
+         const SVGA3dCmdSetMaterial *cmd = (const SVGA3dCmdSetMaterial *)body;
+         dump_SVGA3dCmdSetMaterial(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SETLIGHTDATA:
+      _debug_printf("\tSVGA_3D_CMD_SETLIGHTDATA\n");
+      {
+         const SVGA3dCmdSetLightData *cmd = (const SVGA3dCmdSetLightData *)body;
+         dump_SVGA3dCmdSetLightData(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SETLIGHTENABLED:
+      _debug_printf("\tSVGA_3D_CMD_SETLIGHTENABLED\n");
+      {
+         const SVGA3dCmdSetLightEnabled *cmd = (const SVGA3dCmdSetLightEnabled *)body;
+         dump_SVGA3dCmdSetLightEnabled(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SETVIEWPORT:
+      _debug_printf("\tSVGA_3D_CMD_SETVIEWPORT\n");
+      {
+         const SVGA3dCmdSetViewport *cmd = (const SVGA3dCmdSetViewport *)body;
+         dump_SVGA3dCmdSetViewport(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SETCLIPPLANE:
+      _debug_printf("\tSVGA_3D_CMD_SETCLIPPLANE\n");
+      {
+         const SVGA3dCmdSetClipPlane *cmd = (const SVGA3dCmdSetClipPlane *)body;
+         dump_SVGA3dCmdSetClipPlane(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_CLEAR:
+      _debug_printf("\tSVGA_3D_CMD_CLEAR\n");
+      {
+         const SVGA3dCmdClear *cmd = (const SVGA3dCmdClear *)body;
+         dump_SVGA3dCmdClear(cmd);
+         body = (const uint8_t *)&cmd[1];
+         while(body + sizeof(SVGA3dRect) <= next) {
+            dump_SVGA3dRect((const SVGA3dRect *)body);
+            body += sizeof(SVGA3dRect);
+         }
+      }
+      break;
+   case SVGA_3D_CMD_PRESENT:
+      _debug_printf("\tSVGA_3D_CMD_PRESENT\n");
+      {
+         const SVGA3dCmdPresent *cmd = (const SVGA3dCmdPresent *)body;
+         dump_SVGA3dCmdPresent(cmd);
+         body = (const uint8_t *)&cmd[1];
+         while(body + sizeof(SVGA3dCopyRect) <= next) {
+            dump_SVGA3dCopyRect((const SVGA3dCopyRect *)body);
+            body += sizeof(SVGA3dCopyRect);
+         }
+      }
+      break;
+   case SVGA_3D_CMD_SHADER_DEFINE:
+      _debug_printf("\tSVGA_3D_CMD_SHADER_DEFINE\n");
+      {
+         const SVGA3dCmdDefineShader *cmd = (const SVGA3dCmdDefineShader *)body;
+         dump_SVGA3dCmdDefineShader(cmd);
+         body = (const uint8_t *)&cmd[1];
+         svga_shader_dump((const uint32_t *)body, 
+                      (unsigned)(next - body)/sizeof(uint32_t),
+                      FALSE );
+         body = next;
+      }
+      break;
+   case SVGA_3D_CMD_SHADER_DESTROY:
+      _debug_printf("\tSVGA_3D_CMD_SHADER_DESTROY\n");
+      {
+         const SVGA3dCmdDestroyShader *cmd = (const SVGA3dCmdDestroyShader *)body;
+         dump_SVGA3dCmdDestroyShader(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SET_SHADER:
+      _debug_printf("\tSVGA_3D_CMD_SET_SHADER\n");
+      {
+         const SVGA3dCmdSetShader *cmd = (const SVGA3dCmdSetShader *)body;
+         dump_SVGA3dCmdSetShader(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_SET_SHADER_CONST:
+      _debug_printf("\tSVGA_3D_CMD_SET_SHADER_CONST\n");
+      {
+         const SVGA3dCmdSetShaderConst *cmd = (const SVGA3dCmdSetShaderConst *)body;
+         dump_SVGA3dCmdSetShaderConst(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_DRAW_PRIMITIVES:
+      _debug_printf("\tSVGA_3D_CMD_DRAW_PRIMITIVES\n");
+      {
+         const SVGA3dCmdDrawPrimitives *cmd = (const SVGA3dCmdDrawPrimitives *)body;
+         unsigned i, j;
+         dump_SVGA3dCmdDrawPrimitives(cmd);
+         body = (const uint8_t *)&cmd[1];
+         for(i = 0; i < cmd->numVertexDecls; ++i) {
+            dump_SVGA3dVertexDecl((const SVGA3dVertexDecl *)body);
+            body += sizeof(SVGA3dVertexDecl);
+         }
+         for(j = 0; j < cmd->numRanges; ++j) {
+            dump_SVGA3dPrimitiveRange((const SVGA3dPrimitiveRange *)body);
+            body += sizeof(SVGA3dPrimitiveRange);
+         }
+         while(body + sizeof(SVGA3dVertexDivisor) <= next) {
+            dump_SVGA3dVertexDivisor((const SVGA3dVertexDivisor *)body);
+            body += sizeof(SVGA3dVertexDivisor);
+         }
+      }
+      break;
+   case SVGA_3D_CMD_SETSCISSORRECT:
+      _debug_printf("\tSVGA_3D_CMD_SETSCISSORRECT\n");
+      {
+         const SVGA3dCmdSetScissorRect *cmd = (const SVGA3dCmdSetScissorRect *)body;
+         dump_SVGA3dCmdSetScissorRect(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_BEGIN_QUERY:
+      _debug_printf("\tSVGA_3D_CMD_BEGIN_QUERY\n");
+      {
+         const SVGA3dCmdBeginQuery *cmd = (const SVGA3dCmdBeginQuery *)body;
+         dump_SVGA3dCmdBeginQuery(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_END_QUERY:
+      _debug_printf("\tSVGA_3D_CMD_END_QUERY\n");
+      {
+         const SVGA3dCmdEndQuery *cmd = (const SVGA3dCmdEndQuery *)body;
+         dump_SVGA3dCmdEndQuery(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_WAIT_FOR_QUERY:
+      _debug_printf("\tSVGA_3D_CMD_WAIT_FOR_QUERY\n");
+      {
+         const SVGA3dCmdWaitForQuery *cmd = (const SVGA3dCmdWaitForQuery *)body;
+         dump_SVGA3dCmdWaitForQuery(cmd);
+         body = (const uint8_t *)&cmd[1];
+      }
+      break;
+   case SVGA_3D_CMD_BLIT_SURFACE_TO_SCREEN:
+      _debug_printf("\tSVGA_3D_CMD_BLIT_SURFACE_TO_SCREEN\n");
+      {
+         const SVGA3dCmdBlitSurfaceToScreen *cmd = (const SVGA3dCmdBlitSurfaceToScreen *)body;
+         dump_SVGA3dCmdBlitSurfaceToScreen(cmd);
+         body = (const uint8_t *)&cmd[1];
+         while(body + sizeof(SVGASignedRect) <= next) {
+            dump_SVGASignedRect((const SVGASignedRect *)body);
+            body += sizeof(SVGASignedRect);
+         }
+      }
+      break;
+   default:
+      _debug_printf("\t0x%08x\n", cmd_id);
+      break;
+   }
+
+   while(body + sizeof(uint32_t) <= next) {
+      _debug_printf("\t\t0x%08x\n", *(const uint32_t *)body);
+      body += sizeof(uint32_t);
+   }
+   while(body + sizeof(uint32_t) <= next)
+      _debug_printf("\t\t0x%02x\n", *body++);
 }
 
 
@@ -1432,303 +1764,19 @@ svga_dump_commands(const void *commands, uint32_t size)
          const SVGA3dCmdHeader *header = (const SVGA3dCmdHeader *)next;
          const uint8_t *body = (const uint8_t *)&header[1];
 
-         next = (const uint8_t *)body + header->size;
+         next = body + header->size;
          if(next > last)
             break;
 
-         switch(cmd_id) {
-         case SVGA_3D_CMD_SURFACE_DEFINE:
-            debug_printf("\tSVGA_3D_CMD_SURFACE_DEFINE\n");
-            {
-               const SVGA3dCmdDefineSurface *cmd = (const SVGA3dCmdDefineSurface *)body;
-               dump_SVGA3dCmdDefineSurface(cmd);
-               body = (const uint8_t *)&cmd[1];
-               while(body + sizeof(SVGA3dSize) <= next) {
-                  dump_SVGA3dSize((const SVGA3dSize *)body);
-                  body += sizeof(SVGA3dSize);
-               }
-            }
-            break;
-         case SVGA_3D_CMD_SURFACE_DESTROY:
-            debug_printf("\tSVGA_3D_CMD_SURFACE_DESTROY\n");
-            {
-               const SVGA3dCmdDestroySurface *cmd = (const SVGA3dCmdDestroySurface *)body;
-               dump_SVGA3dCmdDestroySurface(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SURFACE_COPY:
-            debug_printf("\tSVGA_3D_CMD_SURFACE_COPY\n");
-            {
-               const SVGA3dCmdSurfaceCopy *cmd = (const SVGA3dCmdSurfaceCopy *)body;
-               dump_SVGA3dCmdSurfaceCopy(cmd);
-               body = (const uint8_t *)&cmd[1];
-               while(body + sizeof(SVGA3dCopyBox) <= next) {
-                  dump_SVGA3dCopyBox((const SVGA3dCopyBox *)body);
-                  body += sizeof(SVGA3dCopyBox);
-               }
-            }
-            break;
-         case SVGA_3D_CMD_SURFACE_STRETCHBLT:
-            debug_printf("\tSVGA_3D_CMD_SURFACE_STRETCHBLT\n");
-            {
-               const SVGA3dCmdSurfaceStretchBlt *cmd = (const SVGA3dCmdSurfaceStretchBlt *)body;
-               dump_SVGA3dCmdSurfaceStretchBlt(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SURFACE_DMA:
-            debug_printf("\tSVGA_3D_CMD_SURFACE_DMA\n");
-            {
-               const SVGA3dCmdSurfaceDMA *cmd = (const SVGA3dCmdSurfaceDMA *)body;
-               dump_SVGA3dCmdSurfaceDMA(cmd);
-               body = (const uint8_t *)&cmd[1];
-               while(body + sizeof(SVGA3dCopyBox) <= next) {
-                  dump_SVGA3dCopyBox((const SVGA3dCopyBox *)body);
-                  body += sizeof(SVGA3dCopyBox);
-               }
-               while(body + sizeof(SVGA3dCmdSurfaceDMASuffix) <= next) {
-                  dump_SVGA3dCmdSurfaceDMASuffix((const SVGA3dCmdSurfaceDMASuffix *)body);
-                  body += sizeof(SVGA3dCmdSurfaceDMASuffix);
-               }
-            }
-            break;
-         case SVGA_3D_CMD_CONTEXT_DEFINE:
-            debug_printf("\tSVGA_3D_CMD_CONTEXT_DEFINE\n");
-            {
-               const SVGA3dCmdDefineContext *cmd = (const SVGA3dCmdDefineContext *)body;
-               dump_SVGA3dCmdDefineContext(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_CONTEXT_DESTROY:
-            debug_printf("\tSVGA_3D_CMD_CONTEXT_DESTROY\n");
-            {
-               const SVGA3dCmdDestroyContext *cmd = (const SVGA3dCmdDestroyContext *)body;
-               dump_SVGA3dCmdDestroyContext(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SETTRANSFORM:
-            debug_printf("\tSVGA_3D_CMD_SETTRANSFORM\n");
-            {
-               const SVGA3dCmdSetTransform *cmd = (const SVGA3dCmdSetTransform *)body;
-               dump_SVGA3dCmdSetTransform(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SETZRANGE:
-            debug_printf("\tSVGA_3D_CMD_SETZRANGE\n");
-            {
-               const SVGA3dCmdSetZRange *cmd = (const SVGA3dCmdSetZRange *)body;
-               dump_SVGA3dCmdSetZRange(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SETRENDERSTATE:
-            debug_printf("\tSVGA_3D_CMD_SETRENDERSTATE\n");
-            {
-               const SVGA3dCmdSetRenderState *cmd = (const SVGA3dCmdSetRenderState *)body;
-               dump_SVGA3dCmdSetRenderState(cmd);
-               body = (const uint8_t *)&cmd[1];
-               while(body + sizeof(SVGA3dRenderState) <= next) {
-                  dump_SVGA3dRenderState((const SVGA3dRenderState *)body);
-                  body += sizeof(SVGA3dRenderState);
-               }
-            }
-            break;
-         case SVGA_3D_CMD_SETRENDERTARGET:
-            debug_printf("\tSVGA_3D_CMD_SETRENDERTARGET\n");
-            {
-               const SVGA3dCmdSetRenderTarget *cmd = (const SVGA3dCmdSetRenderTarget *)body;
-               dump_SVGA3dCmdSetRenderTarget(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SETTEXTURESTATE:
-            debug_printf("\tSVGA_3D_CMD_SETTEXTURESTATE\n");
-            {
-               const SVGA3dCmdSetTextureState *cmd = (const SVGA3dCmdSetTextureState *)body;
-               dump_SVGA3dCmdSetTextureState(cmd);
-               body = (const uint8_t *)&cmd[1];
-               while(body + sizeof(SVGA3dTextureState) <= next) {
-                  dump_SVGA3dTextureState((const SVGA3dTextureState *)body);
-                  body += sizeof(SVGA3dTextureState);
-               }
-            }
-            break;
-         case SVGA_3D_CMD_SETMATERIAL:
-            debug_printf("\tSVGA_3D_CMD_SETMATERIAL\n");
-            {
-               const SVGA3dCmdSetMaterial *cmd = (const SVGA3dCmdSetMaterial *)body;
-               dump_SVGA3dCmdSetMaterial(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SETLIGHTDATA:
-            debug_printf("\tSVGA_3D_CMD_SETLIGHTDATA\n");
-            {
-               const SVGA3dCmdSetLightData *cmd = (const SVGA3dCmdSetLightData *)body;
-               dump_SVGA3dCmdSetLightData(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SETLIGHTENABLED:
-            debug_printf("\tSVGA_3D_CMD_SETLIGHTENABLED\n");
-            {
-               const SVGA3dCmdSetLightEnabled *cmd = (const SVGA3dCmdSetLightEnabled *)body;
-               dump_SVGA3dCmdSetLightEnabled(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SETVIEWPORT:
-            debug_printf("\tSVGA_3D_CMD_SETVIEWPORT\n");
-            {
-               const SVGA3dCmdSetViewport *cmd = (const SVGA3dCmdSetViewport *)body;
-               dump_SVGA3dCmdSetViewport(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SETCLIPPLANE:
-            debug_printf("\tSVGA_3D_CMD_SETCLIPPLANE\n");
-            {
-               const SVGA3dCmdSetClipPlane *cmd = (const SVGA3dCmdSetClipPlane *)body;
-               dump_SVGA3dCmdSetClipPlane(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_CLEAR:
-            debug_printf("\tSVGA_3D_CMD_CLEAR\n");
-            {
-               const SVGA3dCmdClear *cmd = (const SVGA3dCmdClear *)body;
-               dump_SVGA3dCmdClear(cmd);
-               body = (const uint8_t *)&cmd[1];
-               while(body + sizeof(SVGA3dRect) <= next) {
-                  dump_SVGA3dRect((const SVGA3dRect *)body);
-                  body += sizeof(SVGA3dRect);
-               }
-            }
-            break;
-         case SVGA_3D_CMD_PRESENT:
-            debug_printf("\tSVGA_3D_CMD_PRESENT\n");
-            {
-               const SVGA3dCmdPresent *cmd = (const SVGA3dCmdPresent *)body;
-               dump_SVGA3dCmdPresent(cmd);
-               body = (const uint8_t *)&cmd[1];
-               while(body + sizeof(SVGA3dCopyRect) <= next) {
-                  dump_SVGA3dCopyRect((const SVGA3dCopyRect *)body);
-                  body += sizeof(SVGA3dCopyRect);
-               }
-            }
-            break;
-         case SVGA_3D_CMD_SHADER_DEFINE:
-            debug_printf("\tSVGA_3D_CMD_SHADER_DEFINE\n");
-            {
-               const SVGA3dCmdDefineShader *cmd = (const SVGA3dCmdDefineShader *)body;
-               dump_SVGA3dCmdDefineShader(cmd);
-               body = (const uint8_t *)&cmd[1];
-               svga_shader_dump((const uint32_t *)body, 
-                            (unsigned)(next - body)/sizeof(uint32_t),
-                            FALSE );
-               body = next;
-            }
-            break;
-         case SVGA_3D_CMD_SHADER_DESTROY:
-            debug_printf("\tSVGA_3D_CMD_SHADER_DESTROY\n");
-            {
-               const SVGA3dCmdDestroyShader *cmd = (const SVGA3dCmdDestroyShader *)body;
-               dump_SVGA3dCmdDestroyShader(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SET_SHADER:
-            debug_printf("\tSVGA_3D_CMD_SET_SHADER\n");
-            {
-               const SVGA3dCmdSetShader *cmd = (const SVGA3dCmdSetShader *)body;
-               dump_SVGA3dCmdSetShader(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_SET_SHADER_CONST:
-            debug_printf("\tSVGA_3D_CMD_SET_SHADER_CONST\n");
-            {
-               const SVGA3dCmdSetShaderConst *cmd = (const SVGA3dCmdSetShaderConst *)body;
-               dump_SVGA3dCmdSetShaderConst(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_DRAW_PRIMITIVES:
-            debug_printf("\tSVGA_3D_CMD_DRAW_PRIMITIVES\n");
-            {
-               const SVGA3dCmdDrawPrimitives *cmd = (const SVGA3dCmdDrawPrimitives *)body;
-               unsigned i, j;
-               dump_SVGA3dCmdDrawPrimitives(cmd);
-               body = (const uint8_t *)&cmd[1];
-               for(i = 0; i < cmd->numVertexDecls; ++i) {
-                  dump_SVGA3dVertexDecl((const SVGA3dVertexDecl *)body);
-                  body += sizeof(SVGA3dVertexDecl);
-               }
-               for(j = 0; j < cmd->numRanges; ++j) {
-                  dump_SVGA3dPrimitiveRange((const SVGA3dPrimitiveRange *)body);
-                  body += sizeof(SVGA3dPrimitiveRange);
-               }
-               while(body + sizeof(SVGA3dVertexDivisor) <= next) {
-                  dump_SVGA3dVertexDivisor((const SVGA3dVertexDivisor *)body);
-                  body += sizeof(SVGA3dVertexDivisor);
-               }
-            }
-            break;
-         case SVGA_3D_CMD_SETSCISSORRECT:
-            debug_printf("\tSVGA_3D_CMD_SETSCISSORRECT\n");
-            {
-               const SVGA3dCmdSetScissorRect *cmd = (const SVGA3dCmdSetScissorRect *)body;
-               dump_SVGA3dCmdSetScissorRect(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_BEGIN_QUERY:
-            debug_printf("\tSVGA_3D_CMD_BEGIN_QUERY\n");
-            {
-               const SVGA3dCmdBeginQuery *cmd = (const SVGA3dCmdBeginQuery *)body;
-               dump_SVGA3dCmdBeginQuery(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_END_QUERY:
-            debug_printf("\tSVGA_3D_CMD_END_QUERY\n");
-            {
-               const SVGA3dCmdEndQuery *cmd = (const SVGA3dCmdEndQuery *)body;
-               dump_SVGA3dCmdEndQuery(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         case SVGA_3D_CMD_WAIT_FOR_QUERY:
-            debug_printf("\tSVGA_3D_CMD_WAIT_FOR_QUERY\n");
-            {
-               const SVGA3dCmdWaitForQuery *cmd = (const SVGA3dCmdWaitForQuery *)body;
-               dump_SVGA3dCmdWaitForQuery(cmd);
-               body = (const uint8_t *)&cmd[1];
-            }
-            break;
-         default:
-            debug_printf("\t0x%08x\n", cmd_id);
-            break;
-         }
-
-         while(body + sizeof(uint32_t) <= next) {
-            debug_printf("\t\t0x%08x\n", *(const uint32_t *)body);
-            body += sizeof(uint32_t);
-         }
-         while(body + sizeof(uint32_t) <= next)
-            debug_printf("\t\t0x%02x\n", *body++);
+         svga_dump_command(cmd_id, body, header->size);
       }
       else if(cmd_id == SVGA_CMD_FENCE) {
-         debug_printf("\tSVGA_CMD_FENCE\n");
-         debug_printf("\t\t0x%08x\n", ((const uint32_t *)next)[1]);
+         _debug_printf("\tSVGA_CMD_FENCE\n");
+         _debug_printf("\t\t0x%08x\n", ((const uint32_t *)next)[1]);
          next += 2*sizeof(uint32_t);
       }
       else {
-         debug_printf("\t0x%08x\n", cmd_id);
+         _debug_printf("\t0x%08x\n", cmd_id);
          next += sizeof(uint32_t);
       }
    }
index 69a870208753358129d93fb444b7222d88066f93..ca0154361ccc82bf0816c2821d1303866c3ebf68 100644 (file)
@@ -28,6 +28,9 @@
 
 #include "pipe/p_compiler.h"
 
+void            
+svga_dump_command(uint32_t cmd_id, const void *data, uint32_t size);
+
 void
 svga_dump_commands(const void *commands, uint32_t size);
 
index 288e753296eb5df9a24409184da56e9909152543..0bc0b3ae3179342b64bbee66427bfd7e8ad73ad5 100755 (executable)
@@ -71,14 +71,14 @@ class decl_dumper_t(decl_visitor.decl_visitor_t):
             print '   switch(%s) {' % ("(*cmd)" + self._instance,)
             for name, value in self.decl.values:
                 print '   case %s:' % (name,)
-                print '      debug_printf("\\t\\t%s = %s\\n");' % (self._instance, name)
+                print '      _debug_printf("\\t\\t%s = %s\\n");' % (self._instance, name)
                 print '      break;'
             print '   default:'
-            print '      debug_printf("\\t\\t%s = %%i\\n", %s);' % (self._instance, "(*cmd)" + self._instance)
+            print '      _debug_printf("\\t\\t%s = %%i\\n", %s);' % (self._instance, "(*cmd)" + self._instance)
             print '      break;'
             print '   }'
         else:
-            print '   debug_printf("\\t\\t%s = %%i\\n", %s);' % (self._instance, "(*cmd)" + self._instance)
+            print '   _debug_printf("\\t\\t%s = %%i\\n", %s);' % (self._instance, "(*cmd)" + self._instance)
 
 
 def dump_decl(instance, decl):
@@ -154,7 +154,7 @@ class type_dumper_t(type_visitor.type_visitor_t):
         dump_decl(self.instance, decl)
 
     def print_instance(self, format):
-        print '   debug_printf("\\t\\t%s = %s\\n", %s);' % (self.instance, format, "(*cmd)" + self.instance)
+        print '   _debug_printf("\\t\\t%s = %s\\n", %s);' % (self.instance, format, "(*cmd)" + self.instance)
 
 
 def dump_type(instance, type_):
@@ -202,11 +202,62 @@ cmds = [
     ('SVGA_3D_CMD_END_QUERY', 'SVGA3dCmdEndQuery', (), None),
     ('SVGA_3D_CMD_WAIT_FOR_QUERY', 'SVGA3dCmdWaitForQuery', (), None),
     #('SVGA_3D_CMD_PRESENT_READBACK', None, (), None),
+    ('SVGA_3D_CMD_BLIT_SURFACE_TO_SCREEN', 'SVGA3dCmdBlitSurfaceToScreen', (), 'SVGASignedRect'),
 ]
 
 def dump_cmds():
     print r'''
 void            
+svga_dump_command(uint32_t cmd_id, const void *data, uint32_t size)
+{
+   const uint8_t *body = (const uint8_t *)data;
+   const uint8_t *next = body + size;
+'''
+    print '   switch(cmd_id) {'
+    indexes = 'ijklmn'
+    for id, header, body, footer in cmds:
+        print '   case %s:' % id
+        print '      _debug_printf("\\t%s\\n");' % id
+        print '      {'
+        print '         const %s *cmd = (const %s *)body;' % (header, header)
+        if len(body):
+            print '         unsigned ' + ', '.join(indexes[:len(body)]) + ';'
+        print '         dump_%s(cmd);' % header
+        print '         body = (const uint8_t *)&cmd[1];'
+        for i in range(len(body)):
+            struct, count = body[i]
+            idx = indexes[i]
+            print '         for(%s = 0; %s < cmd->%s; ++%s) {' % (idx, idx, count, idx)
+            print '            dump_%s((const %s *)body);' % (struct, struct)
+            print '            body += sizeof(%s);' % struct
+            print '         }'
+        if footer is not None:
+            print '         while(body + sizeof(%s) <= next) {' % footer
+            print '            dump_%s((const %s *)body);' % (footer, footer)
+            print '            body += sizeof(%s);' % footer
+            print '         }'
+        if id == 'SVGA_3D_CMD_SHADER_DEFINE':
+            print '         svga_shader_dump((const uint32_t *)body,'
+            print '                          (unsigned)(next - body)/sizeof(uint32_t),'
+            print '                          FALSE);'
+            print '         body = next;'
+        print '      }'
+        print '      break;'
+    print '   default:'
+    print '      _debug_printf("\\t0x%08x\\n", cmd_id);'
+    print '      break;'
+    print '   }'
+    print r'''
+   while(body + sizeof(uint32_t) <= next) {
+      _debug_printf("\t\t0x%08x\n", *(const uint32_t *)body);
+      body += sizeof(uint32_t);
+   }
+   while(body + sizeof(uint32_t) <= next)
+      _debug_printf("\t\t0x%02x\n", *body++);
+}
+'''
+    print r'''
+void            
 svga_dump_commands(const void *commands, uint32_t size)
 {
    const uint8_t *next = commands;
@@ -221,59 +272,19 @@ svga_dump_commands(const void *commands, uint32_t size)
          const SVGA3dCmdHeader *header = (const SVGA3dCmdHeader *)next;
          const uint8_t *body = (const uint8_t *)&header[1];
 
-         next = (const uint8_t *)body + header->size;
+         next = body + header->size;
          if(next > last)
             break;
-'''
 
-    print '         switch(cmd_id) {'
-    indexes = 'ijklmn'
-    for id, header, body, footer in cmds:
-        print '         case %s:' % id
-        print '            debug_printf("\\t%s\\n");' % id
-        print '            {'
-        print '               const %s *cmd = (const %s *)body;' % (header, header)
-        if len(body):
-            print '               unsigned ' + ', '.join(indexes[:len(body)]) + ';'
-        print '               dump_%s(cmd);' % header
-        print '               body = (const uint8_t *)&cmd[1];'
-        for i in range(len(body)):
-            struct, count = body[i]
-            idx = indexes[i]
-            print '               for(%s = 0; %s < cmd->%s; ++%s) {' % (idx, idx, count, idx)
-            print '                  dump_%s((const %s *)body);' % (struct, struct)
-            print '                  body += sizeof(%s);' % struct
-            print '               }'
-        if footer is not None:
-            print '               while(body + sizeof(%s) <= next) {' % footer
-            print '                  dump_%s((const %s *)body);' % (footer, footer)
-            print '                  body += sizeof(%s);' % footer
-            print '               }'
-        if id == 'SVGA_3D_CMD_SHADER_DEFINE':
-            print '               sh_svga_dump((const uint32_t *)body, (unsigned)(next - body)/sizeof(uint32_t));'
-            print '               body = next;'
-        print '            }'
-        print '            break;'
-    print '         default:'
-    print '            debug_printf("\\t0x%08x\\n", cmd_id);'
-    print '            break;'
-    print '         }'
-            
-    print r'''
-         while(body + sizeof(uint32_t) <= next) {
-            debug_printf("\t\t0x%08x\n", *(const uint32_t *)body);
-            body += sizeof(uint32_t);
-         }
-         while(body + sizeof(uint32_t) <= next)
-            debug_printf("\t\t0x%02x\n", *body++);
+         svga_dump_command(cmd_id, body, header->size);
       }
       else if(cmd_id == SVGA_CMD_FENCE) {
-         debug_printf("\tSVGA_CMD_FENCE\n");
-         debug_printf("\t\t0x%08x\n", ((const uint32_t *)next)[1]);
+         _debug_printf("\tSVGA_CMD_FENCE\n");
+         _debug_printf("\t\t0x%08x\n", ((const uint32_t *)next)[1]);
          next += 2*sizeof(uint32_t);
       }
       else {
-         debug_printf("\t0x%08x\n", cmd_id);
+         _debug_printf("\t0x%08x\n", cmd_id);
          next += sizeof(uint32_t);
       }
    }
@@ -294,18 +305,18 @@ def main():
     print '#include "svga_shader_dump.h"'
     print '#include "svga3d_reg.h"'
     print
-    print '#include "pipe/p_debug.h"'
+    print '#include "util/u_debug.h"'
     print '#include "svga_dump.h"'
     print
 
     config = parser.config_t(
-        include_paths = ['include'],
+        include_paths = ['../../../include', '../include'],
         compiler = 'gcc',
     )
 
     headers = [
-        'include/svga_types.h', 
-        'include/svga3d_reg.h', 
+        'svga_types.h', 
+        'svga3d_reg.h', 
     ]
 
     decls = parser.parse(headers, config, parser.COMPILATION_MODE.ALL_AT_ONCE)
index b0e7fdf378a03ca45e03950106435c7b7615c467..70e27d86d306df1c492d9cfe96e263ebf00b3698 100644 (file)
@@ -50,16 +50,16 @@ static void dump_op( struct sh_op op, const char *mnemonic )
    assert( op.is_reg == 0 );
 
    if (op.coissue)
-      debug_printf( "+" );
-   debug_printf( "%s", mnemonic );
+      _debug_printf( "+" );
+   _debug_printf( "%s", mnemonic );
    switch (op.control) {
    case 0:
       break;
    case SVGA3DOPCONT_PROJECT:
-      debug_printf( "p" );
+      _debug_printf( "p" );
       break;
    case SVGA3DOPCONT_BIAS:
-      debug_printf( "b" );
+      _debug_printf( "b" );
       break;
    default:
       assert( 0 );
@@ -72,28 +72,28 @@ static void dump_comp_op( struct sh_op op, const char *mnemonic )
    assert( op.is_reg == 0 );
 
    if (op.coissue)
-      debug_printf( "+" );
-   debug_printf( "%s", mnemonic );
+      _debug_printf( "+" );
+   _debug_printf( "%s", mnemonic );
    switch (op.control) {
    case SVGA3DOPCOMP_RESERVED0:
       break;
    case SVGA3DOPCOMP_GT:
-      debug_printf("_gt");
+      _debug_printf("_gt");
       break;
    case SVGA3DOPCOMP_EQ:
-      debug_printf("_eq");
+      _debug_printf("_eq");
       break;
    case SVGA3DOPCOMP_GE:
-      debug_printf("_ge");
+      _debug_printf("_ge");
       break;
    case SVGA3DOPCOMP_LT:
-      debug_printf("_lt");
+      _debug_printf("_lt");
       break;
    case SVGA3DOPCOMPC_NE:
-      debug_printf("_ne");
+      _debug_printf("_ne");
       break;
    case SVGA3DOPCOMP_LE:
-      debug_printf("_le");
+      _debug_printf("_le");
       break;
    case SVGA3DOPCOMP_RESERVED1:
    default:
@@ -109,93 +109,93 @@ static void dump_reg( struct sh_reg reg, struct sh_srcreg *indreg, const struct
 
    switch (sh_reg_type( reg )) {
    case SVGA3DREG_TEMP:
-      debug_printf( "r%u", reg.number );
+      _debug_printf( "r%u", reg.number );
       break;
 
    case SVGA3DREG_INPUT:
-      debug_printf( "v%u", reg.number );
+      _debug_printf( "v%u", reg.number );
       break;
 
    case SVGA3DREG_CONST:
       if (reg.relative) {
          if (sh_srcreg_type( *indreg ) == SVGA3DREG_LOOP)
-            debug_printf( "c[aL+%u]", reg.number );
+            _debug_printf( "c[aL+%u]", reg.number );
          else
-            debug_printf( "c[a%u.x+%u]", indreg->number, reg.number );
+            _debug_printf( "c[a%u.x+%u]", indreg->number, reg.number );
       }
       else
-         debug_printf( "c%u", reg.number );
+         _debug_printf( "c%u", reg.number );
       break;
 
    case SVGA3DREG_ADDR:    /* VS */
    /* SVGA3DREG_TEXTURE */ /* PS */
       if (di->is_ps)
-         debug_printf( "t%u", reg.number );
+         _debug_printf( "t%u", reg.number );
       else
-         debug_printf( "a%u", reg.number );
+         _debug_printf( "a%u", reg.number );
       break;
 
    case SVGA3DREG_RASTOUT:
       switch (reg.number) {
       case 0 /*POSITION*/:
-         debug_printf( "oPos" );
+         _debug_printf( "oPos" );
          break;
       case 1 /*FOG*/:
-         debug_printf( "oFog" );
+         _debug_printf( "oFog" );
          break;
       case 2 /*POINT_SIZE*/:
-         debug_printf( "oPts" );
+         _debug_printf( "oPts" );
          break;
       default:
          assert( 0 );
-         debug_printf( "???" );
+         _debug_printf( "???" );
       }
       break;
 
    case SVGA3DREG_ATTROUT:
       assert( reg.number < 2 );
-      debug_printf( "oD%u", reg.number );
+      _debug_printf( "oD%u", reg.number );
       break;
 
    case SVGA3DREG_TEXCRDOUT:
    /* SVGA3DREG_OUTPUT */
-      debug_printf( "oT%u", reg.number );
+      _debug_printf( "oT%u", reg.number );
       break;
 
    case SVGA3DREG_COLOROUT:
-      debug_printf( "oC%u", reg.number );
+      _debug_printf( "oC%u", reg.number );
       break;
 
    case SVGA3DREG_DEPTHOUT:
-      debug_printf( "oD%u", reg.number );
+      _debug_printf( "oD%u", reg.number );
       break;
 
    case SVGA3DREG_SAMPLER:
-      debug_printf( "s%u", reg.number );
+      _debug_printf( "s%u", reg.number );
       break;
 
    case SVGA3DREG_CONSTBOOL:
       assert( !reg.relative );
-      debug_printf( "b%u", reg.number );
+      _debug_printf( "b%u", reg.number );
       break;
 
    case SVGA3DREG_CONSTINT:
       assert( !reg.relative );
-      debug_printf( "i%u", reg.number );
+      _debug_printf( "i%u", reg.number );
       break;
 
    case SVGA3DREG_LOOP:
       assert( reg.number == 0 );
-      debug_printf( "aL" );
+      _debug_printf( "aL" );
       break;
 
    case SVGA3DREG_MISCTYPE:
       switch (reg.number) {
       case SVGA3DMISCREG_POSITION:
-         debug_printf( "vPos" );
+         _debug_printf( "vPos" );
          break;
       case SVGA3DMISCREG_FACE:
-         debug_printf( "vFace" );
+         _debug_printf( "vFace" );
          break;
       default:
          assert(0);
@@ -204,46 +204,46 @@ static void dump_reg( struct sh_reg reg, struct sh_srcreg *indreg, const struct
       break;
 
    case SVGA3DREG_LABEL:
-      debug_printf( "l%u", reg.number );
+      _debug_printf( "l%u", reg.number );
       break;
 
    case SVGA3DREG_PREDICATE:
-      debug_printf( "p%u", reg.number );
+      _debug_printf( "p%u", reg.number );
       break;
 
 
    default:
       assert( 0 );
-      debug_printf( "???" );
+      _debug_printf( "???" );
    }
 }
 
 static void dump_cdata( struct sh_cdata cdata )
 {
-   debug_printf( "%f, %f, %f, %f", cdata.xyzw[0], cdata.xyzw[1], cdata.xyzw[2], cdata.xyzw[3] );
+   _debug_printf( "%f, %f, %f, %f", cdata.xyzw[0], cdata.xyzw[1], cdata.xyzw[2], cdata.xyzw[3] );
 }
 
 static void dump_idata( struct sh_idata idata )
 {
-   debug_printf( "%d, %d, %d, %d", idata.xyzw[0], idata.xyzw[1], idata.xyzw[2], idata.xyzw[3] );
+   _debug_printf( "%d, %d, %d, %d", idata.xyzw[0], idata.xyzw[1], idata.xyzw[2], idata.xyzw[3] );
 }
 
 static void dump_bdata( boolean bdata )
 {
-   debug_printf( bdata ? "TRUE" : "FALSE" );
+   _debug_printf( bdata ? "TRUE" : "FALSE" );
 }
 
 static void dump_sampleinfo( struct ps_sampleinfo sampleinfo )
 {
    switch (sampleinfo.texture_type) {
    case SVGA3DSAMP_2D:
-      debug_printf( "_2d" );
+      _debug_printf( "_2d" );
       break;
    case SVGA3DSAMP_CUBE:
-      debug_printf( "_cube" );
+      _debug_printf( "_cube" );
       break;
    case SVGA3DSAMP_VOLUME:
-      debug_printf( "_volume" );
+      _debug_printf( "_volume" );
       break;
    default:
       assert( 0 );
@@ -255,46 +255,46 @@ static void dump_usageinfo( struct vs_semantic semantic )
 {
    switch (semantic.usage) {
    case SVGA3D_DECLUSAGE_POSITION:
-      debug_printf("_position" );
+      _debug_printf("_position" );
       break;
    case SVGA3D_DECLUSAGE_BLENDWEIGHT:
-      debug_printf("_blendweight" );
+      _debug_printf("_blendweight" );
       break;
    case SVGA3D_DECLUSAGE_BLENDINDICES:
-      debug_printf("_blendindices" );
+      _debug_printf("_blendindices" );
       break;
    case SVGA3D_DECLUSAGE_NORMAL:
-      debug_printf("_normal" );
+      _debug_printf("_normal" );
       break;
    case SVGA3D_DECLUSAGE_PSIZE:
-      debug_printf("_psize" );
+      _debug_printf("_psize" );
       break;
    case SVGA3D_DECLUSAGE_TEXCOORD:
-      debug_printf("_texcoord");
+      _debug_printf("_texcoord");
       break;
    case SVGA3D_DECLUSAGE_TANGENT:
-      debug_printf("_tangent" );
+      _debug_printf("_tangent" );
       break;
    case SVGA3D_DECLUSAGE_BINORMAL:
-      debug_printf("_binormal" );
+      _debug_printf("_binormal" );
       break;
    case SVGA3D_DECLUSAGE_TESSFACTOR:
-      debug_printf("_tessfactor" );
+      _debug_printf("_tessfactor" );
       break;
    case SVGA3D_DECLUSAGE_POSITIONT:
-      debug_printf("_positiont" );
+      _debug_printf("_positiont" );
       break;
    case SVGA3D_DECLUSAGE_COLOR:
-      debug_printf("_color" );
+      _debug_printf("_color" );
       break;
    case SVGA3D_DECLUSAGE_FOG:
-      debug_printf("_fog" );
+      _debug_printf("_fog" );
       break;
    case SVGA3D_DECLUSAGE_DEPTH:
-      debug_printf("_depth" );
+      _debug_printf("_depth" );
       break;
    case SVGA3D_DECLUSAGE_SAMPLE:
-      debug_printf("_sample");
+      _debug_printf("_sample");
       break;
    default:
       assert( 0 );
@@ -302,7 +302,7 @@ static void dump_usageinfo( struct vs_semantic semantic )
    }
 
    if (semantic.usage_index != 0) {
-      debug_printf("%d", semantic.usage_index );
+      _debug_printf("%d", semantic.usage_index );
    }
 }
 
@@ -316,47 +316,47 @@ static void dump_dstreg( struct sh_dstreg dstreg, const struct dump_info *di )
    assert( (dstreg.modifier & (SVGA3DDSTMOD_SATURATE | SVGA3DDSTMOD_PARTIALPRECISION)) == dstreg.modifier );
 
    if (dstreg.modifier & SVGA3DDSTMOD_SATURATE)
-      debug_printf( "_sat" );
+      _debug_printf( "_sat" );
    if (dstreg.modifier & SVGA3DDSTMOD_PARTIALPRECISION)
-      debug_printf( "_pp" );
+      _debug_printf( "_pp" );
    switch (dstreg.shift_scale) {
    case 0:
       break;
    case 1:
-      debug_printf( "_x2" );
+      _debug_printf( "_x2" );
       break;
    case 2:
-      debug_printf( "_x4" );
+      _debug_printf( "_x4" );
       break;
    case 3:
-      debug_printf( "_x8" );
+      _debug_printf( "_x8" );
       break;
    case 13:
-      debug_printf( "_d8" );
+      _debug_printf( "_d8" );
       break;
    case 14:
-      debug_printf( "_d4" );
+      _debug_printf( "_d4" );
       break;
    case 15:
-      debug_printf( "_d2" );
+      _debug_printf( "_d2" );
       break;
    default:
       assert( 0 );
    }
-   debug_printf( " " );
+   _debug_printf( " " );
 
    u.dstreg = dstreg;
    dump_reg( u.reg, NULL, di );
    if (dstreg.write_mask != SVGA3DWRITEMASK_ALL) {
-      debug_printf( "." );
+      _debug_printf( "." );
       if (dstreg.write_mask & SVGA3DWRITEMASK_0)
-         debug_printf( "x" );
+         _debug_printf( "x" );
       if (dstreg.write_mask & SVGA3DWRITEMASK_1)
-         debug_printf( "y" );
+         _debug_printf( "y" );
       if (dstreg.write_mask & SVGA3DWRITEMASK_2)
-         debug_printf( "z" );
+         _debug_printf( "z" );
       if (dstreg.write_mask & SVGA3DWRITEMASK_3)
-         debug_printf( "w" );
+         _debug_printf( "w" );
    }
 }
 
@@ -372,19 +372,19 @@ static void dump_srcreg( struct sh_srcreg srcreg, struct sh_srcreg *indreg, cons
    case SVGA3DSRCMOD_BIASNEG:
    case SVGA3DSRCMOD_SIGNNEG:
    case SVGA3DSRCMOD_X2NEG:
-      debug_printf( "-" );
+      _debug_printf( "-" );
       break;
    case SVGA3DSRCMOD_ABS:
-      debug_printf( "|" );
+      _debug_printf( "|" );
       break;
    case SVGA3DSRCMOD_ABSNEG:
-      debug_printf( "-|" );
+      _debug_printf( "-|" );
       break;
    case SVGA3DSRCMOD_COMP:
-      debug_printf( "1-" );
+      _debug_printf( "1-" );
       break;
    case SVGA3DSRCMOD_NOT:
-      debug_printf( "!" );
+      _debug_printf( "!" );
    }
 
    u.srcreg = srcreg;
@@ -397,39 +397,39 @@ static void dump_srcreg( struct sh_srcreg srcreg, struct sh_srcreg *indreg, cons
       break;
    case SVGA3DSRCMOD_ABS:
    case SVGA3DSRCMOD_ABSNEG:
-      debug_printf( "|" );
+      _debug_printf( "|" );
       break;
    case SVGA3DSRCMOD_BIAS:
    case SVGA3DSRCMOD_BIASNEG:
-      debug_printf( "_bias" );
+      _debug_printf( "_bias" );
       break;
    case SVGA3DSRCMOD_SIGN:
    case SVGA3DSRCMOD_SIGNNEG:
-      debug_printf( "_bx2" );
+      _debug_printf( "_bx2" );
       break;
    case SVGA3DSRCMOD_X2:
    case SVGA3DSRCMOD_X2NEG:
-      debug_printf( "_x2" );
+      _debug_printf( "_x2" );
       break;
    case SVGA3DSRCMOD_DZ:
-      debug_printf( "_dz" );
+      _debug_printf( "_dz" );
       break;
    case SVGA3DSRCMOD_DW:
-      debug_printf( "_dw" );
+      _debug_printf( "_dw" );
       break;
    default:
       assert( 0 );
    }
    if (srcreg.swizzle_x != 0 || srcreg.swizzle_y != 1 || srcreg.swizzle_z != 2 || srcreg.swizzle_w != 3) {
-      debug_printf( "." );
+      _debug_printf( "." );
       if (srcreg.swizzle_x == srcreg.swizzle_y && srcreg.swizzle_y == srcreg.swizzle_z && srcreg.swizzle_z == srcreg.swizzle_w) {
-         debug_printf( "%c", "xyzw"[srcreg.swizzle_x] );
+         _debug_printf( "%c", "xyzw"[srcreg.swizzle_x] );
       }
       else {
-         debug_printf( "%c", "xyzw"[srcreg.swizzle_x] );
-         debug_printf( "%c", "xyzw"[srcreg.swizzle_y] );
-         debug_printf( "%c", "xyzw"[srcreg.swizzle_z] );
-         debug_printf( "%c", "xyzw"[srcreg.swizzle_w] );
+         _debug_printf( "%c", "xyzw"[srcreg.swizzle_x] );
+         _debug_printf( "%c", "xyzw"[srcreg.swizzle_y] );
+         _debug_printf( "%c", "xyzw"[srcreg.swizzle_z] );
+         _debug_printf( "%c", "xyzw"[srcreg.swizzle_w] );
       }
    }
 }
@@ -447,15 +447,15 @@ svga_shader_dump(
 
    if (do_binary) {
       for (i = 0; i < dwords; i++) 
-         debug_printf("  0x%08x,\n", assem[i]);
+         _debug_printf("  0x%08x,\n", assem[i]);
       
-      debug_printf("\n\n");
+      _debug_printf("\n\n");
    }
 
    di.version.value = *assem++;
    di.is_ps = (di.version.type == SVGA3D_PS_TYPE);
 
-   debug_printf(
+   _debug_printf(
       "%s_%u_%u\n",
       di.is_ps ? "ps" : "vs",
       di.version.major,
@@ -465,7 +465,7 @@ svga_shader_dump(
       struct sh_op op = *(struct sh_op *) assem;
 
       if (assem - start >= dwords) {
-         debug_printf("... ran off end of buffer\n");
+         _debug_printf("... ran off end of buffer\n");
          assert(0);
          return;
       }
@@ -475,7 +475,7 @@ svga_shader_dump(
          {
             struct sh_dcl dcl = *(struct sh_dcl *) assem;
 
-            debug_printf( "dcl" );
+            _debug_printf( "dcl" );
             if (sh_dstreg_type( dcl.reg ) == SVGA3DREG_SAMPLER)
                dump_sampleinfo( dcl.u.ps.sampleinfo );
             else if (di.is_ps) {
@@ -486,7 +486,7 @@ svga_shader_dump(
             else
                dump_usageinfo( dcl.u.vs.semantic );
             dump_dstreg( dcl.reg, &di );
-            debug_printf( "\n" );
+            _debug_printf( "\n" );
             assem += sizeof( struct sh_dcl ) / sizeof( unsigned );
          }
          break;
@@ -495,11 +495,11 @@ svga_shader_dump(
          {
             struct sh_defb defb = *(struct sh_defb *) assem;
 
-            debug_printf( "defb " );
+            _debug_printf( "defb " );
             dump_reg( defb.reg, NULL, &di );
-            debug_printf( ", " );
+            _debug_printf( ", " );
             dump_bdata( defb.data );
-            debug_printf( "\n" );
+            _debug_printf( "\n" );
             assem += sizeof( struct sh_defb ) / sizeof( unsigned );
          }
          break;
@@ -508,11 +508,11 @@ svga_shader_dump(
          {
             struct sh_defi defi = *(struct sh_defi *) assem;
 
-            debug_printf( "defi " );
+            _debug_printf( "defi " );
             dump_reg( defi.reg, NULL, &di );
-            debug_printf( ", " );
+            _debug_printf( ", " );
             dump_idata( defi.idata );
-            debug_printf( "\n" );
+            _debug_printf( "\n" );
             assem += sizeof( struct sh_defi ) / sizeof( unsigned );
          }
          break;
@@ -528,11 +528,11 @@ svga_shader_dump(
          else {
             struct sh_unaryop unaryop = *(struct sh_unaryop *) assem;
             dump_dstreg( unaryop.dst, &di );
-            debug_printf( ", " );
+            _debug_printf( ", " );
             dump_srcreg( unaryop.src, NULL, &di );
             assem += sizeof( struct sh_unaryop ) / sizeof( unsigned );
          }
-         debug_printf( "\n" );
+         _debug_printf( "\n" );
          break;
 
       case SVGA3DOP_TEX:
@@ -549,7 +549,7 @@ svga_shader_dump(
                struct sh_unaryop unaryop = *(struct sh_unaryop *) assem;
 
                dump_dstreg( unaryop.dst, &di );
-               debug_printf( ", " );
+               _debug_printf( ", " );
                dump_srcreg( unaryop.src, NULL, &di );
                assem += sizeof( struct sh_unaryop ) / sizeof( unsigned );
             }
@@ -559,30 +559,30 @@ svga_shader_dump(
 
             dump_op( op, "texld" );
             dump_dstreg( binaryop.dst, &di );
-            debug_printf( ", " );
+            _debug_printf( ", " );
             dump_srcreg( binaryop.src0, NULL, &di );
-            debug_printf( ", " );
+            _debug_printf( ", " );
             dump_srcreg( binaryop.src1, NULL, &di );
             assem += sizeof( struct sh_binaryop ) / sizeof( unsigned );
          }
-         debug_printf( "\n" );
+         _debug_printf( "\n" );
          break;
 
       case SVGA3DOP_DEF:
          {
             struct sh_def def = *(struct sh_def *) assem;
 
-            debug_printf( "def " );
+            _debug_printf( "def " );
             dump_reg( def.reg, NULL, &di );
-            debug_printf( ", " );
+            _debug_printf( ", " );
             dump_cdata( def.cdata );
-            debug_printf( "\n" );
+            _debug_printf( "\n" );
             assem += sizeof( struct sh_def ) / sizeof( unsigned );
          }
          break;
 
       case SVGA3DOP_PHASE:
-         debug_printf( "phase\n" );
+         _debug_printf( "phase\n" );
          assem += sizeof( struct sh_op ) / sizeof( unsigned );
          break;
 
@@ -596,12 +596,12 @@ svga_shader_dump(
          break;
 
       case SVGA3DOP_RET:
-         debug_printf( "ret\n" );
+         _debug_printf( "ret\n" );
          assem += sizeof( struct sh_op ) / sizeof( unsigned );
          break;
 
       case SVGA3DOP_END:
-         debug_printf( "end\n" );
+         _debug_printf( "end\n" );
          finished = TRUE;
          break;
 
@@ -640,14 +640,14 @@ svga_shader_dump(
                }
 
                if (not_first_arg)
-                  debug_printf( ", " );
+                  _debug_printf( ", " );
                else
-                  debug_printf( " " );
+                  _debug_printf( " " );
                dump_srcreg( srcreg, &indreg, &di );
                not_first_arg = TRUE;
             }
 
-            debug_printf( "\n" );
+            _debug_printf( "\n" );
          }
       }
    }
index bf470b46ae1d01d753995e588ccf2c89228e6017..80f4874b78050ec48ef34375c31b3987c20da865 100644 (file)
@@ -94,25 +94,6 @@ trace_surface_unwrap(struct trace_context *tr_ctx,
 }
 
 
-static INLINE void
-trace_context_set_edgeflags(struct pipe_context *_pipe,
-                            const unsigned *bitfield)
-{
-   struct trace_context *tr_ctx = trace_context(_pipe);
-   struct pipe_context *pipe = tr_ctx->pipe;
-
-   trace_dump_call_begin("pipe_context", "set_edgeflags");
-
-   trace_dump_arg(ptr, pipe);
-   /* FIXME: we don't know how big this array is */
-   trace_dump_arg(ptr, bitfield);
-
-   pipe->set_edgeflags(pipe, bitfield);;
-
-   trace_dump_call_end();
-}
-
-
 static INLINE void
 trace_context_draw_block(struct trace_context *tr_ctx, int flag)
 {
@@ -143,10 +124,16 @@ trace_context_draw_block(struct trace_context *tr_ctx, int flag)
          for (k = 0; k < tr_ctx->curr.nr_cbufs; k++)
             if (tr_ctx->draw_rule.surf == tr_ctx->curr.cbufs[k])
                block = TRUE;
-      if (tr_ctx->draw_rule.tex)
+      if (tr_ctx->draw_rule.tex) {
          for (k = 0; k < tr_ctx->curr.num_texs; k++)
             if (tr_ctx->draw_rule.tex == tr_ctx->curr.tex[k])
                block = TRUE;
+         for (k = 0; k < tr_ctx->curr.num_vert_texs; k++) {
+            if (tr_ctx->draw_rule.tex == tr_ctx->curr.vert_tex[k]) {
+               block = TRUE;
+            }
+         }
+      }
 
       if (block)
          tr_ctx->draw_blocked |= (flag | 4);
@@ -192,7 +179,7 @@ trace_context_draw_arrays(struct pipe_context *_pipe,
    trace_dump_arg(uint, start);
    trace_dump_arg(uint, count);
 
-   result = pipe->draw_arrays(pipe, mode, start, count);;
+   result = pipe->draw_arrays(pipe, mode, start, count);
 
    trace_dump_ret(bool, result);
 
@@ -232,7 +219,7 @@ trace_context_draw_elements(struct pipe_context *_pipe,
    trace_dump_arg(uint, start);
    trace_dump_arg(uint, count);
 
-   result = pipe->draw_elements(pipe, indexBuffer, indexSize, mode, start, count);;
+   result = pipe->draw_elements(pipe, indexBuffer, indexSize, mode, start, count);
 
    trace_dump_ret(bool, result);
 
@@ -306,7 +293,7 @@ trace_context_create_query(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(uint, query_type);
 
-   result = pipe->create_query(pipe, query_type);;
+   result = pipe->create_query(pipe, query_type);
 
    trace_dump_ret(ptr, result);
 
@@ -328,7 +315,7 @@ trace_context_destroy_query(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(ptr, query);
 
-   pipe->destroy_query(pipe, query);;
+   pipe->destroy_query(pipe, query);
 
    trace_dump_call_end();
 }
@@ -346,7 +333,7 @@ trace_context_begin_query(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(ptr, query);
 
-   pipe->begin_query(pipe, query);;
+   pipe->begin_query(pipe, query);
 
    trace_dump_call_end();
 }
@@ -385,7 +372,7 @@ trace_context_get_query_result(struct pipe_context *_pipe,
 
    trace_dump_arg(ptr, pipe);
 
-   _result = pipe->get_query_result(pipe, query, wait, presult);;
+   _result = pipe->get_query_result(pipe, query, wait, presult);
    result = *presult;
 
    trace_dump_arg(uint, result);
@@ -410,7 +397,7 @@ trace_context_create_blend_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(blend_state, state);
 
-   result = pipe->create_blend_state(pipe, state);;
+   result = pipe->create_blend_state(pipe, state);
 
    trace_dump_ret(ptr, result);
 
@@ -432,7 +419,7 @@ trace_context_bind_blend_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(ptr, state);
 
-   pipe->bind_blend_state(pipe, state);;
+   pipe->bind_blend_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -450,7 +437,7 @@ trace_context_delete_blend_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(ptr, state);
 
-   pipe->delete_blend_state(pipe, state);;
+   pipe->delete_blend_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -469,7 +456,7 @@ trace_context_create_sampler_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(sampler_state, state);
 
-   result = pipe->create_sampler_state(pipe, state);;
+   result = pipe->create_sampler_state(pipe, state);
 
    trace_dump_ret(ptr, result);
 
@@ -480,19 +467,40 @@ trace_context_create_sampler_state(struct pipe_context *_pipe,
 
 
 static INLINE void
-trace_context_bind_sampler_states(struct pipe_context *_pipe,
-                                  unsigned num_states, void **states)
+trace_context_bind_fragment_sampler_states(struct pipe_context *_pipe,
+                                           unsigned num_states,
+                                           void **states)
+{
+   struct trace_context *tr_ctx = trace_context(_pipe);
+   struct pipe_context *pipe = tr_ctx->pipe;
+
+   trace_dump_call_begin("pipe_context", "bind_fragment_sampler_states");
+
+   trace_dump_arg(ptr, pipe);
+   trace_dump_arg(uint, num_states);
+   trace_dump_arg_array(ptr, states, num_states);
+
+   pipe->bind_fragment_sampler_states(pipe, num_states, states);
+
+   trace_dump_call_end();
+}
+
+
+static INLINE void
+trace_context_bind_vertex_sampler_states(struct pipe_context *_pipe,
+                                         unsigned num_states,
+                                         void **states)
 {
    struct trace_context *tr_ctx = trace_context(_pipe);
    struct pipe_context *pipe = tr_ctx->pipe;
 
-   trace_dump_call_begin("pipe_context", "bind_sampler_states");
+   trace_dump_call_begin("pipe_context", "bind_vertex_sampler_states");
 
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(uint, num_states);
    trace_dump_arg_array(ptr, states, num_states);
 
-   pipe->bind_sampler_states(pipe, num_states, states);;
+   pipe->bind_vertex_sampler_states(pipe, num_states, states);
 
    trace_dump_call_end();
 }
@@ -510,7 +518,7 @@ trace_context_delete_sampler_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(ptr, state);
 
-   pipe->delete_sampler_state(pipe, state);;
+   pipe->delete_sampler_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -529,7 +537,7 @@ trace_context_create_rasterizer_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(rasterizer_state, state);
 
-   result = pipe->create_rasterizer_state(pipe, state);;
+   result = pipe->create_rasterizer_state(pipe, state);
 
    trace_dump_ret(ptr, result);
 
@@ -551,7 +559,7 @@ trace_context_bind_rasterizer_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(ptr, state);
 
-   pipe->bind_rasterizer_state(pipe, state);;
+   pipe->bind_rasterizer_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -569,7 +577,7 @@ trace_context_delete_rasterizer_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(ptr, state);
 
-   pipe->delete_rasterizer_state(pipe, state);;
+   pipe->delete_rasterizer_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -585,7 +593,7 @@ trace_context_create_depth_stencil_alpha_state(struct pipe_context *_pipe,
 
    trace_dump_call_begin("pipe_context", "create_depth_stencil_alpha_state");
 
-   result = pipe->create_depth_stencil_alpha_state(pipe, state);;
+   result = pipe->create_depth_stencil_alpha_state(pipe, state);
 
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(depth_stencil_alpha_state, state);
@@ -610,7 +618,7 @@ trace_context_bind_depth_stencil_alpha_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(ptr, state);
 
-   pipe->bind_depth_stencil_alpha_state(pipe, state);;
+   pipe->bind_depth_stencil_alpha_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -628,7 +636,7 @@ trace_context_delete_depth_stencil_alpha_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(ptr, state);
 
-   pipe->delete_depth_stencil_alpha_state(pipe, state);;
+   pipe->delete_depth_stencil_alpha_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -647,7 +655,7 @@ trace_context_create_fs_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(shader_state, state);
 
-   result = pipe->create_fs_state(pipe, state);;
+   result = pipe->create_fs_state(pipe, state);
 
    trace_dump_ret(ptr, result);
 
@@ -750,7 +758,7 @@ trace_context_bind_vs_state(struct pipe_context *_pipe,
    if (tr_shdr && tr_shdr->replaced)
       state = tr_shdr->replaced;
 
-   pipe->bind_vs_state(pipe, state);;
+   pipe->bind_vs_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -770,7 +778,7 @@ trace_context_delete_vs_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(ptr, state);
 
-   pipe->delete_vs_state(pipe, state);;
+   pipe->delete_vs_state(pipe, state);
 
    trace_dump_call_end();
 
@@ -790,7 +798,7 @@ trace_context_set_blend_color(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(blend_color, state);
 
-   pipe->set_blend_color(pipe, state);;
+   pipe->set_blend_color(pipe, state);
 
    trace_dump_call_end();
 }
@@ -808,7 +816,7 @@ trace_context_set_clip_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(clip_state, state);
 
-   pipe->set_clip_state(pipe, state);;
+   pipe->set_clip_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -880,7 +888,7 @@ trace_context_set_framebuffer_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(framebuffer_state, state);
 
-   pipe->set_framebuffer_state(pipe, state);;
+   pipe->set_framebuffer_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -898,7 +906,7 @@ trace_context_set_polygon_stipple(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(poly_stipple, state);
 
-   pipe->set_polygon_stipple(pipe, state);;
+   pipe->set_polygon_stipple(pipe, state);
 
    trace_dump_call_end();
 }
@@ -916,7 +924,7 @@ trace_context_set_scissor_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(scissor_state, state);
 
-   pipe->set_scissor_state(pipe, state);;
+   pipe->set_scissor_state(pipe, state);
 
    trace_dump_call_end();
 }
@@ -934,16 +942,16 @@ trace_context_set_viewport_state(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(viewport_state, state);
 
-   pipe->set_viewport_state(pipe, state);;
+   pipe->set_viewport_state(pipe, state);
 
    trace_dump_call_end();
 }
 
 
 static INLINE void
-trace_context_set_sampler_textures(struct pipe_context *_pipe,
-                                   unsigned num_textures,
-                                   struct pipe_texture **textures)
+trace_context_set_fragment_sampler_textures(struct pipe_context *_pipe,
+                                            unsigned num_textures,
+                                            struct pipe_texture **textures)
 {
    struct trace_context *tr_ctx = trace_context(_pipe);
    struct trace_texture *tr_tex;
@@ -959,13 +967,44 @@ trace_context_set_sampler_textures(struct pipe_context *_pipe,
    }
    textures = unwrapped_textures;
 
-   trace_dump_call_begin("pipe_context", "set_sampler_textures");
+   trace_dump_call_begin("pipe_context", "set_fragment_sampler_textures");
+
+   trace_dump_arg(ptr, pipe);
+   trace_dump_arg(uint, num_textures);
+   trace_dump_arg_array(ptr, textures, num_textures);
+
+   pipe->set_fragment_sampler_textures(pipe, num_textures, textures);
+
+   trace_dump_call_end();
+}
+
+
+static INLINE void
+trace_context_set_vertex_sampler_textures(struct pipe_context *_pipe,
+                                          unsigned num_textures,
+                                          struct pipe_texture **textures)
+{
+   struct trace_context *tr_ctx = trace_context(_pipe);
+   struct trace_texture *tr_tex;
+   struct pipe_context *pipe = tr_ctx->pipe;
+   struct pipe_texture *unwrapped_textures[PIPE_MAX_VERTEX_SAMPLERS];
+   unsigned i;
+
+   tr_ctx->curr.num_vert_texs = num_textures;
+   for(i = 0; i < num_textures; ++i) {
+      tr_tex = trace_texture(textures[i]);
+      tr_ctx->curr.vert_tex[i] = tr_tex;
+      unwrapped_textures[i] = tr_tex ? tr_tex->texture : NULL;
+   }
+   textures = unwrapped_textures;
+
+   trace_dump_call_begin("pipe_context", "set_vertex_sampler_textures");
 
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(uint, num_textures);
    trace_dump_arg_array(ptr, textures, num_textures);
 
-   pipe->set_sampler_textures(pipe, num_textures, textures);;
+   pipe->set_vertex_sampler_textures(pipe, num_textures, textures);
 
    trace_dump_call_end();
 }
@@ -1024,7 +1063,7 @@ trace_context_set_vertex_elements(struct pipe_context *_pipe,
    trace_dump_struct_array(vertex_element, elements, num_elements);
    trace_dump_arg_end();
 
-   pipe->set_vertex_elements(pipe, num_elements, elements);;
+   pipe->set_vertex_elements(pipe, num_elements, elements);
 
    trace_dump_call_end();
 }
@@ -1085,7 +1124,7 @@ trace_context_surface_fill(struct pipe_context *_pipe,
    trace_dump_arg(uint, width);
    trace_dump_arg(uint, height);
 
-   pipe->surface_fill(pipe, dst, dstx, dsty, width, height, value);;
+   pipe->surface_fill(pipe, dst, dstx, dsty, width, height, value);
 
    trace_dump_call_end();
 }
@@ -1128,7 +1167,7 @@ trace_context_flush(struct pipe_context *_pipe,
    trace_dump_arg(ptr, pipe);
    trace_dump_arg(uint, flags);
 
-   pipe->flush(pipe, flags, fence);;
+   pipe->flush(pipe, flags, fence);
 
    if(fence)
       trace_dump_ret(ptr, *fence);
@@ -1240,7 +1279,6 @@ trace_context_create(struct pipe_screen *_screen,
    tr_ctx->base.winsys = _screen->winsys;
    tr_ctx->base.screen = _screen;
    tr_ctx->base.destroy = trace_context_destroy;
-   tr_ctx->base.set_edgeflags = trace_context_set_edgeflags;
    tr_ctx->base.draw_arrays = trace_context_draw_arrays;
    tr_ctx->base.draw_elements = trace_context_draw_elements;
    tr_ctx->base.draw_range_elements = trace_context_draw_range_elements;
@@ -1253,7 +1291,8 @@ trace_context_create(struct pipe_screen *_screen,
    tr_ctx->base.bind_blend_state = trace_context_bind_blend_state;
    tr_ctx->base.delete_blend_state = trace_context_delete_blend_state;
    tr_ctx->base.create_sampler_state = trace_context_create_sampler_state;
-   tr_ctx->base.bind_sampler_states = trace_context_bind_sampler_states;
+   tr_ctx->base.bind_fragment_sampler_states = trace_context_bind_fragment_sampler_states;
+   tr_ctx->base.bind_vertex_sampler_states = trace_context_bind_vertex_sampler_states;
    tr_ctx->base.delete_sampler_state = trace_context_delete_sampler_state;
    tr_ctx->base.create_rasterizer_state = trace_context_create_rasterizer_state;
    tr_ctx->base.bind_rasterizer_state = trace_context_bind_rasterizer_state;
@@ -1274,7 +1313,8 @@ trace_context_create(struct pipe_screen *_screen,
    tr_ctx->base.set_polygon_stipple = trace_context_set_polygon_stipple;
    tr_ctx->base.set_scissor_state = trace_context_set_scissor_state;
    tr_ctx->base.set_viewport_state = trace_context_set_viewport_state;
-   tr_ctx->base.set_sampler_textures = trace_context_set_sampler_textures;
+   tr_ctx->base.set_fragment_sampler_textures = trace_context_set_fragment_sampler_textures;
+   tr_ctx->base.set_vertex_sampler_textures = trace_context_set_vertex_sampler_textures;
    tr_ctx->base.set_vertex_buffers = trace_context_set_vertex_buffers;
    tr_ctx->base.set_vertex_elements = trace_context_set_vertex_elements;
    if (pipe->surface_copy)
index 6febe4b41143b9a0cc78d7a74bc864a9ce4e71f4..852b480765a03deb12402d543af51ff18ab6943c 100644 (file)
@@ -54,6 +54,9 @@ struct trace_context
       struct trace_texture *tex[PIPE_MAX_SAMPLERS];
       unsigned num_texs;
 
+      struct trace_texture *vert_tex[PIPE_MAX_VERTEX_SAMPLERS];
+      unsigned num_vert_texs;
+
       unsigned nr_cbufs;
       struct trace_texture *cbufs[PIPE_MAX_COLOR_BUFS];
       struct trace_texture *zsbuf;
index bcf6751af4fae5090f12ab94c50ce8b3be3c4a1e..0102cc187636542fd966f8ccd996cdf17fe2ffae 100644 (file)
@@ -43,19 +43,6 @@ void trace_dump_format(enum pipe_format format)
 }
 
 
-void trace_dump_block(const struct pipe_format_block *block)
-{
-   if (!trace_dumping_enabled_locked())
-      return;
-
-   trace_dump_struct_begin("pipe_format_block");
-   trace_dump_member(uint, block, size);
-   trace_dump_member(uint, block, width);
-   trace_dump_member(uint, block, height);
-   trace_dump_struct_end();
-}
-
-
 static void trace_dump_reference(const struct pipe_reference *reference)
 {
    if (!trace_dumping_enabled_locked())
@@ -83,19 +70,15 @@ void trace_dump_template(const struct pipe_texture *templat)
    trace_dump_member(format, templat, format);
 
    trace_dump_member_begin("width");
-   trace_dump_array(uint, templat->width, 1);
+   trace_dump_uint(templat->width0);
    trace_dump_member_end();
 
    trace_dump_member_begin("height");
-   trace_dump_array(uint, templat->height, 1);
+   trace_dump_uint(templat->height0);
    trace_dump_member_end();
 
    trace_dump_member_begin("depth");
-   trace_dump_array(uint, templat->depth, 1);
-   trace_dump_member_end();
-
-   trace_dump_member_begin("block");
-   trace_dump_block(&templat->block);
+   trace_dump_uint(templat->depth0);
    trace_dump_member_end();
 
    trace_dump_member(uint, templat, last_level);
@@ -483,16 +466,9 @@ void trace_dump_transfer(const struct pipe_transfer *state)
 
    trace_dump_struct_begin("pipe_transfer");
 
-   trace_dump_member(format, state, format);
    trace_dump_member(uint, state, width);
    trace_dump_member(uint, state, height);
 
-   trace_dump_member_begin("block");
-   trace_dump_block(&state->block);
-   trace_dump_member_end();
-
-   trace_dump_member(uint, state, nblocksx);
-   trace_dump_member(uint, state, nblocksy);
    trace_dump_member(uint, state, stride);
    trace_dump_member(uint, state, usage);
 
index 05b821adb64318e00e1f1010c12ddafcac3fa1bb..07ad6fbb205ffdd2e38aea50bb311fe53729cfef 100644 (file)
 
 void trace_dump_format(enum pipe_format format);
 
-void trace_dump_block(const struct pipe_format_block *block);
-
 void trace_dump_template(const struct pipe_texture *templat);
 
-
 void trace_dump_rasterizer_state(const struct pipe_rasterizer_state *state);
 
 void trace_dump_poly_stipple(const struct pipe_poly_stipple *state);
index 81e0a6f3b000a539a08909286a8b2e0ba1c13e38..c31b1d869861a95bc83c4cb96e0ed3e483927f5b 100644 (file)
@@ -26,6 +26,7 @@
  **************************************************************************/
 
 
+#include "util/u_format.h"
 #include "util/u_string.h"
 #include "util/u_memory.h"
 #include "util/u_simple_list.h"
@@ -200,10 +201,12 @@ trace_rbug_texture_info(struct trace_rbug *tr_rbug, struct rbug_header *header,
    t = tr_tex->texture;
    rbug_send_texture_info_reply(tr_rbug->con, serial,
                                t->target, t->format,
-                               t->width, t->last_level + 1,
-                               t->height, t->last_level + 1,
-                               t->depth, t->last_level + 1,
-                               t->block.width, t->block.height, t->block.size,
+                               &t->width0, 1,
+                               &t->height0, 1,
+                               &t->depth0, 1,
+                               util_format_get_blockwidth(t->format),
+                               util_format_get_blockheight(t->format),
+                               util_format_get_blocksize(t->format),
                                t->last_level,
                                t->nr_samples,
                                t->tex_usage,
@@ -251,9 +254,12 @@ trace_rbug_texture_read(struct trace_rbug *tr_rbug, struct rbug_header *header,
    map = screen->transfer_map(screen, t);
 
    rbug_send_texture_read_reply(tr_rbug->con, serial,
-                                t->format,
-                                t->block.width, t->block.height, t->block.size,
-                                (uint8_t*)map, t->stride * t->nblocksy,
+                                t->texture->format,
+                                util_format_get_blockwidth(t->texture->format),
+                                util_format_get_blockheight(t->texture->format),
+                                util_format_get_blocksize(t->texture->format),
+                                (uint8_t*)map,
+                                t->stride * util_format_get_nblocksy(t->texture->format, t->height),
                                 t->stride,
                                 NULL);
 
index 7da9bd3866ba8e6b2c82713db3b7b0a0ccc20c47..117503aaff694c8346e14fd53365362a7a2eb849 100644 (file)
@@ -25,6 +25,7 @@
  *
  **************************************************************************/
 
+#include "util/u_format.h"
 #include "util/u_memory.h"
 #include "util/u_simple_list.h"
 
@@ -35,6 +36,7 @@
 #include "tr_screen.h"
 
 #include "pipe/p_inlines.h"
+#include "pipe/p_format.h"
 
 
 static boolean trace = FALSE;
@@ -424,7 +426,7 @@ trace_screen_transfer_unmap(struct pipe_screen *_screen,
    struct pipe_transfer *transfer = tr_trans->transfer;
 
    if(tr_trans->map) {
-      size_t size = transfer->nblocksy * transfer->stride;
+      size_t size = util_format_get_nblocksy(transfer->texture->format, transfer->height) * transfer->stride;
 
       trace_dump_call_begin("pipe_screen", "transfer_write");
 
index c36286f9bee49b1eac55f3c798f5dad8a9b7c41a..f7368bb95b354819322765af1c146aa7deb2766a 100644 (file)
@@ -167,7 +167,7 @@ typedef unsigned char boolean;
 #define ALIGN16_ASSIGN(NAME) NAME##___aligned
 #define ALIGN16_ATTRIB  __attribute__(( aligned( 16 ) ))
 #define ALIGN8_ATTRIB  __attribute__(( aligned( 8 ) ))
-#if __GNUC__ > 4 || (__GNUC__ == 4 &&__GNUC_MINOR__>1)
+#if (__GNUC__ > 4 || (__GNUC__ == 4 &&__GNUC_MINOR__>1)) && !defined(PIPE_ARCH_X86_64)
 #define ALIGN_STACK __attribute__((force_align_arg_pointer))
 #else
 #define ALIGN_STACK
index f6feea5f74d7ac568f15beeaca061d8dd12d7153..064605a4a059189973751f9780aefe64575eae0d 100644 (file)
@@ -53,6 +53,7 @@
 
 #if defined(__GNUC__)
 #define PIPE_CC_GCC
+#define PIPE_CC_GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__)
 #endif
 
 /*
index 5569001e601b12bdbd4c3fdb7e582ef64895a2fc..6c06fb9027041dfdebeb006f5d1404d7b7ef93f1 100644 (file)
@@ -57,14 +57,6 @@ struct pipe_context {
 
    void (*destroy)( struct pipe_context * );
 
-   
-   /* Possible interface for setting edgeflags.  These aren't really
-    * vertex elements, so don't fit there.
-    */
-   void (*set_edgeflags)( struct pipe_context *,
-                          const unsigned *bitfield );
-
-
    /**
     * VBO drawing (return false on fallbacks (temporary??))
     */
@@ -123,7 +115,12 @@ struct pipe_context {
 
    void * (*create_sampler_state)(struct pipe_context *,
                                   const struct pipe_sampler_state *);
-   void   (*bind_sampler_states)(struct pipe_context *, unsigned num, void **);
+   void   (*bind_fragment_sampler_states)(struct pipe_context *,
+                                          unsigned num_samplers,
+                                          void **samplers);
+   void   (*bind_vertex_sampler_states)(struct pipe_context *,
+                                        unsigned num_samplers,
+                                        void **samplers);
    void   (*delete_sampler_state)(struct pipe_context *, void *);
 
    void * (*create_rasterizer_state)(struct pipe_context *,
@@ -145,6 +142,12 @@ struct pipe_context {
                              const struct pipe_shader_state *);
    void   (*bind_vs_state)(struct pipe_context *, void *);
    void   (*delete_vs_state)(struct pipe_context *, void *);
+
+   void * (*create_gs_state)(struct pipe_context *,
+                             const struct pipe_shader_state *);
+   void   (*bind_gs_state)(struct pipe_context *, void *);
+   void   (*delete_gs_state)(struct pipe_context *, void *);
+
    /*@}*/
 
    /**
@@ -173,9 +176,13 @@ struct pipe_context {
    void (*set_viewport_state)( struct pipe_context *,
                                const struct pipe_viewport_state * );
 
-   void (*set_sampler_textures)( struct pipe_context *,
-                                 unsigned num_textures,
-                                 struct pipe_texture ** );
+   void (*set_fragment_sampler_textures)(struct pipe_context *,
+                                         unsigned num_textures,
+                                         struct pipe_texture **);
+
+   void (*set_vertex_sampler_textures)(struct pipe_context *,
+                                       unsigned num_textures,
+                                       struct pipe_texture **);
 
    void (*set_vertex_buffers)( struct pipe_context *,
                                unsigned num_buffers,
index fd14dc8e92d71231b31e69447f59b22d89448d9e..2cda408fecc877a6a48db218db0f9f38064c14a8 100644 (file)
@@ -140,7 +140,8 @@ enum pipe_texture_target {
    PIPE_TEXTURE_1D   = 0,
    PIPE_TEXTURE_2D   = 1,
    PIPE_TEXTURE_3D   = 2,
-   PIPE_TEXTURE_CUBE = 3
+   PIPE_TEXTURE_CUBE = 3,
+   PIPE_MAX_TEXTURE_TYPES
 };
 
 #define PIPE_TEX_FACE_POS_X 0
@@ -320,23 +321,28 @@ enum pipe_transfer_usage {
  */
 #define PIPE_SHADER_VERTEX   0
 #define PIPE_SHADER_FRAGMENT 1
-#define PIPE_SHADER_TYPES    2
+#define PIPE_SHADER_GEOMETRY 2
+#define PIPE_SHADER_TYPES    3
 
 
 /**
  * Primitive types:
  */
-#define PIPE_PRIM_POINTS          0
-#define PIPE_PRIM_LINES           1
-#define PIPE_PRIM_LINE_LOOP       2
-#define PIPE_PRIM_LINE_STRIP      3
-#define PIPE_PRIM_TRIANGLES       4
-#define PIPE_PRIM_TRIANGLE_STRIP  5
-#define PIPE_PRIM_TRIANGLE_FAN    6
-#define PIPE_PRIM_QUADS           7
-#define PIPE_PRIM_QUAD_STRIP      8
-#define PIPE_PRIM_POLYGON         9
-#define PIPE_PRIM_MAX             10
+#define PIPE_PRIM_POINTS               0
+#define PIPE_PRIM_LINES                1
+#define PIPE_PRIM_LINE_LOOP            2
+#define PIPE_PRIM_LINE_STRIP           3
+#define PIPE_PRIM_TRIANGLES            4
+#define PIPE_PRIM_TRIANGLE_STRIP       5
+#define PIPE_PRIM_TRIANGLE_FAN         6
+#define PIPE_PRIM_QUADS                7
+#define PIPE_PRIM_QUAD_STRIP           8
+#define PIPE_PRIM_POLYGON              9
+#define PIPE_PRIM_LINES_ADJACENCY          10
+#define PIPE_PRIM_LINE_STRIP_ADJACENCY    11
+#define PIPE_PRIM_TRIANGLES_ADJACENCY      12
+#define PIPE_PRIM_TRIANGLE_STRIP_ADJACENCY 13
+#define PIPE_PRIM_MAX                      14
 
 
 /**
@@ -390,6 +396,8 @@ enum pipe_transfer_usage {
 #define PIPE_CAP_BLEND_EQUATION_SEPARATE 28
 #define PIPE_CAP_SM3                     29  /*< Shader Model 3 supported */
 #define PIPE_CAP_MAX_PREDICATE_REGISTERS 30
+#define PIPE_CAP_MAX_COMBINED_SAMPLERS   31  /*< Maximum texture image units accessible from vertex
+                                                 and fragment shaders combined */
 
 
 /**
index af2308092015f909f27acdd6f3499a1258a58486..6bfff1cc59ca847c10c891e87087c5c8efb24175 100644 (file)
 extern "C" {
 #endif
 
-/**
- * The pipe_format enum is a 32-bit wide bitfield that encodes all the
- * information needed to uniquely describe a pixel format.
- */
-
-/**
- * Possible format layouts are encoded in the first 2 bits.
- * The interpretation of the remaining 30 bits depends on a particular
- * format layout.
- */
-#define PIPE_FORMAT_LAYOUT_RGBAZS   0
-#define PIPE_FORMAT_LAYOUT_YCBCR    1
-#define PIPE_FORMAT_LAYOUT_DXT      2  /**< XXX temporary? */
-#define PIPE_FORMAT_LAYOUT_MIXED    3
-
-static INLINE uint pf_layout(uint f)  /**< PIPE_FORMAT_LAYOUT_ */
-{
-   return f & 0x3;
-}
-
-/**
- * RGBAZS Format Layout.
- */
-
-/**
- * Format component selectors for RGBAZS & MIXED layout.
- */
-#define PIPE_FORMAT_COMP_R    0
-#define PIPE_FORMAT_COMP_G    1
-#define PIPE_FORMAT_COMP_B    2
-#define PIPE_FORMAT_COMP_A    3
-#define PIPE_FORMAT_COMP_0    4
-#define PIPE_FORMAT_COMP_1    5
-#define PIPE_FORMAT_COMP_Z    6
-#define PIPE_FORMAT_COMP_S    7
-
-/**
- * Format types for RGBAZS layout.
- */
-#define PIPE_FORMAT_TYPE_UNKNOWN 0
-#define PIPE_FORMAT_TYPE_FLOAT   1  /**< 16/32/64-bit/channel formats */
-#define PIPE_FORMAT_TYPE_UNORM   2  /**< uints, normalized to [0,1] */
-#define PIPE_FORMAT_TYPE_SNORM   3  /**< ints, normalized to [-1,1] */
-#define PIPE_FORMAT_TYPE_USCALED 4  /**< uints, not normalized */
-#define PIPE_FORMAT_TYPE_SSCALED 5  /**< ints, not normalized */
-#define PIPE_FORMAT_TYPE_SRGB    6  /**< sRGB colorspace */
-#define PIPE_FORMAT_TYPE_FIXED   7  /**< 16.16 fixed point */
-
-
-/**
- * Because the destination vector is assumed to be RGBA FLOAT, we
- * need to know how to swizzle and expand components from the source
- * vector.
- * Let's take U_A1_R5_G5_B5 as an example. X swizzle is A, X size
- * is 1 bit and type is UNORM. So we take the most significant bit
- * from source vector, convert 0 to 0.0 and 1 to 1.0 and save it
- * in the last component of the destination RGBA component.
- * Next, Y swizzle is R, Y size is 5 and type is UNORM. We normalize
- * those 5 bits into [0.0; 1.0] range and put it into second
- * component of the destination vector. Rinse and repeat for
- * components Z and W.
- * If any of size fields is zero, it means the source format contains
- * less than four components.
- * If any swizzle is 0 or 1, the corresponding destination component
- * should be filled with 0.0 and 1.0, respectively.
- */
-typedef uint pipe_format_rgbazs_t;
-
-static INLINE uint pf_get(pipe_format_rgbazs_t f, uint shift, uint mask)
-{
-   return (f >> shift) & mask;
-}
-
-#define pf_swizzle_x(f)       pf_get(f, 2, 0x7)  /**< PIPE_FORMAT_COMP_ */
-#define pf_swizzle_y(f)       pf_get(f, 5, 0x7)  /**< PIPE_FORMAT_COMP_ */
-#define pf_swizzle_z(f)       pf_get(f, 8, 0x7)  /**< PIPE_FORMAT_COMP_ */
-#define pf_swizzle_w(f)       pf_get(f, 11, 0x7) /**< PIPE_FORMAT_COMP_ */
-#define pf_swizzle_xyzw(f,i)  pf_get(f, 2+((i)*3), 0x7)
-#define pf_size_x(f)          pf_get(f, 14, 0x7) /**< Size of X */
-#define pf_size_y(f)          pf_get(f, 17, 0x7) /**< Size of Y */
-#define pf_size_z(f)          pf_get(f, 20, 0x7) /**< Size of Z */
-#define pf_size_w(f)          pf_get(f, 23, 0x7) /**< Size of W */
-#define pf_size_xyzw(f,i)     pf_get(f, 14+((i)*3), 0x7)
-#define pf_exp2(f)            pf_get(f, 26, 0x7) /**< Scale size by 2 ^ exp2 */
-#define pf_type(f)            pf_get(f, 29, 0x7) /**< PIPE_FORMAT_TYPE_ */
-
-/**
- * Helper macro to encode the above structure into a 32-bit value.
- */
-#define _PIPE_FORMAT_RGBAZS( SWZ, SIZEX, SIZEY, SIZEZ, SIZEW, EXP2, TYPE ) (\
-   (PIPE_FORMAT_LAYOUT_RGBAZS << 0) |\
-   ((SWZ) << 2) |\
-   ((SIZEX) << 14) |\
-   ((SIZEY) << 17) |\
-   ((SIZEZ) << 20) |\
-   ((SIZEW) << 23) |\
-   ((EXP2) << 26) |\
-   ((TYPE) << 29) )
-
-/**
- * Helper macro to encode the swizzle part of the structure above.
- */
-#define _PIPE_FORMAT_SWZ( SWZX, SWZY, SWZZ, SWZW ) (((SWZX) << 0) | ((SWZY) << 3) | ((SWZZ) << 6) | ((SWZW) << 9))
-
-/**
- * Shorthand macro for RGBAZS layout with component sizes in 1-bit units.
- */
-#define _PIPE_FORMAT_RGBAZS_1( SWZ, SIZEX, SIZEY, SIZEZ, SIZEW, TYPE )\
-   _PIPE_FORMAT_RGBAZS( SWZ, SIZEX, SIZEY, SIZEZ, SIZEW, 0, TYPE )
-
-/**
- * Shorthand macro for RGBAZS layout with component sizes in 2-bit units.
- */
-#define _PIPE_FORMAT_RGBAZS_2( SWZ, SIZEX, SIZEY, SIZEZ, SIZEW, TYPE )\
-   _PIPE_FORMAT_RGBAZS( SWZ, SIZEX, SIZEY, SIZEZ, SIZEW, 1, TYPE )
-
-/**
- * Shorthand macro for RGBAZS layout with component sizes in 8-bit units.
- */
-#define _PIPE_FORMAT_RGBAZS_8( SWZ, SIZEX, SIZEY, SIZEZ, SIZEW, TYPE )\
-   _PIPE_FORMAT_RGBAZS( SWZ, SIZEX, SIZEY, SIZEZ, SIZEW, 3, TYPE )
-
-/**
- * Shorthand macro for RGBAZS layout with component sizes in 64-bit units.
- */
-#define _PIPE_FORMAT_RGBAZS_64( SWZ, SIZEX, SIZEY, SIZEZ, SIZEW, TYPE )\
-   _PIPE_FORMAT_RGBAZS( SWZ, SIZEX, SIZEY, SIZEZ, SIZEW, 6, TYPE )
-
-typedef uint pipe_format_mixed_t;
-
-/* NOTE: Use pf_swizzle_* and pf_size_* macros for swizzles and sizes.
- */
-
-#define pf_mixed_sign_x(f)       pf_get( f, 26, 0x1 )       /*< Sign of X */
-#define pf_mixed_sign_y(f)       pf_get( f, 27, 0x1 )       /*< Sign of Y */
-#define pf_mixed_sign_z(f)       pf_get( f, 28, 0x1 )       /*< Sign of Z */
-#define pf_mixed_sign_w(f)       pf_get( f, 29, 0x1 )       /*< Sign of W */
-#define pf_mixed_sign_xyzw(f, i) pf_get( f, 26 + (i), 0x1 )
-#define pf_mixed_normalized(f)   pf_get( f, 30, 0x1 )       /*< Type is NORM (1) or SCALED (0) */
-#define pf_mixed_scale8(f)       pf_get( f, 31, 0x1 )       /*< Scale size by either one (0) or eight (1) */
-
-/**
- * Helper macro to encode the above structure into a 32-bit value.
- */
-#define _PIPE_FORMAT_MIXED( SWZ, SIZEX, SIZEY, SIZEZ, SIZEW, SIGNX, SIGNY, SIGNZ, SIGNW, NORMALIZED, SCALE8 ) (\
-   (PIPE_FORMAT_LAYOUT_MIXED << 0) |\
-   ((SWZ) << 2) |\
-   ((SIZEX) << 14) |\
-   ((SIZEY) << 17) |\
-   ((SIZEZ) << 20) |\
-   ((SIZEW) << 23) |\
-   ((SIGNX) << 26) |\
-   ((SIGNY) << 27) |\
-   ((SIGNZ) << 28) |\
-   ((SIGNW) << 29) |\
-   ((NORMALIZED) << 30) |\
-   ((SCALE8) << 31) )
-
-/**
- * Shorthand macro for common format swizzles.
- */
-#define _PIPE_FORMAT_R001 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_1 )
-#define _PIPE_FORMAT_RG01 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_G, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_1 )
-#define _PIPE_FORMAT_RGB1 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_G, PIPE_FORMAT_COMP_B, PIPE_FORMAT_COMP_1 )
-#define _PIPE_FORMAT_RGBA _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_G, PIPE_FORMAT_COMP_B, PIPE_FORMAT_COMP_A )
-#define _PIPE_FORMAT_ARGB _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_A, PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_G, PIPE_FORMAT_COMP_B )
-#define _PIPE_FORMAT_ABGR _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_A, PIPE_FORMAT_COMP_B, PIPE_FORMAT_COMP_G, PIPE_FORMAT_COMP_R )
-#define _PIPE_FORMAT_BGRA _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_B, PIPE_FORMAT_COMP_G, PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_A )
-#define _PIPE_FORMAT_1RGB _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_1, PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_G, PIPE_FORMAT_COMP_B )
-#define _PIPE_FORMAT_1BGR _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_1, PIPE_FORMAT_COMP_B, PIPE_FORMAT_COMP_G, PIPE_FORMAT_COMP_R )
-#define _PIPE_FORMAT_BGR1 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_B, PIPE_FORMAT_COMP_G, PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_1 )
-#define _PIPE_FORMAT_0000 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0 )
-#define _PIPE_FORMAT_000R _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_R )
-#define _PIPE_FORMAT_RRR1 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_1 )
-#define _PIPE_FORMAT_RRRR _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_R )
-#define _PIPE_FORMAT_RRRG _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_R, PIPE_FORMAT_COMP_G )
-#define _PIPE_FORMAT_Z000 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_Z, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0 )
-#define _PIPE_FORMAT_0Z00 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_Z, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0 )
-#define _PIPE_FORMAT_SZ00 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_S, PIPE_FORMAT_COMP_Z, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0 )
-#define _PIPE_FORMAT_ZS00 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_Z, PIPE_FORMAT_COMP_S, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0 )
-#define _PIPE_FORMAT_S000 _PIPE_FORMAT_SWZ( PIPE_FORMAT_COMP_S, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0, PIPE_FORMAT_COMP_0 )
-
-/**
- * YCBCR Format Layout.
- */
-
-/**
- * This only contains a flag that indicates whether the format is reversed or
- * not.
- */
-typedef uint pipe_format_ycbcr_t;
-
-/**
- * Helper macro to encode the above structure into a 32-bit value.
- */
-#define _PIPE_FORMAT_YCBCR( REV ) (\
-   (PIPE_FORMAT_LAYOUT_YCBCR << 0) |\
-   ((REV) << 2) )
-
-static INLINE uint pf_rev(pipe_format_ycbcr_t f)
-{
-   return (f >> 2) & 0x1;
-}
-
-
-/**
-  * Compresssed format layouts (this will probably change)
-  */
-#define _PIPE_FORMAT_DXT( LEVEL, RSIZE, GSIZE, BSIZE, ASIZE, TYPE ) \
-   ((PIPE_FORMAT_LAYOUT_DXT << 0) | \
-    ((LEVEL) << 2) | \
-    ((RSIZE) << 5) | \
-    ((GSIZE) << 8) | \
-    ((BSIZE) << 11) | \
-    ((ASIZE) << 14) | \
-    ((TYPE) << 29))
-
-
-
 /**
  * Texture/surface image formats (preliminary)
  */
 
 /* KW: Added lots of surface formats to support vertex element layout
- * definitions, and eventually render-to-vertex-buffer.  Could
- * consider making float/int/uint/scaled/normalized a separate
- * parameter, but on the other hand there are special cases like
- * z24s8, compressed textures, ycbcr, etc that won't fit that model.
+ * definitions, and eventually render-to-vertex-buffer.
  */
 
 enum pipe_format {
-   PIPE_FORMAT_NONE                  = _PIPE_FORMAT_RGBAZS_1 ( _PIPE_FORMAT_0000, 0, 0, 0, 0, PIPE_FORMAT_TYPE_UNKNOWN ),
-   PIPE_FORMAT_A8R8G8B8_UNORM        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_ARGB, 1, 1, 1, 1, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_X8R8G8B8_UNORM        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_1RGB, 1, 1, 1, 1, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_B8G8R8A8_UNORM        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_BGRA, 1, 1, 1, 1, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_B8G8R8X8_UNORM        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_BGR1, 1, 1, 1, 1, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_A1R5G5B5_UNORM        = _PIPE_FORMAT_RGBAZS_1 ( _PIPE_FORMAT_ARGB, 1, 5, 5, 5, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_A4R4G4B4_UNORM        = _PIPE_FORMAT_RGBAZS_1 ( _PIPE_FORMAT_ARGB, 4, 4, 4, 4, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R5G6B5_UNORM          = _PIPE_FORMAT_RGBAZS_1 ( _PIPE_FORMAT_RGB1, 5, 6, 5, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_A2B10G10R10_UNORM     = _PIPE_FORMAT_RGBAZS_2 ( _PIPE_FORMAT_ABGR, 1, 5, 5, 5, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_L8_UNORM              = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RRR1, 1, 1, 1, 0, PIPE_FORMAT_TYPE_UNORM ), /**< ubyte luminance */
-   PIPE_FORMAT_A8_UNORM              = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_000R, 0, 0, 0, 1, PIPE_FORMAT_TYPE_UNORM ), /**< ubyte alpha */
-   PIPE_FORMAT_I8_UNORM              = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RRRR, 1, 1, 1, 1, PIPE_FORMAT_TYPE_UNORM ), /**< ubyte intensity */
-   PIPE_FORMAT_A8L8_UNORM            = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RRRG, 1, 1, 1, 1, PIPE_FORMAT_TYPE_UNORM ), /**< ubyte alpha, luminance */
-   PIPE_FORMAT_L16_UNORM             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RRR1, 2, 2, 2, 0, PIPE_FORMAT_TYPE_UNORM ), /**< ushort luminance */
-   PIPE_FORMAT_YCBCR                 = _PIPE_FORMAT_YCBCR( 0 ),
-   PIPE_FORMAT_YCBCR_REV             = _PIPE_FORMAT_YCBCR( 1 ),
-   PIPE_FORMAT_Z16_UNORM             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_Z000, 2, 0, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_Z32_UNORM             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_Z000, 4, 0, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_Z32_FLOAT             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_Z000, 4, 0, 0, 0, PIPE_FORMAT_TYPE_FLOAT ),
-   PIPE_FORMAT_S8Z24_UNORM           = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_SZ00, 1, 3, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_Z24S8_UNORM           = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_ZS00, 3, 1, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_X8Z24_UNORM           = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_0Z00, 1, 3, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_Z24X8_UNORM           = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_Z000, 3, 1, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_S8_UNORM              = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_S000, 1, 0, 0, 0, PIPE_FORMAT_TYPE_UNORM ), /**< ubyte stencil */
-   PIPE_FORMAT_R64_FLOAT             = _PIPE_FORMAT_RGBAZS_64( _PIPE_FORMAT_R001, 1, 0, 0, 0, PIPE_FORMAT_TYPE_FLOAT ),
-   PIPE_FORMAT_R64G64_FLOAT          = _PIPE_FORMAT_RGBAZS_64( _PIPE_FORMAT_RG01, 1, 1, 0, 0, PIPE_FORMAT_TYPE_FLOAT ),
-   PIPE_FORMAT_R64G64B64_FLOAT       = _PIPE_FORMAT_RGBAZS_64( _PIPE_FORMAT_RGB1, 1, 1, 1, 0, PIPE_FORMAT_TYPE_FLOAT ),
-   PIPE_FORMAT_R64G64B64A64_FLOAT    = _PIPE_FORMAT_RGBAZS_64( _PIPE_FORMAT_RGBA, 1, 1, 1, 1, PIPE_FORMAT_TYPE_FLOAT ),
-   PIPE_FORMAT_R32_FLOAT             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 4, 0, 0, 0, PIPE_FORMAT_TYPE_FLOAT ),
-   PIPE_FORMAT_R32G32_FLOAT          = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 4, 4, 0, 0, PIPE_FORMAT_TYPE_FLOAT ),
-   PIPE_FORMAT_R32G32B32_FLOAT       = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 4, 4, 4, 0, PIPE_FORMAT_TYPE_FLOAT ),
-   PIPE_FORMAT_R32G32B32A32_FLOAT    = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 4, 4, 4, 4, PIPE_FORMAT_TYPE_FLOAT ),
-   PIPE_FORMAT_R32_UNORM             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 4, 0, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R32G32_UNORM          = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 4, 4, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R32G32B32_UNORM       = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 4, 4, 4, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R32G32B32A32_UNORM    = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 4, 4, 4, 4, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R32_USCALED           = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 4, 0, 0, 0, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R32G32_USCALED        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 4, 4, 0, 0, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R32G32B32_USCALED     = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 4, 4, 4, 0, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R32G32B32A32_USCALED  = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 4, 4, 4, 4, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R32_SNORM             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 4, 0, 0, 0, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R32G32_SNORM          = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 4, 4, 0, 0, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R32G32B32_SNORM       = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 4, 4, 4, 0, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R32G32B32A32_SNORM    = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 4, 4, 4, 4, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R32_SSCALED           = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 4, 0, 0, 0, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R32G32_SSCALED        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 4, 4, 0, 0, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R32G32B32_SSCALED     = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 4, 4, 4, 0, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R32G32B32A32_SSCALED  = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 4, 4, 4, 4, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R16_UNORM             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 2, 0, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R16G16_UNORM          = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 2, 2, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R16G16B16_UNORM       = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 2, 2, 2, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R16G16B16A16_UNORM    = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 2, 2, 2, 2, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R16_USCALED           = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 2, 0, 0, 0, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R16G16_USCALED        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 2, 2, 0, 0, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R16G16B16_USCALED     = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 2, 2, 2, 0, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R16G16B16A16_USCALED  = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 2, 2, 2, 2, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R16_SNORM             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 2, 0, 0, 0, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R16G16_SNORM          = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 2, 2, 0, 0, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R16G16B16_SNORM       = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 2, 2, 2, 0, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R16G16B16A16_SNORM    = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 2, 2, 2, 2, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R16_SSCALED           = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 2, 0, 0, 0, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R16G16_SSCALED        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 2, 2, 0, 0, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R16G16B16_SSCALED     = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 2, 2, 2, 0, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R16G16B16A16_SSCALED  = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 2, 2, 2, 2, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R8_UNORM              = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 1, 0, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R8G8_UNORM            = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 1, 1, 0, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R8G8B8_UNORM          = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 1, 1, 1, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R8G8B8A8_UNORM        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 1, 1, 1, 1, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R8G8B8X8_UNORM        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 1, 1, 1, 1, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_R8_USCALED            = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 1, 0, 0, 0, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R8G8_USCALED          = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 1, 1, 0, 0, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R8G8B8_USCALED        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 1, 1, 1, 0, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R8G8B8A8_USCALED      = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 1, 1, 1, 1, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R8G8B8X8_USCALED      = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 1, 1, 1, 1, PIPE_FORMAT_TYPE_USCALED ),
-   PIPE_FORMAT_R8_SNORM              = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 1, 0, 0, 0, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R8G8_SNORM            = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 1, 1, 0, 0, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R8G8B8_SNORM          = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 1, 1, 1, 0, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R8G8B8A8_SNORM        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R8G8B8X8_SNORM        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_B6G5R5_SNORM          = _PIPE_FORMAT_RGBAZS_1 ( _PIPE_FORMAT_BGR1, 6, 5, 5, 0, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_A8B8G8R8_SNORM        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_BGRA, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_X8B8G8R8_SNORM        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_BGR1, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SNORM ),
-   PIPE_FORMAT_R8_SSCALED            = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 1, 0, 0, 0, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R8G8_SSCALED          = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 1, 1, 0, 0, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R8G8B8_SSCALED        = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 1, 1, 1, 0, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R8G8B8A8_SSCALED      = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R8G8B8X8_SSCALED      = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SSCALED ),
-   PIPE_FORMAT_R32_FIXED             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_R001, 4, 0, 0, 0, PIPE_FORMAT_TYPE_FIXED ),
-   PIPE_FORMAT_R32G32_FIXED          = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RG01, 4, 4, 0, 0, PIPE_FORMAT_TYPE_FIXED ),
-   PIPE_FORMAT_R32G32B32_FIXED       = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 4, 4, 4, 0, PIPE_FORMAT_TYPE_FIXED ),
-   PIPE_FORMAT_R32G32B32A32_FIXED    = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 4, 4, 4, 4, PIPE_FORMAT_TYPE_FIXED ),
+   PIPE_FORMAT_NONE                  = 0,
+   PIPE_FORMAT_A8R8G8B8_UNORM        = 1,
+   PIPE_FORMAT_X8R8G8B8_UNORM        = 2,
+   PIPE_FORMAT_B8G8R8A8_UNORM        = 3,
+   PIPE_FORMAT_B8G8R8X8_UNORM        = 4,
+   PIPE_FORMAT_A1R5G5B5_UNORM        = 5,
+   PIPE_FORMAT_A4R4G4B4_UNORM        = 6,
+   PIPE_FORMAT_R5G6B5_UNORM          = 7,
+   PIPE_FORMAT_A2B10G10R10_UNORM     = 8,
+   PIPE_FORMAT_L8_UNORM              = 9,    /**< ubyte luminance */
+   PIPE_FORMAT_A8_UNORM              = 10,   /**< ubyte alpha */
+   PIPE_FORMAT_I8_UNORM              = 11,   /**< ubyte intensity */
+   PIPE_FORMAT_A8L8_UNORM            = 12,   /**< ubyte alpha, luminance */
+   PIPE_FORMAT_L16_UNORM             = 13,   /**< ushort luminance */
+   PIPE_FORMAT_YCBCR                 = 14,
+   PIPE_FORMAT_YCBCR_REV             = 15,
+   PIPE_FORMAT_Z16_UNORM             = 16,
+   PIPE_FORMAT_Z32_UNORM             = 17,
+   PIPE_FORMAT_Z32_FLOAT             = 18,
+   PIPE_FORMAT_S8Z24_UNORM           = 19,
+   PIPE_FORMAT_Z24S8_UNORM           = 20,
+   PIPE_FORMAT_X8Z24_UNORM           = 21,
+   PIPE_FORMAT_Z24X8_UNORM           = 22,
+   PIPE_FORMAT_S8_UNORM              = 23,   /**< ubyte stencil */
+   PIPE_FORMAT_R64_FLOAT             = 24,
+   PIPE_FORMAT_R64G64_FLOAT          = 25,
+   PIPE_FORMAT_R64G64B64_FLOAT       = 26,
+   PIPE_FORMAT_R64G64B64A64_FLOAT    = 27,
+   PIPE_FORMAT_R32_FLOAT             = 28,
+   PIPE_FORMAT_R32G32_FLOAT          = 29,
+   PIPE_FORMAT_R32G32B32_FLOAT       = 30,
+   PIPE_FORMAT_R32G32B32A32_FLOAT    = 31,
+   PIPE_FORMAT_R32_UNORM             = 32,
+   PIPE_FORMAT_R32G32_UNORM          = 33,
+   PIPE_FORMAT_R32G32B32_UNORM       = 34,
+   PIPE_FORMAT_R32G32B32A32_UNORM    = 35,
+   PIPE_FORMAT_R32_USCALED           = 36,
+   PIPE_FORMAT_R32G32_USCALED        = 37,
+   PIPE_FORMAT_R32G32B32_USCALED     = 38,
+   PIPE_FORMAT_R32G32B32A32_USCALED  = 39,
+   PIPE_FORMAT_R32_SNORM             = 40,
+   PIPE_FORMAT_R32G32_SNORM          = 41,
+   PIPE_FORMAT_R32G32B32_SNORM       = 42,
+   PIPE_FORMAT_R32G32B32A32_SNORM    = 43,
+   PIPE_FORMAT_R32_SSCALED           = 44,
+   PIPE_FORMAT_R32G32_SSCALED        = 45,
+   PIPE_FORMAT_R32G32B32_SSCALED     = 46,
+   PIPE_FORMAT_R32G32B32A32_SSCALED  = 47,
+   PIPE_FORMAT_R16_UNORM             = 48,
+   PIPE_FORMAT_R16G16_UNORM          = 49,
+   PIPE_FORMAT_R16G16B16_UNORM       = 50,
+   PIPE_FORMAT_R16G16B16A16_UNORM    = 51,
+   PIPE_FORMAT_R16_USCALED           = 52,
+   PIPE_FORMAT_R16G16_USCALED        = 53,
+   PIPE_FORMAT_R16G16B16_USCALED     = 54,
+   PIPE_FORMAT_R16G16B16A16_USCALED  = 55,
+   PIPE_FORMAT_R16_SNORM             = 56,
+   PIPE_FORMAT_R16G16_SNORM          = 57,
+   PIPE_FORMAT_R16G16B16_SNORM       = 58,
+   PIPE_FORMAT_R16G16B16A16_SNORM    = 59,
+   PIPE_FORMAT_R16_SSCALED           = 60,
+   PIPE_FORMAT_R16G16_SSCALED        = 61,
+   PIPE_FORMAT_R16G16B16_SSCALED     = 62,
+   PIPE_FORMAT_R16G16B16A16_SSCALED  = 63,
+   PIPE_FORMAT_R8_UNORM              = 64,
+   PIPE_FORMAT_R8G8_UNORM            = 65,
+   PIPE_FORMAT_R8G8B8_UNORM          = 66,
+   PIPE_FORMAT_R8G8B8A8_UNORM        = 67,
+   PIPE_FORMAT_R8G8B8X8_UNORM        = 68,
+   PIPE_FORMAT_R8_USCALED            = 69,
+   PIPE_FORMAT_R8G8_USCALED          = 70,
+   PIPE_FORMAT_R8G8B8_USCALED        = 71,
+   PIPE_FORMAT_R8G8B8A8_USCALED      = 72,
+   PIPE_FORMAT_R8G8B8X8_USCALED      = 73,
+   PIPE_FORMAT_R8_SNORM              = 74,
+   PIPE_FORMAT_R8G8_SNORM            = 75,
+   PIPE_FORMAT_R8G8B8_SNORM          = 76,
+   PIPE_FORMAT_R8G8B8A8_SNORM        = 77,
+   PIPE_FORMAT_R8G8B8X8_SNORM        = 78,
+   PIPE_FORMAT_B6G5R5_SNORM          = 79,
+   PIPE_FORMAT_A8B8G8R8_SNORM        = 80,
+   PIPE_FORMAT_X8B8G8R8_SNORM        = 81,
+   PIPE_FORMAT_R8_SSCALED            = 82,
+   PIPE_FORMAT_R8G8_SSCALED          = 83,
+   PIPE_FORMAT_R8G8B8_SSCALED        = 84,
+   PIPE_FORMAT_R8G8B8A8_SSCALED      = 85,
+   PIPE_FORMAT_R8G8B8X8_SSCALED      = 86,
+   PIPE_FORMAT_R32_FIXED             = 87,
+   PIPE_FORMAT_R32G32_FIXED          = 88,
+   PIPE_FORMAT_R32G32B32_FIXED       = 89,
+   PIPE_FORMAT_R32G32B32A32_FIXED    = 90,
    /* sRGB formats */
-   PIPE_FORMAT_L8_SRGB               = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RRR1, 1, 1, 1, 0, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_A8L8_SRGB             = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RRRG, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_R8G8B8_SRGB           = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 1, 1, 1, 0, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_R8G8B8A8_SRGB         = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGBA, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_R8G8B8X8_SRGB         = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_RGB1, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_A8R8G8B8_SRGB         = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_ARGB, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_X8R8G8B8_SRGB         = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_1RGB, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_B8G8R8A8_SRGB         = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_BGRA, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_B8G8R8X8_SRGB         = _PIPE_FORMAT_RGBAZS_8 ( _PIPE_FORMAT_BGR1, 1, 1, 1, 1, PIPE_FORMAT_TYPE_SRGB ),
+   PIPE_FORMAT_L8_SRGB               = 91,
+   PIPE_FORMAT_A8L8_SRGB             = 92,
+   PIPE_FORMAT_R8G8B8_SRGB           = 93,
+   PIPE_FORMAT_R8G8B8A8_SRGB         = 94,
+   PIPE_FORMAT_R8G8B8X8_SRGB         = 95,
+   PIPE_FORMAT_A8R8G8B8_SRGB         = 96,
+   PIPE_FORMAT_X8R8G8B8_SRGB         = 97,
+   PIPE_FORMAT_B8G8R8A8_SRGB         = 98,
+   PIPE_FORMAT_B8G8R8X8_SRGB         = 99,
 
    /* mixed formats */
-   PIPE_FORMAT_X8UB8UG8SR8S_NORM     = _PIPE_FORMAT_MIXED( _PIPE_FORMAT_1BGR, 1, 1, 1, 1, 0, 0, 1, 1, 1, 1 ),
-   PIPE_FORMAT_B6UG5SR5S_NORM        = _PIPE_FORMAT_MIXED( _PIPE_FORMAT_BGR1, 6, 5, 5, 0, 0, 1, 1, 0, 1, 0 ),
+   PIPE_FORMAT_X8UB8UG8SR8S_NORM     = 100,
+   PIPE_FORMAT_B6UG5SR5S_NORM        = 101,
 
    /* compressed formats */
-   PIPE_FORMAT_DXT1_RGB              = _PIPE_FORMAT_DXT( 1, 8, 8, 8, 0, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_DXT1_RGBA             = _PIPE_FORMAT_DXT( 1, 8, 8, 8, 8, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_DXT3_RGBA             = _PIPE_FORMAT_DXT( 3, 8, 8, 8, 8, PIPE_FORMAT_TYPE_UNORM ),
-   PIPE_FORMAT_DXT5_RGBA             = _PIPE_FORMAT_DXT( 5, 8, 8, 8, 8, PIPE_FORMAT_TYPE_UNORM ),
+   PIPE_FORMAT_DXT1_RGB              = 102,
+   PIPE_FORMAT_DXT1_RGBA             = 103,
+   PIPE_FORMAT_DXT3_RGBA             = 104,
+   PIPE_FORMAT_DXT5_RGBA             = 105,
 
    /* sRGB, compressed */
-   PIPE_FORMAT_DXT1_SRGB             = _PIPE_FORMAT_DXT( 1, 8, 8, 8, 0, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_DXT1_SRGBA            = _PIPE_FORMAT_DXT( 1, 8, 8, 8, 8, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_DXT3_SRGBA            = _PIPE_FORMAT_DXT( 3, 8, 8, 8, 8, PIPE_FORMAT_TYPE_SRGB ),
-   PIPE_FORMAT_DXT5_SRGBA            = _PIPE_FORMAT_DXT( 5, 8, 8, 8, 8, PIPE_FORMAT_TYPE_SRGB )
+   PIPE_FORMAT_DXT1_SRGB             = 106,
+   PIPE_FORMAT_DXT1_SRGBA            = 107,
+   PIPE_FORMAT_DXT3_SRGBA            = 108,
+   PIPE_FORMAT_DXT5_SRGBA            = 109,
+
+   PIPE_FORMAT_COUNT
 };
 
 /**
@@ -394,224 +174,6 @@ enum pipe_format {
  */
 extern const char *pf_name( enum pipe_format format );
 
-/**
- * Return bits for a particular component.
- * \param comp  component index, starting at 0
- */
-static INLINE uint pf_get_component_bits( enum pipe_format format, uint comp )
-{
-   uint size;
-
-   if (pf_swizzle_x(format) == comp) {
-      size = pf_size_x(format);
-   }
-   else if (pf_swizzle_y(format) == comp) {
-      size = pf_size_y(format);
-   }
-   else if (pf_swizzle_z(format) == comp) {
-      size = pf_size_z(format);
-   }
-   else if (pf_swizzle_w(format) == comp) {
-      size = pf_size_w(format);
-   }
-   else {
-      size = 0;
-   }
-   if (pf_layout( format ) == PIPE_FORMAT_LAYOUT_RGBAZS)
-      return size << pf_exp2( format );
-   return size << (pf_mixed_scale8( format ) * 3);
-}
-
-/**
- * Return total bits needed for the pixel format.
- */
-static INLINE uint pf_get_bits( enum pipe_format format )
-{
-   switch (pf_layout(format)) {
-   case PIPE_FORMAT_LAYOUT_RGBAZS:
-   case PIPE_FORMAT_LAYOUT_MIXED:
-      return
-         pf_get_component_bits( format, PIPE_FORMAT_COMP_0 ) +
-         pf_get_component_bits( format, PIPE_FORMAT_COMP_1 ) +
-         pf_get_component_bits( format, PIPE_FORMAT_COMP_R ) +
-         pf_get_component_bits( format, PIPE_FORMAT_COMP_G ) +
-         pf_get_component_bits( format, PIPE_FORMAT_COMP_B ) +
-         pf_get_component_bits( format, PIPE_FORMAT_COMP_A ) +
-         pf_get_component_bits( format, PIPE_FORMAT_COMP_Z ) +
-         pf_get_component_bits( format, PIPE_FORMAT_COMP_S );
-   case PIPE_FORMAT_LAYOUT_YCBCR:
-      assert( format == PIPE_FORMAT_YCBCR || format == PIPE_FORMAT_YCBCR_REV );
-      /* return effective bits per pixel */
-      return 16; 
-   default:
-      assert( 0 );
-      return 0;
-   }
-}
-
-/**
- * Return bytes per pixel for the given format.
- */
-static INLINE uint pf_get_size( enum pipe_format format )
-{
-   assert(pf_get_bits(format) % 8 == 0);
-   return pf_get_bits(format) / 8;
-}
-
-/**
- * Describe accurately the pixel format.
- * 
- * The chars-per-pixel concept falls apart with compressed and yuv images, where
- * more than one pixel are coded in a single data block. This structure 
- * describes that block.
- * 
- * Simple pixel formats are effectively a 1x1xcpp block.
- */
-struct pipe_format_block
-{
-   /** Block size in bytes */
-   unsigned size;
-   
-   /** Block width in pixels */
-   unsigned width;
-   
-   /** Block height in pixels */
-   unsigned height;
-};
-
-/**
- * Describe pixel format's block.   
- * 
- * @sa http://msdn2.microsoft.com/en-us/library/ms796147.aspx
- */
-static INLINE void 
-pf_get_block(enum pipe_format format, struct pipe_format_block *block)
-{
-   switch(format) {
-   case PIPE_FORMAT_DXT1_RGBA:
-   case PIPE_FORMAT_DXT1_RGB:
-   case PIPE_FORMAT_DXT1_SRGBA:
-   case PIPE_FORMAT_DXT1_SRGB:
-      block->size = 8;
-      block->width = 4;
-      block->height = 4;
-      break;
-   case PIPE_FORMAT_DXT3_RGBA:
-   case PIPE_FORMAT_DXT5_RGBA:
-   case PIPE_FORMAT_DXT3_SRGBA:
-   case PIPE_FORMAT_DXT5_SRGBA:
-      block->size = 16;
-      block->width = 4;
-      block->height = 4;
-      break;
-   case PIPE_FORMAT_YCBCR:
-   case PIPE_FORMAT_YCBCR_REV:
-      block->size = 4; /* 2*cpp */
-      block->width = 2;
-      block->height = 1;
-      break;
-   default:
-      block->size = pf_get_size(format);
-      block->width = 1;
-      block->height = 1;
-      break;
-   }
-}
-
-static INLINE unsigned
-pf_get_nblocksx(const struct pipe_format_block *block, unsigned x)
-{
-   return (x + block->width - 1)/block->width;
-}
-
-static INLINE unsigned
-pf_get_nblocksy(const struct pipe_format_block *block, unsigned y)
-{
-   return (y + block->height - 1)/block->height;
-}
-
-static INLINE unsigned
-pf_get_nblocks(const struct pipe_format_block *block, unsigned width, unsigned height)
-{
-   return pf_get_nblocksx(block, width)*pf_get_nblocksy(block, height);
-}
-
-static INLINE size_t
-pf_get_stride(const struct pipe_format_block *block, unsigned width)
-{
-   return pf_get_nblocksx(block, width)*block->size;
-}
-
-static INLINE size_t
-pf_get_2d_size(const struct pipe_format_block *block, size_t stride, unsigned height)
-{
-   return pf_get_nblocksy(block, height)*stride;
-}
-
-static INLINE boolean 
-pf_is_depth_or_stencil( enum pipe_format format )
-{
-   return (pf_get_component_bits( format, PIPE_FORMAT_COMP_Z ) +
-           pf_get_component_bits( format, PIPE_FORMAT_COMP_S )) != 0;
-}
-
-static INLINE boolean 
-pf_is_depth_and_stencil( enum pipe_format format )
-{
-   return (pf_get_component_bits( format, PIPE_FORMAT_COMP_Z ) != 0 &&
-           pf_get_component_bits( format, PIPE_FORMAT_COMP_S ) != 0);
-}
-
-/** DEPRECATED: For backwards compatibility */
-static INLINE boolean
-pf_is_depth_stencil( enum pipe_format format )
-{
-   return pf_is_depth_or_stencil( format );
-}
-
-static INLINE boolean 
-pf_is_compressed( enum pipe_format format )
-{
-   return pf_layout(format) == PIPE_FORMAT_LAYOUT_DXT ? TRUE : FALSE;
-}
-
-static INLINE boolean 
-pf_is_ycbcr( enum pipe_format format )
-{
-   return pf_layout(format) == PIPE_FORMAT_LAYOUT_YCBCR ? TRUE : FALSE;
-}
-
-static INLINE boolean 
-pf_has_alpha( enum pipe_format format )
-{
-   switch (pf_layout(format)) {
-   case PIPE_FORMAT_LAYOUT_RGBAZS:
-   case PIPE_FORMAT_LAYOUT_MIXED:
-      /* FIXME: pf_get_component_bits( PIPE_FORMAT_A8L8_UNORM, PIPE_FORMAT_COMP_A ) should not return 0 right? */
-      if(format == PIPE_FORMAT_A8_UNORM || 
-         format == PIPE_FORMAT_A8L8_UNORM || 
-         format == PIPE_FORMAT_A8L8_SRGB)
-         return TRUE;
-      return pf_get_component_bits( format, PIPE_FORMAT_COMP_A ) ? TRUE : FALSE;
-   case PIPE_FORMAT_LAYOUT_YCBCR:
-      return FALSE; 
-   case PIPE_FORMAT_LAYOUT_DXT:
-      switch (format) {
-      case PIPE_FORMAT_DXT1_RGBA:
-      case PIPE_FORMAT_DXT3_RGBA:
-      case PIPE_FORMAT_DXT5_RGBA:
-      case PIPE_FORMAT_DXT1_SRGBA:
-      case PIPE_FORMAT_DXT3_SRGBA:
-      case PIPE_FORMAT_DXT5_SRGBA:
-         return TRUE;
-      default:
-         return FALSE;
-      }
-   default:
-      assert( 0 );
-      return FALSE;
-   }
-}
 
 enum pipe_video_chroma_format
 {
index 1f9088b3e9c87f4b5efc26731cae3933aa89697e..c1c7415e023f974370b09e95ab67d129e952a63d 100644 (file)
@@ -51,7 +51,7 @@ pipe_reference_init(struct pipe_reference *reference, unsigned count)
 }
 
 
-static INLINE bool
+static INLINE boolean
 pipe_is_referenced(struct pipe_reference *reference)
 {
    return p_atomic_read(&reference->count) != 0;
@@ -59,30 +59,29 @@ pipe_is_referenced(struct pipe_reference *reference)
 
 
 /**
- * Set 'ptr' to point to 'reference' and update reference counting.
- * The old thing pointed to, if any, will be unreferenced first.
- * 'reference' may be NULL.
+ * Update reference counting.
+ * The old thing pointed to, if any, will be unreferenced.
+ * Both 'ptr' and 'reference' may be NULL.
+ * \return TRUE if the object's refcount hits zero and should be destroyed.
  */
-static INLINE bool
-pipe_reference(struct pipe_reference **ptr, struct pipe_reference *reference)
+static INLINE boolean
+pipe_reference(struct pipe_reference *ptr, struct pipe_reference *reference)
 {
-   bool destroy = FALSE;
+   boolean destroy = FALSE;
 
-   if(*ptr != reference) {
+   if(ptr != reference) {
       /* bump the reference.count first */
       if (reference) {
          assert(pipe_is_referenced(reference));
          p_atomic_inc(&reference->count);
       }
    
-      if (*ptr) {
-         assert(pipe_is_referenced(*ptr));
-         if (p_atomic_dec_zero(&(*ptr)->count)) {
+      if (ptr) {
+         assert(pipe_is_referenced(ptr));
+         if (p_atomic_dec_zero(&ptr->count)) {
             destroy = TRUE;
          }
       }
-   
-      *ptr = reference;
    }
 
    return destroy;
index d4c8aadaf927f698608e1c0b79d934a895a8472d..7b19364b9754fa2426d553af33a0ea1b538cf0c1 100644 (file)
@@ -35,12 +35,6 @@ extern "C" {
 
 #include "p_compiler.h"
 
-struct tgsi_version
-{
-   unsigned MajorVersion  : 8;
-   unsigned MinorVersion  : 8;
-   unsigned Padding       : 16;
-};
 
 struct tgsi_header
 {
@@ -61,26 +55,27 @@ struct tgsi_processor
 #define TGSI_TOKEN_TYPE_DECLARATION    0
 #define TGSI_TOKEN_TYPE_IMMEDIATE      1
 #define TGSI_TOKEN_TYPE_INSTRUCTION    2
+#define TGSI_TOKEN_TYPE_PROPERTY       3
 
 struct tgsi_token
 {
    unsigned Type       : 4;  /**< TGSI_TOKEN_TYPE_x */
    unsigned NrTokens   : 8;  /**< UINT */
-   unsigned Padding    : 19;
-   unsigned Extended   : 1;  /**< BOOL */
+   unsigned Padding    : 20;
 };
 
 enum tgsi_file_type {
-   TGSI_FILE_NULL        =0,
-   TGSI_FILE_CONSTANT    =1,
-   TGSI_FILE_INPUT       =2,
-   TGSI_FILE_OUTPUT      =3,
-   TGSI_FILE_TEMPORARY   =4,
-   TGSI_FILE_SAMPLER     =5,
-   TGSI_FILE_ADDRESS     =6,
-   TGSI_FILE_IMMEDIATE   =7,
-   TGSI_FILE_LOOP        =8,
-   TGSI_FILE_PREDICATE   =9,
+   TGSI_FILE_NULL         =0,
+   TGSI_FILE_CONSTANT     =1,
+   TGSI_FILE_INPUT        =2,
+   TGSI_FILE_OUTPUT       =3,
+   TGSI_FILE_TEMPORARY    =4,
+   TGSI_FILE_SAMPLER      =5,
+   TGSI_FILE_ADDRESS      =6,
+   TGSI_FILE_IMMEDIATE    =7,
+   TGSI_FILE_LOOP         =8,
+   TGSI_FILE_PREDICATE    =9,
+   TGSI_FILE_SYSTEM_VALUE =10,
    TGSI_FILE_COUNT      /**< how many TGSI_FILE_ types */
 };
 
@@ -117,8 +112,7 @@ struct tgsi_declaration
    unsigned Semantic    : 1;  /**< BOOL, any semantic info? */
    unsigned Centroid    : 1;  /**< centroid sampling? */
    unsigned Invariant   : 1;  /**< invariant optimization? */
-   unsigned Padding     : 4;
-   unsigned Extended    : 1;  /**< BOOL */
+   unsigned Padding     : 5;
 };
 
 struct tgsi_declaration_range
@@ -127,20 +121,22 @@ struct tgsi_declaration_range
    unsigned Last    : 16; /**< UINT */
 };
 
-#define TGSI_SEMANTIC_POSITION 0
-#define TGSI_SEMANTIC_COLOR    1
-#define TGSI_SEMANTIC_BCOLOR   2 /**< back-face color */
-#define TGSI_SEMANTIC_FOG      3
-#define TGSI_SEMANTIC_PSIZE    4
-#define TGSI_SEMANTIC_GENERIC  5
-#define TGSI_SEMANTIC_NORMAL   6
-#define TGSI_SEMANTIC_FACE     7
-#define TGSI_SEMANTIC_COUNT    8 /**< number of semantic values */
+#define TGSI_SEMANTIC_POSITION  0
+#define TGSI_SEMANTIC_COLOR     1
+#define TGSI_SEMANTIC_BCOLOR    2 /**< back-face color */
+#define TGSI_SEMANTIC_FOG       3
+#define TGSI_SEMANTIC_PSIZE     4
+#define TGSI_SEMANTIC_GENERIC   5
+#define TGSI_SEMANTIC_NORMAL    6
+#define TGSI_SEMANTIC_FACE      7
+#define TGSI_SEMANTIC_EDGEFLAG  8
+#define TGSI_SEMANTIC_PRIMID    9
+#define TGSI_SEMANTIC_COUNT    10 /**< number of semantic values */
 
 struct tgsi_declaration_semantic
 {
-   unsigned SemanticName   : 8;  /**< one of TGSI_SEMANTIC_x */
-   unsigned SemanticIndex  : 16; /**< UINT */
+   unsigned Name           : 8;  /**< one of TGSI_SEMANTIC_x */
+   unsigned Index          : 16; /**< UINT */
    unsigned Padding        : 8;
 };
 
@@ -151,8 +147,7 @@ struct tgsi_immediate
    unsigned Type       : 4;  /**< TGSI_TOKEN_TYPE_IMMEDIATE */
    unsigned NrTokens   : 8;  /**< UINT */
    unsigned DataType   : 4;  /**< one of TGSI_IMM_x */
-   unsigned Padding    : 15;
-   unsigned Extended   : 1;  /**< BOOL */
+   unsigned Padding    : 16;
 };
 
 union tgsi_immediate_data
@@ -160,6 +155,22 @@ union tgsi_immediate_data
    float Float;
 };
 
+#define TGSI_PROPERTY_GS_INPUT_PRIM          0
+#define TGSI_PROPERTY_GS_OUTPUT_PRIM         1
+#define TGSI_PROPERTY_GS_MAX_VERTICES        2
+#define TGSI_PROPERTY_COUNT                  3
+
+struct tgsi_property {
+   unsigned Type         : 4;  /**< TGSI_TOKEN_TYPE_PROPERTY */
+   unsigned NrTokens     : 8;  /**< UINT */
+   unsigned PropertyName : 8;  /**< one of TGSI_PROPERTY */
+   unsigned Padding      : 12;
+};
+
+struct tgsi_property_data {
+   unsigned Data;
+};
+
 /* TGSI opcodes.  
  * 
  * For more information on semantics of opcodes and
@@ -293,7 +304,7 @@ union tgsi_immediate_data
  * respectively. For a given operation code, those numbers are fixed and are
  * present here only for convenience.
  *
- * If Extended is TRUE, it is now executed.
+ * If Predicate is TRUE, tgsi_instruction_predicate token immediately follows.
  *
  * Saturate controls how are final results in destination registers modified.
  */
@@ -306,12 +317,16 @@ struct tgsi_instruction
    unsigned Saturate   : 2;  /* TGSI_SAT_ */
    unsigned NumDstRegs : 2;  /* UINT */
    unsigned NumSrcRegs : 4;  /* UINT */
-   unsigned Padding    : 3;
-   unsigned Extended   : 1;  /* BOOL */
+   unsigned Predicate  : 1;  /* BOOL */
+   unsigned Label      : 1;
+   unsigned Texture    : 1;
+   unsigned Padding    : 1;
 };
 
 /*
- * If tgsi_instruction::Extended is TRUE, tgsi_instruction_ext follows.
+ * If tgsi_instruction::Label is TRUE, tgsi_instruction_label follows.
+ *
+ * If tgsi_instruction::Texture is TRUE, tgsi_instruction_texture follows.
  * 
  * Then, tgsi_instruction::NumDstRegs of tgsi_dst_register follow.
  * 
@@ -321,42 +336,15 @@ struct tgsi_instruction
  * instruction, including the instruction word.
  */
 
-#define TGSI_INSTRUCTION_EXT_TYPE_LABEL     1
-#define TGSI_INSTRUCTION_EXT_TYPE_TEXTURE   2
-#define TGSI_INSTRUCTION_EXT_TYPE_PREDICATE 3
-
-struct tgsi_instruction_ext
-{
-   unsigned Type       : 4;  /* TGSI_INSTRUCTION_EXT_TYPE_ */
-   unsigned Padding    : 27;
-   unsigned Extended   : 1;  /* BOOL */
-};
-
-/*
- * If tgsi_instruction_ext::Type is TGSI_INSTRUCTION_EXT_TYPE_LABEL, it
- * should be cast to tgsi_instruction_ext_label.
- * 
- * If tgsi_instruction_ext::Type is TGSI_INSTRUCTION_EXT_TYPE_TEXTURE, it
- * should be cast to tgsi_instruction_ext_texture.
- * 
- * If tgsi_instruction_ext::Type is TGSI_INSTRUCTION_EXT_TYPE_PREDICATE, it
- * should be cast to tgsi_instruction_ext_predicate.
- * 
- * If tgsi_instruction_ext::Extended is TRUE, another tgsi_instruction_ext
- * follows.
- */
-
 #define TGSI_SWIZZLE_X      0
 #define TGSI_SWIZZLE_Y      1
 #define TGSI_SWIZZLE_Z      2
 #define TGSI_SWIZZLE_W      3
 
-struct tgsi_instruction_ext_label
+struct tgsi_instruction_label
 {
-   unsigned Type     : 4;    /* TGSI_INSTRUCTION_EXT_TYPE_LABEL */
    unsigned Label    : 24;   /* UINT */
-   unsigned Padding  : 3;
-   unsigned Extended : 1;    /* BOOL */
+   unsigned Padding  : 8;
 };
 
 #define TGSI_TEXTURE_UNKNOWN        0
@@ -370,29 +358,25 @@ struct tgsi_instruction_ext_label
 #define TGSI_TEXTURE_SHADOWRECT     8
 #define TGSI_TEXTURE_COUNT          9
 
-struct tgsi_instruction_ext_texture
+struct tgsi_instruction_texture
 {
-   unsigned Type     : 4;    /* TGSI_INSTRUCTION_EXT_TYPE_TEXTURE */
    unsigned Texture  : 8;    /* TGSI_TEXTURE_ */
-   unsigned Padding  : 19;
-   unsigned Extended : 1;    /* BOOL */
+   unsigned Padding  : 24;
 };
 
 /*
  * For SM3, the following constraint applies.
  *   - Swizzle is either set to identity or replicate.
  */
-struct tgsi_instruction_ext_predicate
+struct tgsi_instruction_predicate
 {
-   unsigned Type     : 4;  /* TGSI_INSTRUCTION_EXT_TYPE_PREDICATE */
+   int      Index    : 16; /* SINT */
    unsigned SwizzleX : 2;  /* TGSI_SWIZZLE_x */
    unsigned SwizzleY : 2;  /* TGSI_SWIZZLE_x */
    unsigned SwizzleZ : 2;  /* TGSI_SWIZZLE_x */
    unsigned SwizzleW : 2;  /* TGSI_SWIZZLE_x */
    unsigned Negate   : 1;  /* BOOL */
-   unsigned SrcIndex : 8;  /* UINT */
-   unsigned Padding  : 10;
-   unsigned Extended : 1;  /* BOOL */
+   unsigned Padding  : 7;
 };
 
 /**
@@ -409,26 +393,24 @@ struct tgsi_instruction_ext_predicate
  * The fetched register components are swizzled according to SwizzleX, SwizzleY,
  * SwizzleZ and SwizzleW.
  *
- * If Extended is TRUE, any further modifications to the source register are
- * made to this temporary storage.
  */
 
 struct tgsi_src_register
 {
    unsigned File        : 4;  /* TGSI_FILE_ */
+   unsigned Indirect    : 1;  /* BOOL */
+   unsigned Dimension   : 1;  /* BOOL */
+   int      Index       : 16; /* SINT */
    unsigned SwizzleX    : 2;  /* TGSI_SWIZZLE_ */
    unsigned SwizzleY    : 2;  /* TGSI_SWIZZLE_ */
    unsigned SwizzleZ    : 2;  /* TGSI_SWIZZLE_ */
    unsigned SwizzleW    : 2;  /* TGSI_SWIZZLE_ */
-   unsigned Negate      : 1;  /* BOOL */
-   unsigned Indirect    : 1;  /* BOOL */
-   unsigned Dimension   : 1;  /* BOOL */
-   int      Index       : 16; /* SINT */
-   unsigned Extended    : 1;  /* BOOL */
+   unsigned Absolute    : 1;    /* BOOL */
+   unsigned Negate      : 1;    /* BOOL */
 };
 
 /**
- * If tgsi_src_register::Extended is TRUE, tgsi_src_register_ext follows.
+ * If tgsi_src_register::Modifier is TRUE, tgsi_src_register_modifier follows.
  * 
  * Then, if tgsi_src_register::Indirect is TRUE, another tgsi_src_register
  * follows.
@@ -436,58 +418,13 @@ struct tgsi_src_register
  * Then, if tgsi_src_register::Dimension is TRUE, tgsi_dimension follows.
  */
 
-#define TGSI_SRC_REGISTER_EXT_TYPE_MOD      1
-
-struct tgsi_src_register_ext
-{
-   unsigned Type     : 4;    /* TGSI_SRC_REGISTER_EXT_TYPE_ */
-   unsigned Padding  : 27;
-   unsigned Extended : 1;    /* BOOL */
-};
-
-/**
- * If tgsi_src_register_ext::Type is TGSI_SRC_REGISTER_EXT_TYPE_MOD,
- * it should be cast to tgsi_src_register_ext_mod.
- * 
- * If tgsi_dst_register_ext::Extended is TRUE, another tgsi_dst_register_ext
- * follows.
- */
-
-
-/**
- * Extra src register modifiers
- *
- * If Complement is TRUE, the source register is modified by subtracting it
- * from 1.0.
- *
- * If Bias is TRUE, the source register is modified by subtracting 0.5 from it.
- *
- * If Scale2X is TRUE, the source register is modified by multiplying it by 2.0.
- *
- * If Absolute is TRUE, the source register is modified by removing the sign.
- *
- * If Negate is TRUE, the source register is modified by negating it.
- */
-
-struct tgsi_src_register_ext_mod
-{
-   unsigned Type         : 4;    /* TGSI_SRC_REGISTER_EXT_TYPE_MOD */
-   unsigned Complement   : 1;    /* BOOL */
-   unsigned Bias         : 1;    /* BOOL */
-   unsigned Scale2X      : 1;    /* BOOL */
-   unsigned Absolute     : 1;    /* BOOL */
-   unsigned Negate       : 1;    /* BOOL */
-   unsigned Padding      : 22;
-   unsigned Extended     : 1;    /* BOOL */
-};
 
 struct tgsi_dimension
 {
    unsigned Indirect    : 1;  /* BOOL */
    unsigned Dimension   : 1;  /* BOOL */
-   unsigned Padding     : 13;
+   unsigned Padding     : 14;
    int      Index       : 16; /* SINT */
-   unsigned Extended    : 1;  /* BOOL */
 };
 
 struct tgsi_dst_register
@@ -497,51 +434,9 @@ struct tgsi_dst_register
    unsigned Indirect    : 1;  /* BOOL */
    unsigned Dimension   : 1;  /* BOOL */
    int      Index       : 16; /* SINT */
-   unsigned Padding     : 5;
-   unsigned Extended    : 1;  /* BOOL */
-};
-
-/*
- * If tgsi_dst_register::Extended is TRUE, tgsi_dst_register_ext follows.
- * 
- * Then, if tgsi_dst_register::Indirect is TRUE, tgsi_src_register follows.
- */
-
-#define TGSI_DST_REGISTER_EXT_TYPE_MODULATE     1
-
-struct tgsi_dst_register_ext
-{
-   unsigned Type     : 4;    /* TGSI_DST_REGISTER_EXT_TYPE_ */
-   unsigned Padding  : 27;
-   unsigned Extended : 1;    /* BOOL */
+   unsigned Padding     : 6;
 };
 
-/**
- * Extra destination register modifiers
- *
- * If tgsi_dst_register_ext::Type is TGSI_DST_REGISTER_EXT_TYPE_MODULATE,
- * it should be cast to tgsi_dst_register_ext_modulate.
- * 
- * If tgsi_dst_register_ext::Extended is TRUE, another tgsi_dst_register_ext
- * follows.
- */
-
-#define TGSI_MODULATE_1X        0
-#define TGSI_MODULATE_2X        1
-#define TGSI_MODULATE_4X        2
-#define TGSI_MODULATE_8X        3
-#define TGSI_MODULATE_HALF      4
-#define TGSI_MODULATE_QUARTER   5
-#define TGSI_MODULATE_EIGHTH    6
-#define TGSI_MODULATE_COUNT     7
-
-struct tgsi_dst_register_ext_modulate
-{
-   unsigned Type     : 4;    /* TGSI_DST_REGISTER_EXT_TYPE_MODULATE */
-   unsigned Modulate : 4;    /* TGSI_MODULATE_ */
-   unsigned Padding  : 23;
-   unsigned Extended : 1;    /* BOOL */
-};
 
 #ifdef __cplusplus
 }
index 287b424e4aca6abd6c783a0dcdfe02236cb02634..60e96b98debb3d90690e3f5e125933285e9d5c1e 100644 (file)
@@ -60,6 +60,7 @@ extern "C" {
 #define PIPE_MAX_COLOR_BUFS        8
 #define PIPE_MAX_CONSTANT         32
 #define PIPE_MAX_SAMPLERS         16
+#define PIPE_MAX_VERTEX_SAMPLERS  16
 #define PIPE_MAX_SHADER_INPUTS    16
 #define PIPE_MAX_SHADER_OUTPUTS   16
 #define PIPE_MAX_TEXTURE_LEVELS   16
@@ -315,14 +316,10 @@ struct pipe_surface
  */
 struct pipe_transfer
 {
-   enum pipe_format format;      /**< PIPE_FORMAT_x */
    unsigned x;                   /**< x offset from start of texture image */
    unsigned y;                   /**< y offset from start of texture image */
    unsigned width;               /**< logical width in pixels */
    unsigned height;              /**< logical height in pixels */
-   struct pipe_format_block block;
-   unsigned nblocksx;            /**< allocated width in blocks */
-   unsigned nblocksy;            /**< allocated height in blocks */
    unsigned stride;              /**< stride in bytes between rows of blocks */
    enum pipe_transfer_usage usage; /**< PIPE_TRANSFER_*  */
 
@@ -343,13 +340,9 @@ struct pipe_texture
    enum pipe_texture_target target; /**< PIPE_TEXTURE_x */
    enum pipe_format format;         /**< PIPE_FORMAT_x */
 
-   unsigned width[PIPE_MAX_TEXTURE_LEVELS];
-   unsigned height[PIPE_MAX_TEXTURE_LEVELS];
-   unsigned depth[PIPE_MAX_TEXTURE_LEVELS];
-
-   struct pipe_format_block block;
-   unsigned nblocksx[PIPE_MAX_TEXTURE_LEVELS]; /**< allocated width in blocks */
-   unsigned nblocksy[PIPE_MAX_TEXTURE_LEVELS]; /**< allocated height in blocks */
+   unsigned width0;
+   unsigned height0;
+   unsigned depth0;
 
    unsigned last_level:8;    /**< Index of last mipmap level present/defined */
 
@@ -399,8 +392,9 @@ pipe_buffer_reference(struct pipe_buffer **ptr, struct pipe_buffer *buf)
 {
    struct pipe_buffer *old_buf = *ptr;
 
-   if (pipe_reference((struct pipe_reference **)ptr, &buf->reference))
+   if (pipe_reference(&(*ptr)->reference, &buf->reference))
       old_buf->screen->buffer_destroy(old_buf);
+   *ptr = buf;
 }
 
 static INLINE void
@@ -408,8 +402,9 @@ pipe_surface_reference(struct pipe_surface **ptr, struct pipe_surface *surf)
 {
    struct pipe_surface *old_surf = *ptr;
 
-   if (pipe_reference((struct pipe_reference **)ptr, &surf->reference))
+   if (pipe_reference(&(*ptr)->reference, &surf->reference))
       old_surf->texture->screen->tex_surface_destroy(old_surf);
+   *ptr = surf;
 }
 
 static INLINE void
@@ -417,8 +412,9 @@ pipe_texture_reference(struct pipe_texture **ptr, struct pipe_texture *tex)
 {
    struct pipe_texture *old_tex = *ptr;
 
-   if (pipe_reference((struct pipe_reference **)ptr, &tex->reference))
+   if (pipe_reference(&(*ptr)->reference, &tex->reference))
       old_tex->screen->texture_destroy(old_tex);
+   *ptr = tex;
 }
 
 
index 4da26d608cf23326ed4c3412125368d2aeba7516..b85f01c2b02c66fb4843ca895f0b8dc733980949 100644 (file)
@@ -56,8 +56,9 @@ pipe_video_surface_reference(struct pipe_video_surface **ptr, struct pipe_video_
 {
    struct pipe_video_surface *old_surf = *ptr;
 
-   if (pipe_reference((struct pipe_reference **)ptr, &surf->reference))
+   if (pipe_reference(&(*ptr)->reference, &surf->reference))
       old_surf->screen->video_surface_destroy(old_surf);
+   *ptr = surf;
 }
 
 struct pipe_video_rect
index 5625ff53cfd77096bdf7e293f1cbe7d7fe03d04c..4b12243ddff553af1b13215e538f165147fda818 100644 (file)
 #include "state_tracker/st_context.h"
 #include "state_tracker/st_cb_fbo.h"
 
+#include "util/u_format.h"
 #include "util/u_memory.h"
 #include "util/u_rect.h"
-
 static struct pipe_surface *
 dri_surface_from_handle(struct drm_api *api,
                        struct pipe_screen *screen,
@@ -62,11 +63,10 @@ dri_surface_from_handle(struct drm_api *api,
    templat.tex_usage |= PIPE_TEXTURE_USAGE_RENDER_TARGET;
    templat.target = PIPE_TEXTURE_2D;
    templat.last_level = 0;
-   templat.depth[0] = 1;
+   templat.depth0 = 1;
    templat.format = format;
-   templat.width[0] = width;
-   templat.height[0] = height;
-   pf_get_block(templat.format, &templat.block);
+   templat.width0 = width;
+   templat.height0 = height;
 
    texture = api->texture_from_shared_handle(api, screen, &templat,
                                              "dri2 buffer", pitch, handle);
index 91615abebee031cb957ed4dffec8f27aed85b836..0b60b5be0593c4297c9a1f7258c5fb1ecbb65005 100644 (file)
@@ -12,6 +12,7 @@
 
 #include "state_tracker/drm_api.h"
 
+#include "util/u_format.h"
 #include "util/u_rect.h"
 
 /*
@@ -114,11 +115,10 @@ drm_create_texture(_EGLDisplay *dpy,
        templat.tex_usage |= PIPE_TEXTURE_USAGE_PRIMARY;
        templat.target = PIPE_TEXTURE_2D;
        templat.last_level = 0;
-       templat.depth[0] = 1;
+       templat.depth0 = 1;
        templat.format = PIPE_FORMAT_A8R8G8B8_UNORM;
-       templat.width[0] = w;
-       templat.height[0] = h;
-       pf_get_block(templat.format, &templat.block);
+       templat.width0 = w;
+       templat.height0 = h;
 
        texture = screen->texture_create(dev->screen,
                                         &templat);
index 3f79cc1a3d74be0afc1790bb29c054613c85f3c3..96b13c2258e859d87fbd6141b59b68d9c6ac3d26 100644 (file)
@@ -46,6 +46,7 @@
 #include "util/u_draw_quad.h"
 #include "util/u_tile.h"
 #include "util/u_math.h"
+#include "util/u_format.h"
 #include "util/u_memory.h"
 #include "tgsi/tgsi_text.h"
 #include "tgsi/tgsi_dump.h"
@@ -80,7 +81,6 @@
 %rename(Stencil) pipe_stencil_state;
 %rename(Alpha) pipe_alpha_state;
 %rename(DepthStencilAlpha) pipe_depth_stencil_alpha_state;
-%rename(FormatBlock) pipe_format_block;
 %rename(Framebuffer) pipe_framebuffer_state;
 %rename(PolyStipple) pipe_poly_stipple;
 %rename(Rasterizer) pipe_rasterizer_state;
index a40aa1e518772fdc6b5ac278f473d6a8ee6d327d..84ce1a41e6d98207f4bef6f16f80770ef3dea918 100644 (file)
@@ -52,11 +52,16 @@ struct st_context {
       cso_set_blend($self->cso, state);
    }
    
-   void set_sampler( unsigned index, const struct pipe_sampler_state *state ) {
+   void set_fragment_sampler( unsigned index, const struct pipe_sampler_state *state ) {
       cso_single_sampler($self->cso, index, state);
       cso_single_sampler_done($self->cso);
    }
 
+   void set_vertex_sampler( unsigned index, const struct pipe_sampler_state *state ) {
+      cso_single_vertex_sampler($self->cso, index, state);
+      cso_single_vertex_sampler_done($self->cso);
+   }
+
    void set_rasterizer( const struct pipe_rasterizer_state *state ) {
       cso_set_rasterizer($self->cso, state);
    }
@@ -103,6 +108,25 @@ struct st_context {
       $self->vs = vs;
    }
 
+   void set_geometry_shader( const struct pipe_shader_state *state ) {
+      void *gs;
+
+      if(!state) {
+         cso_set_geometry_shader_handle($self->cso, NULL);
+         return;
+      }
+
+      gs = $self->pipe->create_gs_state($self->pipe, state);
+      if(!gs)
+         return;
+
+      if(cso_set_geometry_shader_handle($self->cso, gs) != PIPE_OK)
+         return;
+
+      cso_delete_geometry_shader($self->cso, $self->gs);
+      $self->gs = gs;
+   }
+
    /*
     * Parameter-like state (or properties)
     */
@@ -142,14 +166,24 @@ struct st_context {
       cso_set_viewport($self->cso, state);
    }
 
-   void set_sampler_texture(unsigned index,
-                            struct pipe_texture *texture) {
+   void set_fragment_sampler_texture(unsigned index,
+                                     struct pipe_texture *texture) {
+      if(!texture)
+         texture = $self->default_texture;
+      pipe_texture_reference(&$self->fragment_sampler_textures[index], texture);
+      $self->pipe->set_fragment_sampler_textures($self->pipe,
+                                                 PIPE_MAX_SAMPLERS,
+                                                 $self->fragment_sampler_textures);
+   }
+
+   void set_vertex_sampler_texture(unsigned index,
+                                   struct pipe_texture *texture) {
       if(!texture)
          texture = $self->default_texture;
-      pipe_texture_reference(&$self->sampler_textures[index], texture);
-      $self->pipe->set_sampler_textures($self->pipe, 
-                                        PIPE_MAX_SAMPLERS,
-                                        $self->sampler_textures);
+      pipe_texture_reference(&$self->vertex_sampler_textures[index], texture);
+      $self->pipe->set_vertex_sampler_textures($self->pipe,
+                                               PIPE_MAX_VERTEX_SAMPLERS,
+                                               $self->vertex_sampler_textures);
    }
 
    void set_vertex_buffer(unsigned index,
index f16fe5b0ff70182f978dfb2d4fa9f3b9eb43abca..2dc995adb07687ba7e1a1835d8fd36ae51fe9b71 100644 (file)
@@ -112,10 +112,9 @@ struct st_device {
       struct pipe_texture templat;
       memset(&templat, 0, sizeof(templat));
       templat.format = format;
-      pf_get_block(templat.format, &templat.block);
-      templat.width[0] = width;
-      templat.height[0] = height;
-      templat.depth[0] = depth;
+      templat.width0 = width;
+      templat.height0 = height;
+      templat.depth0 = depth;
       templat.last_level = last_level;
       templat.target = target;
       templat.tex_usage = tex_usage;
index 26fb12b387f821886ad9f4357fb7de2336dd949e..68df0093315b737bd5b058e3ed0dc42cd4849ea2 100644 (file)
@@ -152,11 +152,3 @@ enum pipe_format {
    PIPE_FORMAT_DXT5_SRGBA,
 };
 
-
-struct pipe_format_block
-{
-   unsigned size;
-   unsigned width;
-   unsigned height;
-};
-
index 1d513abf3c75489a751cf7de42e4b6299c655c22..761587dc533bbd4b24350a63570026789ec7f02b 100644 (file)
    }
    
    unsigned get_width(unsigned level=0) {
-      return $self->width[level];
+      return u_minify($self->width0, level);
    }
    
    unsigned get_height(unsigned level=0) {
-      return $self->height[level];
+      return u_minify($self->height0, level);
    }
    
    unsigned get_depth(unsigned level=0) {
-      return $self->depth[level];
+      return u_minify($self->depth0, level);
    }
-   
-   unsigned get_nblocksx(unsigned level=0) {
-      return $self->nblocksx[level];
-   }
-   
-   unsigned get_nblocksy(unsigned level=0) {
-      return $self->nblocksy[level];
-   }
-   
+  
    /** Get a surface which is a "view" into a texture */
    struct st_surface *
    get_surface(unsigned face=0, unsigned level=0, unsigned zslice=0)
@@ -88,7 +80,7 @@
          SWIG_exception(SWIG_ValueError, "face out of bounds");
       if(level > $self->last_level)
          SWIG_exception(SWIG_ValueError, "level out of bounds");
-      if(zslice >= $self->depth[level])
+      if(zslice >= u_minify($self->depth0, level))
          SWIG_exception(SWIG_ValueError, "zslice out of bounds");
       
       surface = CALLOC_STRUCT(st_surface);
@@ -126,8 +118,6 @@ struct st_surface
    unsigned format;
    unsigned width;
    unsigned height;
-   unsigned nblocksx;
-   unsigned nblocksy;
    
    ~st_surface() {
       pipe_texture_reference(&$self->texture, NULL);
@@ -142,8 +132,8 @@ struct st_surface
       struct pipe_transfer *transfer;
       unsigned stride;
 
-      stride = pf_get_nblocksx(&texture->block, w) * texture->block.size;
-      *LENGTH = pf_get_nblocksy(&texture->block, h) * stride;
+      stride = util_format_get_stride(texture->format, w);
+      *LENGTH = util_format_get_nblocksy(texture->format, h) * stride;
       *STRING = (char *) malloc(*LENGTH);
       if(!*STRING)
          return;
@@ -169,9 +159,9 @@ struct st_surface
       struct pipe_transfer *transfer;
      
       if(stride == 0)
-         stride = pf_get_nblocksx(&texture->block, w) * texture->block.size;
+         stride = util_format_get_stride(texture->format, w);
       
-      if(LENGTH < pf_get_nblocksy(&texture->block, h) * stride)
+      if(LENGTH < util_format_get_nblocksy(texture->format, h) * stride)
          SWIG_exception(SWIG_ValueError, "offset must be smaller than buffer size");
          
       transfer = screen->get_tex_transfer(screen,
@@ -375,25 +365,13 @@ struct st_surface
    static unsigned
    st_surface_width_get(struct st_surface *surface)
    {
-      return surface->texture->width[surface->level];
+      return u_minify(surface->texture->width0, surface->level);
    }
    
    static unsigned
    st_surface_height_get(struct st_surface *surface)
    {
-      return surface->texture->height[surface->level];
-   }
-
-   static unsigned
-   st_surface_nblocksx_get(struct st_surface *surface)
-   {
-      return surface->texture->nblocksx[surface->level];
-   }
-   
-   static unsigned
-   st_surface_nblocksy_get(struct st_surface *surface)
-   {
-      return surface->texture->nblocksy[surface->level];
+      return u_minify(surface->texture->height0, surface->level);
    }
 %}
 
index 348f2e436831e89186895fc4a140e06ebf36de63..2487af6bd13a3f03025dc63725192d0492aaaf96 100755 (executable)
@@ -99,7 +99,6 @@ struct_factories = {
     "pipe_stencil_state": gallium.Stencil,
     "pipe_alpha_state": gallium.Alpha,
     "pipe_depth_stencil_alpha_state": gallium.DepthStencilAlpha,
-    "pipe_format_block": gallium.FormatBlock,
     #"pipe_framebuffer_state": gallium.Framebuffer,
     "pipe_poly_stipple": gallium.PolyStipple,
     "pipe_rasterizer_state": gallium.Rasterizer,
@@ -279,9 +278,9 @@ class Screen(Object):
     def texture_create(self, templat):
         return self.real.texture_create(
             format = templat.format,
-            width = templat.width[0],
-            height = templat.height[0],
-            depth = templat.depth[0],
+            width = templat.width,
+            height = templat.height,
+            depth = templat.depth,
             last_level = templat.last_level,
             target = templat.target,
             tex_usage = templat.tex_usage,
@@ -307,7 +306,7 @@ class Screen(Object):
     def surface_write(self, surface, data, stride, size):
         if surface is None:
             return
-        assert surface.nblocksy * stride == size 
+#        assert surface.nblocksy * stride == size 
         surface.put_tile_raw(0, 0, surface.width, surface.height, data, stride)
 
     def get_tex_transfer(self, texture, face, level, zslice, usage, x, y, w, h):
@@ -388,9 +387,13 @@ class Context(Object):
     def delete_sampler_state(self, state):
         pass
 
-    def bind_sampler_states(self, num_states, states):
+    def bind_vertex_sampler_states(self, num_states, states):
         for i in range(num_states):
-            self.real.set_sampler(i, states[i])
+            self.real.set_vertex_sampler(i, states[i])
+        
+    def bind_fragment_sampler_states(self, num_states, states):
+        for i in range(num_states):
+            self.real.set_fragment_sampler(i, states[i])
         
     def create_rasterizer_state(self, state):
         return state
@@ -486,9 +489,13 @@ class Context(Object):
     def set_viewport_state(self, state):
         self.real.set_viewport(state)
 
-    def set_sampler_textures(self, num_textures, textures):
+    def set_fragment_sampler_textures(self, num_textures, textures):
+        for i in range(num_textures):
+            self.real.set_fragment_sampler_texture(i, textures[i])
+
+    def set_vertex_sampler_textures(self, num_textures, textures):
         for i in range(num_textures):
-            self.real.set_sampler_texture(i, textures[i])
+            self.real.set_vertex_sampler_texture(i, textures[i])
 
     def set_vertex_buffers(self, num_buffers, buffers):
         self.vbufs = buffers[0:num_buffers]
@@ -508,10 +515,6 @@ class Context(Object):
             self.real.set_vertex_element(i, elements[i])
         self.real.set_vertex_elements(num_elements)
 
-    def set_edgeflags(self, bitfield):
-        # FIXME
-        pass
-    
     def dump_vertices(self, start, count):
         if not self.interpreter.verbosity(2):
             return
diff --git a/src/gallium/state_trackers/python/samples/gs.py b/src/gallium/state_trackers/python/samples/gs.py
new file mode 100644 (file)
index 0000000..1ceead5
--- /dev/null
@@ -0,0 +1,254 @@
+#!/usr/bin/env python
+##########################################################################
+#
+# Copyright 2009 VMware
+# All Rights Reserved.
+#
+# Permission is hereby granted, free of charge, to any person obtaining a
+# copy of this software and associated documentation files (the
+# "Software"), to deal in the Software without restriction, including
+# without limitation the rights to use, copy, modify, merge, publish,
+# distribute, sub license, and/or sell copies of the Software, and to
+# permit persons to whom the Software is furnished to do so, subject to
+# the following conditions:
+#
+# The above copyright notice and this permission notice (including the
+# next paragraph) shall be included in all copies or substantial portions
+# of the Software.
+#
+# THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+# OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+# MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+# IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+# ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+# TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+# SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+#
+##########################################################################
+
+
+from gallium import *
+
+
+def make_image(surface):
+    data = surface.get_tile_rgba8(0, 0, surface.width, surface.height)
+
+    import Image
+    outimage = Image.fromstring('RGBA', (surface.width, surface.height), data, "raw", 'RGBA', 0, 1)
+    return outimage
+
+def save_image(filename, surface):
+    outimage = make_image(surface)
+    outimage.save(filename, "PNG")
+
+def show_image(surface):
+    outimage = make_image(surface)
+
+    import Tkinter as tk
+    from PIL import Image, ImageTk
+    root = tk.Tk()
+
+    root.title('background image')
+
+    image1 = ImageTk.PhotoImage(outimage)
+    w = image1.width()
+    h = image1.height()
+    x = 100
+    y = 100
+    root.geometry("%dx%d+%d+%d" % (w, h, x, y))
+    panel1 = tk.Label(root, image=image1)
+    panel1.pack(side='top', fill='both', expand='yes')
+    panel1.image = image1
+    root.mainloop()
+
+
+def test(dev):
+    ctx = dev.context_create()
+
+    width = 255
+    height = 255
+    minz = 0.0
+    maxz = 1.0
+
+    # disabled blending/masking
+    blend = Blend()
+    blend.rgb_src_factor = PIPE_BLENDFACTOR_ONE
+    blend.alpha_src_factor = PIPE_BLENDFACTOR_ONE
+    blend.rgb_dst_factor = PIPE_BLENDFACTOR_ZERO
+    blend.alpha_dst_factor = PIPE_BLENDFACTOR_ZERO
+    blend.colormask = PIPE_MASK_RGBA
+    ctx.set_blend(blend)
+
+    # depth/stencil/alpha
+    depth_stencil_alpha = DepthStencilAlpha()
+    depth_stencil_alpha.depth.enabled = 1
+    depth_stencil_alpha.depth.writemask = 1
+    depth_stencil_alpha.depth.func = PIPE_FUNC_LESS
+    ctx.set_depth_stencil_alpha(depth_stencil_alpha)
+
+    # rasterizer
+    rasterizer = Rasterizer()
+    rasterizer.front_winding = PIPE_WINDING_CW
+    rasterizer.cull_mode = PIPE_WINDING_NONE
+    rasterizer.scissor = 1
+    ctx.set_rasterizer(rasterizer)
+
+    # viewport
+    viewport = Viewport()
+    scale = FloatArray(4)
+    scale[0] = width / 2.0
+    scale[1] = -height / 2.0
+    scale[2] = (maxz - minz) / 2.0
+    scale[3] = 1.0
+    viewport.scale = scale
+    translate = FloatArray(4)
+    translate[0] = width / 2.0
+    translate[1] = height / 2.0
+    translate[2] = (maxz - minz) / 2.0
+    translate[3] = 0.0
+    viewport.translate = translate
+    ctx.set_viewport(viewport)
+
+    # samplers
+    sampler = Sampler()
+    sampler.wrap_s = PIPE_TEX_WRAP_CLAMP_TO_EDGE
+    sampler.wrap_t = PIPE_TEX_WRAP_CLAMP_TO_EDGE
+    sampler.wrap_r = PIPE_TEX_WRAP_CLAMP_TO_EDGE
+    sampler.min_mip_filter = PIPE_TEX_MIPFILTER_NONE
+    sampler.min_img_filter = PIPE_TEX_MIPFILTER_NEAREST
+    sampler.mag_img_filter = PIPE_TEX_MIPFILTER_NEAREST
+    sampler.normalized_coords = 1
+    ctx.set_sampler(0, sampler)
+
+    # scissor
+    scissor = Scissor()
+    scissor.minx = 0
+    scissor.miny = 0
+    scissor.maxx = width
+    scissor.maxy = height
+    ctx.set_scissor(scissor)
+
+    clip = Clip()
+    clip.nr = 0
+    ctx.set_clip(clip)
+
+    # framebuffer
+    cbuf = dev.texture_create(
+        PIPE_FORMAT_X8R8G8B8_UNORM,
+        width, height,
+        tex_usage=PIPE_TEXTURE_USAGE_DISPLAY_TARGET,
+    ).get_surface()
+    zbuf = dev.texture_create(
+        PIPE_FORMAT_Z16_UNORM,
+        width, height,
+        tex_usage=PIPE_TEXTURE_USAGE_DEPTH_STENCIL,
+    ).get_surface()
+    fb = Framebuffer()
+    fb.width = width
+    fb.height = height
+    fb.nr_cbufs = 1
+    fb.set_cbuf(0, cbuf)
+    fb.set_zsbuf(zbuf)
+    ctx.set_framebuffer(fb)
+    rgba = FloatArray(4);
+    rgba[0] = 0.0
+    rgba[1] = 0.0
+    rgba[2] = 0.0
+    rgba[3] = 0.0
+    ctx.clear(PIPE_CLEAR_COLOR | PIPE_CLEAR_DEPTHSTENCIL, rgba, 1.0, 0xff)
+
+    # vertex shader
+    vs = Shader('''
+        VERT
+        DCL IN[0], POSITION, CONSTANT
+        DCL IN[1], COLOR, CONSTANT
+        DCL OUT[0], POSITION, CONSTANT
+        DCL OUT[1], COLOR, CONSTANT
+        0:MOV OUT[0], IN[0]
+        1:MOV OUT[1], IN[1]
+        2:END
+    ''')
+    ctx.set_vertex_shader(vs)
+
+    gs = Shader('''
+        GEOM
+        PROPERTY GS_INPUT_PRIMITIVE TRIANGLES
+        PROPERTY GS_OUTPUT_PRIMITIVE TRIANGLE_STRIP
+        DCL IN[][0], POSITION, CONSTANT
+        DCL IN[][1], COLOR, CONSTANT
+        DCL OUT[0], POSITION, CONSTANT
+        DCL OUT[1], COLOR, CONSTANT
+        0:MOV OUT[0], IN[0][0]
+        1:MOV OUT[1], IN[0][1]
+        2:EMIT
+        3:MOV OUT[0], IN[1][0]
+        4:MOV OUT[1], IN[1][1]
+        5:EMIT
+        6:MOV OUT[0], IN[2][0]
+        7:MOV OUT[1], IN[2][1]
+        8:EMIT
+        9:ENDPRIM
+        10:END
+    ''')
+    ctx.set_geometry_shader(gs)
+
+    # fragment shader
+    fs = Shader('''
+        FRAG
+        DCL IN[0], COLOR, LINEAR
+        DCL OUT[0], COLOR, CONSTANT
+        0:MOV OUT[0], IN[0]
+        1:END
+    ''')
+    ctx.set_fragment_shader(fs)
+
+    nverts = 3
+    nattrs = 2
+    verts = FloatArray(nverts * nattrs * 4)
+
+    verts[ 0] =   0.0 # x1
+    verts[ 1] =   0.8 # y1
+    verts[ 2] =   0.2 # z1
+    verts[ 3] =   1.0 # w1
+    verts[ 4] =   1.0 # r1
+    verts[ 5] =   0.0 # g1
+    verts[ 6] =   0.0 # b1
+    verts[ 7] =   1.0 # a1
+    verts[ 8] =  -0.8 # x2
+    verts[ 9] =  -0.8 # y2
+    verts[10] =   0.5 # z2
+    verts[11] =   1.0 # w2
+    verts[12] =   0.0 # r2
+    verts[13] =   1.0 # g2
+    verts[14] =   0.0 # b2
+    verts[15] =   1.0 # a2
+    verts[16] =   0.8 # x3
+    verts[17] =  -0.8 # y3
+    verts[18] =   0.8 # z3
+    verts[19] =   1.0 # w3
+    verts[20] =   0.0 # r3
+    verts[21] =   0.0 # g3
+    verts[22] =   1.0 # b3
+    verts[23] =   1.0 # a3
+
+    ctx.draw_vertices(PIPE_PRIM_TRIANGLES,
+                      nverts,
+                      nattrs,
+                      verts)
+
+    ctx.flush()
+
+    show_image(cbuf)
+    #show_image(zbuf)
+    #save_image('cbuf.png', cbuf)
+    #save_image('zbuf.png', zbuf)
+
+
+
+def main():
+    dev = Device()
+    test(dev)
+
+
+if __name__ == '__main__':
+    main()
index b721e0b5750a6ef1cf3ed13f7fbd16440f5ff4f8..af80426dc6cdc90965fe11bcfba68d18aad33ef6 100644 (file)
@@ -118,7 +118,7 @@ def test(dev):
     sampler.min_img_filter = PIPE_TEX_MIPFILTER_NEAREST
     sampler.mag_img_filter = PIPE_TEX_MIPFILTER_NEAREST
     sampler.normalized_coords = 1
-    ctx.set_sampler(0, sampler)
+    ctx.set_fragment_sampler(0, sampler)
 
     # scissor
     scissor = Scissor()
@@ -159,7 +159,7 @@ def test(dev):
     
     # vertex shader
     vs = Shader('''
-        VERT1.1
+        VERT
         DCL IN[0], POSITION, CONSTANT
         DCL IN[1], COLOR, CONSTANT
         DCL OUT[0], POSITION, CONSTANT
@@ -172,7 +172,7 @@ def test(dev):
 
     # fragment shader
     fs = Shader('''
-        FRAG1.1
+        FRAG
         DCL IN[0], COLOR, LINEAR
         DCL OUT[0], COLOR, CONSTANT
         0:MOV OUT[0], IN[0]
index ea7d18738f69a19ba12a9c4d381195ad3a9d94db..d144af2447d1d8c0955f9b8acd5ba0ee1d723694 100644 (file)
@@ -62,8 +62,9 @@ st_device_reference(struct st_device **ptr, struct st_device *st_dev)
 {
    struct st_device *old_dev = *ptr;
 
-   if (pipe_reference((struct pipe_reference **)ptr, &st_dev->reference))
+   if (pipe_reference(&(*ptr)->reference, &st_dev->reference))
       st_device_really_destroy(old_dev);
+   *ptr = st_dev;
 }
 
 
@@ -134,7 +135,9 @@ st_context_destroy(struct st_context *st_ctx)
          st_ctx->pipe->destroy(st_ctx->pipe);
       
       for(i = 0; i < PIPE_MAX_SAMPLERS; ++i)
-         pipe_texture_reference(&st_ctx->sampler_textures[i], NULL);
+         pipe_texture_reference(&st_ctx->fragment_sampler_textures[i], NULL);
+      for(i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; ++i)
+         pipe_texture_reference(&st_ctx->vertex_sampler_textures[i], NULL);
       pipe_texture_reference(&st_ctx->default_texture, NULL);
 
       FREE(st_ctx);
@@ -249,12 +252,9 @@ st_context_create(struct st_device *st_dev)
       memset( &templat, 0, sizeof( templat ) );
       templat.target = PIPE_TEXTURE_2D;
       templat.format = PIPE_FORMAT_A8R8G8B8_UNORM;
-      templat.block.size = 4;
-      templat.block.width = 1;
-      templat.block.height = 1;
-      templat.width[0] = 1;
-      templat.height[0] = 1;
-      templat.depth[0] = 1;
+      templat.width0 = 1;
+      templat.height0 = 1;
+      templat.depth0 = 1;
       templat.last_level = 0;
    
       st_ctx->default_texture = screen->texture_create( screen, &templat );
@@ -264,8 +264,8 @@ st_context_create(struct st_device *st_dev)
                                              0, 0, 0,
                                              PIPE_TRANSFER_WRITE,
                                              0, 0,
-                                             st_ctx->default_texture->width[0],
-                                             st_ctx->default_texture->height[0]);
+                                             st_ctx->default_texture->width0,
+                                             st_ctx->default_texture->height0);
          if (transfer) {
             uint32_t *map;
             map = (uint32_t *) screen->transfer_map(screen, transfer);
@@ -278,9 +278,12 @@ st_context_create(struct st_device *st_dev)
       }
    
       for (i = 0; i < PIPE_MAX_SAMPLERS; i++)
-         pipe_texture_reference(&st_ctx->sampler_textures[i], st_ctx->default_texture);
+         pipe_texture_reference(&st_ctx->fragment_sampler_textures[i], st_ctx->default_texture);
+      for (i = 0; i < PIPE_MAX_VERTEX_SAMPLERS; i++)
+         pipe_texture_reference(&st_ctx->vertex_sampler_textures[i], st_ctx->default_texture);
       
-      cso_set_sampler_textures(st_ctx->cso, PIPE_MAX_SAMPLERS, st_ctx->sampler_textures);
+      cso_set_sampler_textures(st_ctx->cso, PIPE_MAX_SAMPLERS, st_ctx->fragment_sampler_textures);
+      cso_set_vertex_sampler_textures(st_ctx->cso, PIPE_MAX_VERTEX_SAMPLERS, st_ctx->vertex_sampler_textures);
    }
    
    /* vertex shader */
index a246b6a1f25b495e5e3f03d7679fc598f043f60e..f786e1341189ac1b32c15eb26658b0fe6eb6879d 100644 (file)
@@ -57,9 +57,11 @@ struct st_context {
    
    void *vs;
    void *fs;
+   void *gs;
 
    struct pipe_texture *default_texture;
-   struct pipe_texture *sampler_textures[PIPE_MAX_SAMPLERS];
+   struct pipe_texture *fragment_sampler_textures[PIPE_MAX_SAMPLERS];
+   struct pipe_texture *vertex_sampler_textures[PIPE_MAX_VERTEX_SAMPLERS];
    
    unsigned num_vertex_buffers;
    struct pipe_vertex_buffer vertex_buffers[PIPE_MAX_ATTRIBS];
index 53a01891e128759346c06809ecc4fd183641b139..963774142112d37761a6bbae1fc678dc6f3f8a14 100644 (file)
@@ -30,6 +30,7 @@
 #include "pipe/p_format.h"
 #include "pipe/p_state.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_tile.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
@@ -423,7 +424,6 @@ dxt5_rgba_data[] = {
 
 static INLINE void 
 st_sample_dxt_pixel_block(enum pipe_format format, 
-                          const struct pipe_format_block *block,
                           uint8_t *raw,
                           float *rgba, unsigned rgba_stride, 
                           unsigned w, unsigned h)
@@ -462,21 +462,21 @@ st_sample_dxt_pixel_block(enum pipe_format format,
          for(ch = 0; ch < 4; ++ch)
             rgba[y*rgba_stride + x*4 + ch] = (float)(data[i].rgba[y*4*4 + x*4 + ch])/255.0f;
    
-   memcpy(raw, data[i].raw, block->size);
+   memcpy(raw, data[i].raw, util_format_get_blocksize(format));
 }
 
 
 static INLINE void 
 st_sample_generic_pixel_block(enum pipe_format format, 
-                              const struct pipe_format_block *block,
                               uint8_t *raw,
                               float *rgba, unsigned rgba_stride,
                               unsigned w, unsigned h)
 {
    unsigned i;
    unsigned x, y, ch;
+   int blocksize = util_format_get_blocksize(format);
    
-   for(i = 0; i < block->size; ++i)
+   for(i = 0; i < blocksize; ++i)
       raw[i] = (uint8_t)st_random();
    
    
@@ -503,7 +503,6 @@ st_sample_generic_pixel_block(enum pipe_format format,
  */
 void 
 st_sample_pixel_block(enum pipe_format format,
-                      const struct pipe_format_block *block,
                       void *raw,
                       float *rgba, unsigned rgba_stride,
                       unsigned w, unsigned h)
@@ -513,11 +512,11 @@ st_sample_pixel_block(enum pipe_format format,
    case PIPE_FORMAT_DXT1_RGBA:
    case PIPE_FORMAT_DXT3_RGBA:
    case PIPE_FORMAT_DXT5_RGBA:
-      st_sample_dxt_pixel_block(format, block, raw, rgba, rgba_stride, w, h);
+      st_sample_dxt_pixel_block(format, raw, rgba, rgba_stride, w, h);
       break;
 
    default:
-      st_sample_generic_pixel_block(format, block, raw, rgba, rgba_stride, w, h);
+      st_sample_generic_pixel_block(format, raw, rgba, rgba_stride, w, h);
       break;
    }
 }
@@ -528,8 +527,8 @@ st_sample_surface(struct st_surface *surface, float *rgba)
 {
    struct pipe_texture *texture = surface->texture;
    struct pipe_screen *screen = texture->screen;
-   unsigned width = texture->width[surface->level];
-   unsigned height = texture->height[surface->level];
+   unsigned width = u_minify(texture->width0, surface->level);
+   unsigned height = u_minify(texture->height0, surface->level);
    uint rgba_stride = width * 4;
    struct pipe_transfer *transfer;
    void *raw;
@@ -548,18 +547,23 @@ st_sample_surface(struct st_surface *surface, float *rgba)
 
    raw = screen->transfer_map(screen, transfer);
    if (raw) {
-      const struct pipe_format_block *block = &texture->block;
+      enum pipe_format format = texture->format;
       uint x, y;
+      int nblocksx = util_format_get_nblocksx(format, width);
+      int nblocksy = util_format_get_nblocksy(format, height);
+      int blockwidth = util_format_get_blockwidth(format);
+      int blockheight = util_format_get_blockheight(format);
+      int blocksize = util_format_get_blocksize(format);
 
-      for (y = 0; y < transfer->nblocksy; ++y) {
-         for (x = 0; x < transfer->nblocksx; ++x) {
-            st_sample_pixel_block(texture->format,
-                                  block,
-                                  (uint8_t *) raw + y * transfer->stride + x * block->size,
-                                  rgba + y * block->height * rgba_stride + x * block->width * 4,
+
+      for (y = 0; y < nblocksy; ++y) {
+         for (x = 0; x < nblocksx; ++x) {
+            st_sample_pixel_block(format,
+                                  (uint8_t *) raw + y * transfer->stride + x * blocksize,
+                                  rgba + y * blockheight * rgba_stride + x * blockwidth * 4,
                                   rgba_stride,
-                                  MIN2(block->width, width - x*block->width),
-                                  MIN2(block->height, height - y*block->height));
+                                  MIN2(blockwidth, width - x*blockwidth),
+                                  MIN2(blockheight, height - y*blockheight));
          }
       }
 
index 0a27083549fa6dc6281891ec95449a5659188b5b..888114d3021d899c76f564fb92f9cc26442dbedc 100644 (file)
@@ -35,7 +35,6 @@
 
 void 
 st_sample_pixel_block(enum pipe_format format,
-                      const struct pipe_format_block *block,
                       void *raw,
                       float *rgba, unsigned rgba_stride,
                       unsigned w, unsigned h);
index f0abd12e3dcad2127cfe396aabb27067645731a9..a3294e877a62217e938f63bb2d60f6719afbdb9f 100644 (file)
@@ -40,6 +40,7 @@
 #include "pipe/p_format.h"
 #include "pipe/p_context.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 #include "softpipe/sp_winsys.h"
@@ -157,16 +158,6 @@ st_softpipe_user_buffer_create(struct pipe_winsys *winsys,
 }
 
 
-/**
- * Round n up to next multiple.
- */
-static INLINE unsigned
-round_up(unsigned n, unsigned multiple)
-{
-   return (n + multiple - 1) & ~(multiple - 1);
-}
-
-
 static struct pipe_buffer *
 st_softpipe_surface_buffer_create(struct pipe_winsys *winsys,
                                   unsigned width, unsigned height,
@@ -176,13 +167,10 @@ st_softpipe_surface_buffer_create(struct pipe_winsys *winsys,
                                   unsigned *stride)
 {
    const unsigned alignment = 64;
-   struct pipe_format_block block;
-   unsigned nblocksx, nblocksy;
+   unsigned nblocksy;
 
-   pf_get_block(format, &block);
-   nblocksx = pf_get_nblocksx(&block, width);
-   nblocksy = pf_get_nblocksy(&block, height);
-   *stride = round_up(nblocksx * block.size, alignment);
+   nblocksy = util_format_get_nblocksy(format, height);
+   *stride = align(util_format_get_stride(format, width), alignment);
 
    return winsys->buffer_create(winsys, alignment,
                                 usage,
index 202ccfc35097fbf1e364c63fe34342cfa2012b05..b022d073fda815a77dae2be07fcf3bdf6df76a06 100755 (executable)
@@ -47,7 +47,7 @@ for name, value in globals().items():
         formats[value] = name
 
 def is_depth_stencil_format(format):
-    # FIXME: make and use binding to pf_is_depth_stencil
+    # FIXME: make and use binding to util_format_is_depth_or_stencil
     return format in (
         PIPE_FORMAT_Z32_UNORM,
         PIPE_FORMAT_Z24S8_UNORM,
index 7a0006bf660006bc55069dd3a83b68ddc1b2895c..103d7497f48fa922626a93a85957a569d5503e57 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index f7836c85ddba71199de9b1d03c20c8bc98eb80ee..bcb942059639f20246dff215e773ab86084a9461 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index c89cd748a80655907f38057337adcfeba6b96525..b5281975d4a60c3589558ce09818cb30bfc34057 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 6517e3c494572aeb7003b852c08e8b5026ddafd2..d59df76e70b3e2a6d3ae259e622369086dd90130 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 464880ba68f18a03312589744c76fa83d3f561b6..fbb20fa9f6268f47449f5c44c565cc2b6e1d7025 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 2684076f1d61884a449ea9d9da86c7bc4da529da..b511288f4b6b4c4cd2b4f99456a0b7bdb84e4dec 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index ad11e28918abca141d539937d6ffc30ad0722cd0..99a2f96103ac2c67569fd51e5841321197832470 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 4f3aa30d666fbf65444d33d1765d3e467b830918..a54c2623b0a47f2c1ac1a01092cb0202623475a3 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 54c7c64459867a46b617a2d1655fa785ad9ec581..5f5b4be1092c6c10fdc18c48ab570104e65ca370 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 0e78ef86b54cbca46727178eae0abeefebc5befb..6323c4712dc0fe37fad6d6fa18f9b0301232467c 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index e9ee0f81478a84203f214e2830633d8dd7599bd6..740809d22e001844b8fbc2e58b94187474697f0f 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 439acd5bbdec6dfa5756c9f12969fd5fd72e2496..413b9dc39164bb6e8970531a48ffd6d0a62d862d 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index ab21b245dd8cf7c5a126d9e4d9c257f96f81b1d2..b69f2132612a9fea7c9fbaafa2c29709965bfd60 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 969ae73d98efea710483edbdd6759eba47455658..df284f49e71ea7c8f1256637538b91909b60efe1 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 612975e05709422a3081f6bded5da33e90a4ce1e..64af72f381ba5896c1cf9c906d28cbfd190493f4 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index ed158b0fc6923fc564b90f850e94936deb46bcaf..bdd0b0026b9a01a38047791540289bde00e94f9d 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index cc9feef07e1398f8c0b573c73587f5898e0b3f70..f4b611b26abf7803db3fbd1791c0da0ffb75034b 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 695621fdc97c88fe9f7a807208e7bc8640baed8c..d1e9b0b53be4b0c09364bc5eb996fe1b7aee5c46 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 9505bc3c3e32c4b9ec6a91ef4ca072a9cd007c88..1f33fac472789a007e91299214452ccb51d78b29 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index 83441fa8207fb4b2f6b5208581b5f694c8d3b177..673fca139aa93868570e6ef34be4a6a86a2a0c92 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index d6f66c4927ab00453ad7c82b298e12fb09199eeb..6ec8b1184cc37bf9f7d2e712bda85d87708ccdf3 100644 (file)
@@ -1,4 +1,4 @@
-FRAG1.1
+FRAG
 
 DCL IN[0], COLOR, LINEAR
 DCL OUT[0], COLOR
index d60fb38d1ae6c28201341c64a26d82baf2fccc9a..eed6cdd1e6460be4771991ee5463d7d95a949bed 100644 (file)
@@ -96,7 +96,7 @@ def test(dev, name):
     sampler.min_img_filter = PIPE_TEX_MIPFILTER_NEAREST
     sampler.mag_img_filter = PIPE_TEX_MIPFILTER_NEAREST
     sampler.normalized_coords = 1
-    ctx.set_sampler(0, sampler)
+    ctx.set_fragment_sampler(0, sampler)
 
     # scissor
     scissor = Scissor()
@@ -131,7 +131,7 @@ def test(dev, name):
 
     # vertex shader
     vs = Shader('''
-        VERT1.1
+        VERT
         DCL IN[0], POSITION
         DCL IN[1], COLOR
         DCL OUT[0], POSITION
index 7638e96346e1ad8e2390ca8211a9ada22d9f8782..321140e89e8f77469be6efe20752ef81fbac5595 100644 (file)
@@ -1,4 +1,4 @@
-VERT1.1
+VERT
 
 DCL IN[0], POSITION
 DCL OUT[0], POSITION
index 28ce6f9a0cf03330fa789df4f4350563371a9baf..d60ea46b3672a080ae4da43c6d6e38ba8d94d3f0 100644 (file)
@@ -1,4 +1,4 @@
-VERT1.1
+VERT
 
 DCL IN[0], POSITION
 DCL OUT[0], POSITION
index 44ad708119decd052382f7cbc5e320cc787b0732..aa80d6e394c4f39b5f4b9e5522e6c065936487ba 100644 (file)
@@ -1,4 +1,4 @@
-VERT1.1
+VERT
 
 DCL IN[0], POSITION
 DCL OUT[0], POSITION
index d179749de84acb6d8f30705a814cb1d62f9c8543..64d1a494e13f46e637865ae4fa39080354e886a8 100644 (file)
@@ -1,4 +1,4 @@
-VERT1.1
+VERT
 
 DCL IN[0], POSITION
 DCL OUT[0], POSITION
index 472769f259232f1d50de2aa206caffacb7847cf6..41bebd0604a423029f033c06b7ca2c29d4d25213 100644 (file)
@@ -96,7 +96,7 @@ def test(dev, name):
     sampler.min_img_filter = PIPE_TEX_MIPFILTER_NEAREST
     sampler.mag_img_filter = PIPE_TEX_MIPFILTER_NEAREST
     sampler.normalized_coords = 1
-    ctx.set_sampler(0, sampler)
+    ctx.set_fragment_sampler(0, sampler)
 
     # scissor
     scissor = Scissor()
@@ -135,7 +135,7 @@ def test(dev, name):
 
     # fragment shader
     fs = Shader('''
-        FRAG1.1
+        FRAG
         DCL IN[0], COLOR, LINEAR
         DCL OUT[0], COLOR, CONSTANT
         0:MOV OUT[0], IN[0]
index 3ceecbbd3aa6deb0354adbca38bc31805fdaa4ed..df5babb78af0f91cc512977a3a11ddca31a232b2 100755 (executable)
@@ -98,9 +98,10 @@ class TextureTest(TestCase):
         y = 0
         w = dst_surface.width
         h = dst_surface.height
-    
-        stride = dst_surface.nblocksx * dst_texture.block.size
-        size = dst_surface.nblocksy * stride
+
+        # ???
+        stride = pf_get_stride(texture->format, w)
+        size = pf_get_nblocksy(texture->format) * stride
         src_raw = os.urandom(size)
 
         src_surface.put_tile_raw(0, 0, w, h, src_raw, stride)
index 0b76932b6ede275d521235a551932b80a1e08db3..79287f2caceaa115907cd1b9d7a588ea71d13598 100755 (executable)
@@ -144,8 +144,8 @@ class TextureTest(TestCase):
         sampler.normalized_coords = 1
         sampler.min_lod = 0
         sampler.max_lod = PIPE_MAX_TEXTURE_LEVELS - 1
-        ctx.set_sampler(0, sampler)
-        ctx.set_sampler_texture(0, src_texture)
+        ctx.set_fragment_sampler(0, sampler)
+        ctx.set_fragment_sampler_texture(0, src_texture)
 
         #  framebuffer 
         cbuf_tex = dev.texture_create(
@@ -171,7 +171,7 @@ class TextureTest(TestCase):
     
         # vertex shader
         vs = Shader('''
-            VERT1.1
+            VERT
             DCL IN[0], POSITION, CONSTANT
             DCL IN[1], GENERIC, CONSTANT
             DCL OUT[0], POSITION, CONSTANT
@@ -185,7 +185,7 @@ class TextureTest(TestCase):
     
         # fragment shader
         fs = Shader('''
-            FRAG1.1
+            FRAG
             DCL IN[0], GENERIC[0], LINEAR
             DCL OUT[0], COLOR, CONSTANT
             DCL SAMP[0], CONSTANT
index c7b78abbbec4de163a24a25931260359d2832058..520961c8051ef9e98f002c9312827dbfe1bab747 100755 (executable)
@@ -169,7 +169,7 @@ class TextureColorSampleTest(TestCase):
         sampler.normalized_coords = 1
         sampler.min_lod = 0
         sampler.max_lod = PIPE_MAX_TEXTURE_LEVELS - 1
-        ctx.set_sampler(0, sampler)
+        ctx.set_fragment_sampler(0, sampler)
     
         #  texture 
         texture = dev.texture_create(
@@ -189,7 +189,7 @@ class TextureColorSampleTest(TestCase):
             zslice = zslice,
         ).sample_rgba(expected_rgba)
         
-        ctx.set_sampler_texture(0, texture)
+        ctx.set_fragment_sampler_texture(0, texture)
 
         #  framebuffer 
         cbuf_tex = dev.texture_create(
@@ -216,7 +216,7 @@ class TextureColorSampleTest(TestCase):
     
         # vertex shader
         vs = Shader('''
-            VERT1.1
+            VERT
             DCL IN[0], POSITION, CONSTANT
             DCL IN[1], GENERIC, CONSTANT
             DCL OUT[0], POSITION, CONSTANT
@@ -236,7 +236,7 @@ class TextureColorSampleTest(TestCase):
             PIPE_TEXTURE_CUBE: "CUBE",
         }[target]
         fs = Shader('''
-            FRAG1.1
+            FRAG
             DCL IN[0], GENERIC[0], LINEAR
             DCL OUT[0], COLOR, CONSTANT
             DCL SAMP[0], CONSTANT
@@ -359,7 +359,7 @@ class TextureDepthSampleTest(TestCase):
         sampler.normalized_coords = 1
         sampler.min_lod = 0
         sampler.max_lod = PIPE_MAX_TEXTURE_LEVELS - 1
-        ctx.set_sampler(0, sampler)
+        ctx.set_fragment_sampler(0, sampler)
     
         #  texture 
         texture = dev.texture_create(
@@ -379,7 +379,7 @@ class TextureDepthSampleTest(TestCase):
             zslice = zslice,
         ).sample_rgba(expected_rgba)
         
-        ctx.set_sampler_texture(0, texture)
+        ctx.set_fragment_sampler_texture(0, texture)
 
         #  framebuffer 
         cbuf_tex = dev.texture_create(
@@ -415,7 +415,7 @@ class TextureDepthSampleTest(TestCase):
     
         # vertex shader
         vs = Shader('''
-            VERT1.1
+            VERT
             DCL IN[0], POSITION, CONSTANT
             DCL IN[1], GENERIC, CONSTANT
             DCL OUT[0], POSITION, CONSTANT
@@ -435,7 +435,7 @@ class TextureDepthSampleTest(TestCase):
             PIPE_TEXTURE_CUBE: "CUBE",
         }[target]
         fs = Shader('''
-            FRAG1.1
+            FRAG
             DCL IN[0], GENERIC[0], LINEAR
             DCL SAMP[0], CONSTANT
             DCL OUT[0].z, POSITION
index e65b425adf4e28429daa59240e84afe6c6cbfb41..35daca9e498990f726cedf3f6c4cf2c9011e3c83 100755 (executable)
@@ -86,8 +86,9 @@ class TextureTest(TestCase):
         
         surface = texture.get_surface(face, level, zslice)
         
-        stride = surface.nblocksx * texture.block.size
-        size = surface.nblocksy * stride
+        # ???
+        stride = pf_get_stride(texture->format, w)
+        size = pf_get_nblocksy(texture->format) * stride
 
         in_raw = os.urandom(size)
 
index 862cbb03c439673c570105b285ea39ad3656379e..2f984fb7b9afa012d4c6ea47ad9851bcf68512ef 100644 (file)
@@ -38,6 +38,7 @@
 #include "pipe/p_screen.h"
 #include "pipe/p_shader_tokens.h"
 
+#include "util/u_format.h"
 #include "util/u_memory.h"
 
 
@@ -68,10 +69,9 @@ static INLINE struct pipe_texture *create_texture_1d(struct vg_context *ctx,
    templ.target = PIPE_TEXTURE_1D;
    templ.format = PIPE_FORMAT_A8R8G8B8_UNORM;
    templ.last_level = 0;
-   templ.width[0] = color_data_len;
-   templ.height[0] = 1;
-   templ.depth[0] = 1;
-   pf_get_block(PIPE_FORMAT_A8R8G8B8_UNORM, &templ.block);
+   templ.width0 = color_data_len;
+   templ.height0 = 1;
+   templ.depth0 = 1;
    templ.tex_usage = PIPE_TEXTURE_USAGE_SAMPLER;
 
    tex = screen->texture_create(screen, &templ);
@@ -81,7 +81,7 @@ static INLINE struct pipe_texture *create_texture_1d(struct vg_context *ctx,
          screen->get_tex_transfer(screen, tex,
                                   0, 0, 0,
                                   PIPE_TRANSFER_READ_WRITE ,
-                                  0, 0, tex->width[0], tex->height[0]);
+                                  0, 0, tex->width0, tex->height0);
       void *map = screen->transfer_map(screen, transfer);
       memcpy(map, color_data, sizeof(VGint)*color_data_len);
       screen->transfer_unmap(screen, transfer);
index 9a49f2e12d0139bf3eb0167fc7b521b8e0ec30b7..60bed197e28c5aad3ecc74d90c97ff6153f84a18 100644 (file)
@@ -28,7 +28,7 @@
 #define ASM_FILTERS_H
 
 static const char color_matrix_asm[] =
-   "FRAG1.1\n"
+   "FRAG\n"
    "DCL IN[0], GENERIC[0], PERSPECTIVE\n"
    "DCL OUT[0], COLOR, CONSTANT\n"
    "DCL CONST[0..4], CONSTANT\n"
@@ -51,7 +51,7 @@ static const char color_matrix_asm[] =
    "END\n";
 
 static const char convolution_asm[] =
-   "FRAG1.1\n"
+   "FRAG\n"
    "DCL IN[0], GENERIC[0], PERSPECTIVE\n"
    "DCL OUT[0], COLOR, CONSTANT\n"
    "DCL TEMP[0..4], CONSTANT\n"
@@ -78,7 +78,7 @@ static const char convolution_asm[] =
 
 
 static const char lookup_asm[] =
-   "FRAG1.1\n"
+   "FRAG\n"
    "DCL IN[0], GENERIC[0], PERSPECTIVE\n"
    "DCL OUT[0], COLOR, CONSTANT\n"
    "DCL TEMP[0..2], CONSTANT\n"
@@ -103,7 +103,7 @@ static const char lookup_asm[] =
 
 
 static const char lookup_single_asm[] =
-   "FRAG1.1\n"
+   "FRAG\n"
    "DCL IN[0], GENERIC[0], PERSPECTIVE\n"
    "DCL OUT[0], COLOR, CONSTANT\n"
    "DCL TEMP[0..2], CONSTANT\n"
index 218e1d166db8c7e8e4c5b6f1f98a557e03dc532c..903bfc88a4d2fdd39757bc01c3be8ad9fb5c249b 100644 (file)
@@ -29,7 +29,7 @@
 
 
 static const char pass_through_depth_asm[] =
-   "FRAG1.1\n"
+   "FRAG\n"
    "DCL IN[0], POSITION, LINEAR\n"
    "DCL OUT[0].z, POSITION, CONSTANT\n"
    "0: MOV OUT[0].z, IN[0].zzzz\n"
@@ -39,7 +39,7 @@ static const char pass_through_depth_asm[] =
 
 /* Î¼new = Î¼mask */
 static const char set_mask_asm[] =
-   "FRAG1.1\n"
+   "FRAG\n"
    "DCL IN[0], GENERIC[0], PERSPECTIVE\n"
    "DCL SAMP[0], CONSTANT\n"
    "DCL OUT[0], COLOR, CONSTANT\n"
@@ -48,7 +48,7 @@ static const char set_mask_asm[] =
 
 /* Î¼new = 1 â€“ (1 â€“ Î¼mask)*(1 â€“ Î¼prev) */
 static const char union_mask_asm[] =
-   "FRAG1.1\n"
+   "FRAG\n"
    "DCL IN[0], GENERIC[0], PERSPECTIVE\n"
    "DCL IN[1], POSITION, LINEAR\n"
    "DCL CONST[0], CONSTANT\n"
@@ -65,7 +65,7 @@ static const char union_mask_asm[] =
 
 /* Î¼new = Î¼mask *μprev */
 static const char intersect_mask_asm[] =
-   "FRAG1.1\n"
+   "FRAG\n"
    "DCL IN[0], GENERIC[0], PERSPECTIVE\n"
    "DCL IN[1], POSITION, LINEAR\n"
    "DCL CONST[0], CONSTANT\n"
@@ -79,7 +79,7 @@ static const char intersect_mask_asm[] =
 
 /* Î¼new = Î¼prev*(1 â€“ Î¼mask) */
 static const char subtract_mask_asm[] =
-   "FRAG1.1\n"
+   "FRAG\n"
    "DCL IN[0], GENERIC[0], PERSPECTIVE\n"
    "DCL IN[1], POSITION, LINEAR\n"
    "DCL CONST[0], CONSTANT\n"
@@ -94,7 +94,7 @@ static const char subtract_mask_asm[] =
 
 
 static const char vs_plain_asm[] =
-   "VERT1.1\n"
+   "VERT\n"
    "DCL IN[0]\n"
    "DCL OUT[0], POSITION\n"
    "DCL TEMP[0]\n"
@@ -105,7 +105,7 @@ static const char vs_plain_asm[] =
    "3: END\n";
 
 static const char vs_clear_asm[] =
-   "VERT1.1\n"
+   "VERT\n"
    "DCL IN[0]\n"
    "DCL IN[1]\n"
    "DCL OUT[0], POSITION\n"
@@ -120,7 +120,7 @@ static const char vs_clear_asm[] =
 
 
 static const char vs_texture_asm[] =
-   "VERT1.1\n"
+   "VERT\n"
    "DCL IN[0]\n"
    "DCL IN[1]\n"
    "DCL OUT[0], POSITION\n"
index 9a722980d522e4881af6075316e1a10271c0eac7..278ba6d46eb2cdff8a531936e68594ba05c865f5 100644 (file)
@@ -39,6 +39,7 @@
 #include "pipe/p_screen.h"
 #include "pipe/p_inlines.h"
 #include "util/u_blit.h"
+#include "util/u_format.h"
 #include "util/u_tile.h"
 #include "util/u_memory.h"
 #include "util/u_math.h"
@@ -93,8 +94,8 @@ static void vg_copy_texture(struct vg_context *ctx,
    dst_loc[3] = height;
    dst_bounds[0] = 0.f;
    dst_bounds[1] = 0.f;
-   dst_bounds[2] = dst->width[0];
-   dst_bounds[3] = dst->height[0];
+   dst_bounds[2] = dst->width0;
+   dst_bounds[3] = dst->height0;
 
    src_loc[0] = sx;
    src_loc[1] = sy;
@@ -102,8 +103,8 @@ static void vg_copy_texture(struct vg_context *ctx,
    src_loc[3] = height;
    src_bounds[0] = 0.f;
    src_bounds[1] = 0.f;
-   src_bounds[2] = src->width[0];
-   src_bounds[3] = src->height[0];
+   src_bounds[2] = src->width0;
+   src_bounds[3] = src->height0;
 
    vg_bound_rect(src_loc, src_bounds, src_shift);
    vg_bound_rect(dst_loc, dst_bounds, dst_shift);
@@ -270,11 +271,10 @@ struct vg_image * image_create(VGImageFormat format,
    memset(&pt, 0, sizeof(pt));
    pt.target = PIPE_TEXTURE_2D;
    pt.format = pformat;
-   pf_get_block(pformat, &pt.block);
    pt.last_level = 0;
-   pt.width[0] = width;
-   pt.height[0] = height;
-   pt.depth[0] = 1;
+   pt.width0 = width;
+   pt.height0 = height;
+   pt.depth0 = 1;
    pt.tex_usage = PIPE_TEXTURE_USAGE_SAMPLER;
 
    newtex = screen->texture_create(screen, &pt);
@@ -414,7 +414,7 @@ void image_sub_data(struct vg_image *image,
    { /* upload color_data */
       struct pipe_transfer *transfer = screen->get_tex_transfer(
          screen, texture, 0, 0, 0,
-         PIPE_TRANSFER_WRITE, 0, 0, texture->width[0], texture->height[0]);
+         PIPE_TRANSFER_WRITE, 0, 0, texture->width0, texture->height0);
       src += (dataStride * yoffset);
       for (i = 0; i < height; i++) {
          _vega_unpack_float_span_rgba(ctx, width, xoffset, src, dataFormat, temp);
index 24650a37d502f30a8012948387934a8b9c3ad797..42300bb6d573fdbd0c6d5be10a4a7b847ce10ac8 100644 (file)
@@ -36,6 +36,7 @@
 #include "pipe/p_context.h"
 #include "pipe/p_screen.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_memory.h"
 
 struct vg_mask_layer {
@@ -426,7 +427,7 @@ static void mask_using_texture(struct pipe_texture *texture,
    if (!surface)
       return;
    if (!intersect_rectangles(surface->width, surface->height,
-                             texture->width[0], texture->height[0],
+                             texture->width0, texture->height0,
                              x, y, width, height,
                              offsets, loc))
       return;
@@ -491,11 +492,10 @@ struct vg_mask_layer * mask_layer_create(VGint width, VGint height)
       memset(&pt, 0, sizeof(pt));
       pt.target = PIPE_TEXTURE_2D;
       pt.format = PIPE_FORMAT_A8R8G8B8_UNORM;
-      pf_get_block(PIPE_FORMAT_A8R8G8B8_UNORM, &pt.block);
       pt.last_level = 0;
-      pt.width[0] = width;
-      pt.height[0] = height;
-      pt.depth[0] = 1;
+      pt.width0 = width;
+      pt.height0 = height;
+      pt.depth0 = 1;
       pt.tex_usage = PIPE_TEXTURE_USAGE_SAMPLER;
       pt.compressed = 0;
 
@@ -607,8 +607,8 @@ void mask_render_to(struct path *path,
    struct vg_mask_layer *temp_layer;
    VGint width, height;
 
-   width = fb_buffers->alpha_mask->width[0];
-   height = fb_buffers->alpha_mask->width[0];
+   width = fb_buffers->alpha_mask->width0;
+   height = fb_buffers->alpha_mask->width0;
 
    temp_layer = mask_layer_create(width, height);
 
index 04a6ba9cdcd307b4aff905b37c7d186d84c87a8f..cc73771d358704ef24014e4a3da03fc4db9ceca0 100644 (file)
@@ -34,6 +34,7 @@
 #include "pipe/p_compiler.h"
 #include "pipe/p_inlines.h"
 
+#include "util/u_format.h"
 #include "util/u_memory.h"
 #include "util/u_math.h"
 
@@ -151,10 +152,9 @@ static INLINE struct pipe_texture *create_gradient_texture(struct vg_paint *p)
    templ.target = PIPE_TEXTURE_1D;
    templ.format = PIPE_FORMAT_A8R8G8B8_UNORM;
    templ.last_level = 0;
-   templ.width[0] = 1024;
-   templ.height[0] = 1;
-   templ.depth[0] = 1;
-   pf_get_block(PIPE_FORMAT_A8R8G8B8_UNORM, &templ.block);
+   templ.width0 = 1024;
+   templ.height0 = 1;
+   templ.depth0 = 1;
    templ.tex_usage = PIPE_TEXTURE_USAGE_SAMPLER;
 
    tex = screen->texture_create(screen, &templ);
@@ -328,8 +328,8 @@ static INLINE void  paint_pattern_buffer(struct vg_paint *paint, void *buffer)
 
    map[4] = 0.f;
    map[5] = 1.f;
-   map[6] = paint->pattern.texture->width[0];
-   map[7] = paint->pattern.texture->height[0];
+   map[6] = paint->pattern.texture->width0;
+   map[7] = paint->pattern.texture->height0;
    {
       struct matrix mat;
       memcpy(&mat, &ctx->state.vg.fill_paint_to_user_matrix,
index 396c88aa3d25430aadd1927017c483b7f42152cf..64e3a7c5453573189e36202573ebb0bf9b992088 100644 (file)
@@ -35,6 +35,7 @@
 #include "pipe/p_shader_tokens.h"
 
 #include "util/u_draw_quad.h"
+#include "util/u_format.h"
 #include "util/u_simple_shaders.h"
 #include "util/u_memory.h"
 #include "util/u_rect.h"
@@ -56,7 +57,7 @@ static void setup_shaders(struct renderer *ctx)
 {
    struct pipe_context *pipe = ctx->pipe;
    /* fragment shader */
-   ctx->fs = util_make_fragment_tex_shader(pipe);
+   ctx->fs = util_make_fragment_tex_shader(pipe, TGSI_TEXTURE_2D);
 }
 
 static struct pipe_buffer *
@@ -230,13 +231,13 @@ void renderer_draw_texture(struct renderer *r,
    struct pipe_buffer *buf;
    VGfloat s0, t0, s1, t1;
 
-   assert(tex->width[0] != 0);
-   assert(tex->height[0] != 0);
+   assert(tex->width0 != 0);
+   assert(tex->height0 != 0);
 
-   s0 = x1offset / tex->width[0];
-   s1 = x2offset / tex->width[0];
-   t0 = y1offset / tex->height[0];
-   t1 = y2offset / tex->height[0];
+   s0 = x1offset / tex->width0;
+   s1 = x2offset / tex->width0;
+   t0 = y1offset / tex->height0;
+   t1 = y2offset / tex->height0;
 
    cso_save_vertex_shader(r->cso);
    /* shaders */
@@ -276,10 +277,10 @@ void renderer_copy_texture(struct renderer *ctx,
    struct pipe_framebuffer_state fb;
    float s0, t0, s1, t1;
 
-   assert(src->width[0] != 0);
-   assert(src->height[0] != 0);
-   assert(dst->width[0] != 0);
-   assert(dst->height[0] != 0);
+   assert(src->width0 != 0);
+   assert(src->height0 != 0);
+   assert(dst->width0 != 0);
+   assert(dst->height0 != 0);
 
 #if 0
    debug_printf("copy texture [%f, %f, %f, %f], [%f, %f, %f, %f]\n",
@@ -287,10 +288,10 @@ void renderer_copy_texture(struct renderer *ctx,
 #endif
 
 #if 1
-   s0 = sx1 / src->width[0];
-   s1 = sx2 / src->width[0];
-   t0 = sy1 / src->height[0];
-   t1 = sy2 / src->height[0];
+   s0 = sx1 / src->width0;
+   s1 = sx2 / src->width0;
+   t0 = sy1 / src->height0;
+   t1 = sy2 / src->height0;
 #else
    s0 = 0;
    s1 = 1;
@@ -445,10 +446,9 @@ void renderer_copy_surface(struct renderer *ctx,
    texTemp.target = PIPE_TEXTURE_2D;
    texTemp.format = src->format;
    texTemp.last_level = 0;
-   texTemp.width[0] = srcW;
-   texTemp.height[0] = srcH;
-   texTemp.depth[0] = 1;
-   pf_get_block(src->format, &texTemp.block);
+   texTemp.width0 = srcW;
+   texTemp.height0 = srcH;
+   texTemp.depth0 = 1;
 
    tex = screen->texture_create(screen, &texTemp);
    if (!tex)
@@ -570,13 +570,13 @@ void renderer_texture_quad(struct renderer *r,
    struct pipe_buffer *buf;
    VGfloat s0, t0, s1, t1;
 
-   assert(tex->width[0] != 0);
-   assert(tex->height[0] != 0);
+   assert(tex->width0 != 0);
+   assert(tex->height0 != 0);
 
-   s0 = x1offset / tex->width[0];
-   s1 = x2offset / tex->width[0];
-   t0 = y1offset / tex->height[0];
-   t1 = y2offset / tex->height[0];
+   s0 = x1offset / tex->width0;
+   s1 = x2offset / tex->width0;
+   t0 = y1offset / tex->height0;
+   t1 = y2offset / tex->height0;
 
    cso_save_vertex_shader(r->cso);
    /* shaders */
index fd0831fab1f70ee27b598b975f4c03ee9d342c46..f620075d0bc95a34e3a78a6298494a151e25f8d5 100644 (file)
@@ -97,7 +97,7 @@ static INLINE struct tgsi_token *tokens_from_assembly(const char *txt, int num_t
 
 /*
 static const char max_shader_preamble[] =
-   "FRAG1.1\n"
+   "FRAG\n"
    "DCL IN[0], POSITION, LINEAR\n"
    "DCL IN[1], GENERIC[0], PERSPECTIVE\n"
    "DCL OUT[0], COLOR, CONSTANT\n"
@@ -168,7 +168,7 @@ create_preamble(char *txt,
    --end_temp;
    --end_sampler;
 
-   sprintf(txt, "FRAG1.1\n");
+   sprintf(txt, "FRAG\n");
 
    if (declare_input) {
       sprintf(txt + strlen(txt), "DCL IN[0], POSITION, LINEAR\n");
index c4da01e52ccbb07c721b36acd1417557406a89f5..e5039132758c8893d7b0089d8336ca8ee73245fd 100644 (file)
 #include "pipe/p_context.h"
 #include "pipe/p_inlines.h"
 #include "pipe/p_screen.h"
+#include "util/u_format.h"
 #include "util/u_memory.h"
 #include "util/u_math.h"
+#include "util/u_rect.h"
 
 static struct pipe_texture *
 create_texture(struct pipe_context *pipe, enum pipe_format format,
@@ -50,13 +52,12 @@ create_texture(struct pipe_context *pipe, enum pipe_format format,
    }
 
    templ.target = PIPE_TEXTURE_2D;
-   pf_get_block(templ.format, &templ.block);
-   templ.width[0] = width;
-   templ.height[0] = height;
-   templ.depth[0] = 1;
+   templ.width0 = width;
+   templ.height0 = height;
+   templ.depth0 = 1;
    templ.last_level = 0;
 
-   if (pf_get_component_bits(format, PIPE_FORMAT_COMP_S)) {
+   if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_ZS, 1)) {
       templ.tex_usage = PIPE_TEXTURE_USAGE_DEPTH_STENCIL;
    } else {
       templ.tex_usage = (PIPE_TEXTURE_USAGE_DISPLAY_TARGET |
index 00e07647062aa15aa978fe8f09d7938dd335180f..03575ca3ddd56da66a3b1847071f69e9ae55e190 100644 (file)
@@ -474,6 +474,7 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
                                   VGfloat rgba[][4])
 {
    VGint i;
+   union util_color uc;
 
    switch (dataFormat) {
    case VG_sRGBX_8888: {
@@ -486,8 +487,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          b = (*src >>  8) & 0xff;
          a = 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
    }
@@ -502,8 +506,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          b = (*src >>  8) & 0xff;
          a = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -519,8 +526,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          b = (*src >>  8) & 0xff;
          a = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -536,8 +546,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          clr[2] = ((*src >>  0) & 31)/31.;
          clr[3] = 1.f;
 
-         util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                         rgba[i]);
+         util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
    }
@@ -552,8 +565,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          clr[2] = ((*src >>  1) & 31)/31.;
          clr[3] = ((*src >>  0) & 1)/1.;
 
-         util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                         rgba[i]);
+         util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
    }
@@ -568,8 +584,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          clr[2] = ((*src >>  4) & 15)/15.;
          clr[3] = ((*src >>  0) & 15)/15.;
 
-         util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                         rgba[i]);
+         util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
    }
@@ -578,8 +597,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
       VGubyte *src = (VGubyte *)data;
       src += offset;
       for (i = 0; i < n; ++i) {
-         util_pack_color_ub(0xff, 0xff, 0xff, *src, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                         rgba[i]);
+         util_pack_color_ub(0xff, 0xff, 0xff, *src, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
    }
@@ -594,8 +616,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          b = (*src >>  8) & 0xff;
          a = 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
    }
@@ -610,8 +635,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          b = (*src >>  8) & 0xff;
          a = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -627,8 +655,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          b = (*src >>  8) & 0xff;
          a = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -638,8 +669,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
       VGubyte *src = (VGubyte *)data;
       src += offset;
       for (i = 0; i < n; ++i) {
-         util_pack_color_ub(0xff, 0xff, 0xff, *src, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                         rgba[i]);
+         util_pack_color_ub(0xff, 0xff, 0xff, *src, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
    }
@@ -648,8 +682,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
       VGubyte *src = (VGubyte *)data;
       src += offset;
       for (i = 0; i < n; ++i) {
-         util_pack_color_ub(0xff, 0xff, 0xff, *src, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(0xff, 0xff, 0xff, *src, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
    }
@@ -667,8 +704,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
             clr[2] = clr[0];
             clr[3] = 1.f;
 
-            util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i+j]);
+            util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+            rgba[i+j][0] = uc.f[0];
+            rgba[i+j][1] = uc.f[1];
+            rgba[i+j][2] = uc.f[2];
+            rgba[i+j][3] = uc.f[3];
          }
          ++src;
       }
@@ -688,8 +728,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
             clr[2] = 0.f;
             clr[3] = (((*src) & (1<<shift)) >> shift);
 
-            util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i+j]);
+            util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+            rgba[i+j][0] = uc.f[0];
+            rgba[i+j][1] = uc.f[1];
+            rgba[i+j][2] = uc.f[2];
+            rgba[i+j][3] = uc.f[3];
          }
          ++src;
       }
@@ -715,8 +758,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
             clr[2] = 0.f;
             clr[3] = ((*src) & (bitter)) >> shift;
 
-            util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i +j]);
+            util_pack_color(clr, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+            rgba[i+j][0] = uc.f[0];
+            rgba[i+j][1] = uc.f[1];
+            rgba[i+j][2] = uc.f[2];
+            rgba[i+j][3] = uc.f[3];
          }
          ++src;
       }
@@ -735,8 +781,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          g = (*src >>  8) & 0xff;
          b = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -752,8 +801,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          g = (*src >>  8) & 0xff;
          b = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -775,8 +827,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          g = (*src >>  8) & 0xff;
          b = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -792,8 +847,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          g = (*src >>  8) & 0xff;
          b = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -811,8 +869,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          r = (*src >>  8) & 0xff;
          a = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -828,8 +889,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          r = (*src >>  8) & 0xff;
          a = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -853,8 +917,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          r = (*src >>  8) & 0xff;
          a = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -870,8 +937,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          r = (*src >>  8) & 0xff;
          a = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -889,8 +959,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          g = (*src >>  8) & 0xff;
          r = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -906,8 +979,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          g = (*src >>  8) & 0xff;
          r = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -929,8 +1005,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          g = (*src >>  8) & 0xff;
          r = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
@@ -946,8 +1025,11 @@ void _vega_unpack_float_span_rgba(struct vg_context *ctx,
          g = (*src >>  8) & 0xff;
          r = (*src >>  0) & 0xff;
 
-         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT,
-                            rgba[i]);
+         util_pack_color_ub(r, g, b, a, PIPE_FORMAT_R32G32B32A32_FLOAT, &uc);
+         rgba[i][0] = uc.f[0];
+         rgba[i][1] = uc.f[1];
+         rgba[i][2] = uc.f[2];
+         rgba[i][3] = uc.f[3];
          ++src;
       }
       return;
index 6d095019815395a7899ebe333690ac396ac2be10..129a6298a7771a97b4086faf09740f007731ea1a 100644 (file)
@@ -30,6 +30,7 @@
 #include "main/context.h"
 #include "pipe/p_format.h"
 #include "pipe/p_screen.h"
+#include "util/u_format.h"
 #include "state_tracker/st_context.h"
 #include "state_tracker/st_public.h"
 
@@ -267,15 +268,13 @@ stw_framebuffer_allocate(
       enum pipe_format colorFormat, depthFormat, stencilFormat;
 
       colorFormat = pfi->color_format;
-      
-      assert(pf_layout( pfi->depth_stencil_format ) == PIPE_FORMAT_LAYOUT_RGBAZS );
-   
-      if(pf_get_component_bits( pfi->depth_stencil_format, PIPE_FORMAT_COMP_Z ))
+
+      if(util_format_get_component_bits(pfi->depth_stencil_format, UTIL_FORMAT_COLORSPACE_ZS, 0))
          depthFormat = pfi->depth_stencil_format;
       else
          depthFormat = PIPE_FORMAT_NONE;
    
-      if(pf_get_component_bits( pfi->depth_stencil_format, PIPE_FORMAT_COMP_S ))
+      if(util_format_get_component_bits(pfi->depth_stencil_format, UTIL_FORMAT_COLORSPACE_ZS, 1))
          stencilFormat = pfi->depth_stencil_format;
       else
          stencilFormat = PIPE_FORMAT_NONE;
index 7abe5d9f7fa19a9cadaefc1dcccc31f6d295198b..54cc36141290385106d9d9d52e9b90ef641d6d57 100644 (file)
@@ -32,6 +32,7 @@
 #include "pipe/p_defines.h"
 #include "pipe/p_screen.h"
 
+#include "util/u_format.h"
 #include "util/u_debug.h"
 
 #include "stw_icd.h"
@@ -132,14 +133,12 @@ stw_pixelformat_add(
    if(stw_dev->pixelformat_extended_count >= STW_MAX_PIXELFORMATS)
       return;
 
-   assert(pf_layout( color->format ) == PIPE_FORMAT_LAYOUT_RGBAZS );
-   assert(pf_get_component_bits( color->format, PIPE_FORMAT_COMP_R ) == color->bits.red );
-   assert(pf_get_component_bits( color->format, PIPE_FORMAT_COMP_G ) == color->bits.green );
-   assert(pf_get_component_bits( color->format, PIPE_FORMAT_COMP_B ) == color->bits.blue );
-   assert(pf_get_component_bits( color->format, PIPE_FORMAT_COMP_A ) == color->bits.alpha );
-   assert(pf_layout( depth->format ) == PIPE_FORMAT_LAYOUT_RGBAZS );
-   assert(pf_get_component_bits( depth->format, PIPE_FORMAT_COMP_Z ) == depth->bits.depth );
-   assert(pf_get_component_bits( depth->format, PIPE_FORMAT_COMP_S ) == depth->bits.stencil );
+   assert(util_format_get_component_bits(color->format, UTIL_FORMAT_COLORSPACE_RGB, 0) == color->bits.red);
+   assert(util_format_get_component_bits(color->format, UTIL_FORMAT_COLORSPACE_RGB, 1) == color->bits.green);
+   assert(util_format_get_component_bits(color->format, UTIL_FORMAT_COLORSPACE_RGB, 2) == color->bits.blue);
+   assert(util_format_get_component_bits(color->format, UTIL_FORMAT_COLORSPACE_RGB, 3) == color->bits.alpha);
+   assert(util_format_get_component_bits(depth->format, UTIL_FORMAT_COLORSPACE_ZS, 0) == depth->bits.depth);
+   assert(util_format_get_component_bits(depth->format, UTIL_FORMAT_COLORSPACE_ZS, 1) == depth->bits.stencil);
    
    pfi = &stw_dev->pixelformats[stw_dev->pixelformat_extended_count];
    
index a5975aad51501090547808b884a6fa0dcd27eea5..1c248a629e681e24625a98125bb6d0323492d761 100644 (file)
@@ -423,8 +423,6 @@ bind_samplers(struct exa_context *exa, int op,
 
 
 
-
-
 static INLINE boolean matrix_from_pict_transform(PictTransform *trans, float *matrix)
 {
    if (!trans)
index 337449a7451c63acc099ca96f686f7713e78b320..c776faa53f8faafeab7e77d0575a7d795f3fd385 100644 (file)
@@ -50,6 +50,7 @@
 #endif
 
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_rect.h"
 
 #ifdef HAVE_LIBKMS
@@ -196,11 +197,10 @@ crtc_load_cursor_argb_ga3d(xf86CrtcPtr crtc, CARD32 * image)
        templat.tex_usage |= PIPE_TEXTURE_USAGE_PRIMARY;
        templat.target = PIPE_TEXTURE_2D;
        templat.last_level = 0;
-       templat.depth[0] = 1;
+       templat.depth0 = 1;
        templat.format = PIPE_FORMAT_A8R8G8B8_UNORM;
-       templat.width[0] = 64;
-       templat.height[0] = 64;
-       pf_get_block(templat.format, &templat.block);
+       templat.width0 = 64;
+       templat.height0 = 64;
 
        crtcp->cursor_tex = ms->screen->texture_create(ms->screen,
                                                       &templat);
@@ -216,7 +216,7 @@ crtc_load_cursor_argb_ga3d(xf86CrtcPtr crtc, CARD32 * image)
                                            PIPE_TRANSFER_WRITE,
                                            0, 0, 64, 64);
     ptr = ms->screen->transfer_map(ms->screen, transfer);
-    util_copy_rect(ptr, &crtcp->cursor_tex->block,
+    util_copy_rect(ptr, crtcp->cursor_tex->format,
                   transfer->stride, 0, 0,
                   64, 64, (void*)image, 64 * 4, 0, 0);
     ms->screen->transfer_unmap(ms->screen, transfer);
index 4fa47548a43f2dc664ebc6c9072f9bedf87e7dc6..fd82f4fa1d1db4d31aca8003aac5db16446f6c82 100644 (file)
@@ -40,6 +40,7 @@
 #include "pipe/p_state.h"
 #include "pipe/p_inlines.h"
 
+#include "util/u_format.h"
 #include "util/u_rect.h"
 
 /* Make all the #if cases in the code esier to read */
@@ -92,7 +93,7 @@ dri2_do_create_buffer(DrawablePtr pDraw, DRI2BufferPtr buffer, unsigned int form
     case 9:
 #endif
        if (exa_priv->depth_stencil_tex &&
-           !pf_is_depth_stencil(exa_priv->depth_stencil_tex->format))
+           !util_format_is_depth_or_stencil(exa_priv->depth_stencil_tex->format))
            exa_priv->depth_stencil_tex = NULL;
         /* Fall through */
     case DRI2BufferDepth:
@@ -108,10 +109,9 @@ dri2_do_create_buffer(DrawablePtr pDraw, DRI2BufferPtr buffer, unsigned int form
            else
                template.format = ms->ds_depth_bits_last ?
                    PIPE_FORMAT_S8Z24_UNORM : PIPE_FORMAT_Z24S8_UNORM;
-           pf_get_block(template.format, &template.block);
-           template.width[0] = pDraw->width;
-           template.height[0] = pDraw->height;
-           template.depth[0] = 1;
+           template.width0 = pDraw->width;
+           template.height0 = pDraw->height;
+           template.depth0 = 1;
            template.last_level = 0;
            template.tex_usage = PIPE_TEXTURE_USAGE_DEPTH_STENCIL |
                PIPE_TEXTURE_USAGE_DISPLAY_TARGET;
index 17a032ee87e6778c6e15348d2006ea1e0966326e..d5c005ebadd7de0030bbedb677d3bd2f74ef661b 100644 (file)
@@ -43,6 +43,7 @@
 #include "pipe/p_state.h"
 #include "pipe/p_inlines.h"
 
+#include "util/u_format.h"
 #include "util/u_rect.h"
 #include "util/u_math.h"
 #include "util/u_debug.h"
@@ -202,7 +203,7 @@ ExaDownloadFromScreen(PixmapPtr pPix, int x,  int y, int w,  int h, char *dst,
                  x, y, w, h, dst_pitch);
 #endif
 
-    util_copy_rect((unsigned char*)dst, &priv->tex->block, dst_pitch, 0, 0,
+    util_copy_rect((unsigned char*)dst, priv->tex->format, dst_pitch, 0, 0,
                   w, h, exa->scrn->transfer_map(exa->scrn, transfer),
                   transfer->stride, 0, 0);
 
@@ -242,7 +243,7 @@ ExaUploadToScreen(PixmapPtr pPix, int x, int y, int w, int h, char *src,
 #endif
 
     util_copy_rect(exa->scrn->transfer_map(exa->scrn, transfer),
-                  &priv->tex->block, transfer->stride, 0, 0, w, h,
+                  priv->tex->format, transfer->stride, 0, 0, w, h,
                   (unsigned char*)src, src_pitch, 0, 0);
 
     exa->scrn->transfer_unmap(exa->scrn, transfer);
@@ -274,8 +275,8 @@ ExaPrepareAccess(PixmapPtr pPix, int index)
            PIPE_REFERENCED_FOR_WRITE)
            exa->pipe->flush(exa->pipe, 0, NULL);
 
-        assert(pPix->drawable.width <= priv->tex->width[0]);
-        assert(pPix->drawable.height <= priv->tex->height[0]);
+        assert(pPix->drawable.width <= priv->tex->width0);
+        assert(pPix->drawable.height <= priv->tex->height0);
 
        priv->map_transfer =
            exa->scrn->get_tex_transfer(exa->scrn, priv->tex, 0, 0, 0,
@@ -531,8 +532,8 @@ ExaCopy(PixmapPtr pDstPixmap, int srcX, int srcY, int dstX, int dstY,
                            dstX, dstY,
                            srcX, srcY,
                            width, height,
-                           exa->copy.src_texture->width[0],
-                           exa->copy.src_texture->height[0]);
+                           exa->copy.src_texture->width0,
+                           exa->copy.src_texture->height0);
    }
 }
 
@@ -874,8 +875,8 @@ ExaModifyPixmapHeader(PixmapPtr pPixmap, int width, int height,
        
        if (priv->tex)
           debug_printf("  ==> old texture %dx%d\n",
-                       priv->tex->width[0]
-                       priv->tex->height[0]);
+                       priv->tex->width0
+                       priv->tex->height0);
     }
 
 
@@ -903,8 +904,8 @@ ExaModifyPixmapHeader(PixmapPtr pPixmap, int width, int height,
     /* Deal with screen resize */
     if ((exa->accel || priv->flags) &&
         (!priv->tex ||
-         !size_match(width, priv->tex->width[0]) ||
-         !size_match(height, priv->tex->height[0]) ||
+         !size_match(width, priv->tex->width0) ||
+         !size_match(height, priv->tex->height0) ||
          priv->tex_flags != priv->flags)) {
        struct pipe_texture *texture = NULL;
        struct pipe_texture template;
@@ -912,17 +913,16 @@ ExaModifyPixmapHeader(PixmapPtr pPixmap, int width, int height,
        memset(&template, 0, sizeof(template));
        template.target = PIPE_TEXTURE_2D;
        exa_get_pipe_format(depth, &template.format, &bitsPerPixel, &priv->picture_format);
-       pf_get_block(template.format, &template.block);
         if (ROUND_UP_TEXTURES && priv->flags == 0) {
-           template.width[0] = util_next_power_of_two(width);
-           template.height[0] = util_next_power_of_two(height);
+           template.width0 = util_next_power_of_two(width);
+           template.height0 = util_next_power_of_two(height);
         }
         else {
-           template.width[0] = width;
-           template.height[0] = height;
+           template.width0 = width;
+           template.height0 = height;
         }
 
-       template.depth[0] = 1;
+       template.depth0 = 1;
        template.last_level = 0;
        template.tex_usage = PIPE_TEXTURE_USAGE_RENDER_TARGET | priv->flags;
        priv->tex_flags = priv->flags;
@@ -937,12 +937,12 @@ ExaModifyPixmapHeader(PixmapPtr pPixmap, int width, int height,
            src_surf = xorg_gpu_surface(exa->pipe->screen, priv);
             if (exa->pipe->surface_copy) {
                exa->pipe->surface_copy(exa->pipe, dst_surf, 0, 0, src_surf,
-                                       0, 0, min(width, texture->width[0]),
-                                       min(height, texture->height[0]));
+                                       0, 0, min(width, texture->width0),
+                                       min(height, texture->height0));
             } else {
                util_surface_copy(exa->pipe, FALSE, dst_surf, 0, 0, src_surf,
-                                 0, 0, min(width, texture->width[0]),
-                                 min(height, texture->height[0]));
+                                 0, 0, min(width, texture->width0),
+                                 min(height, texture->height0));
             }
            exa->scrn->tex_surface_destroy(dst_surf);
            exa->scrn->tex_surface_destroy(src_surf);
@@ -975,8 +975,8 @@ xorg_exa_set_texture(PixmapPtr pPixmap, struct  pipe_texture *tex)
     if (!priv)
        return FALSE;
 
-    if (pPixmap->drawable.width != tex->width[0] ||
-       pPixmap->drawable.height != tex->height[0])
+    if (pPixmap->drawable.width != tex->width0 ||
+       pPixmap->drawable.height != tex->height0)
        return FALSE;
 
     pipe_texture_reference(&priv->tex, tex);
@@ -998,10 +998,9 @@ xorg_exa_create_root_texture(ScrnInfoPtr pScrn,
     memset(&template, 0, sizeof(template));
     template.target = PIPE_TEXTURE_2D;
     exa_get_pipe_format(depth, &template.format, &bitsPerPixel, &dummy);
-    pf_get_block(template.format, &template.block);
-    template.width[0] = width;
-    template.height[0] = height;
-    template.depth[0] = 1;
+    template.width0 = width;
+    template.height0 = height;
+    template.depth0 = 1;
     template.last_level = 0;
     template.tex_usage |= PIPE_TEXTURE_USAGE_RENDER_TARGET;
     template.tex_usage |= PIPE_TEXTURE_USAGE_PRIMARY;
index 13a9840bdddf76d15111b78b7883abb7d283c364..89b794a09acb4b0e744eb8741f0d0cac9575f05f 100644 (file)
@@ -47,22 +47,22 @@ static void
 print_fs_traits(int fs_traits)
 {
    const char *strings[] = {
-      "FS_COMPOSITE",       // = 1 << 0,
-      "FS_MASK",            // = 1 << 1,
-      "FS_SOLID_FILL",      // = 1 << 2,
-      "FS_LINGRAD_FILL",    // = 1 << 3,
-      "FS_RADGRAD_FILL",    // = 1 << 4,
-      "FS_CA_FULL",         // = 1 << 5, /* src.rgba * mask.rgba */
-      "FS_CA_SRCALPHA",     // = 1 << 6, /* src.aaaa * mask.rgba */
-      "FS_YUV",             // = 1 << 7,
-      "FS_SRC_REPEAT_NONE", // = 1 << 8,
-      "FS_MASK_REPEAT_NONE",// = 1 << 9,
-      "FS_SRC_SWIZZLE_RGB", // = 1 << 10,
-      "FS_MASK_SWIZZLE_RGB",// = 1 << 11,
-      "FS_SRC_SET_ALPHA",   // = 1 << 12,
-      "FS_MASK_SET_ALPHA",  // = 1 << 13,
-      "FS_SRC_LUMINANCE",   // = 1 << 14,
-      "FS_MASK_LUMINANCE",  // = 1 << 15,
+      "FS_COMPOSITE",       /* = 1 << 0 */
+      "FS_MASK",            /* = 1 << 1 */
+      "FS_SOLID_FILL",      /* = 1 << 2 */
+      "FS_LINGRAD_FILL",    /* = 1 << 3 */
+      "FS_RADGRAD_FILL",    /* = 1 << 4 */
+      "FS_CA_FULL",         /* = 1 << 5 - src.rgba * mask.rgba */
+      "FS_CA_SRCALPHA",     /* = 1 << 6 - src.aaaa * mask.rgba */
+      "FS_YUV",             /* = 1 << 7 */
+      "FS_SRC_REPEAT_NONE", /* = 1 << 8 */
+      "FS_MASK_REPEAT_NONE",/* = 1 << 9 */
+      "FS_SRC_SWIZZLE_RGB", /* = 1 << 10 */
+      "FS_MASK_SWIZZLE_RGB",/* = 1 << 11 */
+      "FS_SRC_SET_ALPHA",   /* = 1 << 12 */
+      "FS_MASK_SET_ALPHA",  /* = 1 << 13 */
+      "FS_SRC_LUMINANCE",   /* = 1 << 14 */
+      "FS_MASK_LUMINANCE",  /* = 1 << 15 */
    };
    int i, k;
    debug_printf("%s: ", __func__);
index cbb84a8c0da6e844d28b861c5152cd63e25590c7..ba15f8a7845b30e256aa6c2bfbaa15a7749a292e 100644 (file)
@@ -5,6 +5,7 @@
 
 #include "cso_cache/cso_context.h"
 #include "util/u_draw_quad.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 #include "util/u_rect.h"
@@ -170,14 +171,14 @@ add_vertex_data1(struct xorg_renderer *r,
       map_point(src_matrix, pt3[0], pt3[1], &pt3[0], &pt3[1]);
    }
 
-   s0 =  pt0[0] / src->width[0];
-   s1 =  pt1[0] / src->width[0];
-   s2 =  pt2[0] / src->width[0];
-   s3 =  pt3[0] / src->width[0];
-   t0 =  pt0[1] / src->height[0];
-   t1 =  pt1[1] / src->height[0];
-   t2 =  pt2[1] / src->height[0];
-   t3 =  pt3[1] / src->height[0];
+   s0 =  pt0[0] / src->width0;
+   s1 =  pt1[0] / src->width0;
+   s2 =  pt2[0] / src->width0;
+   s3 =  pt3[0] / src->width0;
+   t0 =  pt0[1] / src->height0;
+   t1 =  pt1[1] / src->height0;
+   t2 =  pt2[1] / src->height0;
+   t3 =  pt3[1] / src->height0;
 
    /* 1st vertex */
    add_vertex_1tex(r, dstX, dstY, s0, t0);
@@ -248,15 +249,15 @@ add_vertex_data2(struct xorg_renderer *r,
       map_point(mask_matrix, mpt1[0], mpt1[1], &mpt1[0], &mpt1[1]);
    }
 
-   src_s0 = spt0[0] / src->width[0];
-   src_t0 = spt0[1] / src->height[0];
-   src_s1 = spt1[0] / src->width[0];
-   src_t1 = spt1[1] / src->height[0];
+   src_s0 = spt0[0] / src->width0;
+   src_t0 = spt0[1] / src->height0;
+   src_s1 = spt1[0] / src->width0;
+   src_t1 = spt1[1] / src->height0;
 
-   mask_s0 = mpt0[0] / mask->width[0];
-   mask_t0 = mpt0[1] / mask->height[0];
-   mask_s1 = mpt1[0] / mask->width[0];
-   mask_t1 = mpt1[1] / mask->height[0];
+   mask_s0 = mpt0[0] / mask->width0;
+   mask_t0 = mpt0[1] / mask->height0;
+   mask_s1 = mpt1[0] / mask->width0;
+   mask_t1 = mpt1[1] / mask->height0;
 
    /* 1st vertex */
    add_vertex_2tex(r, dstX, dstY,
@@ -286,10 +287,10 @@ setup_vertex_data_yuv(struct xorg_renderer *r,
    spt1[0] = srcX + srcW;
    spt1[1] = srcY + srcH;
 
-   s0 = spt0[0] / tex[0]->width[0];
-   t0 = spt0[1] / tex[0]->height[0];
-   s1 = spt1[0] / tex[0]->width[0];
-   t1 = spt1[1] / tex[0]->height[0];
+   s0 = spt0[0] / tex[0]->width0;
+   t0 = spt0[1] / tex[0]->height0;
+   s1 = spt1[0] / tex[0]->width0;
+   t1 = spt1[1] / tex[0]->height0;
 
    /* 1st vertex */
    add_vertex_1tex(r, dstX, dstY, s0, t0);
@@ -509,10 +510,9 @@ renderer_clone_texture(struct xorg_renderer *r,
    templ.target = PIPE_TEXTURE_2D;
    templ.format = format;
    templ.last_level = 0;
-   templ.width[0] = src->width[0];
-   templ.height[0] = src->height[0];
-   templ.depth[0] = 1;
-   pf_get_block(format, &templ.block);
+   templ.width0 = src->width0;
+   templ.height0 = src->height0;
+   templ.depth0 = 1;
    templ.tex_usage = PIPE_TEXTURE_USAGE_SAMPLER;
 
    pt = screen->texture_create(screen, &templ);
@@ -533,13 +533,13 @@ renderer_clone_texture(struct xorg_renderer *r,
                 ps_tex, /* dest */
                 0, 0, /* destx/y */
                 ps_read,
-                0, 0, src->width[0], src->height[0]);
+                0, 0, src->width0, src->height0);
       } else {
           util_surface_copy(pipe, FALSE,
                 ps_tex, /* dest */
                 0, 0, /* destx/y */
                 ps_read,
-                0, 0, src->width[0], src->height[0]);
+                0, 0, src->width0, src->height0);
       }
       pipe_surface_reference(&ps_read, NULL);
       pipe_surface_reference(&ps_tex, NULL);
index 19c50051a85a23f3e7a2c902614558d55009ef62..6b5a41a372724cba2ba69c3ac456f386b9e4dd07 100644 (file)
@@ -13,6 +13,8 @@
 #include "pipe/p_screen.h"
 #include "pipe/p_inlines.h"
 
+#include "util/u_format.h"
+
 /*XXX get these from pipe's texture limits */
 #define IMAGE_MAX_WIDTH                2048
 #define IMAGE_MAX_HEIGHT       2048
@@ -167,10 +169,9 @@ create_component_texture(struct pipe_context *pipe,
    templ.target = PIPE_TEXTURE_2D;
    templ.format = PIPE_FORMAT_L8_UNORM;
    templ.last_level = 0;
-   templ.width[0] = width;
-   templ.height[0] = height;
-   templ.depth[0] = 1;
-   pf_get_block(PIPE_FORMAT_L8_UNORM, &templ.block);
+   templ.width0 = width;
+   templ.height0 = height;
+   templ.depth0 = 1;
    templ.tex_usage = PIPE_TEXTURE_USAGE_SAMPLER;
 
    tex = screen->texture_create(screen, &templ);
@@ -183,18 +184,18 @@ check_yuv_textures(struct xorg_xv_port_priv *priv,  int width, int height)
 {
    struct pipe_texture **dst = priv->yuv[priv->current_set];
    if (!dst[0] ||
-       dst[0]->width[0] != width ||
-       dst[0]->height[0] != height) {
+       dst[0]->width0 != width ||
+       dst[0]->height0 != height) {
       pipe_texture_reference(&dst[0], NULL);
    }
    if (!dst[1] ||
-       dst[1]->width[0] != width ||
-       dst[1]->height[0] != height) {
+       dst[1]->width0 != width ||
+       dst[1]->height0 != height) {
       pipe_texture_reference(&dst[1], NULL);
    }
    if (!dst[2] ||
-       dst[2]->width[0] != width ||
-       dst[2]->height[0] != height) {
+       dst[2]->width0 != width ||
+       dst[2]->height0 != height) {
       pipe_texture_reference(&dst[2], NULL);
    }
 
index bf9038f356e4d4eb11e5edd2499138343068fcc4..0e39a390c69b2296110621566cb3e3d8c275b867 100644 (file)
@@ -1,8 +1,8 @@
 /**************************************************************************
- * 
+ *
  * Copyright 2009 Younes Manton.
  * All Rights Reserved.
- * 
+ *
  * Permission is hereby granted, free of charge, to any person obtaining a
  * copy of this software and associated documentation files (the
  * "Software"), to deal in the Software without restriction, including
  * distribute, sub license, and/or sell copies of the Software, and to
  * permit persons to whom the Software is furnished to do so, subject to
  * the following conditions:
- * 
+ *
  * The above copyright notice and this permission notice (including the
  * next paragraph) shall be included in all copies or substantial portions
  * of the Software.
- * 
+ *
  * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
  * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
  * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
@@ -22,7 +22,7 @@
  * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
  * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
  * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- * 
+ *
  **************************************************************************/
 
 #include <assert.h>
@@ -103,10 +103,9 @@ CreateOrResizeBackBuffer(struct pipe_video_context *vpipe, unsigned int width, u
    /* XXX: Needs to match the drawable's format? */
    template.format = PIPE_FORMAT_X8R8G8B8_UNORM;
    template.last_level = 0;
-   template.width[0] = width;
-   template.height[0] = height;
-   template.depth[0] = 1;
-   pf_get_block(template.format, &template.block);
+   template.width0 = width;
+   template.height0 = height;
+   template.depth0 = 1;
    template.tex_usage = PIPE_TEXTURE_USAGE_DISPLAY_TARGET;
 
    tex = vpipe->screen->texture_create(vpipe->screen, &template);
index 9f7b383d2d37d3ebf9a8494e78766f014d5ddd1a..66b73a8bf93140bec38c103b31160a6df0d49c8c 100644 (file)
@@ -58,6 +58,11 @@ if env['dri']:
                        'intel/SConscript',
                ])
 
+       if 'i965' in env['winsys']:
+               SConscript([
+                       'i965/SConscript',
+               ])
+
        if 'radeon' in env['winsys']:
                SConscript([
                        'radeon/SConscript',
diff --git a/src/gallium/winsys/drm/i965/Makefile b/src/gallium/winsys/drm/i965/Makefile
new file mode 100644 (file)
index 0000000..d8feef6
--- /dev/null
@@ -0,0 +1,12 @@
+# src/gallium/winsys/drm/intel/Makefile
+TOP = ../../../../..
+include $(TOP)/configs/current
+
+SUBDIRS = gem $(GALLIUM_STATE_TRACKERS_DIRS)
+
+default install clean:
+       @for dir in $(SUBDIRS) ; do \
+               if [ -d $$dir ] ; then \
+                       (cd $$dir && $(MAKE) $@) || exit 1; \
+               fi \
+       done
diff --git a/src/gallium/winsys/drm/i965/SConscript b/src/gallium/winsys/drm/i965/SConscript
new file mode 100644 (file)
index 0000000..50d7b75
--- /dev/null
@@ -0,0 +1,7 @@
+Import('*')
+
+SConscript(['gem/SConscript',])
+
+if 'mesa' in env['statetrackers']:
+
+    SConscript(['dri/SConscript'])
diff --git a/src/gallium/winsys/drm/i965/dri/Makefile b/src/gallium/winsys/drm/i965/dri/Makefile
new file mode 100644 (file)
index 0000000..f7e81ee
--- /dev/null
@@ -0,0 +1,26 @@
+TOP = ../../../../../..
+include $(TOP)/configs/current
+
+LIBNAME = i965_dri.so
+
+PIPE_DRIVERS = \
+       $(TOP)/src/gallium/state_trackers/dri/libdridrm.a \
+       $(TOP)/src/gallium/winsys/drm/i965/gem/libi965drm.a \
+       $(TOP)/src/gallium/drivers/trace/libtrace.a \
+       $(TOP)/src/gallium/drivers/softpipe/libsoftpipe.a \
+       $(TOP)/src/gallium/drivers/identity/libidentity.a \
+       $(TOP)/src/gallium/drivers/i965/libi965.a
+
+
+DRIVER_SOURCES =
+
+C_SOURCES = \
+       $(COMMON_GALLIUM_SOURCES) \
+       $(DRIVER_SOURCES)
+
+include ../../Makefile.template
+
+DRI_LIB_DEPS += -ldrm_intel
+
+symlinks: $(TOP)/$(LIB_DIR)/gallium
+       @rm -f $(TOP)/$(LIB_DIR)/gallium/i965_dri.so
diff --git a/src/gallium/winsys/drm/i965/dri/SConscript b/src/gallium/winsys/drm/i965/dri/SConscript
new file mode 100644 (file)
index 0000000..233ef46
--- /dev/null
@@ -0,0 +1,19 @@
+Import('*')
+
+env = drienv.Clone()
+
+env.ParseConfig('pkg-config --cflags --libs libdrm_intel')
+
+drivers = [
+    st_dri,
+    i965drm,
+    i965,
+    trace,
+]
+
+env.LoadableModule(
+    target ='i965_dri.so',
+    source = COMMON_GALLIUM_SOURCES,
+    LIBS = drivers + mesa + auxiliaries + env['LIBS'],
+    SHLIBPREFIX = '',
+)
diff --git a/src/gallium/winsys/drm/i965/egl/Makefile b/src/gallium/winsys/drm/i965/egl/Makefile
new file mode 100644 (file)
index 0000000..a1b32eb
--- /dev/null
@@ -0,0 +1,29 @@
+TOP = ../../../../../..
+GALLIUMDIR = ../../../..
+include $(TOP)/configs/current
+
+LIBNAME = EGL_i965.so
+
+PIPE_DRIVERS = \
+       $(TOP)/src/gallium/state_trackers/egl/libegldrm.a \
+       $(GALLIUMDIR)/winsys/drm/i965/gem/libi965drm.a \
+       $(TOP)/src/gallium/drivers/softpipe/libsoftpipe.a \
+       $(TOP)/src/gallium/drivers/trace/libtrace.a \
+       $(TOP)/src/gallium/drivers/i965/libi965.a
+
+DRIVER_SOURCES =
+
+C_SOURCES = \
+       $(COMMON_GALLIUM_SOURCES) \
+       $(DRIVER_SOURCES)
+
+DRIVER_EXTRAS = -ldrm_intel
+
+ASM_SOURCES = 
+
+DRIVER_DEFINES = -I../gem $(shell pkg-config libdrm --atleast-version=2.3.1 \
+                               && echo "-DDRM_VBLANK_FLIP=DRM_VBLANK_FLIP")
+
+include ../../Makefile.template
+
+symlinks:
diff --git a/src/gallium/winsys/drm/i965/gem/Makefile b/src/gallium/winsys/drm/i965/gem/Makefile
new file mode 100644 (file)
index 0000000..6a7497b
--- /dev/null
@@ -0,0 +1,14 @@
+TOP = ../../../../../..
+include $(TOP)/configs/current
+
+LIBNAME = i965drm
+
+C_SOURCES = \
+       i965_drm_buffer.c \
+       i965_drm_api.c
+
+LIBRARY_INCLUDES = $(shell pkg-config libdrm --cflags-only-I)
+
+LIBRARY_DEFINES = $(shell pkg-config libdrm --cflags-only-other)
+
+include ../../../../Makefile.template
diff --git a/src/gallium/winsys/drm/i965/gem/SConscript b/src/gallium/winsys/drm/i965/gem/SConscript
new file mode 100644 (file)
index 0000000..6256ec6
--- /dev/null
@@ -0,0 +1,15 @@
+Import('*')
+
+env = drienv.Clone()
+
+i965drm_sources = [
+    'i965_drm_api.c',
+    'i965_drm_buffer.c',
+]
+
+i965drm = env.ConvenienceLibrary(
+    target ='i965drm',
+    source = i965drm_sources,
+)
+
+Export('i965drm')
diff --git a/src/gallium/winsys/drm/i965/gem/i965_drm_api.c b/src/gallium/winsys/drm/i965/gem/i965_drm_api.c
new file mode 100644 (file)
index 0000000..fc9678d
--- /dev/null
@@ -0,0 +1,243 @@
+
+#include <stdio.h>
+#include "state_tracker/drm_api.h"
+
+#include "i965_drm_winsys.h"
+#include "util/u_memory.h"
+
+#include "i965/brw_context.h"        /* XXX: shouldn't be doing this */
+#include "i965/brw_screen.h"         /* XXX: shouldn't be doing this */
+
+#include "trace/tr_drm.h"
+
+/*
+ * Helper functions
+ */
+
+
+static void
+i965_libdrm_get_device_id(unsigned int *device_id)
+{
+   char path[512];
+   FILE *file;
+   void *shutup_gcc;
+
+   /*
+    * FIXME: Fix this up to use a drm ioctl or whatever.
+    */
+
+   snprintf(path, sizeof(path), "/sys/class/drm/card0/device/device");
+   file = fopen(path, "r");
+   if (!file) {
+      return;
+   }
+
+   shutup_gcc = fgets(path, sizeof(path), file);
+   sscanf(path, "%x", device_id);
+   fclose(file);
+}
+
+static struct i965_libdrm_buffer *
+i965_libdrm_buffer_from_handle(struct i965_libdrm_winsys *idws,
+                               const char* name, unsigned handle)
+{
+   struct i965_libdrm_buffer *buf = CALLOC_STRUCT(i965_libdrm_buffer);
+   uint32_t swizzle = 0;
+
+   if (BRW_DUMP)
+      debug_printf("%s\n", __FUNCTION__);
+
+   if (!buf)
+      return NULL;
+   pipe_reference_init(&buf->base.reference, 1);
+   buf->bo = drm_intel_bo_gem_create_from_name(idws->gem, name, handle);
+   buf->base.size = buf->bo->size;
+   buf->base.sws = &idws->base;
+   buf->flinked = TRUE;
+   buf->flink = handle;
+
+
+   if (!buf->bo)
+      goto err;
+
+   drm_intel_bo_get_tiling(buf->bo, &buf->tiling, &swizzle);
+   if (buf->tiling != 0)
+      buf->map_gtt = TRUE;
+
+   return buf;
+
+err:
+   FREE(buf);
+   return NULL;
+}
+
+
+/*
+ * Exported functions
+ */
+
+
+static struct pipe_texture *
+i965_libdrm_texture_from_shared_handle(struct drm_api *api,
+                                       struct pipe_screen *screen,
+                                       struct pipe_texture *template,
+                                       const char* name,
+                                       unsigned pitch,
+                                       unsigned handle)
+{
+   /* XXX: this is silly -- there should be a way to get directly from
+    * the "drm_api" struct to ourselves, without peering into
+    * unrelated code:
+    */
+   struct i965_libdrm_winsys *idws = i965_libdrm_winsys(brw_screen(screen)->sws);
+   struct i965_libdrm_buffer *buffer;
+
+   if (BRW_DUMP)
+      debug_printf("%s %s pitch %d handle 0x%x\n", __FUNCTION__,
+                  name, pitch, handle);
+
+   buffer = i965_libdrm_buffer_from_handle(idws, name, handle);
+   if (!buffer)
+      return NULL;
+
+   return brw_texture_blanket_winsys_buffer(screen, template, pitch,
+                                           buffer->tiling,
+                                           &buffer->base);
+}
+
+
+static boolean
+i965_libdrm_shared_handle_from_texture(struct drm_api *api,
+                                       struct pipe_screen *screen,
+                                       struct pipe_texture *texture,
+                                       unsigned *pitch,
+                                       unsigned *handle)
+{
+   struct i965_libdrm_buffer *buf = NULL;
+   struct brw_winsys_buffer *buffer = NULL;
+
+   if (BRW_DUMP)
+      debug_printf("%s\n", __FUNCTION__);
+
+   if (!brw_texture_get_winsys_buffer(texture, &buffer, pitch))
+      return FALSE;
+
+   buf = i965_libdrm_buffer(buffer);
+   if (!buf->flinked) {
+      if (drm_intel_bo_flink(buf->bo, &buf->flink))
+         return FALSE;
+      buf->flinked = TRUE;
+   }
+
+   *handle = buf->flink;
+
+   if (BRW_DUMP)
+      debug_printf("   -> pitch %d handle 0x%x\n", *pitch, *handle);
+
+   return TRUE;
+}
+
+static boolean
+i965_libdrm_local_handle_from_texture(struct drm_api *api,
+                                      struct pipe_screen *screen,
+                                      struct pipe_texture *texture,
+                                      unsigned *pitch,
+                                      unsigned *handle)
+{
+   struct brw_winsys_buffer *buffer = NULL;
+
+   if (BRW_DUMP)
+      debug_printf("%s\n", __FUNCTION__);
+
+   if (!brw_texture_get_winsys_buffer(texture, &buffer, pitch))
+      return FALSE;
+
+   *handle = i965_libdrm_buffer(buffer)->bo->handle;
+
+   if (BRW_DUMP)
+      debug_printf("   -> pitch %d handle 0x%x\n", *pitch, *handle);
+
+   return TRUE;
+}
+
+static void
+i965_libdrm_winsys_destroy(struct brw_winsys_screen *iws)
+{
+   struct i965_libdrm_winsys *idws = i965_libdrm_winsys(iws);
+
+   if (BRW_DUMP)
+      debug_printf("%s\n", __FUNCTION__);
+
+   drm_intel_bufmgr_destroy(idws->gem);
+
+   FREE(idws);
+}
+
+static struct pipe_screen *
+i965_libdrm_create_screen(struct drm_api *api, int drmFD,
+                          struct drm_create_screen_arg *arg)
+{
+   struct i965_libdrm_winsys *idws;
+   unsigned int deviceID;
+
+   debug_printf("%s\n", __FUNCTION__);
+
+   if (arg != NULL) {
+      switch(arg->mode) {
+      case DRM_CREATE_NORMAL:
+         break;
+      default:
+         return NULL;
+      }
+   }
+
+   idws = CALLOC_STRUCT(i965_libdrm_winsys);
+   if (!idws)
+      return NULL;
+
+   i965_libdrm_get_device_id(&deviceID);
+
+   i965_libdrm_winsys_init_buffer_functions(idws);
+
+   idws->fd = drmFD;
+   idws->id = deviceID;
+
+   idws->base.destroy = i965_libdrm_winsys_destroy;
+
+   idws->gem = drm_intel_bufmgr_gem_init(idws->fd, BRW_BATCH_SIZE);
+   drm_intel_bufmgr_gem_enable_reuse(idws->gem);
+
+   idws->send_cmd = !debug_get_bool_option("BRW_NO_HW", FALSE);
+
+   return brw_create_screen(&idws->base, deviceID);
+}
+
+static struct pipe_context *
+i965_libdrm_create_context(struct drm_api *api, struct pipe_screen *screen)
+{
+   return brw_create_context(screen);
+}
+
+static void
+destroy(struct drm_api *api)
+{
+   if (BRW_DUMP)
+      debug_printf("%s\n", __FUNCTION__);
+
+}
+
+struct drm_api i965_libdrm_api =
+{
+   .create_context = i965_libdrm_create_context,
+   .create_screen = i965_libdrm_create_screen,
+   .texture_from_shared_handle = i965_libdrm_texture_from_shared_handle,
+   .shared_handle_from_texture = i965_libdrm_shared_handle_from_texture,
+   .local_handle_from_texture = i965_libdrm_local_handle_from_texture,
+   .destroy = destroy,
+};
+
+struct drm_api *
+drm_api_create()
+{
+   return trace_drm_create(&i965_libdrm_api);
+}
diff --git a/src/gallium/winsys/drm/i965/gem/i965_drm_buffer.c b/src/gallium/winsys/drm/i965/gem/i965_drm_buffer.c
new file mode 100644 (file)
index 0000000..a4a72b3
--- /dev/null
@@ -0,0 +1,427 @@
+
+#include "i965_drm_winsys.h"
+#include "util/u_memory.h"
+
+#include "i915_drm.h"
+#include "intel_bufmgr.h"
+
+
+
+const char *names[BRW_BUFFER_TYPE_MAX] = {
+   "TEXTURE",
+   "SCANOUT",
+   "VERTEX",
+   "CURBE",
+   "QUERY",
+   "SHADER_CONSTANTS",
+   "WM_SCRATCH",
+   "BATCH",
+   "GENERAL_STATE",
+   "SURFACE_STATE",
+   "PIXEL",
+   "GENERIC",
+};
+
+const char *usages[BRW_USAGE_MAX] = {
+   "STATE",
+   "QUERY_RESULT",
+   "RENDER_TARGET",
+   "DEPTH_BUFFER",
+   "BLIT_SOURCE",
+   "BLIT_DEST",
+   "SAMPLER",
+   "VERTEX",
+   "SCRATCH"
+};
+
+
+const char *data_types[BRW_DATA_MAX] =
+{
+   "GS: CC_VP",
+   "GS: CC_UNIT",
+   "GS: WM_PROG",
+   "GS: SAMPLER_DEFAULT_COLOR",
+   "GS: SAMPLER",
+   "GS: WM_UNIT",
+   "GS: SF_PROG",
+   "GS: SF_VP",
+   "GS: SF_UNIT",
+   "GS: VS_UNIT",
+   "GS: VS_PROG",
+   "GS: GS_UNIT",
+   "GS: GS_PROG",
+   "GS: CLIP_VP",
+   "GS: CLIP_UNIT",
+   "GS: CLIP_PROG",
+   "SS: SURFACE",
+   "SS: SURF_BIND",
+   "CONSTANT DATA",
+   "BATCH DATA",
+   "(untyped)"
+};
+
+static enum pipe_error 
+i965_libdrm_bo_alloc(struct brw_winsys_screen *sws,
+                     enum brw_buffer_type type,
+                     unsigned size,
+                     unsigned alignment,
+                     struct brw_winsys_buffer **bo_out)
+{
+   struct i965_libdrm_winsys *idws = i965_libdrm_winsys(sws);
+   struct i965_libdrm_buffer *buf;
+
+   if (BRW_DUMP)
+      debug_printf("%s type %s sz %d align %d\n",
+                  __FUNCTION__, names[type], size, alignment );
+
+   buf = CALLOC_STRUCT(i965_libdrm_buffer);
+   if (!buf)
+      return PIPE_ERROR_OUT_OF_MEMORY;
+
+   switch (type) {
+   case BRW_BUFFER_TYPE_TEXTURE:
+/* case BRW_BUFFER_TYPE_SCANOUT:*/
+   case BRW_BUFFER_TYPE_VERTEX:
+   case BRW_BUFFER_TYPE_CURBE:
+   case BRW_BUFFER_TYPE_QUERY:
+   case BRW_BUFFER_TYPE_SHADER_CONSTANTS:
+   case BRW_BUFFER_TYPE_SHADER_SCRATCH:
+   case BRW_BUFFER_TYPE_BATCH:
+   case BRW_BUFFER_TYPE_GENERAL_STATE:
+   case BRW_BUFFER_TYPE_SURFACE_STATE:
+   case BRW_BUFFER_TYPE_PIXEL:
+   case BRW_BUFFER_TYPE_GENERIC:
+      break;
+   case BRW_BUFFER_TYPE_SCANOUT:
+      buf->map_gtt = TRUE;
+      break;
+   default:
+      assert(0);
+      break;
+   }
+
+   buf->bo = drm_intel_bo_alloc(idws->gem, 
+                                names[type], 
+                                size, 
+                                alignment);
+
+   if (!buf->bo)
+      goto err;
+
+   pipe_reference_init(&buf->base.reference, 1);
+   buf->base.size = size;
+   buf->base.sws = sws;
+
+   *bo_out = &buf->base;
+   return PIPE_OK;
+
+err:
+   assert(0);
+   FREE(buf);
+   return PIPE_ERROR_OUT_OF_MEMORY;
+}
+
+static void 
+i965_libdrm_bo_destroy(struct brw_winsys_buffer *buffer)
+{
+   struct i965_libdrm_buffer *buf = i965_libdrm_buffer(buffer);
+
+   if (BRW_DUMP)
+      debug_printf("%s\n", __FUNCTION__);
+
+   drm_intel_bo_unreference(buf->bo);
+   FREE(buffer);
+}
+
+static enum pipe_error
+i965_libdrm_bo_emit_reloc(struct brw_winsys_buffer *buffer,
+                          enum brw_buffer_usage usage,
+                          unsigned delta,
+                          unsigned offset,
+                          struct brw_winsys_buffer *buffer2)
+{
+   struct i965_libdrm_buffer *buf = i965_libdrm_buffer(buffer);
+   struct i965_libdrm_buffer *buf2 = i965_libdrm_buffer(buffer2);
+   int read, write;
+   int ret;
+
+   if (BRW_DUMP)
+      debug_printf("%s buf %p offset %x delta %x buf2 %p/%s/%s\n",
+                  __FUNCTION__, (void *)buffer, 
+                  offset, delta,
+                  (void *)buffer2, names[buf2->data_type], usages[usage]);
+
+   switch (usage) {
+   case BRW_USAGE_STATE:
+      read = I915_GEM_DOMAIN_INSTRUCTION;
+      write = 0;
+      break;
+   case BRW_USAGE_QUERY_RESULT:
+      read = I915_GEM_DOMAIN_INSTRUCTION;
+      write = I915_GEM_DOMAIN_INSTRUCTION;
+      break;
+   case BRW_USAGE_RENDER_TARGET:
+      read = I915_GEM_DOMAIN_RENDER;
+      write = 0;
+      break;
+   case BRW_USAGE_DEPTH_BUFFER:
+      read = I915_GEM_DOMAIN_RENDER;
+      write = I915_GEM_DOMAIN_RENDER;
+      break;
+   case BRW_USAGE_BLIT_SOURCE:
+      read = 0;
+      write = I915_GEM_DOMAIN_RENDER;
+      break;
+   case BRW_USAGE_BLIT_DEST:
+      read = I915_GEM_DOMAIN_RENDER;
+      write = I915_GEM_DOMAIN_RENDER;
+      break;
+   case BRW_USAGE_SAMPLER:
+      read = I915_GEM_DOMAIN_SAMPLER;
+      write = 0;
+      break;
+   case BRW_USAGE_VERTEX:
+      read = I915_GEM_DOMAIN_VERTEX;
+      write = 0;
+      break;
+   case BRW_USAGE_SCRATCH:
+      read = 0;
+      write = 0;
+      break;
+   default:
+      assert(0);
+      return -1;
+   }
+
+   /* Needed??
+   ((uint32_t *)buf->bo->virtual)[offset/4] = (delta +
+                                              buf2->bo->offset);
+    */
+
+   ret = dri_bo_emit_reloc( buf->bo, read, write, delta, offset, buf2->bo );
+   if (ret)
+      return -1;
+
+   return 0;
+}
+
+static enum pipe_error 
+i965_libdrm_bo_exec(struct brw_winsys_buffer *buffer,
+                    unsigned bytes_used)
+{
+   struct i965_libdrm_buffer *buf = i965_libdrm_buffer(buffer);
+   struct i965_libdrm_winsys *idws = i965_libdrm_winsys(buffer->sws);
+   int ret;
+
+   if (BRW_DUMP)
+      debug_printf("execute buffer %p, bytes %d\n", (void *)buffer, bytes_used);
+
+   if (idws->send_cmd) {
+      ret = dri_bo_exec(buf->bo, bytes_used, NULL, 0, 0);
+      if (ret)
+         return PIPE_ERROR;
+   }
+
+   return PIPE_OK;
+}
+
+static enum pipe_error
+i965_libdrm_bo_subdata(struct brw_winsys_buffer *buffer,
+                       enum brw_buffer_data_type data_type,
+                       size_t offset,
+                       size_t size,
+                       const void *data,
+                       const struct brw_winsys_reloc *reloc,
+                       unsigned nr_reloc)
+{
+   struct i965_libdrm_buffer *buf = i965_libdrm_buffer(buffer);
+   struct i965_libdrm_winsys *idws = i965_libdrm_winsys(buffer->sws);
+   int ret, i;
+
+   (void)data_type;
+
+   if (BRW_DUMP)
+      debug_printf("%s buf %p off %d sz %d %s relocs: %d\n", 
+                  __FUNCTION__, 
+                  (void *)buffer, offset, size, 
+                  data_types[data_type],
+                  nr_reloc);
+
+   if (BRW_DUMP)
+      brw_dump_data( idws->id,
+                    data_type,
+                    buf->bo->offset + offset, 
+                    data, size );
+
+   /* XXX: use bo_map_gtt/memcpy/unmap_gtt under some circumstances???
+    */
+   ret = drm_intel_bo_subdata(buf->bo, offset, size, (void*)data);
+   if (ret)
+      return PIPE_ERROR;
+  
+   for (i = 0; i < nr_reloc; i++) {
+      i965_libdrm_bo_emit_reloc(buffer, reloc[i].usage, reloc[i].delta,
+                                reloc[i].offset, reloc[i].bo);
+   }
+
+   return PIPE_OK;
+}
+
+static boolean 
+i965_libdrm_bo_is_busy(struct brw_winsys_buffer *buffer)
+{
+   struct i965_libdrm_buffer *buf = i965_libdrm_buffer(buffer);
+   boolean ret;
+
+   if (BRW_DUMP)
+      debug_printf("%s %p\n", __FUNCTION__, (void *)buffer);
+
+   ret = drm_intel_bo_busy(buf->bo);
+
+   if (BRW_DUMP)
+      debug_printf("  --> %d\n", ret);
+
+   return ret;
+}
+
+static boolean 
+i965_libdrm_bo_references(struct brw_winsys_buffer *a,
+                          struct brw_winsys_buffer *b)
+{
+   struct i965_libdrm_buffer *bufa = i965_libdrm_buffer(a);
+   struct i965_libdrm_buffer *bufb = i965_libdrm_buffer(b);
+   boolean ret;
+
+   if (BRW_DUMP)
+      debug_printf("%s %p %p\n", __FUNCTION__, (void *)a, (void *)b);
+
+   ret = drm_intel_bo_references(bufa->bo, bufb->bo);
+
+   if (BRW_DUMP)
+      debug_printf("  --> %d\n", ret);
+
+   return ret;
+}
+
+/* XXX: couldn't this be handled by returning true/false on
+ * bo_emit_reloc?
+ */
+static enum pipe_error
+i965_libdrm_check_aperture_space(struct brw_winsys_screen *iws,
+                                 struct brw_winsys_buffer **buffers,
+                                 unsigned count)
+{
+   static drm_intel_bo *bos[128];
+   int i;
+   int ret;
+
+   if (BRW_DUMP)
+      debug_printf("%s\n", __FUNCTION__);
+
+   if (count > Elements(bos)) {
+      assert(0);
+      return FALSE;
+   }
+
+   for (i = 0; i < count; i++)
+      bos[i] = i965_libdrm_buffer(buffers[i])->bo;
+
+   /* XXX: converting from ??? to pipe_error:
+    */
+   ret = dri_bufmgr_check_aperture_space(bos, count);
+
+   if (BRW_DUMP)
+      debug_printf("  --> %d (ok == %d)\n", ret, PIPE_OK);
+
+   return ret;
+}
+
+static void *
+i965_libdrm_bo_map(struct brw_winsys_buffer *buffer,
+                   enum brw_buffer_data_type data_type,
+                   unsigned offset,
+                   unsigned length,
+                   boolean write,
+                   boolean discard,
+                   boolean flush_explicit)
+{
+   struct i965_libdrm_buffer *buf = i965_libdrm_buffer(buffer);
+   int ret;
+
+
+   if (BRW_DUMP)
+      debug_printf("%s %p %s %s\n", __FUNCTION__, (void *)buffer, 
+                  write ? "read/write" : "read",
+                  write ? data_types[data_type] : "");
+
+   if (!buf->map_count) {
+      if (buf->map_gtt) {
+         ret = drm_intel_gem_bo_map_gtt(buf->bo);
+         if (ret)
+            return NULL;
+      }
+      else {
+         ret = drm_intel_bo_map(buf->bo, write);
+         if (ret)
+            return NULL;
+      }
+   }
+
+   buf->data_type = data_type;
+   buf->map_count++;
+   return buf->bo->virtual;
+}
+
+static void
+i965_libdrm_bo_flush_range(struct brw_winsys_buffer *buffer,
+                           unsigned offset,
+                           unsigned length)
+{
+   struct i965_libdrm_buffer *buf = i965_libdrm_buffer(buffer);
+   struct i965_libdrm_winsys *idws = i965_libdrm_winsys(buffer->sws);
+
+   if (BRW_DUMP)
+      debug_printf("%s %s offset %d len %d\n", __FUNCTION__,
+                  data_types[buf->data_type],
+                  offset, length);
+
+   if (BRW_DUMP)
+      brw_dump_data( idws->id,
+                    buf->data_type,
+                    buf->bo->offset + offset, 
+                    buf->bo->virtual + offset, 
+                    length );
+}
+
+static void 
+i965_libdrm_bo_unmap(struct brw_winsys_buffer *buffer)
+{
+   struct i965_libdrm_buffer *buf = i965_libdrm_buffer(buffer);
+
+   if (BRW_DUMP)
+      debug_printf("%s\n", __FUNCTION__);
+
+   if (--buf->map_count > 0)
+      return;
+
+   if (buf->map_gtt)
+      drm_intel_gem_bo_unmap_gtt(buf->bo);
+   else
+      drm_intel_bo_unmap(buf->bo);
+}
+
+void
+i965_libdrm_winsys_init_buffer_functions(struct i965_libdrm_winsys *idws)
+{
+   idws->base.bo_alloc             = i965_libdrm_bo_alloc;
+   idws->base.bo_destroy           = i965_libdrm_bo_destroy;
+   idws->base.bo_emit_reloc        = i965_libdrm_bo_emit_reloc;
+   idws->base.bo_exec              = i965_libdrm_bo_exec;
+   idws->base.bo_subdata           = i965_libdrm_bo_subdata;
+   idws->base.bo_is_busy           = i965_libdrm_bo_is_busy;
+   idws->base.bo_references        = i965_libdrm_bo_references;
+   idws->base.check_aperture_space = i965_libdrm_check_aperture_space;
+   idws->base.bo_map               = i965_libdrm_bo_map;
+   idws->base.bo_flush_range       = i965_libdrm_bo_flush_range;
+   idws->base.bo_unmap             = i965_libdrm_bo_unmap;
+}
diff --git a/src/gallium/winsys/drm/i965/gem/i965_drm_winsys.h b/src/gallium/winsys/drm/i965/gem/i965_drm_winsys.h
new file mode 100644 (file)
index 0000000..c6a7d4a
--- /dev/null
@@ -0,0 +1,64 @@
+
+#ifndef INTEL_DRM_WINSYS_H
+#define INTEL_DRM_WINSYS_H
+
+#include "i965/brw_winsys.h"
+
+#include "drm.h"
+#include "intel_bufmgr.h"
+
+
+
+/*
+ * Winsys
+ */
+
+
+struct i965_libdrm_winsys
+{
+   struct brw_winsys_screen base;
+   drm_intel_bufmgr *gem;
+
+   boolean send_cmd;
+
+   int fd; /**< Drm file discriptor */
+
+   unsigned id;
+};
+
+static INLINE struct i965_libdrm_winsys *
+i965_libdrm_winsys(struct brw_winsys_screen *iws)
+{
+   return (struct i965_libdrm_winsys *)iws;
+}
+
+struct i965_libdrm_winsys *i965_libdrm_winsys_create(int fd, unsigned pci_id);
+
+void i965_libdrm_winsys_init_buffer_functions(struct i965_libdrm_winsys *idws);
+
+
+/* Buffer.  
+ */
+struct i965_libdrm_buffer {
+   struct brw_winsys_buffer base;
+
+   drm_intel_bo *bo;
+
+   void *ptr;
+   unsigned map_count;
+   unsigned data_type;         /* valid while mapped */
+   unsigned tiling;
+
+   boolean map_gtt;
+   boolean flinked;
+   unsigned flink;
+};
+
+static INLINE struct i965_libdrm_buffer *
+i965_libdrm_buffer(struct brw_winsys_buffer *buffer)
+{
+   return (struct i965_libdrm_buffer *)buffer;
+}
+
+
+#endif
diff --git a/src/gallium/winsys/drm/i965/xlib/Makefile b/src/gallium/winsys/drm/i965/xlib/Makefile
new file mode 100644 (file)
index 0000000..0efa0ca
--- /dev/null
@@ -0,0 +1,97 @@
+# src/gallium/winsys/xlib/Makefile
+
+# This makefile produces a "stand-alone" libGL.so which is based on
+# Xlib (no DRI HW acceleration)
+
+
+TOP = ../../../../../..
+include $(TOP)/configs/current
+
+
+GL_MAJOR = 1
+GL_MINOR = 5
+GL_TINY = 0$(MESA_MAJOR)0$(MESA_MINOR)0$(MESA_TINY)
+
+
+INCLUDE_DIRS = \
+       -I$(TOP)/include \
+       -I$(TOP)/src/mesa \
+       -I$(TOP)/src/mesa/main \
+       -I$(TOP)/src/gallium/include \
+       -I$(TOP)/src/gallium/drivers \
+       -I$(TOP)/src/gallium/drivers/i965 \
+       -I$(TOP)/src/gallium/drivers/i965/include \
+       -I$(TOP)/src/gallium/state_trackers/glx/xlib \
+       -I$(TOP)/src/gallium/auxiliary \
+       -I/usr/include/drm
+
+XLIB_WINSYS_SOURCES = \
+       xlib_i965.c \
+
+
+
+XLIB_WINSYS_OBJECTS = $(XLIB_WINSYS_SOURCES:.c=.o)
+
+
+
+LIBS = \
+       $(TOP)/src/gallium/drivers/i965/libi965.a \
+       $(TOP)/src/gallium/drivers/trace/libtrace.a \
+       $(TOP)/src/gallium/state_trackers/glx/xlib/libxlib.a \
+       $(TOP)/src/mesa/libglapi.a \
+       $(TOP)/src/mesa/libmesagallium.a \
+       $(GALLIUM_AUXILIARIES) 
+
+#      $(TOP)/src/gallium/drivers/i965/lib/libi9xx.a \
+
+.SUFFIXES : .cpp
+
+.c.o:
+       $(CC) -c $(INCLUDE_DIRS) $(DEFINES) $(CFLAGS) $< -o $@
+
+.cpp.o:
+       $(CXX) -c $(INCLUDE_DIRS) $(DEFINES) $(CXXFLAGS) $< -o $@
+
+
+
+default: $(TOP)/$(LIB_DIR)/gallium $(TOP)/$(LIB_DIR)/gallium/$(GL_LIB_NAME)
+
+$(TOP)/$(LIB_DIR)/gallium:
+       @ mkdir -p $(TOP)/$(LIB_DIR)/gallium
+
+# Make the libGL.so library
+$(TOP)/$(LIB_DIR)/gallium/$(GL_LIB_NAME): $(XLIB_WINSYS_OBJECTS) $(LIBS) Makefile
+       $(TOP)/bin/mklib -o $(GL_LIB) \
+               -linker "$(CC)" \
+               -major $(GL_MAJOR) -minor $(GL_MINOR) -patch $(GL_TINY) \
+               -install $(TOP)/$(LIB_DIR)/gallium \
+               $(MKLIB_OPTIONS) $(XLIB_WINSYS_OBJECTS) \
+               -Wl,--start-group $(LIBS) -Wl,--end-group $(GL_LIB_DEPS)
+
+
+depend: $(XLIB_WINSYS_SOURCES)
+       @ echo "running $(MKDEP)"
+       @ rm -f depend  # workaround oops on gutsy?!?
+       @ touch depend
+       $(MKDEP) $(MKDEP_OPTIONS) $(DEFINES) $(INCLUDE_DIRS) $(XLIB_WINSYS_SOURCES) \
+               > /dev/null 2>/dev/null
+
+
+install: default
+       $(INSTALL) -d $(INSTALL_DIR)/include/GL
+       $(INSTALL) -d $(INSTALL_DIR)/$(LIB_DIR)
+       $(INSTALL) -m 644 $(TOP)/include/GL/*.h $(INSTALL_DIR)/include/GL
+       @if [ -e $(TOP)/$(LIB_DIR)/$(GL_LIB_NAME) ]; then \
+               $(MINSTALL) $(TOP)/$(LIB_DIR)/libGL* $(INSTALL_DIR)/$(LIB_DIR); \
+       fi
+
+
+# Emacs tags
+tags:
+       etags `find . -name \*.[ch]` $(TOP)/include/GL/*.h
+
+clean:
+       -rm -f *.o
+
+
+include depend
diff --git a/src/gallium/winsys/drm/i965/xlib/xlib_i965.c b/src/gallium/winsys/drm/i965/xlib/xlib_i965.c
new file mode 100644 (file)
index 0000000..d2b9a1a
--- /dev/null
@@ -0,0 +1,522 @@
+/**************************************************************************
+ * 
+ * Copyright 2007 Tungsten Graphics, Inc., Bismarck, ND., USA
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
+ * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
+ * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
+ * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
+ * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR 
+ * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE 
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * 
+ **************************************************************************/
+
+/*
+ * Authors:
+ *   Keith Whitwell
+ *   Brian Paul
+ */
+
+
+#include "util/u_memory.h"
+#include "util/u_math.h"
+#include "pipe/p_error.h"
+#include "pipe/p_context.h"
+
+#include "xm_winsys.h"
+
+#include "i965/brw_winsys.h"
+#include "i965/brw_screen.h"
+#include "i965/brw_reg.h"
+#include "i965/brw_structs_dump.h"
+
+#define MAX_VRAM (128*1024*1024)
+
+
+
+extern int brw_disasm (FILE *file, 
+                       const struct brw_instruction *inst,
+                       unsigned count );
+
+extern int intel_decode(const uint32_t *data, 
+                        int count,
+                        uint32_t hw_offset,
+                        uint32_t devid);
+
+struct xlib_brw_buffer
+{
+   struct brw_winsys_buffer base;
+   char *virtual;
+   unsigned offset;
+   unsigned type;
+   int map_count;
+   boolean modified;
+};
+
+
+/**
+ * Subclass of brw_winsys_screen for Xlib winsys
+ */
+struct xlib_brw_winsys
+{
+   struct brw_winsys_screen base;
+   struct brw_chipset chipset;
+
+   unsigned size;
+   unsigned used;
+};
+
+static struct xlib_brw_winsys *
+xlib_brw_winsys( struct brw_winsys_screen *screen )
+{
+   return (struct xlib_brw_winsys *)screen;
+}
+
+
+static struct xlib_brw_buffer *
+xlib_brw_buffer( struct brw_winsys_buffer *buffer )
+{
+   return (struct xlib_brw_buffer *)buffer;
+}
+
+
+
+const char *names[BRW_BUFFER_TYPE_MAX] = {
+   "TEXTURE",
+   "SCANOUT",
+   "VERTEX",
+   "CURBE",
+   "QUERY",
+   "SHADER_CONSTANTS",
+   "WM_SCRATCH",
+   "BATCH",
+   "GENERAL_STATE",
+   "SURFACE_STATE",
+   "PIXEL",
+   "GENERIC",
+};
+
+const char *usages[BRW_USAGE_MAX] = {
+   "STATE",
+   "QUERY_RESULT",
+   "RENDER_TARGET",
+   "DEPTH_BUFFER",
+   "BLIT_SOURCE",
+   "BLIT_DEST",
+   "SAMPLER",
+   "VERTEX",
+   "SCRATCH"
+};
+
+
+const char *data_types[BRW_DATA_MAX] =
+{
+   "GS: CC_VP",
+   "GS: CC_UNIT",
+   "GS: WM_PROG",
+   "GS: SAMPLER_DEFAULT_COLOR",
+   "GS: SAMPLER",
+   "GS: WM_UNIT",
+   "GS: SF_PROG",
+   "GS: SF_VP",
+   "GS: SF_UNIT",
+   "GS: VS_UNIT",
+   "GS: VS_PROG",
+   "GS: GS_UNIT",
+   "GS: GS_PROG",
+   "GS: CLIP_VP",
+   "GS: CLIP_UNIT",
+   "GS: CLIP_PROG",
+   "SS: SURFACE",
+   "SS: SURF_BIND",
+   "CONSTANT DATA",
+   "BATCH DATA",
+   "(untyped)"
+};
+
+
+static enum pipe_error
+xlib_brw_bo_alloc( struct brw_winsys_screen *sws,
+                   enum brw_buffer_type type,
+                   unsigned size,
+                   unsigned alignment,
+                   struct brw_winsys_buffer **bo_out )
+{
+   struct xlib_brw_winsys *xbw = xlib_brw_winsys(sws);
+   struct xlib_brw_buffer *buf;
+
+   if (BRW_DEBUG & DEBUG_WINSYS)
+      debug_printf("%s type %s sz %d align %d\n",
+                   __FUNCTION__, names[type], size, alignment );
+
+   buf = CALLOC_STRUCT(xlib_brw_buffer);
+   if (!buf)
+      return PIPE_ERROR_OUT_OF_MEMORY;
+
+   pipe_reference_init(&buf->base.reference, 1);
+
+   buf->offset = align(xbw->used, alignment);
+   buf->type = type;
+   buf->virtual = MALLOC(size);
+   buf->base.size = size;
+   buf->base.sws = sws;
+
+   xbw->used = align(xbw->used, alignment) + size;
+   if (xbw->used > MAX_VRAM)
+      goto err;
+
+   /* XXX: possibly rentrant call to bo_destroy:
+    */
+   bo_reference(bo_out, &buf->base);
+   return PIPE_OK;
+
+err:
+   assert(0);
+   FREE(buf->virtual);
+   FREE(buf);
+   return PIPE_ERROR_OUT_OF_MEMORY;
+}
+
+static void 
+xlib_brw_bo_destroy( struct brw_winsys_buffer *buffer )
+{
+   struct xlib_brw_buffer *buf = xlib_brw_buffer(buffer);
+
+   FREE(buf);
+}
+
+static int 
+xlib_brw_bo_emit_reloc( struct brw_winsys_buffer *buffer,
+                        enum brw_buffer_usage usage,
+                        unsigned delta,
+                        unsigned offset,
+                        struct brw_winsys_buffer *buffer2)
+{
+   struct xlib_brw_buffer *buf = xlib_brw_buffer(buffer);
+   struct xlib_brw_buffer *buf2 = xlib_brw_buffer(buffer2);
+
+   if (BRW_DEBUG & DEBUG_WINSYS)
+      debug_printf("%s buf %p offset %x val %x + %x buf2 %p/%s/%s\n",
+                   __FUNCTION__, (void *)buffer, offset,
+                   buf2->offset, delta,
+                   (void *)buffer2, names[buf2->type], usages[usage]);
+
+   *(uint32_t *)(buf->virtual + offset) = buf2->offset + delta;
+
+   return 0;
+}
+
+static int 
+xlib_brw_bo_exec( struct brw_winsys_buffer *buffer,
+                    unsigned bytes_used )
+{
+   if (BRW_DEBUG & DEBUG_WINSYS)
+      debug_printf("execute buffer %p, bytes %d\n", (void *)buffer, bytes_used);
+
+   return 0;
+}
+
+
+
+
+static int
+xlib_brw_bo_subdata(struct brw_winsys_buffer *buffer,
+                    enum brw_buffer_data_type data_type,
+                    size_t offset,
+                    size_t size,
+                    const void *data,
+                    const struct brw_winsys_reloc *reloc,
+                    unsigned nr_relocs)
+{
+   struct xlib_brw_buffer *buf = xlib_brw_buffer(buffer);
+   struct xlib_brw_winsys *xbw = xlib_brw_winsys(buffer->sws);
+   unsigned i;
+
+   if (BRW_DEBUG & DEBUG_WINSYS)
+      debug_printf("%s buf %p off %d sz %d %s relocs: %d\n", 
+                   __FUNCTION__, 
+                   (void *)buffer, offset, size, 
+                   data_types[data_type],
+                   nr_relocs);
+
+   assert(buf->base.size >= offset + size);
+   memcpy(buf->virtual + offset, data, size);
+
+   /* Apply the relocations:
+    */
+   for (i = 0; i < nr_relocs; i++) {
+      if (BRW_DEBUG & DEBUG_WINSYS)
+         debug_printf("\treloc[%d] usage %s off %d value %x+%x\n", 
+                      i, usages[reloc[i].usage], reloc[i].offset,
+                      xlib_brw_buffer(reloc[i].bo)->offset, reloc[i].delta);
+
+      *(unsigned *)(buf->virtual + offset + reloc[i].offset) = 
+         xlib_brw_buffer(reloc[i].bo)->offset + reloc[i].delta;
+   }
+
+   if (BRW_DUMP)
+      brw_dump_data( xbw->chipset.pci_id,
+                    data_type,
+                    buf->offset + offset, 
+                    buf->virtual + offset, size );
+
+
+   return 0;
+}
+
+
+static boolean 
+xlib_brw_bo_is_busy(struct brw_winsys_buffer *buffer)
+{
+   if (BRW_DEBUG & DEBUG_WINSYS)
+      debug_printf("%s %p\n", __FUNCTION__, (void *)buffer);
+   return TRUE;
+}
+
+static boolean 
+xlib_brw_bo_references(struct brw_winsys_buffer *a,
+                         struct brw_winsys_buffer *b)
+{
+   if (BRW_DEBUG & DEBUG_WINSYS)
+      debug_printf("%s %p %p\n", __FUNCTION__, (void *)a, (void *)b);
+   return TRUE;
+}
+
+static enum pipe_error
+xlib_brw_check_aperture_space( struct brw_winsys_screen *iws,
+                                struct brw_winsys_buffer **buffers,
+                                unsigned count )
+{
+   unsigned tot_size = 0;
+   unsigned i;
+
+   for (i = 0; i < count; i++)
+      tot_size += buffers[i]->size;
+
+   if (BRW_DEBUG & DEBUG_WINSYS)
+      debug_printf("%s %d bufs, tot_size: %d kb\n", 
+                   __FUNCTION__, count, 
+                   (tot_size + 1023) / 1024);
+
+   return PIPE_OK;
+}
+
+static void *
+xlib_brw_bo_map(struct brw_winsys_buffer *buffer,
+                enum brw_buffer_data_type data_type,
+                unsigned offset,
+                unsigned length,
+                boolean write,
+                boolean discard,
+                boolean explicit)
+{
+   struct xlib_brw_buffer *buf = xlib_brw_buffer(buffer);
+
+   if (BRW_DEBUG & DEBUG_WINSYS)
+      debug_printf("%s %p %s %s\n", __FUNCTION__, (void *)buffer, 
+                   write ? "read/write" : "read",
+                   write ? data_types[data_type] : "");
+
+   if (write)
+      buf->modified = 1;
+
+   buf->map_count++;
+   return buf->virtual;
+}
+
+
+static void
+xlib_brw_bo_flush_range( struct brw_winsys_buffer *buffer,
+                         unsigned offset,
+                         unsigned length )
+{
+}
+
+
+static void 
+xlib_brw_bo_unmap(struct brw_winsys_buffer *buffer)
+{
+   struct xlib_brw_buffer *buf = xlib_brw_buffer(buffer);
+
+   if (BRW_DEBUG & DEBUG_WINSYS)
+      debug_printf("%s %p\n", __FUNCTION__, (void *)buffer);
+
+   --buf->map_count;
+   assert(buf->map_count >= 0);
+
+   if (buf->map_count == 0 &&
+       buf->modified) {
+
+      buf->modified = 0;
+      
+      /* Consider dumping new buffer contents here, using the
+       * flush-range info to minimize verbosity.
+       */
+   }
+}
+
+
+static void
+xlib_brw_bo_wait_idle( struct brw_winsys_buffer *buffer )
+{
+}
+
+
+static void
+xlib_brw_winsys_destroy( struct brw_winsys_screen *sws )
+{
+   struct xlib_brw_winsys *xbw = xlib_brw_winsys(sws);
+
+   FREE(xbw);
+}
+
+static struct brw_winsys_screen *
+xlib_create_brw_winsys_screen( void )
+{
+   struct xlib_brw_winsys *ws;
+
+   ws = CALLOC_STRUCT(xlib_brw_winsys);
+   if (!ws)
+      return NULL;
+
+   ws->used = 0;
+
+   ws->base.destroy              = xlib_brw_winsys_destroy;
+   ws->base.bo_alloc             = xlib_brw_bo_alloc;
+   ws->base.bo_destroy           = xlib_brw_bo_destroy;
+   ws->base.bo_emit_reloc        = xlib_brw_bo_emit_reloc;
+   ws->base.bo_exec              = xlib_brw_bo_exec;
+   ws->base.bo_subdata           = xlib_brw_bo_subdata;
+   ws->base.bo_is_busy           = xlib_brw_bo_is_busy;
+   ws->base.bo_references        = xlib_brw_bo_references;
+   ws->base.check_aperture_space = xlib_brw_check_aperture_space;
+   ws->base.bo_map               = xlib_brw_bo_map;
+   ws->base.bo_flush_range       = xlib_brw_bo_flush_range;
+   ws->base.bo_unmap             = xlib_brw_bo_unmap;
+   ws->base.bo_wait_idle         = xlib_brw_bo_wait_idle;
+
+   return &ws->base;
+}
+
+
+/***********************************************************************
+ * Implementation of Xlib co-state-tracker's winsys interface
+ */
+
+static void
+xlib_i965_display_surface(struct xmesa_buffer *xm_buffer,
+                          struct pipe_surface *surf)
+{
+   struct brw_surface *surface = brw_surface(surf);
+   struct xlib_brw_buffer *bo = xlib_brw_buffer(surface->bo);
+
+   if (BRW_DEBUG & DEBUG_WINSYS)
+      debug_printf("%s offset %x+%x sz %dx%d\n", __FUNCTION__, 
+                   bo->offset,
+                   surface->draw_offset,
+                   surf->width,
+                   surf->height);
+}
+
+static void
+xlib_i965_flush_frontbuffer(struct pipe_screen *screen,
+                           struct pipe_surface *surf,
+                           void *context_private)
+{
+   xlib_i965_display_surface(NULL, surf);
+}
+
+
+static struct pipe_screen *
+xlib_create_i965_screen( void )
+{
+   struct brw_winsys_screen *winsys;
+   struct pipe_screen *screen;
+
+   winsys = xlib_create_brw_winsys_screen();
+   if (winsys == NULL)
+      return NULL;
+
+   screen = brw_create_screen(winsys, PCI_CHIP_GM45_GM);
+   if (screen == NULL)
+      goto fail;
+
+   xlib_brw_winsys(winsys)->chipset = brw_screen(screen)->chipset;
+
+   screen->flush_frontbuffer = xlib_i965_flush_frontbuffer;
+   return screen;
+
+fail:
+   if (winsys)
+      winsys->destroy( winsys );
+
+   return NULL;
+}
+
+
+static struct pipe_context *
+xlib_create_i965_context( struct pipe_screen *screen,
+                          void *context_private )
+{
+   struct pipe_context *pipe;
+   
+   pipe = brw_create_context(screen);
+   if (pipe == NULL)
+      goto fail;
+
+   pipe->priv = context_private;
+   return pipe;
+
+fail:
+   /* Free stuff here */
+   return NULL;
+}
+
+
+
+
+struct xm_driver xlib_i965_driver = 
+{
+   .create_pipe_screen = xlib_create_i965_screen,
+   .create_pipe_context = xlib_create_i965_context,
+   .display_surface = xlib_i965_display_surface
+};
+
+
+/* Register this driver at library load: 
+ */
+static void _init( void ) __attribute__((constructor));
+static void _init( void )
+{
+   xmesa_set_driver( &xlib_i965_driver );
+}
+
+
+
+/***********************************************************************
+ *
+ * Butt-ugly hack to convince the linker not to throw away public GL
+ * symbols (they are all referenced from getprocaddress, I guess).
+ */
+extern void (*linker_foo(const unsigned char *procName))();
+extern void (*glXGetProcAddress(const unsigned char *procName))();
+
+extern void (*linker_foo(const unsigned char *procName))()
+{
+   return glXGetProcAddress(procName);
+}
diff --git a/src/gallium/winsys/drm/i965/xorg/Makefile b/src/gallium/winsys/drm/i965/xorg/Makefile
new file mode 100644 (file)
index 0000000..d91d000
--- /dev/null
@@ -0,0 +1,57 @@
+TARGET     = modesetting_drv.so
+CFILES     = $(wildcard ./*.c)
+OBJECTS    = $(patsubst ./%.c,./%.o,$(CFILES))
+TOP        = ../../../../../..
+
+include $(TOP)/configs/current
+
+INCLUDES = \
+       $(shell pkg-config --cflags-only-I pixman-1 xorg-server libdrm xproto) \
+       -I../gem \
+       -I$(TOP)/src/gallium/include \
+       -I$(TOP)/src/gallium/drivers \
+       -I$(TOP)/src/gallium/auxiliary \
+       -I$(TOP)/src/mesa \
+       -I$(TOP)/include \
+       -I$(TOP)/src/egl/main
+
+LIBS = \
+       $(TOP)/src/gallium/state_trackers/xorg/libxorgtracker.a \
+       $(TOP)/src/gallium/winsys/drm/i965/gem/libi965drm.a \
+       $(TOP)/src/gallium/drivers/i965/libi965.a \
+       $(TOP)/src/gallium/drivers/trace/libtrace.a \
+       $(TOP)/src/gallium/drivers/softpipe/libsoftpipe.a \
+       $(GALLIUM_AUXILIARIES)
+
+DRIVER_DEFINES = \
+       -DHAVE_CONFIG_H
+
+
+#############################################
+
+
+
+all default: $(TARGET)
+
+$(TARGET): $(OBJECTS) Makefile $(TOP)/src/gallium/state_trackers/xorg/libxorgtracker.a $(LIBS)
+       $(TOP)/bin/mklib -noprefix -o $@ \
+       $(OBJECTS) $(LIBS) $(shell pkg-config --libs libdrm) -ldrm_intel
+
+clean:
+       rm -rf $(OBJECTS) $(TARGET)
+
+install:
+       $(INSTALL) -d $(DESTDIR)/$(XORG_DRIVER_INSTALL_DIR)
+       $(MINSTALL) -m 755 $(TARGET) $(DESTDIR)/$(XORG_DRIVER_INSTALL_DIR)
+
+
+##############################################
+
+
+.c.o:
+       $(CC) -c $(CFLAGS) $(INCLUDES) $(DRIVER_DEFINES) $< -o $@
+
+
+##############################################
+
+.PHONY = all clean install
diff --git a/src/gallium/winsys/drm/i965/xorg/intel_xorg.c b/src/gallium/winsys/drm/i965/xorg/intel_xorg.c
new file mode 100644 (file)
index 0000000..ac691cb
--- /dev/null
@@ -0,0 +1,147 @@
+/*
+ * Copyright 2008 Tungsten Graphics, Inc., Cedar Park, Texas.
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ *
+ * Author: Alan Hourihane <alanh@tungstengraphics.com>
+ * Author: Jakob Bornecrantz <wallbraker@gmail.com>
+ *
+ */
+
+#include "../../../../state_trackers/xorg/xorg_winsys.h"
+
+static void intel_xorg_identify(int flags);
+static Bool intel_xorg_pci_probe(DriverPtr driver,
+                                int entity_num,
+                                struct pci_device *device,
+                                intptr_t match_data);
+
+static const struct pci_id_match intel_xorg_device_match[] = {
+    {0x8086, PCI_MATCH_ANY, PCI_MATCH_ANY, PCI_MATCH_ANY, 0, 0, 0},
+    {0, 0, 0},
+};
+
+static SymTabRec intel_xorg_chipsets[] = {
+    {PCI_MATCH_ANY, "Intel Graphics Device"},
+    {-1, NULL}
+};
+
+static PciChipsets intel_xorg_pci_devices[] = {
+    {PCI_MATCH_ANY, PCI_MATCH_ANY, NULL},
+    {-1, -1, NULL}
+};
+
+static XF86ModuleVersionInfo intel_xorg_version = {
+    "modesetting",
+    MODULEVENDORSTRING,
+    MODINFOSTRING1,
+    MODINFOSTRING2,
+    XORG_VERSION_CURRENT,
+    0, 1, 0, /* major, minor, patch */
+    ABI_CLASS_VIDEODRV,
+    ABI_VIDEODRV_VERSION,
+    MOD_CLASS_VIDEODRV,
+    {0, 0, 0, 0}
+};
+
+/*
+ * Xorg driver exported structures
+ */
+
+_X_EXPORT DriverRec modesetting = {
+    1,
+    "modesetting",
+    intel_xorg_identify,
+    NULL,
+    xorg_tracker_available_options,
+    NULL,
+    0,
+    NULL,
+    intel_xorg_device_match,
+    intel_xorg_pci_probe
+};
+
+static MODULESETUPPROTO(intel_xorg_setup);
+
+_X_EXPORT XF86ModuleData modesettingModuleData = {
+    &intel_xorg_version,
+    intel_xorg_setup,
+    NULL
+};
+
+/*
+ * Xorg driver functions
+ */
+
+static pointer
+intel_xorg_setup(pointer module, pointer opts, int *errmaj, int *errmin)
+{
+    static Bool setupDone = 0;
+
+    /* This module should be loaded only once, but check to be sure.
+     */
+    if (!setupDone) {
+       setupDone = 1;
+       xf86AddDriver(&modesetting, module, HaveDriverFuncs);
+
+       /*
+        * The return value must be non-NULL on success even though there
+        * is no TearDownProc.
+        */
+       return (pointer) 1;
+    } else {
+       if (errmaj)
+           *errmaj = LDR_ONCEONLY;
+       return NULL;
+    }
+}
+
+static void
+intel_xorg_identify(int flags)
+{
+    xf86PrintChipsets("modesetting", "Driver for Modesetting Kernel Drivers",
+                     intel_xorg_chipsets);
+}
+
+static Bool
+intel_xorg_pci_probe(DriverPtr driver,
+         int entity_num, struct pci_device *device, intptr_t match_data)
+{
+    ScrnInfoPtr scrn = NULL;
+    EntityInfoPtr entity;
+
+    scrn = xf86ConfigPciEntity(scrn, 0, entity_num, intel_xorg_pci_devices,
+                              NULL, NULL, NULL, NULL, NULL);
+    if (scrn != NULL) {
+       scrn->driverVersion = 1;
+       scrn->driverName = "i965";
+       scrn->name = "modesetting";
+       scrn->Probe = NULL;
+
+       entity = xf86GetEntityInfo(entity_num);
+
+       /* Use all the functions from the xorg tracker */
+       xorg_tracker_set_functions(scrn);
+    }
+    return scrn != NULL;
+}
index c0ecd9680e2cdcad3f813a694d1e8ff1f4fd7f9c..26aae4122eb0283e8d88be3573b2c2fdc9d15b8c 100644 (file)
@@ -24,4 +24,3 @@ DRI_LIB_DEPS += -ldrm_intel
 
 symlinks: $(TOP)/$(LIB_DIR)/gallium
        @rm -f $(TOP)/$(LIB_DIR)/gallium/i965_dri.so
-       ln -s i915_dri.so $(TOP)/$(LIB_DIR)/gallium/i965_dri.so
index e70bfe7b44e336911701a9c69db734d748236702..e8b58742ab721b592499df5a8ec09e095559abaf 100644 (file)
@@ -39,11 +39,12 @@ intel_drm_fence_reference(struct intel_winsys *iws,
    struct intel_drm_fence *old = (struct intel_drm_fence *)*ptr;
    struct intel_drm_fence *f = (struct intel_drm_fence *)fence;
 
-   if (pipe_reference((struct pipe_reference**)ptr, &f->reference)) {
+   if (pipe_reference(&((struct intel_drm_fence *)(*ptr))->reference, &f->reference)) {
       if (old->bo)
          drm_intel_bo_unreference(old->bo);
       FREE(old);
    }
+   *ptr = fence;
 }
 
 static int
index 317dc44d22f805b3f0a0e9136034374b4207ed1b..7106a06492d934f3f5306c07a9115d543a037944 100644 (file)
@@ -1,5 +1,6 @@
 #include "pipe/p_context.h"
 #include "pipe/p_state.h"
+#include "util/u_format.h"
 #include "util/u_memory.h"
 
 #include "nouveau_drm_api.h"
@@ -24,11 +25,10 @@ dri_surface_from_handle(struct drm_api *api, struct pipe_screen *pscreen,
        tmpl.tex_usage = PIPE_TEXTURE_USAGE_PRIMARY;
        tmpl.target = PIPE_TEXTURE_2D;
        tmpl.last_level = 0;
-       tmpl.depth[0] = 1;
+       tmpl.depth0 = 1;
        tmpl.format = format;
-       tmpl.width[0] = width;
-       tmpl.height[0] = height;
-       pf_get_block(tmpl.format, &tmpl.block);
+       tmpl.width0 = width;
+       tmpl.height0 = height;
 
        pt = api->texture_from_shared_handle(api, pscreen, &tmpl,
                                             "front buffer", pitch, handle);
@@ -247,7 +247,7 @@ nouveau_drm_handle_from_pt(struct drm_api *api, struct pipe_screen *pscreen,
                return false;
 
        *handle = mt->bo->handle;
-       *stride = mt->base.nblocksx[0] * mt->base.block.size;
+       *stride = util_format_get_stride(mt->base.format, mt->base.width0);
        return true;
 }
 
index 42a6f4abc21898c86c6c1e94efdd4eaf3e9debc3..860cbb6dbf8af3c8e40d932f8a142246f70ac24b 100644 (file)
@@ -7,8 +7,7 @@ LIBNAME = radeonwinsys
 C_SOURCES = \
        radeon_buffer.c \
        radeon_drm.c \
-       radeon_r300.c \
-       radeon_winsys_softpipe.c
+       radeon_r300.c
 
 LIBRARY_INCLUDES = -I$(TOP)/src/gallium/drivers/r300 \
                   $(shell pkg-config libdrm --cflags-only-I)
index 2ad68e403fe359d55c9c5a9a3556c20d3dad1e39..f4e9c397bdf46a06635e8a916d47a355daeae266 100644 (file)
@@ -6,7 +6,6 @@ radeon_sources = [
     'radeon_buffer.c',
     'radeon_drm.c',
     'radeon_r300.c',
-    'radeon_winsys_softpipe.c',
 ]
 
 env.Append(CPPPATH = '#/src/gallium/drivers/r300')
index 81cd9dc4fb1fe17d687c2185e36f97669e246bd8..d2367b245a2b490b7568b07a586bb4d98947c28d 100644 (file)
 #include "radeon_bo_gem.h"
 #include "softpipe/sp_texture.h"
 #include "r300_context.h"
+#include "util/u_format.h"
+#include "util/u_math.h"
 #include <X11/Xutil.h>
+
 struct radeon_vl_context
 {
     Display *display;
@@ -113,17 +116,13 @@ static struct pipe_buffer *radeon_surface_buffer_create(struct pipe_winsys *ws,
                                                         unsigned tex_usage,
                                                         unsigned *stride)
 {
-    struct pipe_format_block block;
-    unsigned nblocksx, nblocksy, size;
-
-    pf_get_block(format, &block);
-
-    nblocksx = pf_get_nblocksx(&block, width);
-    nblocksy = pf_get_nblocksy(&block, height);
-
     /* Radeons enjoy things in multiples of 32. */
     /* XXX this can be 32 when POT */
-    *stride = (nblocksx * block.size + 63) & ~63;
+    const unsigned alignment = 64;
+    unsigned nblocksy, size;
+
+    nblocksy = util_format_get_nblocksy(format, height);
+    *stride = align(util_format_get_stride(format, width), alignment);
     size = *stride * nblocksy;
 
     return radeon_buffer_create(ws, 64, usage, size);
@@ -142,10 +141,15 @@ static void *radeon_buffer_map(struct pipe_winsys *ws,
                                struct pipe_buffer *buffer,
                                unsigned flags)
 {
+    struct radeon_winsys_priv *priv = ((struct radeon_winsys *)ws)->priv;
     struct radeon_pipe_buffer *radeon_buffer =
         (struct radeon_pipe_buffer*)buffer;
     int write = 0;
 
+    if (radeon_bo_is_referenced_by_cs(radeon_buffer->bo, priv->cs)) {
+        priv->flush_cb(priv->flush_data);
+    }
+
     if (flags & PIPE_BUFFER_USAGE_DONTBLOCK) {
         uint32_t domain;
 
@@ -317,13 +321,10 @@ struct pipe_surface *radeon_surface_from_handle(struct radeon_context *radeon_co
     memset(&tmpl, 0, sizeof(tmpl));
     tmpl.tex_usage = PIPE_TEXTURE_USAGE_DISPLAY_TARGET;
     tmpl.target = PIPE_TEXTURE_2D;
-    tmpl.width[0] = w;
-    tmpl.height[0] = h;
-    tmpl.depth[0] = 1;
+    tmpl.width0 = w;
+    tmpl.height0 = h;
+    tmpl.depth0 = 1;
     tmpl.format = format;
-    pf_get_block(tmpl.format, &tmpl.block);
-    tmpl.nblocksx[0] = pf_get_nblocksx(&tmpl.block, w);
-    tmpl.nblocksy[0] = pf_get_nblocksy(&tmpl.block, h);
 
     pt = pipe_screen->texture_blanket(pipe_screen, &tmpl, &pitch, pb);
     if (pt == NULL) {
index f5153b06af535aae6c8a046ed95139c57a74134b..d7f17564a9f75b780711bf30fd824ac360242393 100644 (file)
@@ -45,6 +45,8 @@
 
 #include "radeon_drm.h"
 
+#include "radeon_winsys.h"
+
 struct radeon_pipe_buffer {
     struct pipe_buffer  base;
     struct radeon_bo    *bo;
@@ -66,14 +68,10 @@ struct radeon_winsys_priv {
 
     /* Current CS. */
     struct radeon_cs* cs;
-};
-
-struct radeon_winsys {
-    /* Parent class. */
-    struct pipe_winsys base;
 
-    /* This corresponds to void* radeon_winsys in r300_winsys. */
-    struct radeon_winsys_priv* priv;
+    /* Flush CB */
+    void (*flush_cb)(void *);
+    void *flush_data;
 };
 
 struct radeon_winsys* radeon_pipe_winsys(int fb);
index 69f14e54f2672ba97ec248c7fd19798460343c9b..05194fc52a2d1ad7371515ae349c545ef3eb44c0 100644 (file)
  *      Joakim Sindholt <opensource@zhasha.com>
  */
 
+#include "softpipe/sp_winsys.h"
+
 #include "radeon_drm.h"
 
+/* Helper function to do the ioctls needed for setup and init. */
+static void do_ioctls(int fd, struct radeon_winsys* winsys)
+{
+    struct drm_radeon_gem_info gem_info = {0};
+    struct drm_radeon_info info = {0};
+    int target = 0;
+    int retval;
+
+    info.value = (unsigned long)&target;
+
+    /* We do things in a specific order here.
+     *
+     * First, the PCI ID. This is essential and should return usable numbers
+     * for all Radeons. If this fails, we probably got handed an FD for some
+     * non-Radeon card.
+     *
+     * The GB and Z pipe requests should always succeed, but they might not
+     * return sensical values for all chipsets, but that's alright because
+     * the pipe drivers already know that.
+     *
+     * The GEM info is actually bogus on the kernel side, as well as our side
+     * (see radeon_gem_info_ioctl in radeon_gem.c) but that's alright because
+     * we don't actually use the info for anything yet.
+     * XXX update the above when we can safely use vram_size instead of vram_visible */
+    info.request = RADEON_INFO_DEVICE_ID;
+    retval = drmCommandWriteRead(fd, DRM_RADEON_INFO, &info, sizeof(info));
+    if (retval) {
+        fprintf(stderr, "%s: Failed to get PCI ID, "
+                "error number %d\n", __FUNCTION__, retval);
+        exit(1);
+    }
+    winsys->pci_id = target;
+
+    info.request = RADEON_INFO_NUM_GB_PIPES;
+    retval = drmCommandWriteRead(fd, DRM_RADEON_INFO, &info, sizeof(info));
+    if (retval) {
+        fprintf(stderr, "%s: Failed to get GB pipe count, "
+                "error number %d\n", __FUNCTION__, retval);
+        exit(1);
+    }
+    winsys->gb_pipes = target;
+
+    info.request = RADEON_INFO_NUM_Z_PIPES;
+    retval = drmCommandWriteRead(fd, DRM_RADEON_INFO, &info, sizeof(info));
+    if (retval) {
+        fprintf(stderr, "%s: Failed to get Z pipe count, "
+                "error number %d\n", __FUNCTION__, retval);
+        exit(1);
+    }
+    winsys->z_pipes = target;
+
+    retval = drmCommandWriteRead(fd, DRM_RADEON_GEM_INFO,
+            &gem_info, sizeof(gem_info));
+    if (retval) {
+        fprintf(stderr, "%s: Failed to get MM info, error number %d\n",
+                __FUNCTION__, retval);
+        exit(1);
+    }
+    winsys->gart_size = gem_info.gart_size;
+    /* XXX */
+    winsys->vram_size = gem_info.vram_visible;
+}
+
+/* Guess at whether this chipset should use r300g.
+ *
+ * I believe that this check is valid, but I haven't been exhaustive. */
+static boolean is_r3xx(int pciid)
+{
+    return (pciid > 0x3150) && (pciid < 0x796f);
+}
+
 /* Create a pipe_screen. */
 struct pipe_screen* radeon_create_screen(struct drm_api* api,
                                          int drmFB,
                                          struct drm_create_screen_arg *arg)
 {
-    struct radeon_winsys* winsys = radeon_pipe_winsys(drmFB);
+    struct radeon_winsys* rwinsys = radeon_pipe_winsys(drmFB);
+    do_ioctls(drmFB, rwinsys);
 
-    if (debug_get_bool_option("RADEON_SOFTPIPE", FALSE)) {
-        return softpipe_create_screen((struct pipe_winsys*)winsys);
+    if (!is_r3xx(rwinsys->pci_id) ||
+        debug_get_bool_option("RADEON_SOFTPIPE", FALSE)) {
+        return softpipe_create_screen((struct pipe_winsys*)rwinsys);
     } else {
-        struct r300_winsys* r300 = radeon_create_r300_winsys(drmFB, winsys);
-        FREE(winsys);
-        return r300_create_screen(r300);
+        radeon_setup_winsys(drmFB, rwinsys);
+        return r300_create_screen(rwinsys);
     }
 }
 
@@ -51,11 +125,13 @@ struct pipe_screen* radeon_create_screen(struct drm_api* api,
 struct pipe_context* radeon_create_context(struct drm_api* api,
                                            struct pipe_screen* screen)
 {
-    if (debug_get_bool_option("RADEON_SOFTPIPE", FALSE)) {
-        return radeon_create_softpipe(screen->winsys);
+    struct radeon_winsys* rwinsys = (struct radeon_winsys*)screen->winsys;
+
+    if (!is_r3xx(rwinsys->pci_id) ||
+        debug_get_bool_option("RADEON_SOFTPIPE", FALSE)) {
+        return softpipe_create(screen);
     } else {
-        return r300_create_context(screen,
-                                   (struct r300_winsys*)screen->winsys);
+        return r300_create_context(screen, rwinsys);
     }
 }
 
@@ -130,7 +206,7 @@ static boolean radeon_shared_handle_from_texture(struct drm_api *api,
     int retval, fd;
     struct drm_gem_flink flink;
     struct radeon_pipe_buffer* radeon_buffer;
-    struct pipe_buffer *buffer;
+    struct pipe_buffer *buffer = NULL;
 
     if (!radeon_buffer_from_texture(api, texture, &buffer, stride)) {
         return FALSE;
@@ -163,7 +239,7 @@ static boolean radeon_local_handle_from_texture(struct drm_api *api,
                                                 unsigned *stride,
                                                 unsigned *handle)
 {
-    struct pipe_buffer *buffer;
+    struct pipe_buffer *buffer = NULL;
     if (!radeon_buffer_from_texture(api, texture, &buffer, stride)) {
         return FALSE;
     }
index 9a789ec1a45c2cb676edad451d3e283ee4aa933c..bf0e78138d75e829150386cfd9bfab9f89a3fe7f 100644 (file)
@@ -44,7 +44,6 @@
 
 #include "radeon_buffer.h"
 #include "radeon_r300.h"
-#include "radeon_winsys_softpipe.h"
 
 /* XXX */
 #include "r300_screen.h"
index 7ea5d1fb4e76da4d1c082458249c3d3b9e509070..0875ee41cbfd76d0311258cd5871331f44bff330 100644 (file)
 
 #include "radeon_r300.h"
 
-static void radeon_r300_set_flush_cb(struct r300_winsys *winsys,
-                                    void (*flush_cb)(void *),
-                                    void *data)
+static void radeon_set_flush_cb(struct radeon_winsys *winsys,
+                                void (*flush_cb)(void *),
+                                void *data)
 {
-    struct radeon_winsys_priv* priv =
-        (struct radeon_winsys_priv*)winsys->radeon_winsys;
-
-    radeon_cs_space_set_flush(priv->cs, flush_cb,
-                             data);
+    winsys->priv->flush_cb = flush_cb;
+    winsys->priv->flush_data = data;
+    radeon_cs_space_set_flush(winsys->priv->cs, flush_cb, data);
 }
 
-static boolean radeon_r300_add_buffer(struct r300_winsys* winsys,
-                                      struct pipe_buffer* pbuffer,
-                                      uint32_t rd,
-                                      uint32_t wd)
+static boolean radeon_add_buffer(struct radeon_winsys* winsys,
+                                 struct pipe_buffer* pbuffer,
+                                 uint32_t rd,
+                                 uint32_t wd)
 {
-    struct radeon_winsys_priv* priv =
-        (struct radeon_winsys_priv*)winsys->radeon_winsys;
     struct radeon_bo* bo = ((struct radeon_pipe_buffer*)pbuffer)->bo;
 
-    radeon_cs_space_add_persistent_bo(priv->cs, bo, rd, wd);
+    radeon_cs_space_add_persistent_bo(winsys->priv->cs, bo, rd, wd);
     return TRUE;
 }
 
-static boolean radeon_r300_validate(struct r300_winsys* winsys)
+static boolean radeon_validate(struct radeon_winsys* winsys)
 {
-    struct radeon_winsys_priv* priv =
-        (struct radeon_winsys_priv*)winsys->radeon_winsys;
-
-    if (radeon_cs_space_check(priv->cs) < 0) {
+    if (radeon_cs_space_check(winsys->priv->cs) < 0) {
         return FALSE;
     }
 
@@ -59,45 +52,37 @@ static boolean radeon_r300_validate(struct r300_winsys* winsys)
     return TRUE;
 }
 
-static boolean radeon_r300_check_cs(struct r300_winsys* winsys, int size)
+static boolean radeon_check_cs(struct radeon_winsys* winsys, int size)
 {
-    /* XXX check size here, lazy ass! */
-    /* XXX also validate buffers */
-    return TRUE;
+    struct radeon_cs* cs = winsys->priv->cs;
+
+    return radeon_validate(winsys) && cs->cdw + size <= cs->ndw;
 }
 
-static void radeon_r300_begin_cs(struct r300_winsys* winsys,
-                                 int size,
-                                 const char* file,
-                                 const char* function,
-                                 int line)
+static void radeon_begin_cs(struct radeon_winsys* winsys,
+                            int size,
+                            const char* file,
+                            const char* function,
+                            int line)
 {
-    struct radeon_winsys_priv* priv =
-        (struct radeon_winsys_priv*)winsys->radeon_winsys;
-
-    radeon_cs_begin(priv->cs, size, file, function, line);
+    radeon_cs_begin(winsys->priv->cs, size, file, function, line);
 }
 
-static void radeon_r300_write_cs_dword(struct r300_winsys* winsys,
-                                       uint32_t dword)
+static void radeon_write_cs_dword(struct radeon_winsys* winsys,
+                                  uint32_t dword)
 {
-    struct radeon_winsys_priv* priv =
-        (struct radeon_winsys_priv*)winsys->radeon_winsys;
-
-    radeon_cs_write_dword(priv->cs, dword);
+    radeon_cs_write_dword(winsys->priv->cs, dword);
 }
 
-static void radeon_r300_write_cs_reloc(struct r300_winsys* winsys,
-                                       struct pipe_buffer* pbuffer,
-                                       uint32_t rd,
-                                       uint32_t wd,
-                                       uint32_t flags)
+static void radeon_write_cs_reloc(struct radeon_winsys* winsys,
+                                  struct pipe_buffer* pbuffer,
+                                  uint32_t rd,
+                                  uint32_t wd,
+                                  uint32_t flags)
 {
-    struct radeon_winsys_priv* priv =
-        (struct radeon_winsys_priv*)winsys->radeon_winsys;
     int retval = 0;
 
-    retval = radeon_cs_write_reloc(priv->cs,
+    retval = radeon_cs_write_reloc(winsys->priv->cs,
             ((struct radeon_pipe_buffer*)pbuffer)->bo, rd, wd, flags);
 
     if (retval) {
@@ -106,132 +91,60 @@ static void radeon_r300_write_cs_reloc(struct r300_winsys* winsys,
     }
 }
 
-static void radeon_r300_reset_bos(struct r300_winsys *winsys)
+static void radeon_reset_bos(struct radeon_winsys *winsys)
 {
-    struct radeon_winsys_priv* priv =
-        (struct radeon_winsys_priv*)winsys->radeon_winsys;
-    radeon_cs_space_reset_bos(priv->cs);
+    radeon_cs_space_reset_bos(winsys->priv->cs);
 }
 
-static void radeon_r300_end_cs(struct r300_winsys* winsys,
-                               const char* file,
-                               const char* function,
-                               int line)
+static void radeon_end_cs(struct radeon_winsys* winsys,
+                          const char* file,
+                          const char* function,
+                          int line)
 {
-    struct radeon_winsys_priv* priv =
-        (struct radeon_winsys_priv*)winsys->radeon_winsys;
-
-    radeon_cs_end(priv->cs, file, function, line);
+    radeon_cs_end(winsys->priv->cs, file, function, line);
 }
 
-static void radeon_r300_flush_cs(struct r300_winsys* winsys)
+static void radeon_flush_cs(struct radeon_winsys* winsys)
 {
-    struct radeon_winsys_priv* priv =
-        (struct radeon_winsys_priv*)winsys->radeon_winsys;
     int retval;
 
     /* Emit the CS. */
-    retval = radeon_cs_emit(priv->cs);
+    retval = radeon_cs_emit(winsys->priv->cs);
     if (retval) {
         debug_printf("radeon: Bad CS, dumping...\n");
-        radeon_cs_print(priv->cs, stderr);
+        radeon_cs_print(winsys->priv->cs, stderr);
     }
 
     /* Reset CS.
      * Someday, when we care about performance, we should really find a way
      * to rotate between two or three CS objects so that the GPU can be
      * spinning through one CS while another one is being filled. */
-    radeon_cs_erase(priv->cs);
+    radeon_cs_erase(winsys->priv->cs);
 }
 
-/* Helper function to do the ioctls needed for setup and init. */
-static void do_ioctls(struct r300_winsys* winsys, int fd)
+void
+radeon_setup_winsys(int fd, struct radeon_winsys* winsys)
 {
-    struct drm_radeon_gem_info gem_info = {0};
-    struct drm_radeon_info info = {0};
-    int target = 0;
-    int retval;
-
-    info.value = (unsigned long)&target;
-
-    /* First, get the number of pixel pipes */
-    info.request = RADEON_INFO_NUM_GB_PIPES;
-    retval = drmCommandWriteRead(fd, DRM_RADEON_INFO, &info, sizeof(info));
-    if (retval) {
-        fprintf(stderr, "%s: Failed to get GB pipe count, "
-                "error number %d\n", __FUNCTION__, retval);
-        exit(1);
-    }
-    winsys->gb_pipes = target;
-
-    /* get Z pipes */
-    info.request = RADEON_INFO_NUM_Z_PIPES;
-    retval = drmCommandWriteRead(fd, DRM_RADEON_INFO, &info, sizeof(info));
-    if (retval) {
-        fprintf(stderr, "%s: Failed to get GB pipe count, "
-                "error number %d\n", __FUNCTION__, retval);
-        exit(1);
-    }
-    winsys->z_pipes = target;
-
-    /* Then, get PCI ID */
-    info.request = RADEON_INFO_DEVICE_ID;
-    retval = drmCommandWriteRead(fd, DRM_RADEON_INFO, &info, sizeof(info));
-    if (retval) {
-        fprintf(stderr, "%s: Failed to get PCI ID, "
-                "error number %d\n", __FUNCTION__, retval);
-        exit(1);
-    }
-    winsys->pci_id = target;
-
-    /* Finally, retrieve MM info */
-    retval = drmCommandWriteRead(fd, DRM_RADEON_GEM_INFO,
-            &gem_info, sizeof(gem_info));
-    if (retval) {
-        fprintf(stderr, "%s: Failed to get MM info, error number %d\n",
-                __FUNCTION__, retval);
-        exit(1);
-    }
-    winsys->gart_size = gem_info.gart_size;
-    /* XXX */
-    winsys->vram_size = gem_info.vram_visible;
-}
-
-struct r300_winsys*
-radeon_create_r300_winsys(int fd, struct radeon_winsys* old_winsys)
-{
-    struct r300_winsys* winsys = CALLOC_STRUCT(r300_winsys);
-    struct radeon_winsys_priv* priv;
-
-    if (winsys == NULL) {
-        return NULL;
-    }
-
-    priv = old_winsys->priv;
-
-    do_ioctls(winsys, fd);
+    struct radeon_winsys_priv* priv = winsys->priv;
 
     priv->csm = radeon_cs_manager_gem_ctor(fd);
 
+    /* Size limit on IBs is 64 kibibytes. */
     priv->cs = radeon_cs_create(priv->csm, 1024 * 64 / 4);
     radeon_cs_set_limit(priv->cs,
             RADEON_GEM_DOMAIN_GTT, winsys->gart_size);
     radeon_cs_set_limit(priv->cs,
             RADEON_GEM_DOMAIN_VRAM, winsys->vram_size);
 
-    winsys->add_buffer = radeon_r300_add_buffer;
-    winsys->validate = radeon_r300_validate;
-
-    winsys->check_cs = radeon_r300_check_cs;
-    winsys->begin_cs = radeon_r300_begin_cs;
-    winsys->write_cs_dword = radeon_r300_write_cs_dword;
-    winsys->write_cs_reloc = radeon_r300_write_cs_reloc;
-    winsys->end_cs = radeon_r300_end_cs;
-    winsys->flush_cs = radeon_r300_flush_cs;
-    winsys->reset_bos = radeon_r300_reset_bos;
-    winsys->set_flush_cb = radeon_r300_set_flush_cb;
-
-    memcpy(winsys, old_winsys, sizeof(struct radeon_winsys));
-
-    return winsys;
+    winsys->add_buffer = radeon_add_buffer;
+    winsys->validate = radeon_validate;
+
+    winsys->check_cs = radeon_check_cs;
+    winsys->begin_cs = radeon_begin_cs;
+    winsys->write_cs_dword = radeon_write_cs_dword;
+    winsys->write_cs_reloc = radeon_write_cs_reloc;
+    winsys->end_cs = radeon_end_cs;
+    winsys->flush_cs = radeon_flush_cs;
+    winsys->reset_bos = radeon_reset_bos;
+    winsys->set_flush_cb = radeon_set_flush_cb;
 }
index 775d7937fd85e002c01be668c9693a3a6916260a..cfbdb302661eebb9368bc62726cf4460c4ddfbc4 100644 (file)
@@ -34,9 +34,6 @@
 
 #include "radeon_buffer.h"
 
-struct radeon_winsys;
-
-struct r300_winsys*
-radeon_create_r300_winsys(int fd, struct radeon_winsys* old_winsys);
+void radeon_setup_winsys(int fd, struct radeon_winsys* winsys);
 
 #endif /* RADEON_R300_H */
diff --git a/src/gallium/winsys/drm/radeon/core/radeon_winsys.h b/src/gallium/winsys/drm/radeon/core/radeon_winsys.h
new file mode 100644 (file)
index 0000000..9edc9e0
--- /dev/null
@@ -0,0 +1,105 @@
+/*
+ * Copyright Â© 2009 Corbin Simpson
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
+ * OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
+ * NON-INFRINGEMENT. IN NO EVENT SHALL THE COPYRIGHT HOLDERS, AUTHORS
+ * AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
+ * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
+ * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE
+ * USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ */
+/*
+ * Authors:
+ *      Corbin Simpson <MostAwesomeDude@gmail.com>
+ */
+#ifndef RADEON_WINSYS_H
+#define RADEON_WINSYS_H
+
+#include "pipe/internal/p_winsys_screen.h"
+
+struct radeon_winsys_priv;
+
+struct radeon_winsys {
+    /* Parent class. */
+    struct pipe_winsys base;
+
+    /* Winsys private */
+    struct radeon_winsys_priv* priv;
+
+    /* PCI ID */
+    uint32_t pci_id;
+
+    /* GB pipe count */
+    uint32_t gb_pipes;
+
+    /* Z pipe count (rv530 only) */
+    uint32_t z_pipes;
+
+    /* GART size. */
+    uint32_t gart_size;
+
+    /* VRAM size. */
+    uint32_t vram_size;
+
+    /* Add a pipe_buffer to the list of buffer objects to validate. */
+    boolean (*add_buffer)(struct radeon_winsys* winsys,
+                          struct pipe_buffer* pbuffer,
+                          uint32_t rd,
+                          uint32_t wd);
+
+    /* Revalidate all currently setup pipe_buffers.
+     * Returns TRUE if a flush is required. */
+    boolean (*validate)(struct radeon_winsys* winsys);
+
+    /* Check to see if there's room for commands. */
+    boolean (*check_cs)(struct radeon_winsys* winsys, int size);
+
+    /* Start a command emit. */
+    void (*begin_cs)(struct radeon_winsys* winsys,
+                     int size,
+                     const char* file,
+                     const char* function,
+                     int line);
+
+    /* Write a dword to the command buffer. */
+    void (*write_cs_dword)(struct radeon_winsys* winsys, uint32_t dword);
+
+    /* Write a relocated dword to the command buffer. */
+    void (*write_cs_reloc)(struct radeon_winsys* winsys,
+                           struct pipe_buffer* bo,
+                           uint32_t rd,
+                           uint32_t wd,
+                           uint32_t flags);
+
+    /* Finish a command emit. */
+    void (*end_cs)(struct radeon_winsys* winsys,
+                   const char* file,
+                   const char* function,
+                   int line);
+
+    /* Flush the CS. */
+    void (*flush_cs)(struct radeon_winsys* winsys);
+
+    /* winsys flush - callback from winsys when flush required */
+    void (*set_flush_cb)(struct radeon_winsys *winsys,
+                        void (*flush_cb)(void *), void *data);
+
+    void (*reset_bos)(struct radeon_winsys *winsys);
+};
+
+#endif
diff --git a/src/gallium/winsys/drm/radeon/core/radeon_winsys_softpipe.c b/src/gallium/winsys/drm/radeon/core/radeon_winsys_softpipe.c
deleted file mode 100644 (file)
index f038bfa..0000000
+++ /dev/null
@@ -1,41 +0,0 @@
-/**************************************************************************
- * 
- * Copyright 2006 Tungsten Graphics, Inc., Bismarck, ND., USA
- * All Rights Reserved.
- * 
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sub license, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- * 
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
- * IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT. IN NO EVENT SHALL
- * THE COPYRIGHT HOLDERS, AUTHORS AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM,
- * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR 
- * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE 
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- * 
- * 
- **************************************************************************/
-/*
- * Authors: Keith Whitwell <keithw-at-tungstengraphics-dot-com>
- */
-
-#include "radeon_winsys_softpipe.h"
-
-struct pipe_context *radeon_create_softpipe(struct pipe_winsys* winsys)
-{
-    struct pipe_screen *pipe_screen;
-
-    pipe_screen = softpipe_create_screen(winsys);
-
-    return softpipe_create(pipe_screen);
-}
diff --git a/src/gallium/winsys/drm/radeon/core/radeon_winsys_softpipe.h b/src/gallium/winsys/drm/radeon/core/radeon_winsys_softpipe.h
deleted file mode 100644 (file)
index 04740e4..0000000
+++ /dev/null
@@ -1,44 +0,0 @@
-/* 
- * Copyright Â© 2008 Jérôme Glisse
- * All Rights Reserved.
- * 
- * Permission is hereby granted, free of charge, to any person obtaining
- * a copy of this software and associated documentation files (the
- * "Software"), to deal in the Software without restriction, including
- * without limitation the rights to use, copy, modify, merge, publish,
- * distribute, sub license, and/or sell copies of the Software, and to
- * permit persons to whom the Software is furnished to do so, subject to
- * the following conditions:
- * 
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
- * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES
- * OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
- * NON-INFRINGEMENT. IN NO EVENT SHALL THE COPYRIGHT HOLDERS, AUTHORS
- * AND/OR ITS SUPPLIERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
- * LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE,
- * ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE 
- * USE OR OTHER DEALINGS IN THE SOFTWARE.
- *
- * The above copyright notice and this permission notice (including the
- * next paragraph) shall be included in all copies or substantial portions
- * of the Software.
- */
-/*
- * Authors:
- *      Jérôme Glisse <glisse@freedesktop.org>
- */
-#ifndef RADEON_WINSYS_SOFTPIPE_H
-#define RADEON_WINSYS_SOFTPIPE_H
-
-#include <stdio.h>
-
-#include "pipe/p_defines.h"
-#include "pipe/p_format.h"
-
-#include "softpipe/sp_winsys.h"
-
-#include "util/u_memory.h"
-
-struct pipe_context *radeon_create_softpipe(struct pipe_winsys* winsys);
-
-#endif
index 9fa16dab24cc79c1c9bc848840d7323b3f6c1378..0eb1b3988f36acfb2df83730de48e3de845a0ce8 100644 (file)
@@ -1,11 +1,16 @@
-TARGET     = modesetting_drv.so
-CFILES     = $(wildcard ./*.c)
-OBJECTS    = $(patsubst ./%.c,./%.o,$(CFILES))
-GALLIUMDIR = ../../../..
 TOP        = ../../../../../..
 
+
+GALLIUMDIR = $(TOP)/src/gallium
+
+TARGET     = radeong_drv.so
+
+CFILES     = $(wildcard ./*.c)
+
 include ${TOP}/configs/current
 
+OBJECTS    = $(patsubst ./%.c,./%.o,$(CFILES))
+
 CFLAGS = -DHAVE_CONFIG_H \
          -g -Wall -Wimplicit-function-declaration -fPIC \
          $(shell pkg-config --cflags pixman-1 xorg-server libdrm xproto) \
@@ -24,16 +29,21 @@ LIBS = \
        $(TOP)/src/gallium/drivers/softpipe/libsoftpipe.a \
        $(GALLIUM_AUXILIARIES)
 
+TARGET_STAGING = $(TOP)/$(LIB_DIR)/gallium/$(TARGET)
 #############################################
 
+all default: $(TARGET) $(TARGET_STAGING)
 
-
-all default: $(TARGET)
-
-$(TARGET): $(OBJECTS) Makefile $(GALLIUMDIR)/state_trackers/xorg/libxorgtracker.a
+$(TARGET): $(OBJECTS) Makefile $(GALLIUMDIR)/state_trackers/xorg/libxorgtracker.a $(LIBS)
        $(TOP)/bin/mklib -noprefix -o $@ \
        $(OBJECTS) $(LIBS) $(shell pkg-config --libs libdrm) -ldrm_radeon
 
+$(TOP)/$(LIB_DIR)/gallium:
+       mkdir -p $@
+
+$(TARGET_STAGING): $(TARGET) $(TOP)/$(LIB_DIR)/gallium
+       $(INSTALL) $(TARGET) $(TOP)/$(LIB_DIR)/gallium
+
 clean:
        rm -rf $(OBJECTS) $(TARGET)
 
index 837f2aa8fec5106021e384eeff00f3783afba05c..bb76cc03499d61efe3c947e86ab26718c102819a 100644 (file)
@@ -53,7 +53,7 @@ static PciChipsets radeon_xorg_pci_devices[] = {
 };
 
 static XF86ModuleVersionInfo radeon_xorg_version = {
-    "modesetting",
+    "radeong",
     MODULEVENDORSTRING,
     MODINFOSTRING1,
     MODINFOSTRING2,
@@ -69,9 +69,9 @@ static XF86ModuleVersionInfo radeon_xorg_version = {
  * Xorg driver exported structures
  */
 
-_X_EXPORT DriverRec modesetting = {
+_X_EXPORT DriverRec radeong = {
     1,
-    "modesetting",
+    "radeong",
     radeon_xorg_identify,
     NULL,
     xorg_tracker_available_options,
@@ -84,7 +84,7 @@ _X_EXPORT DriverRec modesetting = {
 
 static MODULESETUPPROTO(radeon_xorg_setup);
 
-_X_EXPORT XF86ModuleData modesettingModuleData = {
+_X_EXPORT XF86ModuleData radeongModuleData = {
     &radeon_xorg_version,
     radeon_xorg_setup,
     NULL
@@ -103,7 +103,7 @@ radeon_xorg_setup(pointer module, pointer opts, int *errmaj, int *errmin)
      */
     if (!setupDone) {
        setupDone = 1;
-       xf86AddDriver(&modesetting, module, HaveDriverFuncs);
+       xf86AddDriver(&radeong, module, HaveDriverFuncs);
 
        /*
         * The return value must be non-NULL on success even though there
@@ -120,7 +120,7 @@ radeon_xorg_setup(pointer module, pointer opts, int *errmaj, int *errmin)
 static void
 radeon_xorg_identify(int flags)
 {
-    xf86PrintChipsets("modesetting", "Driver for Modesetting Kernel Drivers",
+    xf86PrintChipsets("radeong", "Driver for Radeon Gallium with KMS",
                      radeon_xorg_chipsets);
 }
 
@@ -135,8 +135,8 @@ radeon_xorg_pci_probe(DriverPtr driver,
                               NULL, NULL, NULL, NULL, NULL);
     if (scrn != NULL) {
        scrn->driverVersion = 1;
-       scrn->driverName = "radeon";
-       scrn->name = "modesetting";
+       scrn->driverName = "radeong";
+       scrn->name = "radeong";
        scrn->Probe = NULL;
 
        entity = xf86GetEntityInfo(entity_num);
index 64eb32f8b94ef8f52356ef27af5c99409f50e03c..5f1b9ad577093dd49099fc6d02f46fb0728970a2 100644 (file)
@@ -47,7 +47,7 @@ vmw_svga_winsys_surface_reference(struct vmw_svga_winsys_surface **pdst,
    src_ref = src ? &src->refcnt : NULL;
    dst_ref = dst ? &dst->refcnt : NULL;
 
-   if (pipe_reference(&dst_ref, src_ref)) {
+   if (pipe_reference(dst_ref, src_ref)) {
       vmw_ioctl_surface_destroy(dst->screen, dst->sid);
 #ifdef DEBUG
       /* to detect dangling pointers */
index d02f82504781b794a3c91c6aceeaeaf1778ee8ae..599973ce127ce3caacb57e6afe29c0b516dc6de8 100644 (file)
@@ -41,6 +41,7 @@
 #include "pipe/p_state.h"
 #include "pipe/internal/p_winsys_screen.h"
 #include "util/u_memory.h"
+#include "util/u_math.h"
 #include "softpipe/sp_winsys.h"
 #include "softpipe/sp_texture.h"
 
@@ -138,17 +139,6 @@ lookup_context(_EGLContext *ctx)
 }
 
 
-static unsigned int
-bitcount(unsigned int n)
-{
-   unsigned int bits;
-   for (bits = 0; n > 0; n = n >> 1) {
-      bits += (n & 1);
-   }
-   return bits;
-}
-
-
 /**
  * Create the EGLConfigs.  (one per X visual)
  */
@@ -174,9 +164,9 @@ create_configs(struct xlib_egl_display *xdpy, _EGLDisplay *disp)
    for (i = 0; i < num_visuals; i++) {
       _EGLConfig *config = calloc(1, sizeof(_EGLConfig));
       int id = i + 1;
-      int rbits = bitcount(visInfo[i].red_mask);
-      int gbits = bitcount(visInfo[i].green_mask);
-      int bbits = bitcount(visInfo[i].blue_mask);
+      int rbits = util_bitcount(visInfo[i].red_mask);
+      int gbits = util_bitcount(visInfo[i].green_mask);
+      int bbits = util_bitcount(visInfo[i].blue_mask);
       int abits = bbits == 8 ? 8 : 0;
       int zbits = 24;
       int sbits = 8;
index 79ff2cc985d1b5120389fb7b1904ab923c0bb2a1..6ee3ede38cb0d772cb1c5ca4d96dcebfcf89ef15 100644 (file)
@@ -38,6 +38,7 @@
 #include "pipe/internal/p_winsys_screen.h"
 #include "pipe/p_state.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -71,16 +72,6 @@ sw_pipe_buffer(struct pipe_buffer *b)
 }
 
 
-/**
- * Round n up to next multiple.
- */
-static INLINE unsigned
-round_up(unsigned n, unsigned multiple)
-{
-   return (n + multiple - 1) & ~(multiple - 1);
-}
-
-
 static const char *
 get_name(struct pipe_winsys *pws)
 {
@@ -170,13 +161,10 @@ surface_buffer_create(struct pipe_winsys *winsys,
                       unsigned *stride)
 {
    const unsigned alignment = 64;
-   struct pipe_format_block block;
-   unsigned nblocksx, nblocksy;
+   unsigned nblocksy;
 
-   pf_get_block(format, &block);
-   nblocksx = pf_get_nblocksx(&block, width);
-   nblocksy = pf_get_nblocksy(&block, height);
-   *stride = round_up(nblocksx * block.size, alignment);
+   nblocksy = util_format_get_nblocksy(format, height);
+   *stride = align(util_format_get_stride(format, width), alignment);
 
    return winsys->buffer_create(winsys, alignment,
                                 usage,
index 08067aad64c4c3f3cc1964da3db457cc0a62b274..f15bcd37b501ec11a144b166d955049c93996896 100644 (file)
@@ -30,6 +30,7 @@
 #include <pipe/internal/p_winsys_screen.h>
 #include <pipe/p_state.h>
 #include <pipe/p_inlines.h>
+#include <util/u_format.h>
 #include <util/u_memory.h>
 #include <util/u_math.h>
 #include <softpipe/sp_winsys.h>
@@ -138,13 +139,10 @@ static struct pipe_buffer* xsp_surface_buffer_create
 )
 {
    const unsigned int ALIGNMENT = 1;
-   struct pipe_format_block block;
-   unsigned nblocksx, nblocksy;
+   unsigned nblocksy;
 
-   pf_get_block(format, &block);
-   nblocksx = pf_get_nblocksx(&block, width);
-   nblocksy = pf_get_nblocksy(&block, height);
-   *stride = align(nblocksx * block.size, ALIGNMENT);
+   nblocksy = util_format_get_nblocksy(format, height);
+   *stride = align(util_format_get_stride(format, width), ALIGNMENT);
 
    return pws->buffer_create(pws, ALIGNMENT, usage,
                              *stride * nblocksy);
index 8f556daf04a79d149187c176f55095a978e3313f..74f6b2fd4751590c34ab53b578c0bc904843f149 100644 (file)
@@ -45,5 +45,5 @@ if env['platform'] == 'windows':
     env.SharedLibrary(
         target ='opengl32',
         source = sources,
-        LIBS = wgl + glapi + mesa + drivers + auxiliaries + env['LIBS'],
+        LIBS = wgl + glapi + mesa + drivers + auxiliaries + glsl + env['LIBS'],
     )
index e8bc0f55ac4d81118167c895f1ef2d0bf498251e..7d076be3a31c5264cb3d2db74069651a0e19c230 100644 (file)
@@ -39,6 +39,7 @@
 #include "pipe/p_format.h"
 #include "pipe/p_context.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 #include "llvmpipe/lp_winsys.h"
@@ -49,7 +50,6 @@
 struct gdi_llvmpipe_displaytarget
 {
    enum pipe_format format;
-   struct pipe_format_block block;
    unsigned width;
    unsigned height;
    unsigned stride;
@@ -118,16 +118,6 @@ gdi_llvmpipe_displaytarget_destroy(struct llvmpipe_winsys *winsys,
 }
 
 
-/**
- * Round n up to next multiple.
- */
-static INLINE unsigned
-round_up(unsigned n, unsigned multiple)
-{
-   return (n + multiple - 1) & ~(multiple - 1);
-}
-
-
 static struct llvmpipe_displaytarget *
 gdi_llvmpipe_displaytarget_create(struct llvmpipe_winsys *winsys,
                                   enum pipe_format format,
@@ -147,10 +137,10 @@ gdi_llvmpipe_displaytarget_create(struct llvmpipe_winsys *winsys,
    gdt->width = width;
    gdt->height = height;
 
-   bpp = pf_get_bits(format);
-   cpp = pf_get_size(format);
+   bpp = util_format_get_blocksizebits(format);
+   cpp = util_format_get_blocksize(format);
    
-   gdt->stride = round_up(width * cpp, alignment);
+   gdt->stride = align(width * cpp, alignment);
    gdt->size = gdt->stride * height;
    
    gdt->data = align_malloc(gdt->size, alignment);
index 5e0ccf32f48f77bbf6711158505ffebc4b39e892..2ad794c3f0fcfca1d1887d5a5dfe517a2dd3c7f1 100644 (file)
@@ -42,6 +42,7 @@
 #include "pipe/p_format.h"
 #include "pipe/p_context.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 #include "softpipe/sp_winsys.h"
@@ -151,16 +152,6 @@ gdi_softpipe_user_buffer_create(struct pipe_winsys *winsys,
 }
 
 
-/**
- * Round n up to next multiple.
- */
-static INLINE unsigned
-round_up(unsigned n, unsigned multiple)
-{
-   return (n + multiple - 1) & ~(multiple - 1);
-}
-
-
 static struct pipe_buffer *
 gdi_softpipe_surface_buffer_create(struct pipe_winsys *winsys,
                                    unsigned width, unsigned height,
@@ -170,13 +161,10 @@ gdi_softpipe_surface_buffer_create(struct pipe_winsys *winsys,
                                    unsigned *stride)
 {
    const unsigned alignment = 64;
-   struct pipe_format_block block;
-   unsigned nblocksx, nblocksy;
+   unsigned nblocksy;
 
-   pf_get_block(format, &block);
-   nblocksx = pf_get_nblocksx(&block, width);
-   nblocksy = pf_get_nblocksy(&block, height);
-   *stride = round_up(nblocksx * block.size, alignment);
+   nblocksy = util_format_get_nblocksy(format, height);
+   *stride = align(util_format_get_stride(format, width), alignment);
 
    return winsys->buffer_create(winsys, alignment,
                                 usage,
@@ -283,10 +271,10 @@ gdi_softpipe_present(struct pipe_screen *screen,
 
     memset(&bmi, 0, sizeof(BITMAPINFO));
     bmi.bmiHeader.biSize = sizeof(BITMAPINFOHEADER);
-    bmi.bmiHeader.biWidth = texture->stride[surface->level] / pf_get_size(surface->format);
+    bmi.bmiHeader.biWidth = texture->stride[surface->level] / util_format_get_blocksize(surface->format);
     bmi.bmiHeader.biHeight= -(long)surface->height;
     bmi.bmiHeader.biPlanes = 1;
-    bmi.bmiHeader.biBitCount = pf_get_bits(surface->format);
+    bmi.bmiHeader.biBitCount = util_format_get_blocksizebits(surface->format);
     bmi.bmiHeader.biCompression = BI_RGB;
     bmi.bmiHeader.biSizeImage = 0;
     bmi.bmiHeader.biXPelsPerMeter = 0;
index 5f8dc2ab241ada3e5e8d560ac393fa9f0f0f04f7..47ae0519a4baa5fcdab7b3abd14cdf3daacb1fd3 100644 (file)
@@ -45,6 +45,7 @@
 #include "pipe/p_format.h"
 #include "pipe/p_context.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 
@@ -277,15 +278,6 @@ xm_user_buffer_create(struct pipe_winsys *pws, void *ptr, unsigned bytes)
 
 
 
-/**
- * Round n up to next multiple.
- */
-static INLINE unsigned
-round_up(unsigned n, unsigned multiple)
-{
-   return (n + multiple - 1) & ~(multiple - 1);
-}
-
 static struct pipe_buffer *
 xm_surface_buffer_create(struct pipe_winsys *winsys,
                          unsigned width, unsigned height,
@@ -295,18 +287,15 @@ xm_surface_buffer_create(struct pipe_winsys *winsys,
                          unsigned *stride)
 {
    const unsigned alignment = 64;
-   struct pipe_format_block block;
-   unsigned nblocksx, nblocksy;
+   unsigned nblocksy;
 
-   pf_get_block(format, &block);
-   nblocksx = pf_get_nblocksx(&block, width);
-   nblocksy = pf_get_nblocksy(&block, height);
-   *stride = round_up(nblocksx * block.size, alignment);
+   nblocksy = util_format_get_nblocksy(format, height);
+   *stride = align(util_format_get_stride(format, width), alignment);
 
    return winsys->buffer_create(winsys, alignment,
                                 usage,
                                 /* XXX a bit of a hack */
-                                *stride * round_up(nblocksy, TILE_SIZE));
+                                *stride * align(nblocksy, TILE_SIZE));
 }
 
 
index 3dd15e099b043278e5722640a900a11e73beb0b0..2a434b5fd216c45488715cb759b0d2bec9e4ba7c 100644 (file)
@@ -44,6 +44,7 @@
 #include "pipe/p_format.h"
 #include "pipe/p_context.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 #include "llvmpipe/lp_winsys.h"
@@ -58,7 +59,6 @@
 struct xm_displaytarget
 {
    enum pipe_format format;
-   struct pipe_format_block block;
    unsigned width;
    unsigned height;
    unsigned stride;
@@ -262,10 +262,10 @@ xm_llvmpipe_display(struct xmesa_buffer *xm_buffer,
    {
       if (xm_dt->tempImage == NULL)
       {
-         assert(xm_dt->block.width == 1);
-         assert(xm_dt->block.height == 1);
+         assert(util_format_get_blockwidth(xm_dt->format) == 1);
+         assert(util_format_get_blockheight(xm_dt->format) == 1);
          alloc_shm_ximage(xm_dt, xm_buffer,
-                          xm_dt->stride / xm_dt->block.size,
+                          xm_dt->stride / util_format_get_blocksize(xm_dt->format),
                           xm_dt->height);
       }
 
@@ -321,7 +321,7 @@ xm_displaytarget_create(struct llvmpipe_winsys *winsys,
                         unsigned *stride)
 {
    struct xm_displaytarget *xm_dt = CALLOC_STRUCT(xm_displaytarget);
-   unsigned nblocksx, nblocksy, size;
+   unsigned nblocksy, size;
 
    xm_dt = CALLOC_STRUCT(xm_displaytarget);
    if(!xm_dt)
@@ -331,10 +331,8 @@ xm_displaytarget_create(struct llvmpipe_winsys *winsys,
    xm_dt->width = width;
    xm_dt->height = height;
 
-   pf_get_block(format, &xm_dt->block);
-   nblocksx = pf_get_nblocksx(&xm_dt->block, width);
-   nblocksy = pf_get_nblocksy(&xm_dt->block, height);
-   xm_dt->stride = align(nblocksx * xm_dt->block.size, alignment);
+   nblocksy = util_format_get_nblocksy(format, height);
+   xm_dt->stride = align(util_format_get_stride(format, width), alignment);
    size = xm_dt->stride * nblocksy;
 
 #ifdef USE_XSHM
index 260b39e2a0fb1dd93922baa8175590ff8b2a21bd..f7c0099584e8f77cead739623b443007b81609c5 100644 (file)
@@ -42,6 +42,7 @@
 #include "pipe/p_format.h"
 #include "pipe/p_context.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_memory.h"
 #include "softpipe/sp_winsys.h"
@@ -254,10 +255,10 @@ xlib_softpipe_display_surface(struct xmesa_buffer *b,
    {
       if (xm_buf->tempImage == NULL) 
       {
-         assert(surf->texture->block.width == 1);
-         assert(surf->texture->block.height == 1);
+         assert(util_format_get_blockwidth(surf->texture->format) == 1);
+         assert(util_format_get_blockheight(surf->texture->format) == 1);
          alloc_shm_ximage(xm_buf, b, spt->stride[surf->level] /
-                          surf->texture->block.size, surf->height);
+                          util_format_get_blocksize(surf->texture->format), surf->height);
       }
 
       ximage = xm_buf->tempImage;
@@ -360,13 +361,10 @@ xm_surface_buffer_create(struct pipe_winsys *winsys,
                          unsigned *stride)
 {
    const unsigned alignment = 64;
-   struct pipe_format_block block;
-   unsigned nblocksx, nblocksy, size;
+   unsigned nblocksy, size;
 
-   pf_get_block(format, &block);
-   nblocksx = pf_get_nblocksx(&block, width);
-   nblocksy = pf_get_nblocksy(&block, height);
-   *stride = align(nblocksx * block.size, alignment);
+   nblocksy = util_format_get_nblocksy(format, height);
+   *stride = align(util_format_get_stride(format, width), alignment);
    size = *stride * nblocksy;
 
 #ifdef USE_XSHM
diff --git a/src/glsl/Makefile b/src/glsl/Makefile
new file mode 100644 (file)
index 0000000..ca7f2d2
--- /dev/null
@@ -0,0 +1,15 @@
+# src/glsl/Makefile
+
+TOP = ../..
+
+include $(TOP)/configs/current
+
+SUBDIRS = pp cl apps
+
+default install clean:
+       @for dir in $(SUBDIRS) ; do \
+               if [ -d $$dir ] ; then \
+                       (cd $$dir && $(MAKE) $@) || exit 1; \
+               fi \
+       done
+
diff --git a/src/glsl/Makefile.template b/src/glsl/Makefile.template
new file mode 100644 (file)
index 0000000..974987a
--- /dev/null
@@ -0,0 +1,50 @@
+# src/glsl/Makefile.template
+
+# Template makefile for glsl libraries.
+#
+# Usage:
+#   The minimum that the including makefile needs to define
+#   is TOP, LIBNAME and one of of the *_SOURCES.
+#
+# Optional defines:
+#   LIBRARY_INCLUDES are appended to the list of includes directories.
+#   LIBRARY_DEFINES is not used for makedepend, but for compilation.
+
+
+### Basic defines ###
+
+OBJECTS = $(C_SOURCES:.c=.o)
+
+INCLUDES = \
+       -I. \
+       $(LIBRARY_INCLUDES)
+
+
+##### TARGETS #####
+
+default: depend lib$(LIBNAME).a
+
+lib$(LIBNAME).a: $(OBJECTS) Makefile $(TOP)/src/glsl/Makefile.template
+       $(MKLIB) -o $(LIBNAME) -static $(OBJECTS)
+
+depend: $(C_SOURCES)
+       rm -f depend
+       touch depend
+       $(MKDEP) $(MKDEP_OPTIONS) $(INCLUDES) $(C_SOURCES) 2> /dev/null
+
+# Remove .o and backup files
+clean:
+       rm -f $(OBJECTS) lib$(LIBNAME).a depend depend.bak
+
+# Dummy target
+install:
+       @echo -n ""
+
+
+##### RULES #####
+
+.c.o:
+       $(CC) -c $(INCLUDES) $(CFLAGS) $(LIBRARY_DEFINES) $< -o $@
+
+-include depend
+
diff --git a/src/glsl/SConscript b/src/glsl/SConscript
new file mode 100644 (file)
index 0000000..101487d
--- /dev/null
@@ -0,0 +1,69 @@
+import common
+
+Import('*')
+
+env = env.Clone()
+
+sources = [
+    'pp/sl_pp_context.c',
+    'pp/sl_pp_define.c',
+    'pp/sl_pp_dict.c',
+    'pp/sl_pp_error.c',
+    'pp/sl_pp_expression.c',
+    'pp/sl_pp_extension.c',
+    'pp/sl_pp_if.c',
+    'pp/sl_pp_line.c',
+    'pp/sl_pp_macro.c',
+    'pp/sl_pp_pragma.c',
+    'pp/sl_pp_process.c',
+    'pp/sl_pp_purify.c',
+    'pp/sl_pp_token.c',
+    'pp/sl_pp_token_util.c',
+    'pp/sl_pp_version.c',
+    'cl/sl_cl_parse.c',
+]
+
+glsl = env.StaticLibrary(
+    target = 'glsl',
+    source = sources,
+)
+
+Export('glsl')
+
+env = env.Clone()
+
+if env['platform'] == 'windows':
+    env.PrependUnique(LIBS = [
+        'user32',
+    ])
+
+env.Prepend(LIBS = [glsl])
+
+env.Program(
+    target = 'purify',
+    source = ['apps/purify.c'],
+)
+
+env.Program(
+    target = 'tokenise',
+    source = ['apps/tokenise.c'],
+)
+
+env.Program(
+    target = 'version',
+    source = ['apps/version.c'],
+)
+
+env.Program(
+    target = 'process',
+    source = ['apps/process.c'],
+)
+
+glsl_compile = env.Program(
+    target = 'compile',
+    source = ['apps/compile.c'],
+)
+
+if env['platform'] == common.default_platform:
+    # Only export the GLSL compiler when building for the host platform
+    Export('glsl_compile')
diff --git a/src/glsl/apps/Makefile b/src/glsl/apps/Makefile
new file mode 100644 (file)
index 0000000..39a0df7
--- /dev/null
@@ -0,0 +1,43 @@
+# src/glsl/apps/Makefile
+
+TOP = ../../..
+
+include $(TOP)/configs/current
+
+LIBS = \
+       $(TOP)/src/glsl/pp/libglslpp.a \
+       $(TOP)/src/glsl/cl/libglslcl.a
+
+SOURCES = \
+       compile.c \
+       process.c \
+       purify.c \
+       tokenise.c \
+       version.c
+
+APPS = $(SOURCES:%.c=%)
+
+INCLUDES = -I.
+
+
+##### RULES #####
+
+.SUFFIXES:
+.SUFFIXES: .c
+
+.c:
+       $(APP_CC) $(INCLUDES) $(CFLAGS) $(LDFLAGS) $< $(LIBS) -o $@
+
+.c.o:
+       $(APP_CC) -c $(INCLUDES) $(CFLAGS) $(DEFINES) $< -o $@
+
+
+##### TARGETS #####
+
+default: $(APPS)
+
+install:
+
+clean:
+       -rm -f $(APPS)
+       -rm -f *.o
diff --git a/src/glsl/apps/compile.c b/src/glsl/apps/compile.c
new file mode 100644 (file)
index 0000000..c9a830b
--- /dev/null
@@ -0,0 +1,191 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <assert.h>
+#include "../pp/sl_pp_public.h"
+#include "../cl/sl_cl_parse.h"
+
+
+static void
+usage(void)
+{
+   printf("Usage:\n");
+   printf("  compile fragment|vertex <source> <output>\n");
+}
+
+int
+main(int argc,
+     char *argv[])
+{
+   FILE *in;
+   long size;
+   char *inbuf;
+   struct sl_pp_purify_options options;
+   char errmsg[100] = "";
+   struct sl_pp_context *context;
+   unsigned int version;
+   FILE *out;
+   unsigned char *outbytes;
+   unsigned int cboutbytes;
+   unsigned int shader_type;
+
+   if (argc != 4) {
+      usage();
+      return 1;
+   }
+
+   if (!strcmp(argv[1], "fragment")) {
+      shader_type = 1;
+   } else if (!strcmp(argv[1], "vertex")) {
+      shader_type = 2;
+   } else {
+      usage();
+      return 1;
+   }
+
+   in = fopen(argv[2], "rb");
+   if (!in) {
+      printf("Could not open `%s' for read.\n", argv[2]);
+      usage();
+      return 1;
+   }
+
+   fseek(in, 0, SEEK_END);
+   size = ftell(in);
+   fseek(in, 0, SEEK_SET);
+
+   out = fopen(argv[3], "w");
+   if (!out) {
+      fclose(in);
+      printf("Could not open `%s' for write.\n", argv[3]);
+      usage();
+      return 1;
+   }
+
+   inbuf = malloc(size + 1);
+   if (!inbuf) {
+      fprintf(out, "$OOMERROR\n");
+
+      fclose(out);
+      fclose(in);
+      printf("Out of memory.\n");
+      return 0;
+   }
+
+   if (fread(inbuf, 1, size, in) != size) {
+      fprintf(out, "$READERROR\n");
+
+      free(inbuf);
+      fclose(out);
+      fclose(in);
+      printf("Could not read from `%s'.\n", argv[2]);
+      return 0;
+   }
+   inbuf[size] = '\0';
+
+   fclose(in);
+
+   memset(&options, 0, sizeof(options));
+
+   context = sl_pp_context_create(inbuf, &options);
+   if (!context) {
+      fprintf(out, "$CONTEXERROR\n");
+
+      free(inbuf);
+      fclose(out);
+      printf("Could not create parse context.\n");
+      return 0;
+   }
+
+   if (sl_pp_version(context, &version)) {
+      fprintf(out, "$ERROR: `%s'\n", sl_pp_context_error_message(context));
+
+      printf("Error: %s\n", sl_pp_context_error_message(context));
+      sl_pp_context_destroy(context);
+      free(inbuf);
+      fclose(out);
+      return 0;
+   }
+
+   if (sl_pp_context_add_extension(context, "ARB_draw_buffers", "GL_ARB_draw_buffers") ||
+       sl_pp_context_add_extension(context, "ARB_texture_rectangle", "GL_ARB_texture_rectangle")) {
+      fprintf(out, "$ERROR: `%s'\n", sl_pp_context_error_message(context));
+
+      printf("Error: %s\n", sl_pp_context_error_message(context));
+      sl_pp_context_destroy(context);
+      free(inbuf);
+      fclose(out);
+      return 0;
+   }
+
+   if (sl_cl_compile(context, shader_type, 1, &outbytes, &cboutbytes, errmsg, sizeof(errmsg)) == 0) {
+      unsigned int i;
+      unsigned int line = 0;
+
+      fprintf(out, "\n/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE FOLLOWING FILE: */");
+      fprintf(out, "\n/* %s */", argv[2]);
+      fprintf(out, "\n\n");
+
+      for (i = 0; i < cboutbytes; i++) {
+         unsigned int a;
+
+         if (outbytes[i] < 10) {
+            a = 1;
+         } else if (outbytes[i] < 100) {
+            a = 2;
+         } else {
+            a = 3;
+         }
+         if (i < cboutbytes - 1) {
+            a++;
+         }
+         if (line + a >= 100) {
+            fprintf (out, "\n");
+            line = 0;
+         }
+         line += a;
+         fprintf (out, "%u", outbytes[i]);
+         if (i < cboutbytes - 1) {
+            fprintf (out, ",");
+         }
+      }
+      fprintf (out, "\n");
+      free(outbytes);
+   } else {
+      fprintf(out, "$SYNTAXERROR: `%s'\n", errmsg);
+
+      printf("Error: %s\n", errmsg);
+   }
+
+   sl_pp_context_destroy(context);
+   free(inbuf);
+   fclose(out);
+   return 0;
+}
diff --git a/src/glsl/apps/process.c b/src/glsl/apps/process.c
new file mode 100644 (file)
index 0000000..5698902
--- /dev/null
@@ -0,0 +1,381 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <assert.h>
+#include "../pp/sl_pp_public.h"
+
+
+int
+main(int argc,
+     char *argv[])
+{
+   FILE *in;
+   long size;
+   char *inbuf;
+   struct sl_pp_purify_options options;
+   struct sl_pp_context *context;
+   unsigned int version;
+   struct sl_pp_token_info *outtokens;
+   FILE *out;
+   unsigned int i;
+
+   if (argc != 3) {
+      return 1;
+   }
+
+   in = fopen(argv[1], "rb");
+   if (!in) {
+      return 1;
+   }
+
+   fseek(in, 0, SEEK_END);
+   size = ftell(in);
+   fseek(in, 0, SEEK_SET);
+
+   out = fopen(argv[2], "wb");
+   if (!out) {
+      fclose(in);
+      return 1;
+   }
+
+   inbuf = malloc(size + 1);
+   if (!inbuf) {
+      fprintf(out, "$OOMERROR\n");
+
+      fclose(out);
+      fclose(in);
+      return 1;
+   }
+
+   if (fread(inbuf, 1, size, in) != size) {
+      fprintf(out, "$READERROR\n");
+
+      free(inbuf);
+      fclose(out);
+      fclose(in);
+      return 1;
+   }
+   inbuf[size] = '\0';
+
+   fclose(in);
+
+   memset(&options, 0, sizeof(options));
+
+   context = sl_pp_context_create(inbuf, &options);
+   if (!context) {
+      fprintf(out, "$CONTEXERROR\n");
+
+      free(inbuf);
+      fclose(out);
+      return 1;
+   }
+
+   if (sl_pp_version(context, &version)) {
+      fprintf(out, "$ERROR: `%s'\n", sl_pp_context_error_message(context));
+
+      sl_pp_context_destroy(context);
+      free(inbuf);
+      fclose(out);
+      return -1;
+   }
+
+   if (sl_pp_context_add_extension(context, "ARB_draw_buffers", "GL_ARB_draw_buffers") ||
+       sl_pp_context_add_extension(context, "ARB_texture_rectangle", "GL_ARB_texture_rectangle")) {
+      fprintf(out, "$ERROR: `%s'\n", sl_pp_context_error_message(context));
+
+      printf("Error: %s\n", sl_pp_context_error_message(context));
+      sl_pp_context_destroy(context);
+      free(inbuf);
+      fclose(out);
+      return 0;
+   }
+
+   if (sl_pp_context_add_predefined(context, "__GLSL_PP_PREDEFINED_MACRO_TEST", "1")) {
+      fprintf(out, "$ERROR: `%s'\n", sl_pp_context_error_message(context));
+
+      printf("Error: %s\n", sl_pp_context_error_message(context));
+      sl_pp_context_destroy(context);
+      free(inbuf);
+      fclose(out);
+      return 0;
+   }
+
+   if (sl_pp_process(context, &outtokens)) {
+      fprintf(out, "$ERROR: `%s'\n", sl_pp_context_error_message(context));
+
+      sl_pp_context_destroy(context);
+      free(inbuf);
+      fclose(out);
+      return -1;
+   }
+
+   free(inbuf);
+
+   for (i = 0; outtokens[i].token != SL_PP_EOF; i++) {
+      switch (outtokens[i].token) {
+      case SL_PP_NEWLINE:
+         fprintf(out, "\n");
+         break;
+
+      case SL_PP_COMMA:
+         fprintf(out, ", ");
+         break;
+
+      case SL_PP_SEMICOLON:
+         fprintf(out, "; ");
+         break;
+
+      case SL_PP_LBRACE:
+         fprintf(out, "{ ");
+         break;
+
+      case SL_PP_RBRACE:
+         fprintf(out, "} ");
+         break;
+
+      case SL_PP_LPAREN:
+         fprintf(out, "( ");
+         break;
+
+      case SL_PP_RPAREN:
+         fprintf(out, ") ");
+         break;
+
+      case SL_PP_LBRACKET:
+         fprintf(out, "[ ");
+         break;
+
+      case SL_PP_RBRACKET:
+         fprintf(out, "] ");
+         break;
+
+      case SL_PP_DOT:
+         fprintf(out, ". ");
+         break;
+
+      case SL_PP_INCREMENT:
+         fprintf(out, "++ ");
+         break;
+
+      case SL_PP_ADDASSIGN:
+         fprintf(out, "+= ");
+         break;
+
+      case SL_PP_PLUS:
+         fprintf(out, "+ ");
+         break;
+
+      case SL_PP_DECREMENT:
+         fprintf(out, "-- ");
+         break;
+
+      case SL_PP_SUBASSIGN:
+         fprintf(out, "-= ");
+         break;
+
+      case SL_PP_MINUS:
+         fprintf(out, "- ");
+         break;
+
+      case SL_PP_BITNOT:
+         fprintf(out, "~ ");
+         break;
+
+      case SL_PP_NOTEQUAL:
+         fprintf(out, "!= ");
+         break;
+
+      case SL_PP_NOT:
+         fprintf(out, "! ");
+         break;
+
+      case SL_PP_MULASSIGN:
+         fprintf(out, "*= ");
+         break;
+
+      case SL_PP_STAR:
+         fprintf(out, "* ");
+         break;
+
+      case SL_PP_DIVASSIGN:
+         fprintf(out, "/= ");
+         break;
+
+      case SL_PP_SLASH:
+         fprintf(out, "/ ");
+         break;
+
+      case SL_PP_MODASSIGN:
+         fprintf(out, "%%= ");
+         break;
+
+      case SL_PP_MODULO:
+         fprintf(out, "%% ");
+         break;
+
+      case SL_PP_LSHIFTASSIGN:
+         fprintf(out, "<<= ");
+         break;
+
+      case SL_PP_LSHIFT:
+         fprintf(out, "<< ");
+         break;
+
+      case SL_PP_LESSEQUAL:
+         fprintf(out, "<= ");
+         break;
+
+      case SL_PP_LESS:
+         fprintf(out, "< ");
+         break;
+
+      case SL_PP_RSHIFTASSIGN:
+         fprintf(out, ">>= ");
+         break;
+
+      case SL_PP_RSHIFT:
+         fprintf(out, ">> ");
+         break;
+
+      case SL_PP_GREATEREQUAL:
+         fprintf(out, ">= ");
+         break;
+
+      case SL_PP_GREATER:
+         fprintf(out, "> ");
+         break;
+
+      case SL_PP_EQUAL:
+         fprintf(out, "== ");
+         break;
+
+      case SL_PP_ASSIGN:
+         fprintf(out, "= ");
+         break;
+
+      case SL_PP_AND:
+         fprintf(out, "&& ");
+         break;
+
+      case SL_PP_BITANDASSIGN:
+         fprintf(out, "&= ");
+         break;
+
+      case SL_PP_BITAND:
+         fprintf(out, "& ");
+         break;
+
+      case SL_PP_XOR:
+         fprintf(out, "^^ ");
+         break;
+
+      case SL_PP_BITXORASSIGN:
+         fprintf(out, "^= ");
+         break;
+
+      case SL_PP_BITXOR:
+         fprintf(out, "^ ");
+         break;
+
+      case SL_PP_OR:
+         fprintf(out, "|| ");
+         break;
+
+      case SL_PP_BITORASSIGN:
+         fprintf(out, "|= ");
+         break;
+
+      case SL_PP_BITOR:
+         fprintf(out, "| ");
+         break;
+
+      case SL_PP_QUESTION:
+         fprintf(out, "? ");
+         break;
+
+      case SL_PP_COLON:
+         fprintf(out, ": ");
+         break;
+
+      case SL_PP_IDENTIFIER:
+         fprintf(out, "%s ", sl_pp_context_cstr(context, outtokens[i].data.identifier));
+         break;
+
+      case SL_PP_UINT:
+         fprintf(out, "%s ", sl_pp_context_cstr(context, outtokens[i].data._uint));
+         break;
+
+      case SL_PP_FLOAT:
+         fprintf(out, "%s ", sl_pp_context_cstr(context, outtokens[i].data._float));
+         break;
+
+      case SL_PP_OTHER:
+         fprintf(out, "%c", outtokens[i].data.other);
+         break;
+
+      case SL_PP_PRAGMA_OPTIMIZE:
+         fprintf(out, "#pragma optimize(%s)", outtokens[i].data.pragma ? "on" : "off");
+         break;
+
+      case SL_PP_PRAGMA_DEBUG:
+         fprintf(out, "#pragma debug(%s)", outtokens[i].data.pragma ? "on" : "off");
+         break;
+
+      case SL_PP_EXTENSION_REQUIRE:
+         fprintf(out, "#extension %s : require", sl_pp_context_cstr(context, outtokens[i].data.extension));
+         break;
+
+      case SL_PP_EXTENSION_ENABLE:
+         fprintf(out, "#extension %s : enable", sl_pp_context_cstr(context, outtokens[i].data.extension));
+         break;
+
+      case SL_PP_EXTENSION_WARN:
+         fprintf(out, "#extension %s : warn", sl_pp_context_cstr(context, outtokens[i].data.extension));
+         break;
+
+      case SL_PP_EXTENSION_DISABLE:
+         fprintf(out, "#extension %s : disable", sl_pp_context_cstr(context, outtokens[i].data.extension));
+         break;
+
+      case SL_PP_LINE:
+         fprintf(out, "#line %u %u", outtokens[i].data.line.lineno, outtokens[i].data.line.fileno);
+         break;
+
+      default:
+         assert(0);
+      }
+   }
+
+   sl_pp_context_destroy(context);
+   free(outtokens);
+   fclose(out);
+
+   return 0;
+}
diff --git a/src/glsl/apps/purify.c b/src/glsl/apps/purify.c
new file mode 100644 (file)
index 0000000..8c01f4f
--- /dev/null
@@ -0,0 +1,105 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include "../pp/sl_pp_public.h"
+
+
+int
+main(int argc,
+     char *argv[])
+{
+   FILE *in;
+   long size;
+   char *inbuf;
+   struct sl_pp_purify_options options;
+   char *outbuf;
+   char errmsg[100] = "";
+   unsigned int errline = 0;
+   FILE *out;
+
+   if (argc != 3) {
+      return 1;
+   }
+
+   in = fopen(argv[1], "rb");
+   if (!in) {
+      return 1;
+   }
+
+   fseek(in, 0, SEEK_END);
+   size = ftell(in);
+   fseek(in, 0, SEEK_SET);
+
+   out = fopen(argv[2], "wb");
+   if (!out) {
+      fclose(in);
+      return 1;
+   }
+
+   inbuf = malloc(size + 1);
+   if (!inbuf) {
+      fprintf(out, "$OOMERROR\n");
+
+      fclose(out);
+      fclose(in);
+      return 1;
+   }
+
+   if (fread(inbuf, 1, size, in) != size) {
+      fprintf(out, "$READERROR\n");
+
+      free(inbuf);
+      fclose(out);
+      fclose(in);
+      return 1;
+   }
+   inbuf[size] = '\0';
+
+   fclose(in);
+
+   memset(&options, 0, sizeof(options));
+
+   if (sl_pp_purify(inbuf, &options, &outbuf, errmsg, sizeof(errmsg), &errline)) {
+      fprintf(out, "$PURIFYERROR %u: %s\n", errline, errmsg);
+
+      free(inbuf);
+      fclose(out);
+      return 1;
+   }
+
+   free(inbuf);
+
+   fwrite(outbuf, 1, strlen(outbuf), out);
+
+   free(outbuf);
+   fclose(out);
+
+   return 0;
+}
diff --git a/src/glsl/apps/tokenise.c b/src/glsl/apps/tokenise.c
new file mode 100644 (file)
index 0000000..9ff7315
--- /dev/null
@@ -0,0 +1,333 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <assert.h>
+#include "../pp/sl_pp_public.h"
+
+
+int
+main(int argc,
+     char *argv[])
+{
+   FILE *in;
+   long size;
+   char *inbuf;
+   struct sl_pp_purify_options options;
+   struct sl_pp_context *context;
+   struct sl_pp_token_info *tokens;
+   FILE *out;
+   unsigned int i;
+
+   if (argc != 3) {
+      return 1;
+   }
+
+   in = fopen(argv[1], "rb");
+   if (!in) {
+      return 1;
+   }
+
+   fseek(in, 0, SEEK_END);
+   size = ftell(in);
+   fseek(in, 0, SEEK_SET);
+
+   out = fopen(argv[2], "wb");
+   if (!out) {
+      fclose(in);
+      return 1;
+   }
+
+   inbuf = malloc(size + 1);
+   if (!inbuf) {
+      fprintf(out, "$OOMERROR\n");
+
+      fclose(out);
+      fclose(in);
+      return 1;
+   }
+
+   if (fread(inbuf, 1, size, in) != size) {
+      fprintf(out, "$READERROR\n");
+
+      free(inbuf);
+      fclose(out);
+      fclose(in);
+      return 1;
+   }
+   inbuf[size] = '\0';
+
+   fclose(in);
+
+   memset(&options, 0, sizeof(options));
+
+   context = sl_pp_context_create(inbuf, &options);
+   if (!context) {
+      fprintf(out, "$CONTEXERROR\n");
+
+      free(inbuf);
+      fclose(out);
+      return 1;
+   }
+
+   if (sl_pp_tokenise(context, &tokens)) {
+      fprintf(out, "$ERROR: `%s'\n", sl_pp_context_error_message(context));
+
+      sl_pp_context_destroy(context);
+      free(inbuf);
+      fclose(out);
+      return 1;
+   }
+
+   free(inbuf);
+
+   for (i = 0; tokens[i].token != SL_PP_EOF; i++) {
+      switch (tokens[i].token) {
+      case SL_PP_WHITESPACE:
+         break;
+
+      case SL_PP_NEWLINE:
+         fprintf(out, "\n");
+         break;
+
+      case SL_PP_HASH:
+         fprintf(out, "# ");
+         break;
+
+      case SL_PP_COMMA:
+         fprintf(out, ", ");
+         break;
+
+      case SL_PP_SEMICOLON:
+         fprintf(out, "; ");
+         break;
+
+      case SL_PP_LBRACE:
+         fprintf(out, "{ ");
+         break;
+
+      case SL_PP_RBRACE:
+         fprintf(out, "} ");
+         break;
+
+      case SL_PP_LPAREN:
+         fprintf(out, "( ");
+         break;
+
+      case SL_PP_RPAREN:
+         fprintf(out, ") ");
+         break;
+
+      case SL_PP_LBRACKET:
+         fprintf(out, "[ ");
+         break;
+
+      case SL_PP_RBRACKET:
+         fprintf(out, "] ");
+         break;
+
+      case SL_PP_DOT:
+         fprintf(out, ". ");
+         break;
+
+      case SL_PP_INCREMENT:
+         fprintf(out, "++ ");
+         break;
+
+      case SL_PP_ADDASSIGN:
+         fprintf(out, "+= ");
+         break;
+
+      case SL_PP_PLUS:
+         fprintf(out, "+ ");
+         break;
+
+      case SL_PP_DECREMENT:
+         fprintf(out, "-- ");
+         break;
+
+      case SL_PP_SUBASSIGN:
+         fprintf(out, "-= ");
+         break;
+
+      case SL_PP_MINUS:
+         fprintf(out, "- ");
+         break;
+
+      case SL_PP_BITNOT:
+         fprintf(out, "~ ");
+         break;
+
+      case SL_PP_NOTEQUAL:
+         fprintf(out, "!= ");
+         break;
+
+      case SL_PP_NOT:
+         fprintf(out, "! ");
+         break;
+
+      case SL_PP_MULASSIGN:
+         fprintf(out, "*= ");
+         break;
+
+      case SL_PP_STAR:
+         fprintf(out, "* ");
+         break;
+
+      case SL_PP_DIVASSIGN:
+         fprintf(out, "/= ");
+         break;
+
+      case SL_PP_SLASH:
+         fprintf(out, "/ ");
+         break;
+
+      case SL_PP_MODASSIGN:
+         fprintf(out, "%%= ");
+         break;
+
+      case SL_PP_MODULO:
+         fprintf(out, "%% ");
+         break;
+
+      case SL_PP_LSHIFTASSIGN:
+         fprintf(out, "<<= ");
+         break;
+
+      case SL_PP_LSHIFT:
+         fprintf(out, "<< ");
+         break;
+
+      case SL_PP_LESSEQUAL:
+         fprintf(out, "<= ");
+         break;
+
+      case SL_PP_LESS:
+         fprintf(out, "< ");
+         break;
+
+      case SL_PP_RSHIFTASSIGN:
+         fprintf(out, ">>= ");
+         break;
+
+      case SL_PP_RSHIFT:
+         fprintf(out, ">> ");
+         break;
+
+      case SL_PP_GREATEREQUAL:
+         fprintf(out, ">= ");
+         break;
+
+      case SL_PP_GREATER:
+         fprintf(out, "> ");
+         break;
+
+      case SL_PP_EQUAL:
+         fprintf(out, "== ");
+         break;
+
+      case SL_PP_ASSIGN:
+         fprintf(out, "= ");
+         break;
+
+      case SL_PP_AND:
+         fprintf(out, "&& ");
+         break;
+
+      case SL_PP_BITANDASSIGN:
+         fprintf(out, "&= ");
+         break;
+
+      case SL_PP_BITAND:
+         fprintf(out, "& ");
+         break;
+
+      case SL_PP_XOR:
+         fprintf(out, "^^ ");
+         break;
+
+      case SL_PP_BITXORASSIGN:
+         fprintf(out, "^= ");
+         break;
+
+      case SL_PP_BITXOR:
+         fprintf(out, "^ ");
+         break;
+
+      case SL_PP_OR:
+         fprintf(out, "|| ");
+         break;
+
+      case SL_PP_BITORASSIGN:
+         fprintf(out, "|= ");
+         break;
+
+      case SL_PP_BITOR:
+         fprintf(out, "| ");
+         break;
+
+      case SL_PP_QUESTION:
+         fprintf(out, "? ");
+         break;
+
+      case SL_PP_COLON:
+         fprintf(out, ": ");
+         break;
+
+      case SL_PP_IDENTIFIER:
+         fprintf(out, "%s ", sl_pp_context_cstr(context, tokens[i].data.identifier));
+         break;
+
+      case SL_PP_UINT:
+         fprintf(out, "(%s) ", sl_pp_context_cstr(context, tokens[i].data._uint));
+         break;
+
+      case SL_PP_FLOAT:
+         fprintf(out, "(%s) ", sl_pp_context_cstr(context, tokens[i].data._float));
+         break;
+
+      case SL_PP_OTHER:
+         if (tokens[i].data.other == '\'') {
+            fprintf(out, "'\\'' ");
+         } else {
+            fprintf(out, "'%c' ", tokens[i].data.other);
+         }
+         break;
+
+      default:
+         assert(0);
+      }
+   }
+
+   sl_pp_context_destroy(context);
+   free(tokens);
+   fclose(out);
+
+   return 0;
+}
diff --git a/src/glsl/apps/version.c b/src/glsl/apps/version.c
new file mode 100644 (file)
index 0000000..40a4a06
--- /dev/null
@@ -0,0 +1,115 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+#include <assert.h>
+#include "../pp/sl_pp_public.h"
+
+
+int
+main(int argc,
+     char *argv[])
+{
+   FILE *in;
+   long size;
+   char *inbuf;
+   struct sl_pp_purify_options options;
+   struct sl_pp_context *context;
+   unsigned int version;
+   FILE *out;
+
+   if (argc != 3) {
+      return 1;
+   }
+
+   in = fopen(argv[1], "rb");
+   if (!in) {
+      return 1;
+   }
+
+   fseek(in, 0, SEEK_END);
+   size = ftell(in);
+   fseek(in, 0, SEEK_SET);
+
+   out = fopen(argv[2], "wb");
+   if (!out) {
+      fclose(in);
+      return 1;
+   }
+
+   inbuf = malloc(size + 1);
+   if (!inbuf) {
+      fprintf(out, "$OOMERROR\n");
+
+      fclose(out);
+      fclose(in);
+      return 1;
+   }
+
+   if (fread(inbuf, 1, size, in) != size) {
+      fprintf(out, "$READERROR\n");
+
+      free(inbuf);
+      fclose(out);
+      fclose(in);
+      return 1;
+   }
+   inbuf[size] = '\0';
+
+   fclose(in);
+
+   memset(&options, 0, sizeof(options));
+
+   context = sl_pp_context_create(inbuf, &options);
+   if (!context) {
+      fprintf(out, "$CONTEXERROR\n");
+
+      free(inbuf);
+      fclose(out);
+      return 1;
+   }
+
+   if (sl_pp_version(context, &version)) {
+      fprintf(out, "$ERROR: `%s'\n", sl_pp_context_error_message(context));
+
+      sl_pp_context_destroy(context);
+      free(inbuf);
+      fclose(out);
+      return -1;
+   }
+
+   sl_pp_context_destroy(context);
+   free(inbuf);
+
+   fprintf(out, "%u\n", version);
+
+   fclose(out);
+
+   return 0;
+}
diff --git a/src/glsl/cl/Makefile b/src/glsl/cl/Makefile
new file mode 100644 (file)
index 0000000..04a52df
--- /dev/null
@@ -0,0 +1,13 @@
+#src/glsl/cl/Makefile
+
+TOP = ../../..
+
+include $(TOP)/configs/current
+
+LIBNAME = glslcl
+
+C_SOURCES = \
+       sl_cl_parse.c
+
+include ../Makefile.template
+
diff --git a/src/glsl/cl/sl_cl_parse.c b/src/glsl/cl/sl_cl_parse.c
new file mode 100644 (file)
index 0000000..5dddf43
--- /dev/null
@@ -0,0 +1,2823 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <assert.h>
+#include <stdlib.h>
+#include <string.h>
+#include "../pp/sl_pp_public.h"
+#include "sl_cl_parse.h"
+
+
+/* revision number - increment after each change affecting emitted output */
+#define REVISION                                   5
+
+/* external declaration (or precision or invariant stmt) */
+#define EXTERNAL_NULL                              0
+#define EXTERNAL_FUNCTION_DEFINITION               1
+#define EXTERNAL_DECLARATION                       2
+#define DEFAULT_PRECISION                          3
+#define INVARIANT_STMT                             4
+
+/* precision */
+#define PRECISION_DEFAULT                          0
+#define PRECISION_LOW                              1
+#define PRECISION_MEDIUM                           2
+#define PRECISION_HIGH                             3
+
+/* declaration */
+#define DECLARATION_FUNCTION_PROTOTYPE             1
+#define DECLARATION_INIT_DECLARATOR_LIST           2
+
+/* function type */
+#define FUNCTION_ORDINARY                          0
+#define FUNCTION_CONSTRUCTOR                       1
+#define FUNCTION_OPERATOR                          2
+
+/* function call type */
+#define FUNCTION_CALL_NONARRAY                     0
+#define FUNCTION_CALL_ARRAY                        1
+
+/* operator type */
+#define OPERATOR_ADDASSIGN                         1
+#define OPERATOR_SUBASSIGN                         2
+#define OPERATOR_MULASSIGN                         3
+#define OPERATOR_DIVASSIGN                         4
+/*#define OPERATOR_MODASSIGN                         5*/
+/*#define OPERATOR_LSHASSIGN                         6*/
+/*#define OPERATOR_RSHASSIGN                         7*/
+/*#define OPERATOR_ORASSIGN                          8*/
+/*#define OPERATOR_XORASSIGN                         9*/
+/*#define OPERATOR_ANDASSIGN                         10*/
+#define OPERATOR_LOGICALXOR                        11
+/*#define OPERATOR_BITOR                             12*/
+/*#define OPERATOR_BITXOR                            13*/
+/*#define OPERATOR_BITAND                            14*/
+#define OPERATOR_LESS                              15
+#define OPERATOR_GREATER                           16
+#define OPERATOR_LESSEQUAL                         17
+#define OPERATOR_GREATEREQUAL                      18
+/*#define OPERATOR_LSHIFT                            19*/
+/*#define OPERATOR_RSHIFT                            20*/
+#define OPERATOR_MULTIPLY                          21
+#define OPERATOR_DIVIDE                            22
+/*#define OPERATOR_MODULUS                           23*/
+#define OPERATOR_INCREMENT                         24
+#define OPERATOR_DECREMENT                         25
+#define OPERATOR_PLUS                              26
+#define OPERATOR_MINUS                             27
+/*#define OPERATOR_COMPLEMENT                        28*/
+#define OPERATOR_NOT                               29
+
+/* init declarator list */
+#define DECLARATOR_NONE                            0
+#define DECLARATOR_NEXT                            1
+
+/* variable declaration */
+#define VARIABLE_NONE                              0
+#define VARIABLE_IDENTIFIER                        1
+#define VARIABLE_INITIALIZER                       2
+#define VARIABLE_ARRAY_EXPLICIT                    3
+#define VARIABLE_ARRAY_UNKNOWN                     4
+
+/* type qualifier */
+#define TYPE_QUALIFIER_NONE                        0
+#define TYPE_QUALIFIER_CONST                       1
+#define TYPE_QUALIFIER_ATTRIBUTE                   2
+#define TYPE_QUALIFIER_VARYING                     3
+#define TYPE_QUALIFIER_UNIFORM                     4
+#define TYPE_QUALIFIER_FIXEDOUTPUT                 5
+#define TYPE_QUALIFIER_FIXEDINPUT                  6
+
+/* invariant qualifier */
+#define TYPE_VARIANT                               90
+#define TYPE_INVARIANT                             91
+
+/* centroid qualifier */
+#define TYPE_CENTER                                95
+#define TYPE_CENTROID                              96
+
+/* type specifier */
+#define TYPE_SPECIFIER_VOID                        0
+#define TYPE_SPECIFIER_BOOL                        1
+#define TYPE_SPECIFIER_BVEC2                       2
+#define TYPE_SPECIFIER_BVEC3                       3
+#define TYPE_SPECIFIER_BVEC4                       4
+#define TYPE_SPECIFIER_INT                         5
+#define TYPE_SPECIFIER_IVEC2                       6
+#define TYPE_SPECIFIER_IVEC3                       7
+#define TYPE_SPECIFIER_IVEC4                       8
+#define TYPE_SPECIFIER_FLOAT                       9
+#define TYPE_SPECIFIER_VEC2                        10
+#define TYPE_SPECIFIER_VEC3                        11
+#define TYPE_SPECIFIER_VEC4                        12
+#define TYPE_SPECIFIER_MAT2                        13
+#define TYPE_SPECIFIER_MAT3                        14
+#define TYPE_SPECIFIER_MAT4                        15
+#define TYPE_SPECIFIER_SAMPLER1D                   16
+#define TYPE_SPECIFIER_SAMPLER2D                   17
+#define TYPE_SPECIFIER_SAMPLER3D                   18
+#define TYPE_SPECIFIER_SAMPLERCUBE                 19
+#define TYPE_SPECIFIER_SAMPLER1DSHADOW             20
+#define TYPE_SPECIFIER_SAMPLER2DSHADOW             21
+#define TYPE_SPECIFIER_SAMPLER2DRECT               22
+#define TYPE_SPECIFIER_SAMPLER2DRECTSHADOW         23
+#define TYPE_SPECIFIER_STRUCT                      24
+#define TYPE_SPECIFIER_TYPENAME                    25
+
+/* OpenGL 2.1 */
+#define TYPE_SPECIFIER_MAT23                       26
+#define TYPE_SPECIFIER_MAT32                       27
+#define TYPE_SPECIFIER_MAT24                       28
+#define TYPE_SPECIFIER_MAT42                       29
+#define TYPE_SPECIFIER_MAT34                       30
+#define TYPE_SPECIFIER_MAT43                       31
+
+/* type specifier array */
+#define TYPE_SPECIFIER_NONARRAY                    0
+#define TYPE_SPECIFIER_ARRAY                       1
+
+/* structure field */
+#define FIELD_NONE                                 0
+#define FIELD_NEXT                                 1
+#define FIELD_ARRAY                                2
+
+/* operation */
+#define OP_END                                     0
+#define OP_BLOCK_BEGIN_NO_NEW_SCOPE                1
+#define OP_BLOCK_BEGIN_NEW_SCOPE                   2
+#define OP_DECLARE                                 3
+#define OP_ASM                                     4
+#define OP_BREAK                                   5
+#define OP_CONTINUE                                6
+#define OP_DISCARD                                 7
+#define OP_RETURN                                  8
+#define OP_EXPRESSION                              9
+#define OP_IF                                      10
+#define OP_WHILE                                   11
+#define OP_DO                                      12
+#define OP_FOR                                     13
+#define OP_PUSH_VOID                               14
+#define OP_PUSH_BOOL                               15
+#define OP_PUSH_INT                                16
+#define OP_PUSH_FLOAT                              17
+#define OP_PUSH_IDENTIFIER                         18
+#define OP_SEQUENCE                                19
+#define OP_ASSIGN                                  20
+#define OP_ADDASSIGN                               21
+#define OP_SUBASSIGN                               22
+#define OP_MULASSIGN                               23
+#define OP_DIVASSIGN                               24
+/*#define OP_MODASSIGN                               25*/
+/*#define OP_LSHASSIGN                               26*/
+/*#define OP_RSHASSIGN                               27*/
+/*#define OP_ORASSIGN                                28*/
+/*#define OP_XORASSIGN                               29*/
+/*#define OP_ANDASSIGN                               30*/
+#define OP_SELECT                                  31
+#define OP_LOGICALOR                               32
+#define OP_LOGICALXOR                              33
+#define OP_LOGICALAND                              34
+/*#define OP_BITOR                                   35*/
+/*#define OP_BITXOR                                  36*/
+/*#define OP_BITAND                                  37*/
+#define OP_EQUAL                                   38
+#define OP_NOTEQUAL                                39
+#define OP_LESS                                    40
+#define OP_GREATER                                 41
+#define OP_LESSEQUAL                               42
+#define OP_GREATEREQUAL                            43
+/*#define OP_LSHIFT                                  44*/
+/*#define OP_RSHIFT                                  45*/
+#define OP_ADD                                     46
+#define OP_SUBTRACT                                47
+#define OP_MULTIPLY                                48
+#define OP_DIVIDE                                  49
+/*#define OP_MODULUS                                 50*/
+#define OP_PREINCREMENT                            51
+#define OP_PREDECREMENT                            52
+#define OP_PLUS                                    53
+#define OP_MINUS                                   54
+/*#define OP_COMPLEMENT                              55*/
+#define OP_NOT                                     56
+#define OP_SUBSCRIPT                               57
+#define OP_CALL                                    58
+#define OP_FIELD                                   59
+#define OP_POSTINCREMENT                           60
+#define OP_POSTDECREMENT                           61
+#define OP_PRECISION                               62
+#define OP_METHOD                                  63
+
+/* parameter qualifier */
+#define PARAM_QUALIFIER_IN                         0
+#define PARAM_QUALIFIER_OUT                        1
+#define PARAM_QUALIFIER_INOUT                      2
+
+/* function parameter */
+#define PARAMETER_NONE                             0
+#define PARAMETER_NEXT                             1
+
+/* function parameter array presence */
+#define PARAMETER_ARRAY_NOT_PRESENT                0
+#define PARAMETER_ARRAY_PRESENT                    1
+
+
+struct parse_dict {
+   int _void;
+   int _float;
+   int _int;
+   int _bool;
+   int vec2;
+   int vec3;
+   int vec4;
+   int bvec2;
+   int bvec3;
+   int bvec4;
+   int ivec2;
+   int ivec3;
+   int ivec4;
+   int mat2;
+   int mat3;
+   int mat4;
+   int mat2x3;
+   int mat3x2;
+   int mat2x4;
+   int mat4x2;
+   int mat3x4;
+   int mat4x3;
+   int sampler1D;
+   int sampler2D;
+   int sampler3D;
+   int samplerCube;
+   int sampler1DShadow;
+   int sampler2DShadow;
+   int sampler2DRect;
+   int sampler2DRectShadow;
+
+   int invariant;
+
+   int centroid;
+
+   int precision;
+   int lowp;
+   int mediump;
+   int highp;
+
+   int _const;
+   int attribute;
+   int varying;
+   int uniform;
+   int __fixed_output;
+   int __fixed_input;
+
+   int in;
+   int out;
+   int inout;
+
+   int _struct;
+
+   int __constructor;
+   int __operator;
+   int ___asm;
+
+   int _if;
+   int _else;
+   int _for;
+   int _while;
+   int _do;
+
+   int _continue;
+   int _break;
+   int _return;
+   int discard;
+
+   int _false;
+   int _true;
+};
+
+
+struct parse_context {
+   struct sl_pp_context *context;
+
+   struct parse_dict dict;
+
+   struct sl_pp_token_info *tokens;
+   unsigned int tokens_read;
+   unsigned int tokens_cap;
+
+   unsigned char *out_buf;
+   unsigned int out_cap;
+
+   unsigned int shader_type;
+   unsigned int parsing_builtin;
+
+   char error[256];
+   int process_error;
+};
+
+
+struct parse_state {
+   unsigned int in;
+   unsigned int out;
+};
+
+
+static __inline unsigned int
+_emit(struct parse_context *ctx,
+      unsigned int *out,
+      unsigned char b)
+{
+   if (*out == ctx->out_cap) {
+      ctx->out_cap += 4096;
+      ctx->out_buf = (unsigned char *)realloc(ctx->out_buf, ctx->out_cap * sizeof(unsigned char));
+   }
+   ctx->out_buf[*out] = b;
+   return (*out)++;
+}
+
+
+static void
+_update(struct parse_context *ctx,
+        unsigned int out,
+        unsigned char b)
+{
+   ctx->out_buf[out] = b;
+}
+
+
+static void
+_error(struct parse_context *ctx,
+       const char *msg)
+{
+   if (ctx->error[0] == '\0') {
+      strcpy(ctx->error, msg);
+   }
+}
+
+
+static const struct sl_pp_token_info *
+_fetch_token(struct parse_context *ctx,
+             unsigned int pos)
+{
+   if (ctx->process_error) {
+      return NULL;
+   }
+
+   while (pos >= ctx->tokens_read) {
+      if (ctx->tokens_read == ctx->tokens_cap) {
+         ctx->tokens_cap += 1024;
+         ctx->tokens = realloc(ctx->tokens,
+                               ctx->tokens_cap * sizeof(struct sl_pp_token_info));
+         if (!ctx->tokens) {
+            _error(ctx, "out of memory");
+            ctx->process_error = 1;
+            return NULL;
+         }
+      }
+      if (sl_pp_process_get(ctx->context, &ctx->tokens[ctx->tokens_read])) {
+         _error(ctx, sl_pp_context_error_message(ctx->context));
+         ctx->process_error = 1;
+         return NULL;
+      }
+      switch (ctx->tokens[ctx->tokens_read].token) {
+      case SL_PP_COMMA:
+      case SL_PP_SEMICOLON:
+      case SL_PP_LBRACE:
+      case SL_PP_RBRACE:
+      case SL_PP_LPAREN:
+      case SL_PP_RPAREN:
+      case SL_PP_LBRACKET:
+      case SL_PP_RBRACKET:
+      case SL_PP_DOT:
+      case SL_PP_INCREMENT:
+      case SL_PP_ADDASSIGN:
+      case SL_PP_PLUS:
+      case SL_PP_DECREMENT:
+      case SL_PP_SUBASSIGN:
+      case SL_PP_MINUS:
+      case SL_PP_BITNOT:
+      case SL_PP_NOTEQUAL:
+      case SL_PP_NOT:
+      case SL_PP_MULASSIGN:
+      case SL_PP_STAR:
+      case SL_PP_DIVASSIGN:
+      case SL_PP_SLASH:
+      case SL_PP_MODASSIGN:
+      case SL_PP_MODULO:
+      case SL_PP_LSHIFTASSIGN:
+      case SL_PP_LSHIFT:
+      case SL_PP_LESSEQUAL:
+      case SL_PP_LESS:
+      case SL_PP_RSHIFTASSIGN:
+      case SL_PP_RSHIFT:
+      case SL_PP_GREATEREQUAL:
+      case SL_PP_GREATER:
+      case SL_PP_EQUAL:
+      case SL_PP_ASSIGN:
+      case SL_PP_AND:
+      case SL_PP_BITANDASSIGN:
+      case SL_PP_BITAND:
+      case SL_PP_XOR:
+      case SL_PP_BITXORASSIGN:
+      case SL_PP_BITXOR:
+      case SL_PP_OR:
+      case SL_PP_BITORASSIGN:
+      case SL_PP_BITOR:
+      case SL_PP_QUESTION:
+      case SL_PP_COLON:
+      case SL_PP_IDENTIFIER:
+      case SL_PP_UINT:
+      case SL_PP_FLOAT:
+      case SL_PP_EOF:
+         ctx->tokens_read++;
+         break;
+      }
+   }
+   return &ctx->tokens[pos];
+}
+
+
+static int
+_parse_token(struct parse_context *ctx,
+             enum sl_pp_token token,
+             struct parse_state *ps)
+{
+   const struct sl_pp_token_info *input = _fetch_token(ctx, ps->in);
+
+   if (input && input->token == token) {
+      ps->in++;
+      return 0;
+   }
+   return -1;
+}
+
+
+static int
+_parse_id(struct parse_context *ctx,
+          int id,
+          struct parse_state *ps)
+{
+   const struct sl_pp_token_info *input = _fetch_token(ctx, ps->in);
+
+   if (input && input->token == SL_PP_IDENTIFIER && input->data.identifier == id) {
+      ps->in++;
+      return 0;
+   }
+   return -1;
+}
+
+
+static int
+_parse_identifier(struct parse_context *ctx,
+                  struct parse_state *ps)
+{
+   const struct sl_pp_token_info *input = _fetch_token(ctx, ps->in);
+
+   if (input && input->token == SL_PP_IDENTIFIER) {
+      const char *cstr = sl_pp_context_cstr(ctx->context, input->data.identifier);
+
+      do {
+         _emit(ctx, &ps->out, *cstr);
+      } while (*cstr++);
+      ps->in++;
+      return 0;
+   }
+   return -1;
+}
+
+
+static int
+_parse_float(struct parse_context *ctx,
+             struct parse_state *ps)
+{
+   const struct sl_pp_token_info *input = _fetch_token(ctx, ps->in);
+
+   if (input && input->token == SL_PP_FLOAT) {
+      const char *cstr = sl_pp_context_cstr(ctx->context, input->data._float);
+
+      _emit(ctx, &ps->out, 1);
+      do {
+         _emit(ctx, &ps->out, *cstr);
+      } while (*cstr++);
+      ps->in++;
+      return 0;
+   }
+   return -1;
+}
+
+
+static int
+_parse_uint(struct parse_context *ctx,
+            struct parse_state *ps)
+{
+   const struct sl_pp_token_info *input = _fetch_token(ctx, ps->in);
+
+   if (input && input->token == SL_PP_UINT) {
+      const char *cstr = sl_pp_context_cstr(ctx->context, input->data._uint);
+
+      _emit(ctx, &ps->out, 1);
+      do {
+         _emit(ctx, &ps->out, *cstr);
+      } while (*cstr++);
+      ps->in++;
+      return 0;
+   }
+   return -1;
+}
+
+
+/**************************************/
+
+
+static int
+_parse_unary_expression(struct parse_context *ctx,
+                        struct parse_state *ps);
+
+static int
+_parse_conditional_expression(struct parse_context *ctx,
+                              struct parse_state *ps);
+
+
+static int
+_parse_constant_expression(struct parse_context *ctx,
+                           struct parse_state *ps);
+
+
+static int
+_parse_primary_expression(struct parse_context *ctx,
+                          struct parse_state *ps);
+
+
+static int
+_parse_statement(struct parse_context *ctx,
+                 struct parse_state *ps);
+
+
+static int
+_parse_type_specifier(struct parse_context *ctx,
+                      struct parse_state *ps);
+
+
+static int
+_parse_declaration(struct parse_context *ctx,
+                   struct parse_state *ps);
+
+
+static int
+_parse_statement_list(struct parse_context *ctx,
+                      struct parse_state *ps);
+
+
+static int
+_parse_assignment_expression(struct parse_context *ctx,
+                             struct parse_state *ps);
+
+
+static int
+_parse_precision(struct parse_context *ctx,
+                 struct parse_state *ps);
+
+
+static int
+_parse_overriden_operator(struct parse_context *ctx,
+                          struct parse_state *ps)
+{
+   unsigned int op;
+
+   if (_parse_token(ctx, SL_PP_INCREMENT, ps) == 0) {
+      op = OPERATOR_INCREMENT;
+   } else if (_parse_token(ctx, SL_PP_ADDASSIGN, ps) == 0) {
+      op = OPERATOR_ADDASSIGN;
+   } else if (_parse_token(ctx, SL_PP_PLUS, ps) == 0) {
+      op = OPERATOR_PLUS;
+   } else if (_parse_token(ctx, SL_PP_DECREMENT, ps) == 0) {
+      op = OPERATOR_DECREMENT;
+   } else if (_parse_token(ctx, SL_PP_SUBASSIGN, ps) == 0) {
+      op = OPERATOR_SUBASSIGN;
+   } else if (_parse_token(ctx, SL_PP_MINUS, ps) == 0) {
+      op = OPERATOR_MINUS;
+   } else if (_parse_token(ctx, SL_PP_NOT, ps) == 0) {
+      op = OPERATOR_NOT;
+   } else if (_parse_token(ctx, SL_PP_MULASSIGN, ps) == 0) {
+      op = OPERATOR_MULASSIGN;
+   } else if (_parse_token(ctx, SL_PP_STAR, ps) == 0) {
+      op = OPERATOR_MULTIPLY;
+   } else if (_parse_token(ctx, SL_PP_DIVASSIGN, ps) == 0) {
+      op = OPERATOR_DIVASSIGN;
+   } else if (_parse_token(ctx, SL_PP_SLASH, ps) == 0) {
+      op = OPERATOR_DIVIDE;
+   } else if (_parse_token(ctx, SL_PP_LESSEQUAL, ps) == 0) {
+      op = OPERATOR_LESSEQUAL;
+   } else if (_parse_token(ctx, SL_PP_LESS, ps) == 0) {
+      op = OPERATOR_LESS;
+   } else if (_parse_token(ctx, SL_PP_GREATEREQUAL, ps) == 0) {
+      op = OPERATOR_GREATEREQUAL;
+   } else if (_parse_token(ctx, SL_PP_GREATER, ps) == 0) {
+      op = OPERATOR_GREATER;
+   } else if (_parse_token(ctx, SL_PP_XOR, ps) == 0) {
+      op = OPERATOR_LOGICALXOR;
+   } else {
+      return -1;
+   }
+
+   _emit(ctx, &ps->out, op);
+   return 0;
+}
+
+
+static int
+_parse_function_decl_identifier(struct parse_context *ctx,
+                                struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e = _emit(ctx, &p.out, 0);
+
+   if (ctx->parsing_builtin && _parse_id(ctx, ctx->dict.__constructor, &p) == 0) {
+      _update(ctx, e, FUNCTION_CONSTRUCTOR);
+      *ps = p;
+      return 0;
+   }
+
+   if (ctx->parsing_builtin && _parse_id(ctx, ctx->dict.__operator, &p) == 0) {
+      _update(ctx, e, FUNCTION_OPERATOR);
+      if (_parse_overriden_operator(ctx, &p) == 0) {
+         *ps = p;
+         return 0;
+      }
+      return -1;
+   }
+
+   if (_parse_identifier(ctx, &p) == 0) {
+      _update(ctx, e, FUNCTION_ORDINARY);
+      *ps = p;
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_invariant_qualifier(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   if (_parse_id(ctx, ctx->dict.invariant, ps)) {
+      return -1;
+   }
+   _emit(ctx, &ps->out, TYPE_INVARIANT);
+   return 0;
+}
+
+
+static int
+_parse_centroid_qualifier(struct parse_context *ctx,
+                          struct parse_state *ps)
+{
+   if (_parse_id(ctx, ctx->dict.centroid, ps)) {
+      return -1;
+   }
+   _emit(ctx, &ps->out, TYPE_CENTROID);
+   return 0;
+}
+
+
+static int
+_parse_type_qualifier(struct parse_context *ctx,
+                      struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   const struct sl_pp_token_info *input = _fetch_token(ctx, p.in);
+   unsigned int e = _emit(ctx, &p.out, 0);
+   int id;
+
+   if (!input || input->token != SL_PP_IDENTIFIER) {
+      return -1;
+   }
+   id = input->data.identifier;
+
+   if (id == ctx->dict._const) {
+      _update(ctx, e, TYPE_QUALIFIER_CONST);
+   } else if (ctx->shader_type == 2 && id == ctx->dict.attribute) {
+      _update(ctx, e, TYPE_QUALIFIER_ATTRIBUTE);
+   } else if (id == ctx->dict.varying) {
+      _update(ctx, e, TYPE_QUALIFIER_VARYING);
+   } else if (id == ctx->dict.uniform) {
+      _update(ctx, e, TYPE_QUALIFIER_UNIFORM);
+   } else if (ctx->parsing_builtin && id == ctx->dict.__fixed_output) {
+      _update(ctx, e, TYPE_QUALIFIER_FIXEDOUTPUT);
+   } else if (ctx->parsing_builtin && id == ctx->dict.__fixed_input) {
+      _update(ctx, e, TYPE_QUALIFIER_FIXEDINPUT);
+   } else {
+      return -1;
+   }
+   _parse_token(ctx, SL_PP_IDENTIFIER, &p);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_struct_declarator(struct parse_context *ctx,
+                         struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e;
+
+   if (_parse_identifier(ctx, &p)) {
+      return -1;
+   }
+   e = _emit(ctx, &p.out, FIELD_NONE);
+   *ps = p;
+
+   if (_parse_token(ctx, SL_PP_LBRACKET, &p)) {
+      return 0;
+   }
+   if (_parse_constant_expression(ctx, &p)) {
+      _error(ctx, "expected constant integral expression");
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_RBRACKET, &p)) {
+      _error(ctx, "expected `]'");
+      return -1;
+   }
+   _update(ctx, e, FIELD_ARRAY);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_struct_declarator_list(struct parse_context *ctx,
+                              struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_struct_declarator(ctx, &p)) {
+      return -1;
+   }
+
+   for (;;) {
+      *ps = p;
+      _emit(ctx, &p.out, FIELD_NEXT);
+      if (_parse_token(ctx, SL_PP_COMMA, &p)) {
+         return 0;
+      }
+      if (_parse_struct_declarator(ctx, &p)) {
+         return 0;
+      }
+   }
+}
+
+
+static int
+_parse_struct_declaration(struct parse_context *ctx,
+                          struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_type_specifier(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_struct_declarator_list(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_SEMICOLON, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, FIELD_NONE);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_struct_declaration_list(struct parse_context *ctx,
+                               struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_struct_declaration(ctx, &p)) {
+      return -1;
+   }
+
+   for (;;) {
+      *ps = p;
+      _emit(ctx, &p.out, FIELD_NEXT);
+      if (_parse_struct_declaration(ctx, &p)) {
+         return 0;
+      }
+   }
+}
+
+
+static int
+_parse_struct_specifier(struct parse_context *ctx,
+                        struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_id(ctx, ctx->dict._struct, &p)) {
+      return -1;
+   }
+   if (_parse_identifier(ctx, &p)) {
+      _emit(ctx, &p.out, '\0');
+   }
+   if (_parse_token(ctx, SL_PP_LBRACE, &p)) {
+      _error(ctx, "expected `{'");
+      return -1;
+   }
+   if (_parse_struct_declaration_list(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_RBRACE, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, FIELD_NONE);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_type_specifier_nonarray(struct parse_context *ctx,
+                               struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e = _emit(ctx, &p.out, 0);
+   const struct sl_pp_token_info *input;
+   int id;
+
+   if (_parse_struct_specifier(ctx, &p) == 0) {
+      _update(ctx, e, TYPE_SPECIFIER_STRUCT);
+      *ps = p;
+      return 0;
+   }
+
+   input = _fetch_token(ctx, p.in);
+   if (!input || input->token != SL_PP_IDENTIFIER) {
+      return -1;
+   }
+   id = input->data.identifier;
+
+   if (id == ctx->dict._void) {
+      _update(ctx, e, TYPE_SPECIFIER_VOID);
+   } else if (id == ctx->dict._float) {
+      _update(ctx, e, TYPE_SPECIFIER_FLOAT);
+   } else if (id == ctx->dict._int) {
+      _update(ctx, e, TYPE_SPECIFIER_INT);
+   } else if (id == ctx->dict._bool) {
+      _update(ctx, e, TYPE_SPECIFIER_BOOL);
+   } else if (id == ctx->dict.vec2) {
+      _update(ctx, e, TYPE_SPECIFIER_VEC2);
+   } else if (id == ctx->dict.vec3) {
+      _update(ctx, e, TYPE_SPECIFIER_VEC3);
+   } else if (id == ctx->dict.vec4) {
+      _update(ctx, e, TYPE_SPECIFIER_VEC4);
+   } else if (id == ctx->dict.bvec2) {
+      _update(ctx, e, TYPE_SPECIFIER_BVEC2);
+   } else if (id == ctx->dict.bvec3) {
+      _update(ctx, e, TYPE_SPECIFIER_BVEC3);
+   } else if (id == ctx->dict.bvec4) {
+      _update(ctx, e, TYPE_SPECIFIER_BVEC4);
+   } else if (id == ctx->dict.ivec2) {
+      _update(ctx, e, TYPE_SPECIFIER_IVEC2);
+   } else if (id == ctx->dict.ivec3) {
+      _update(ctx, e, TYPE_SPECIFIER_IVEC3);
+   } else if (id == ctx->dict.ivec4) {
+      _update(ctx, e, TYPE_SPECIFIER_IVEC4);
+   } else if (id == ctx->dict.mat2) {
+      _update(ctx, e, TYPE_SPECIFIER_MAT2);
+   } else if (id == ctx->dict.mat3) {
+      _update(ctx, e, TYPE_SPECIFIER_MAT3);
+   } else if (id == ctx->dict.mat4) {
+      _update(ctx, e, TYPE_SPECIFIER_MAT4);
+   } else if (id == ctx->dict.mat2x3) {
+      _update(ctx, e, TYPE_SPECIFIER_MAT23);
+   } else if (id == ctx->dict.mat3x2) {
+      _update(ctx, e, TYPE_SPECIFIER_MAT32);
+   } else if (id == ctx->dict.mat2x4) {
+      _update(ctx, e, TYPE_SPECIFIER_MAT24);
+   } else if (id == ctx->dict.mat4x2) {
+      _update(ctx, e, TYPE_SPECIFIER_MAT42);
+   } else if (id == ctx->dict.mat3x4) {
+      _update(ctx, e, TYPE_SPECIFIER_MAT34);
+   } else if (id == ctx->dict.mat4x3) {
+      _update(ctx, e, TYPE_SPECIFIER_MAT43);
+   } else if (id == ctx->dict.sampler1D) {
+      _update(ctx, e, TYPE_SPECIFIER_SAMPLER1D);
+   } else if (id == ctx->dict.sampler2D) {
+      _update(ctx, e, TYPE_SPECIFIER_SAMPLER2D);
+   } else if (id == ctx->dict.sampler3D) {
+      _update(ctx, e, TYPE_SPECIFIER_SAMPLER3D);
+   } else if (id == ctx->dict.samplerCube) {
+      _update(ctx, e, TYPE_SPECIFIER_SAMPLERCUBE);
+   } else if (id == ctx->dict.sampler1DShadow) {
+      _update(ctx, e, TYPE_SPECIFIER_SAMPLER1DSHADOW);
+   } else if (id == ctx->dict.sampler2DShadow) {
+      _update(ctx, e, TYPE_SPECIFIER_SAMPLER2DSHADOW);
+   } else if (id == ctx->dict.sampler2DRect) {
+      _update(ctx, e, TYPE_SPECIFIER_SAMPLER2DRECT);
+   } else if (id == ctx->dict.sampler2DRectShadow) {
+      _update(ctx, e, TYPE_SPECIFIER_SAMPLER2DRECTSHADOW);
+   } else if (_parse_identifier(ctx, &p) == 0) {
+      _update(ctx, e, TYPE_SPECIFIER_TYPENAME);
+      *ps = p;
+      return 0;
+   } else {
+      return -1;
+   }
+
+   _parse_token(ctx, SL_PP_IDENTIFIER, &p);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_type_specifier_array(struct parse_context *ctx,
+                            struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_token(ctx, SL_PP_LBRACKET, &p)) {
+      return -1;
+   }
+   if (_parse_constant_expression(ctx, &p)) {
+      _error(ctx, "expected constant integral expression");
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_RBRACKET, &p)) {
+      _error(ctx, "expected `]'");
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_type_specifier(struct parse_context *ctx,
+                      struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e;
+
+   if (_parse_type_specifier_nonarray(ctx, &p)) {
+      return -1;
+   }
+
+   e = _emit(ctx, &p.out, TYPE_SPECIFIER_ARRAY);
+   if (_parse_type_specifier_array(ctx, &p)) {
+      _update(ctx, e, TYPE_SPECIFIER_NONARRAY);
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_fully_specified_type(struct parse_context *ctx,
+                            struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_invariant_qualifier(ctx, &p)) {
+      _emit(ctx, &p.out, TYPE_VARIANT);
+   }
+   if (_parse_centroid_qualifier(ctx, &p)) {
+      _emit(ctx, &p.out, TYPE_CENTER);
+   }
+   if (_parse_type_qualifier(ctx, &p)) {
+      _emit(ctx, &p.out, TYPE_QUALIFIER_NONE);
+   }
+   if (_parse_precision(ctx, &p)) {
+      _emit(ctx, &p.out, PRECISION_DEFAULT);
+   }
+   if (_parse_type_specifier(ctx, &p)) {
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_function_header(struct parse_context *ctx,
+                       struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_fully_specified_type(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_function_decl_identifier(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_LPAREN, &p)) {
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_parameter_qualifier(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   unsigned int e = _emit(ctx, &ps->out, PARAM_QUALIFIER_IN);
+
+   if (_parse_id(ctx, ctx->dict.out, ps) == 0) {
+      _update(ctx, e, PARAM_QUALIFIER_OUT);
+   } else if (_parse_id(ctx, ctx->dict.inout, ps) == 0) {
+      _update(ctx, e, PARAM_QUALIFIER_INOUT);
+   } else {
+      _parse_id(ctx, ctx->dict.in, ps);
+   }
+   return 0;
+}
+
+
+static int
+_parse_function_identifier(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   struct parse_state p;
+   unsigned int e;
+
+   if (_parse_identifier(ctx, ps)) {
+      return -1;
+   }
+   e = _emit(ctx, &ps->out, FUNCTION_CALL_NONARRAY);
+
+   p = *ps;
+   if (_parse_token(ctx, SL_PP_LBRACKET, &p)) {
+      return 0;
+   }
+   if (_parse_constant_expression(ctx, &p)) {
+      _error(ctx, "expected constant integral expression");
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_RBRACKET, &p)) {
+      _error(ctx, "expected `]'");
+      return -1;
+   }
+   _update(ctx, e, FUNCTION_CALL_ARRAY);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_function_call_header(struct parse_context *ctx,
+                            struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_function_identifier(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_LPAREN, &p)) {
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_assign_expression(struct parse_context *ctx,
+                         struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int op;
+
+   if (_parse_unary_expression(ctx, &p)) {
+      return -1;
+   }
+
+   if (_parse_token(ctx, SL_PP_ASSIGN, &p) == 0) {
+      op = OP_ASSIGN;
+   } else if (_parse_token(ctx, SL_PP_MULASSIGN, &p) == 0) {
+      op = OP_MULASSIGN;
+   } else if (_parse_token(ctx, SL_PP_DIVASSIGN, &p) == 0) {
+      op = OP_DIVASSIGN;
+   } else if (_parse_token(ctx, SL_PP_ADDASSIGN, &p) == 0) {
+      op = OP_ADDASSIGN;
+   } else if (_parse_token(ctx, SL_PP_SUBASSIGN, &p) == 0) {
+      op = OP_SUBASSIGN;
+   } else {
+      return -1;
+   }
+
+   if (_parse_assignment_expression(ctx, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, op);
+
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_assignment_expression(struct parse_context *ctx,
+                             struct parse_state *ps)
+{
+   if (_parse_assign_expression(ctx, ps) == 0) {
+      return 0;
+   }
+
+   if (_parse_conditional_expression(ctx, ps) == 0) {
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_function_call_header_with_parameters(struct parse_context *ctx,
+                                            struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_function_call_header(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_assignment_expression(ctx, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_END);
+   for (;;) {
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_COMMA, &p)) {
+         return 0;
+      }
+      if (_parse_assignment_expression(ctx, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, OP_END);
+   }
+}
+
+
+static int
+_parse_function_call_header_no_parameters(struct parse_context *ctx,
+                                          struct parse_state *ps)
+{
+   if (_parse_function_call_header(ctx, ps)) {
+      return -1;
+   }
+   _parse_id(ctx, ctx->dict._void, ps);
+   return 0;
+}
+
+
+static int
+_parse_function_call_generic(struct parse_context *ctx,
+                             struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_function_call_header_with_parameters(ctx, &p) == 0) {
+      if (_parse_token(ctx, SL_PP_RPAREN, &p) == 0) {
+         *ps = p;
+         return 0;
+      }
+      _error(ctx, "expected `)'");
+      return -1;
+   }
+
+   p = *ps;
+   if (_parse_function_call_header_no_parameters(ctx, &p) == 0) {
+      if (_parse_token(ctx, SL_PP_RPAREN, &p) == 0) {
+         *ps = p;
+         return 0;
+      }
+      _error(ctx, "expected `)'");
+      return -1;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_method_call(struct parse_context *ctx,
+                   struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   _emit(ctx, &p.out, OP_METHOD);
+   if (_parse_identifier(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_DOT, &p)) {
+      return -1;
+   }
+   if (_parse_function_call_generic(ctx, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_END);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_regular_function_call(struct parse_context *ctx,
+                             struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   _emit(ctx, &p.out, OP_CALL);
+   if (_parse_function_call_generic(ctx, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_END);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_function_call(struct parse_context *ctx,
+                     struct parse_state *ps)
+{
+   if (_parse_regular_function_call(ctx, ps) == 0) {
+      return 0;
+   }
+
+   if (_parse_method_call(ctx, ps) == 0) {
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_expression(struct parse_context *ctx,
+                  struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_assignment_expression(ctx, &p)) {
+      return -1;
+   }
+
+   for (;;) {
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_COMMA, &p)) {
+         return 0;
+      }
+      if (_parse_assignment_expression(ctx, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, OP_SEQUENCE);
+   }
+}
+
+
+static int
+_parse_postfix_expression(struct parse_context *ctx,
+                          struct parse_state *ps)
+{
+   struct parse_state p;
+
+   if (_parse_function_call(ctx, ps)) {
+      if (_parse_primary_expression(ctx, ps)) {
+         return -1;
+      }
+   }
+
+   for (p = *ps;;) {
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_INCREMENT, &p) == 0) {
+         _emit(ctx, &p.out, OP_POSTINCREMENT);
+      } else if (_parse_token(ctx, SL_PP_DECREMENT, &p) == 0) {
+         _emit(ctx, &p.out, OP_POSTDECREMENT);
+      } else if (_parse_token(ctx, SL_PP_LBRACKET, &p) == 0) {
+         if (_parse_expression(ctx, &p)) {
+            _error(ctx, "expected an integral expression");
+            return -1;
+         }
+         if (_parse_token(ctx, SL_PP_RBRACKET, &p)) {
+            _error(ctx, "expected `]'");
+            return -1;
+         }
+         _emit(ctx, &p.out, OP_SUBSCRIPT);
+      } else if (_parse_token(ctx, SL_PP_DOT, &p) == 0) {
+         _emit(ctx, &p.out, OP_FIELD);
+         if (_parse_identifier(ctx, &p)) {
+            return 0;
+         }
+      } else {
+         return 0;
+      }
+   }
+}
+
+
+static int
+_parse_unary_expression(struct parse_context *ctx,
+                        struct parse_state *ps)
+{
+   struct parse_state p;
+   unsigned int op;
+
+   if (_parse_postfix_expression(ctx, ps) == 0) {
+      return 0;
+   }
+
+   p = *ps;
+   if (_parse_token(ctx, SL_PP_INCREMENT, &p) == 0) {
+      op = OP_PREINCREMENT;
+   } else if (_parse_token(ctx, SL_PP_DECREMENT, &p) == 0) {
+      op = OP_PREDECREMENT;
+   } else if (_parse_token(ctx, SL_PP_PLUS, &p) == 0) {
+      op = OP_PLUS;
+   } else if (_parse_token(ctx, SL_PP_MINUS, &p) == 0) {
+      op = OP_MINUS;
+   } else if (_parse_token(ctx, SL_PP_NOT, &p) == 0) {
+      op = OP_NOT;
+   } else {
+      return -1;
+   }
+
+   if (_parse_unary_expression(ctx, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, op);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_multiplicative_expression(struct parse_context *ctx,
+                                 struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_unary_expression(ctx, &p)) {
+      return -1;
+   }
+   for (;;) {
+      unsigned int op;
+
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_STAR, &p) == 0) {
+         op = OP_MULTIPLY;
+      } else if (_parse_token(ctx, SL_PP_SLASH, &p) == 0) {
+         op = OP_DIVIDE;
+      } else {
+         return 0;
+      }
+      if (_parse_unary_expression(ctx, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, op);
+   }
+}
+
+
+static int
+_parse_additive_expression(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_multiplicative_expression(ctx, &p)) {
+      return -1;
+   }
+   for (;;) {
+      unsigned int op;
+
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_PLUS, &p) == 0) {
+         op = OP_ADD;
+      } else if (_parse_token(ctx, SL_PP_MINUS, &p) == 0) {
+         op = OP_SUBTRACT;
+      } else {
+         return 0;
+      }
+      if (_parse_multiplicative_expression(ctx, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, op);
+   }
+}
+
+
+static int
+_parse_relational_expression(struct parse_context *ctx,
+                             struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_additive_expression(ctx, &p)) {
+      return -1;
+   }
+   for (;;) {
+      unsigned int op;
+
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_LESS, &p) == 0) {
+         op = OP_LESS;
+      } else if (_parse_token(ctx, SL_PP_GREATER, &p) == 0) {
+         op = OP_GREATER;
+      } else if (_parse_token(ctx, SL_PP_LESSEQUAL, &p) == 0) {
+         op = OP_LESSEQUAL;
+      } else if (_parse_token(ctx, SL_PP_GREATEREQUAL, &p) == 0) {
+         op = OP_GREATEREQUAL;
+      } else {
+         return 0;
+      }
+      if (_parse_additive_expression(ctx, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, op);
+   }
+}
+
+
+static int
+_parse_equality_expression(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_relational_expression(ctx, &p)) {
+      return -1;
+   }
+   for (;;) {
+      unsigned int op;
+
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_EQUAL, &p) == 0) {
+         op = OP_EQUAL;
+      } else if (_parse_token(ctx, SL_PP_NOTEQUAL, &p) == 0) {
+         op = OP_NOTEQUAL;
+      } else {
+         return 0;
+      }
+      if (_parse_relational_expression(ctx, &p)) {
+         return -1;
+      }
+      _emit(ctx, &p.out, op);
+   }
+}
+
+
+static int
+_parse_logical_and_expression(struct parse_context *ctx,
+                              struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_equality_expression(ctx, &p)) {
+      return -1;
+   }
+   for (;;) {
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_AND, &p)) {
+         return 0;
+      }
+      if (_parse_equality_expression(ctx, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, OP_LOGICALAND);
+   }
+}
+
+
+static int
+_parse_logical_xor_expression(struct parse_context *ctx,
+                              struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_logical_and_expression(ctx, &p)) {
+      return -1;
+   }
+   for (;;) {
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_XOR, &p)) {
+         return 0;
+      }
+      if (_parse_logical_and_expression(ctx, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, OP_LOGICALXOR);
+   }
+}
+
+
+static int
+_parse_logical_or_expression(struct parse_context *ctx,
+                             struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_logical_xor_expression(ctx, &p)) {
+      return -1;
+   }
+   for (;;) {
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_OR, &p)) {
+         return 0;
+      }
+      if (_parse_logical_xor_expression(ctx, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, OP_LOGICALOR);
+   }
+}
+
+
+static int
+_parse_conditional_expression(struct parse_context *ctx,
+                              struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_logical_or_expression(ctx, &p)) {
+      return -1;
+   }
+   for (;;) {
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_QUESTION, &p)) {
+         return 0;
+      }
+      if (_parse_expression(ctx, &p)) {
+         return 0;
+      }
+      if (_parse_token(ctx, SL_PP_COLON, &p)) {
+         return 0;
+      }
+      if (_parse_conditional_expression(ctx, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, OP_SELECT);
+   }
+}
+
+
+static int
+_parse_constant_expression(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   if (_parse_conditional_expression(ctx, ps)) {
+      return -1;
+   }
+   _emit(ctx, &ps->out, OP_END);
+   return 0;
+}
+
+
+static int
+_parse_parameter_declarator_array(struct parse_context *ctx,
+                                  struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_token(ctx, SL_PP_LBRACKET, &p)) {
+      return -1;
+   }
+   if (_parse_constant_expression(ctx, &p)) {
+      _error(ctx, "expected constant integral expression");
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_RBRACKET, &p)) {
+      _error(ctx, "expected `]'");
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_parameter_declarator(struct parse_context *ctx,
+                            struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e;
+
+   if (_parse_type_specifier(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_identifier(ctx, &p)) {
+      return -1;
+   }
+   e = _emit(ctx, &p.out, PARAMETER_ARRAY_PRESENT);
+   if (_parse_parameter_declarator_array(ctx, &p)) {
+      _update(ctx, e, PARAMETER_ARRAY_NOT_PRESENT);
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_parameter_type_specifier_array(struct parse_context *ctx,
+                                      struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_token(ctx, SL_PP_LBRACKET, &p)) {
+      return -1;
+   }
+   if (_parse_constant_expression(ctx, &p)) {
+      _error(ctx, "expected constant integral expression");
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_RBRACKET, &p)) {
+      _error(ctx, "expected `]'");
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_parameter_type_specifier(struct parse_context *ctx,
+                                struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e;
+
+   if (_parse_type_specifier(ctx, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, '\0');
+
+   e = _emit(ctx, &p.out, PARAMETER_ARRAY_PRESENT);
+   if (_parse_parameter_type_specifier_array(ctx, &p)) {
+      _update(ctx, e, PARAMETER_ARRAY_NOT_PRESENT);
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_parameter_declaration(struct parse_context *ctx,
+                             struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e = _emit(ctx, &p.out, PARAMETER_NEXT);
+
+   (void) e;
+
+   if (_parse_type_qualifier(ctx, &p)) {
+      _emit(ctx, &p.out, TYPE_QUALIFIER_NONE);
+   }
+   _parse_parameter_qualifier(ctx, &p);
+   if (_parse_precision(ctx, &p)) {
+      _emit(ctx, &p.out, PRECISION_DEFAULT);
+   }
+   if (_parse_parameter_declarator(ctx, &p) == 0) {
+      *ps = p;
+      return 0;
+   }
+   if (_parse_parameter_type_specifier(ctx, &p) == 0) {
+      *ps = p;
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_function_header_with_parameters(struct parse_context *ctx,
+                                       struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_function_header(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_parameter_declaration(ctx, &p)) {
+      return -1;
+   }
+
+   for (;;) {
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_COMMA, &p)) {
+         return 0;
+      }
+      if (_parse_parameter_declaration(ctx, &p)) {
+         return 0;
+      }
+   }
+}
+
+
+static int
+_parse_function_declarator(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   if (_parse_function_header_with_parameters(ctx, ps) == 0) {
+      return 0;
+   }
+
+   if (_parse_function_header(ctx, ps) == 0) {
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_function_prototype(struct parse_context *ctx,
+                          struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_function_header(ctx, &p) == 0) {
+      if (_parse_id(ctx, ctx->dict._void, &p) == 0) {
+         if (_parse_token(ctx, SL_PP_RPAREN, &p) == 0) {
+            _emit(ctx, &p.out, PARAMETER_NONE);
+            *ps = p;
+            return 0;
+         }
+         _error(ctx, "expected `)'");
+         return -1;
+      }
+   }
+
+   p = *ps;
+   if (_parse_function_declarator(ctx, &p) == 0) {
+      if (_parse_token(ctx, SL_PP_RPAREN, &p) == 0) {
+         _emit(ctx, &p.out, PARAMETER_NONE);
+         *ps = p;
+         return 0;
+      }
+      _error(ctx, "expected `)'");
+      return -1;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_precision(struct parse_context *ctx,
+                 struct parse_state *ps)
+{
+   const struct sl_pp_token_info *input = _fetch_token(ctx, ps->in);
+   int id;
+   unsigned int precision;
+
+   if (!input || input->token != SL_PP_IDENTIFIER) {
+      return -1;
+   }
+   id = input->data.identifier;
+
+   if (id == ctx->dict.lowp) {
+      precision = PRECISION_LOW;
+   } else if (id == ctx->dict.mediump) {
+      precision = PRECISION_MEDIUM;
+   } else if (id == ctx->dict.highp) {
+      precision = PRECISION_HIGH;
+   } else {
+      return -1;
+   }
+
+   _parse_token(ctx, SL_PP_IDENTIFIER, ps);
+   _emit(ctx, &ps->out, precision);
+   return 0;
+}
+
+
+static int
+_parse_prectype(struct parse_context *ctx,
+                 struct parse_state *ps)
+{
+   const struct sl_pp_token_info *input = _fetch_token(ctx, ps->in);
+   int id;
+   unsigned int type;
+
+   if (!input || input->token != SL_PP_IDENTIFIER) {
+      return -1;
+   }
+   id = input->data.identifier;
+
+   if (id == ctx->dict._int) {
+      type = TYPE_SPECIFIER_INT;
+   } else if (id == ctx->dict._float) {
+      type = TYPE_SPECIFIER_FLOAT;
+   } else if (id == ctx->dict.sampler1D) {
+      type = TYPE_SPECIFIER_SAMPLER1D;
+   } else if (id == ctx->dict.sampler2D) {
+      type = TYPE_SPECIFIER_SAMPLER2D;
+   } else if (id == ctx->dict.sampler3D) {
+      type = TYPE_SPECIFIER_SAMPLER3D;
+   } else if (id == ctx->dict.samplerCube) {
+      type = TYPE_SPECIFIER_SAMPLERCUBE;
+   } else if (id == ctx->dict.sampler1DShadow) {
+      type = TYPE_SPECIFIER_SAMPLER1DSHADOW;
+   } else if (id == ctx->dict.sampler2DShadow) {
+      type = TYPE_SPECIFIER_SAMPLER2DSHADOW;
+   } else if (id == ctx->dict.sampler2DRect) {
+      type = TYPE_SPECIFIER_SAMPLER2DRECT;
+   } else if (id == ctx->dict.sampler2DRectShadow) {
+      type = TYPE_SPECIFIER_SAMPLER2DRECTSHADOW;
+   } else {
+      return -1;
+   }
+
+   _parse_token(ctx, SL_PP_IDENTIFIER, ps);
+   _emit(ctx, &ps->out, type);
+   return 0;
+}
+
+
+static int
+_parse_precision_stmt(struct parse_context *ctx,
+                      struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_id(ctx, ctx->dict.precision, &p)) {
+      return -1;
+   }
+   if (_parse_precision(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_prectype(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_SEMICOLON, &p)) {
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_floatconstant(struct parse_context *ctx,
+                     struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   _emit(ctx, &p.out, OP_PUSH_FLOAT);
+   if (_parse_float(ctx, &p)) {
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_intconstant(struct parse_context *ctx,
+                   struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   _emit(ctx, &p.out, OP_PUSH_INT);
+   if (_parse_uint(ctx, &p)) {
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_boolconstant(struct parse_context *ctx,
+                    struct parse_state *ps)
+{
+   if (_parse_id(ctx, ctx->dict._false, ps) == 0) {
+      _emit(ctx, &ps->out, OP_PUSH_BOOL);
+      _emit(ctx, &ps->out, 2);  /* radix */
+      _emit(ctx, &ps->out, '0');
+      _emit(ctx, &ps->out, '\0');
+      return 0;
+   }
+
+   if (_parse_id(ctx, ctx->dict._true, ps) == 0) {
+      _emit(ctx, &ps->out, OP_PUSH_BOOL);
+      _emit(ctx, &ps->out, 2);  /* radix */
+      _emit(ctx, &ps->out, '1');
+      _emit(ctx, &ps->out, '\0');
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_variable_identifier(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   _emit(ctx, &p.out, OP_PUSH_IDENTIFIER);
+   if (_parse_identifier(ctx, &p)) {
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_primary_expression(struct parse_context *ctx,
+                          struct parse_state *ps)
+{
+   struct parse_state p;
+
+   if (_parse_floatconstant(ctx, ps) == 0) {
+      return 0;
+   }
+   if (_parse_boolconstant(ctx, ps) == 0) {
+      return 0;
+   }
+   if (_parse_intconstant(ctx, ps) == 0) {
+      return 0;
+   }
+   if (_parse_variable_identifier(ctx, ps) == 0) {
+      return 0;
+   }
+
+   p = *ps;
+   if (_parse_token(ctx, SL_PP_LPAREN, &p)) {
+      return -1;
+   }
+   if (_parse_expression(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_RPAREN, &p)) {
+      return -1;
+   }
+
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_asm_argument(struct parse_context *ctx,
+                    struct parse_state *ps)
+{
+   if (_parse_variable_identifier(ctx, ps) == 0) {
+      struct parse_state p = *ps;
+
+      if (_parse_token(ctx, SL_PP_DOT, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, OP_FIELD);
+      if (_parse_identifier(ctx, &p)) {
+         return 0;
+      }
+      *ps = p;
+      return 0;
+   }
+
+   if (_parse_floatconstant(ctx, ps) == 0) {
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_asm_arguments(struct parse_context *ctx,
+                     struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_asm_argument(ctx, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_END);
+
+   for (;;) {
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_COMMA, &p)) {
+         return 0;
+      }
+      if (_parse_asm_argument(ctx, &p)) {
+         return 0;
+      }
+      _emit(ctx, &p.out, OP_END);
+   }
+}
+
+
+static int
+_parse_asm_statement(struct parse_context *ctx,
+                     struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_id(ctx, ctx->dict.___asm, &p)) {
+      return -1;
+   }
+   if (_parse_identifier(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_asm_arguments(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_SEMICOLON, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_END);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_selection_statement(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   _emit(ctx, &p.out, OP_IF);
+   if (_parse_id(ctx, ctx->dict._if, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_LPAREN, &p)) {
+      _error(ctx, "expected `('");
+      return -1;
+   }
+   if (_parse_expression(ctx, &p)) {
+      _error(ctx, "expected an expression");
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_RPAREN, &p)) {
+      _error(ctx, "expected `)'");
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_END);
+   if (_parse_statement(ctx, &p)) {
+      return -1;
+   }
+
+   *ps = p;
+   if (_parse_id(ctx, ctx->dict._else, &p) == 0) {
+      if (_parse_statement(ctx, &p) == 0) {
+         *ps = p;
+         return 0;
+      }
+   }
+
+   _emit(ctx, &ps->out, OP_EXPRESSION);
+   _emit(ctx, &ps->out, OP_PUSH_VOID);
+   _emit(ctx, &ps->out, OP_END);
+   return 0;
+}
+
+
+static int
+_parse_expression_statement(struct parse_context *ctx,
+                            struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_expression(ctx, &p)) {
+      _emit(ctx, &p.out, OP_PUSH_VOID);
+   }
+   if (_parse_token(ctx, SL_PP_SEMICOLON, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_END);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_for_init_statement(struct parse_context *ctx,
+                          struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e = _emit(ctx, &p.out, OP_EXPRESSION);
+
+   if (_parse_expression_statement(ctx, &p) == 0) {
+      *ps = p;
+      return 0;
+   }
+
+   if (_parse_declaration(ctx, &p) == 0) {
+      _update(ctx, e, OP_DECLARE);
+      *ps = p;
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_initializer(struct parse_context *ctx,
+                   struct parse_state *ps)
+{
+   if (_parse_assignment_expression(ctx, ps) == 0) {
+      _emit(ctx, &ps->out, OP_END);
+      return 0;
+   }
+   return -1;
+}
+
+
+static int
+_parse_condition_initializer(struct parse_context *ctx,
+                             struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   _emit(ctx, &p.out, OP_DECLARE);
+   _emit(ctx, &p.out, DECLARATION_INIT_DECLARATOR_LIST);
+   if (_parse_fully_specified_type(ctx, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, VARIABLE_IDENTIFIER);
+   if (_parse_identifier(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_ASSIGN, &p)) {
+      _error(ctx, "expected `='");
+      return -1;
+   }
+   _emit(ctx, &p.out, VARIABLE_INITIALIZER);
+   if (_parse_initializer(ctx, &p)) {
+      _error(ctx, "expected an initialiser");
+      return -1;
+   }
+   _emit(ctx, &p.out, DECLARATOR_NONE);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_condition(struct parse_context *ctx,
+                 struct parse_state *ps)
+{
+   struct parse_state p;
+
+   if (_parse_condition_initializer(ctx, ps) == 0) {
+      return 0;
+   }
+
+   p = *ps;
+   _emit(ctx, &p.out, OP_EXPRESSION);
+   if (_parse_expression(ctx, &p) == 0) {
+      _emit(ctx, &p.out, OP_END);
+      *ps = p;
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_for_rest_statement(struct parse_context *ctx,
+                          struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_condition(ctx, &p)) {
+      _emit(ctx, &p.out, OP_EXPRESSION);
+      _emit(ctx, &p.out, OP_PUSH_BOOL);
+      _emit(ctx, &p.out, 2);
+      _emit(ctx, &p.out, '1');
+      _emit(ctx, &p.out, '\0');
+      _emit(ctx, &p.out, OP_END);
+   }
+   if (_parse_token(ctx, SL_PP_SEMICOLON, &p)) {
+      return -1;
+   }
+   if (_parse_expression(ctx, &p)) {
+      _emit(ctx, &p.out, OP_PUSH_VOID);
+   }
+   _emit(ctx, &p.out, OP_END);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_iteration_statement(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_id(ctx, ctx->dict._while, &p) == 0) {
+      _emit(ctx, &p.out, OP_WHILE);
+      if (_parse_token(ctx, SL_PP_LPAREN, &p)) {
+         _error(ctx, "expected `('");
+         return -1;
+      }
+      if (_parse_condition(ctx, &p)) {
+         _error(ctx, "expected an expression");
+         return -1;
+      }
+      if (_parse_token(ctx, SL_PP_RPAREN, &p)) {
+         _error(ctx, "expected `)'");
+         return -1;
+      }
+      if (_parse_statement(ctx, &p)) {
+         return -1;
+      }
+      *ps = p;
+      return 0;
+   }
+
+   if (_parse_id(ctx, ctx->dict._do, &p) == 0) {
+      _emit(ctx, &p.out, OP_DO);
+      if (_parse_statement(ctx, &p)) {
+         return -1;
+      }
+      if (_parse_id(ctx, ctx->dict._while, &p)) {
+         return -1;
+      }
+      if (_parse_token(ctx, SL_PP_LPAREN, &p)) {
+         _error(ctx, "expected `('");
+         return -1;
+      }
+      if (_parse_expression(ctx, &p)) {
+         _error(ctx, "expected an expression");
+         return -1;
+      }
+      if (_parse_token(ctx, SL_PP_RPAREN, &p)) {
+         _error(ctx, "expected `)'");
+         return -1;
+      }
+      _emit(ctx, &p.out, OP_END);
+      if (_parse_token(ctx, SL_PP_SEMICOLON, &p)) {
+         _error(ctx, "expected `;'");
+         return -1;
+      }
+      *ps = p;
+      return 0;
+   }
+
+   if (_parse_id(ctx, ctx->dict._for, &p) == 0) {
+      _emit(ctx, &p.out, OP_FOR);
+      if (_parse_token(ctx, SL_PP_LPAREN, &p)) {
+         _error(ctx, "expected `('");
+         return -1;
+      }
+      if (_parse_for_init_statement(ctx, &p)) {
+         return -1;
+      }
+      if (_parse_for_rest_statement(ctx, &p)) {
+         return -1;
+      }
+      if (_parse_token(ctx, SL_PP_RPAREN, &p)) {
+         _error(ctx, "expected `)'");
+         return -1;
+      }
+      if (_parse_statement(ctx, &p)) {
+         return -1;
+      }
+      *ps = p;
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_jump_statement(struct parse_context *ctx,
+                      struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e = _emit(ctx, &p.out, 0);
+
+   if (_parse_id(ctx, ctx->dict._continue, &p) == 0) {
+      _update(ctx, e, OP_CONTINUE);
+   } else if (_parse_id(ctx, ctx->dict._break, &p) == 0) {
+      _update(ctx, e, OP_BREAK);
+   } else if (_parse_id(ctx, ctx->dict._return, &p) == 0) {
+      _update(ctx, e, OP_RETURN);
+      if (_parse_expression(ctx, &p)) {
+         _emit(ctx, &p.out, OP_PUSH_VOID);
+      }
+      _emit(ctx, &p.out, OP_END);
+   } else if (ctx->shader_type == 1 && _parse_id(ctx, ctx->dict.discard, &p) == 0) {
+      _update(ctx, e, OP_DISCARD);
+   } else {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_SEMICOLON, &p)) {
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_simple_statement(struct parse_context *ctx,
+                        struct parse_state *ps)
+{
+   struct parse_state p;
+   unsigned int e;
+
+   if (_parse_selection_statement(ctx, ps) == 0) {
+      return 0;
+   }
+
+   if (_parse_iteration_statement(ctx, ps) == 0) {
+      return 0;
+   }
+
+   if (_parse_jump_statement(ctx, ps) == 0) {
+      return 0;
+   }
+
+   p = *ps;
+   e = _emit(ctx, &p.out, OP_EXPRESSION);
+   if (_parse_expression_statement(ctx, &p) == 0) {
+      *ps = p;
+      return 0;
+   }
+
+   if (_parse_precision_stmt(ctx, &p) == 0) {
+      _update(ctx, e, OP_PRECISION);
+      *ps = p;
+      return 0;
+   }
+
+   if (ctx->parsing_builtin && _parse_asm_statement(ctx, &p) == 0) {
+      _update(ctx, e, OP_ASM);
+      *ps = p;
+      return 0;
+   }
+
+   if (_parse_declaration(ctx, &p) == 0) {
+      _update(ctx, e, OP_DECLARE);
+      *ps = p;
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_compound_statement(struct parse_context *ctx,
+                          struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_token(ctx, SL_PP_LBRACE, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_BLOCK_BEGIN_NEW_SCOPE);
+   _parse_statement_list(ctx, &p);
+   if (_parse_token(ctx, SL_PP_RBRACE, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_END);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_statement(struct parse_context *ctx,
+                 struct parse_state *ps)
+{
+   if (_parse_compound_statement(ctx, ps) == 0) {
+      return 0;
+   }
+
+   if (_parse_simple_statement(ctx, ps) == 0) {
+      return 0;
+   }
+
+   return -1;
+}
+
+
+static int
+_parse_statement_list(struct parse_context *ctx,
+                      struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_statement(ctx, &p)) {
+      return -1;
+   }
+
+   for (;;) {
+      *ps = p;
+      if (_parse_statement(ctx, &p)) {
+         return 0;
+      }
+   }
+}
+
+
+static int
+_parse_compound_statement_no_new_scope(struct parse_context *ctx,
+                                       struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_token(ctx, SL_PP_LBRACE, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_BLOCK_BEGIN_NO_NEW_SCOPE);
+   _parse_statement_list(ctx, &p);
+   if (_parse_token(ctx, SL_PP_RBRACE, &p)) {
+      return -1;
+   }
+   _emit(ctx, &p.out, OP_END);
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_function_definition(struct parse_context *ctx,
+                           struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_function_prototype(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_compound_statement_no_new_scope(ctx, &p)) {
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_invariant_stmt(struct parse_context *ctx,
+                      struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_id(ctx, ctx->dict.invariant, &p)) {
+      return -1;
+   }
+   if (_parse_identifier(ctx, &p)) {
+      return -1;
+   }
+   if (_parse_token(ctx, SL_PP_SEMICOLON, &p)) {
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_single_declaration(struct parse_context *ctx,
+                          struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e;
+
+   if (_parse_fully_specified_type(ctx, &p)) {
+      return -1;
+   }
+
+   e = _emit(ctx, &p.out, VARIABLE_IDENTIFIER);
+   if (_parse_identifier(ctx, &p)) {
+      _update(ctx, e, VARIABLE_NONE);
+      *ps = p;
+      return 0;
+   }
+
+   e = _emit(ctx, &p.out, VARIABLE_NONE);
+   *ps = p;
+
+   if (_parse_token(ctx, SL_PP_ASSIGN, &p) == 0) {
+      _update(ctx, e, VARIABLE_INITIALIZER);
+      if (_parse_initializer(ctx, &p) == 0) {
+         *ps = p;
+         return 0;
+      }
+      _error(ctx, "expected an initialiser");
+      return -1;
+   }
+   p = *ps;
+
+   if (_parse_token(ctx, SL_PP_LBRACKET, &p) == 0) {
+      if (_parse_constant_expression(ctx, &p)) {
+         _update(ctx, e, VARIABLE_ARRAY_UNKNOWN);
+      } else {
+         _update(ctx, e, VARIABLE_ARRAY_EXPLICIT);
+      }
+      if (_parse_token(ctx, SL_PP_RBRACKET, &p) == 0) {
+         *ps = p;
+         return 0;
+      }
+      _error(ctx, "expected `]'");
+      return -1;
+   }
+   return 0;
+}
+
+
+static int
+_parse_init_declarator_list(struct parse_context *ctx,
+                            struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+
+   if (_parse_single_declaration(ctx, &p)) {
+      return -1;
+   }
+
+   for (;;) {
+      unsigned int e;
+
+      *ps = p;
+      if (_parse_token(ctx, SL_PP_COMMA, &p)) {
+         break;
+      }
+      _emit(ctx, &p.out, DECLARATOR_NEXT);
+      _emit(ctx, &p.out, VARIABLE_IDENTIFIER);
+      if (_parse_identifier(ctx, &p)) {
+         break;
+      }
+
+      e = _emit(ctx, &p.out, VARIABLE_NONE);
+      *ps = p;
+
+      if (_parse_token(ctx, SL_PP_ASSIGN, &p) == 0) {
+         if (_parse_initializer(ctx, &p) == 0) {
+            _update(ctx, e, VARIABLE_INITIALIZER);
+            *ps = p;
+            continue;
+         }
+         _error(ctx, "expected an initialiser");
+         break;
+      }
+      p = *ps;
+
+      if (_parse_token(ctx, SL_PP_LBRACKET, &p) == 0) {
+         unsigned int arr;
+
+         if (_parse_constant_expression(ctx, &p)) {
+            arr = VARIABLE_ARRAY_UNKNOWN;
+         } else {
+            arr = VARIABLE_ARRAY_EXPLICIT;
+         }
+         if (_parse_token(ctx, SL_PP_RBRACKET, &p) == 0) {
+            _update(ctx, e, arr);
+            *ps = p;
+            continue;
+         }
+         _error(ctx, "expected `]'");
+         break;
+      }
+      p = *ps;
+   }
+
+   _emit(ctx, &ps->out, DECLARATOR_NONE);
+   return 0;
+}
+
+
+static int
+_parse_declaration(struct parse_context *ctx,
+                   struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e = _emit(ctx, &p.out, DECLARATION_FUNCTION_PROTOTYPE);
+
+   if (_parse_function_prototype(ctx, &p)) {
+      if (_parse_init_declarator_list(ctx, &p)) {
+         return -1;
+      }
+      _update(ctx, e, DECLARATION_INIT_DECLARATOR_LIST);
+   }
+   if (_parse_token(ctx, SL_PP_SEMICOLON, &p)) {
+      _error(ctx, "expected `;'");
+      return -1;
+   }
+   *ps = p;
+   return 0;
+}
+
+
+static int
+_parse_external_declaration(struct parse_context *ctx,
+                            struct parse_state *ps)
+{
+   struct parse_state p = *ps;
+   unsigned int e = _emit(ctx, &p.out, 0);
+
+   if (_parse_precision_stmt(ctx, &p) == 0) {
+      _update(ctx, e, DEFAULT_PRECISION);
+      *ps = p;
+      return 0;
+   }
+
+   if (_parse_function_definition(ctx, &p) == 0) {
+      _update(ctx, e, EXTERNAL_FUNCTION_DEFINITION);
+      *ps = p;
+      return 0;
+   }
+
+   if (_parse_invariant_stmt(ctx, &p) == 0) {
+      _update(ctx, e, INVARIANT_STMT);
+      *ps = p;
+      return 0;
+   }
+
+   if (_parse_declaration(ctx, &p) == 0) {
+      _update(ctx, e, EXTERNAL_DECLARATION);
+      *ps = p;
+      return 0;
+   }
+
+   _error(ctx, "expected an identifier");
+   return -1;
+}
+
+
+static int
+_parse_translation_unit(struct parse_context *ctx,
+                        struct parse_state *ps)
+{
+   _emit(ctx, &ps->out, REVISION);
+   if (_parse_external_declaration(ctx, ps)) {
+      return -1;
+   }
+   while (_parse_external_declaration(ctx, ps) == 0) {
+   }
+   _emit(ctx, &ps->out, EXTERNAL_NULL);
+   if (_parse_token(ctx, SL_PP_EOF, ps)) {
+      return -1;
+   }
+   return 0;
+}
+
+
+#define ADD_NAME_STR(CTX, NAME, STR)\
+   do {\
+      (CTX).dict.NAME = sl_pp_context_add_unique_str((CTX).context, (STR));\
+      if ((CTX).dict.NAME == -1) {\
+         return -1;\
+      }\
+   } while (0)
+
+#define ADD_NAME(CTX, NAME) ADD_NAME_STR(CTX, NAME, #NAME)
+
+
+int
+sl_cl_compile(struct sl_pp_context *context,
+              unsigned int shader_type,
+              unsigned int parsing_builtin,
+              unsigned char **output,
+              unsigned int *cboutput,
+              char *error,
+              unsigned int cberror)
+{
+   struct parse_context ctx;
+   struct parse_state ps;
+
+   ctx.context = context;
+
+   ADD_NAME_STR(ctx, _void, "void");
+   ADD_NAME_STR(ctx, _float, "float");
+   ADD_NAME_STR(ctx, _int, "int");
+   ADD_NAME_STR(ctx, _bool, "bool");
+   ADD_NAME(ctx, vec2);
+   ADD_NAME(ctx, vec3);
+   ADD_NAME(ctx, vec4);
+   ADD_NAME(ctx, bvec2);
+   ADD_NAME(ctx, bvec3);
+   ADD_NAME(ctx, bvec4);
+   ADD_NAME(ctx, ivec2);
+   ADD_NAME(ctx, ivec3);
+   ADD_NAME(ctx, ivec4);
+   ADD_NAME(ctx, mat2);
+   ADD_NAME(ctx, mat3);
+   ADD_NAME(ctx, mat4);
+   ADD_NAME(ctx, mat2x3);
+   ADD_NAME(ctx, mat3x2);
+   ADD_NAME(ctx, mat2x4);
+   ADD_NAME(ctx, mat4x2);
+   ADD_NAME(ctx, mat3x4);
+   ADD_NAME(ctx, mat4x3);
+   ADD_NAME(ctx, sampler1D);
+   ADD_NAME(ctx, sampler2D);
+   ADD_NAME(ctx, sampler3D);
+   ADD_NAME(ctx, samplerCube);
+   ADD_NAME(ctx, sampler1DShadow);
+   ADD_NAME(ctx, sampler2DShadow);
+   ADD_NAME(ctx, sampler2DRect);
+   ADD_NAME(ctx, sampler2DRectShadow);
+
+   ADD_NAME(ctx, invariant);
+
+   ADD_NAME(ctx, centroid);
+
+   ADD_NAME(ctx, precision);
+   ADD_NAME(ctx, lowp);
+   ADD_NAME(ctx, mediump);
+   ADD_NAME(ctx, highp);
+
+   ADD_NAME_STR(ctx, _const, "const");
+   ADD_NAME(ctx, attribute);
+   ADD_NAME(ctx, varying);
+   ADD_NAME(ctx, uniform);
+   ADD_NAME(ctx, __fixed_output);
+   ADD_NAME(ctx, __fixed_input);
+
+   ADD_NAME(ctx, in);
+   ADD_NAME(ctx, out);
+   ADD_NAME(ctx, inout);
+
+   ADD_NAME_STR(ctx, _struct, "struct");
+
+   ADD_NAME(ctx, __constructor);
+   ADD_NAME(ctx, __operator);
+   ADD_NAME_STR(ctx, ___asm, "__asm");
+
+   ADD_NAME_STR(ctx, _if, "if");
+   ADD_NAME_STR(ctx, _else, "else");
+   ADD_NAME_STR(ctx, _for, "for");
+   ADD_NAME_STR(ctx, _while, "while");
+   ADD_NAME_STR(ctx, _do, "do");
+
+   ADD_NAME_STR(ctx, _continue, "continue");
+   ADD_NAME_STR(ctx, _break, "break");
+   ADD_NAME_STR(ctx, _return, "return");
+   ADD_NAME(ctx, discard);
+
+   ADD_NAME_STR(ctx, _false, "false");
+   ADD_NAME_STR(ctx, _true, "true");
+
+   ctx.out_buf = NULL;
+   ctx.out_cap = 0;
+
+   ctx.shader_type = shader_type;
+   ctx.parsing_builtin = 1;
+
+   ctx.error[0] = '\0';
+   ctx.process_error = 0;
+
+   ctx.tokens_cap = 1024;
+   ctx.tokens_read = 0;
+   ctx.tokens = malloc(ctx.tokens_cap * sizeof(struct sl_pp_token_info));
+   if (!ctx.tokens) {
+      strncpy(error, "out of memory", cberror);
+      return -1;
+   }
+
+   ps.in = 0;
+   ps.out = 0;
+
+   if (_parse_translation_unit(&ctx, &ps)) {
+      strncpy(error, ctx.error, cberror);
+      free(ctx.tokens);
+      return -1;
+   }
+
+   *output = ctx.out_buf;
+   *cboutput = ps.out;
+   free(ctx.tokens);
+   return 0;
+}
diff --git a/src/glsl/cl/sl_cl_parse.h b/src/glsl/cl/sl_cl_parse.h
new file mode 100644 (file)
index 0000000..dd5791d
--- /dev/null
@@ -0,0 +1,40 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef SL_CL_PARSE_H
+#define SL_CL_PARSE_H
+
+int
+sl_cl_compile(struct sl_pp_context *context,
+              unsigned int shader_type,
+              unsigned int parsing_builtin,
+              unsigned char **output,
+              unsigned int *cboutput,
+              char *error,
+              unsigned int cberror);
+
+#endif /* SL_CL_PARSE_H */
diff --git a/src/glsl/pp/Makefile b/src/glsl/pp/Makefile
new file mode 100644 (file)
index 0000000..fda1c42
--- /dev/null
@@ -0,0 +1,27 @@
+#src/glsl/pp/Makefile
+
+TOP = ../../..
+
+include $(TOP)/configs/current
+
+LIBNAME = glslpp
+
+C_SOURCES = \
+       sl_pp_context.c \
+       sl_pp_define.c \
+       sl_pp_dict.c \
+       sl_pp_error.c \
+       sl_pp_expression.c \
+       sl_pp_extension.c \
+       sl_pp_if.c \
+       sl_pp_line.c \
+       sl_pp_macro.c \
+       sl_pp_pragma.c \
+       sl_pp_process.c \
+       sl_pp_purify.c \
+       sl_pp_token.c \
+       sl_pp_token_util.c \
+       sl_pp_version.c
+
+include ../Makefile.template
+
diff --git a/src/glsl/pp/sl_pp_context.c b/src/glsl/pp/sl_pp_context.c
new file mode 100644 (file)
index 0000000..74a9bdd
--- /dev/null
@@ -0,0 +1,182 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_public.h"
+#include "sl_pp_context.h"
+
+
+struct sl_pp_context *
+sl_pp_context_create(const char *input,
+                     const struct sl_pp_purify_options *options)
+{
+   struct sl_pp_context *context;
+
+   context = calloc(1, sizeof(struct sl_pp_context));
+   if (!context) {
+      return NULL;
+   }
+
+   if (sl_pp_dict_init(context)) {
+      sl_pp_context_destroy(context);
+      return NULL;
+   }
+
+   context->getc_buf_capacity = 64;
+   context->getc_buf = malloc(context->getc_buf_capacity * sizeof(char));
+   if (!context->getc_buf) {
+      sl_pp_context_destroy(context);
+      return NULL;
+   }
+
+   if (sl_pp_token_buffer_init(&context->tokens, context)) {
+      sl_pp_context_destroy(context);
+      return NULL;
+   }
+
+   context->macro_tail = &context->macro;
+   context->if_ptr = SL_PP_MAX_IF_NESTING;
+   context->if_value = 1;
+   memset(context->error_msg, 0, sizeof(context->error_msg));
+   context->error_line = 1;
+   context->line = 1;
+   context->file = 0;
+
+   sl_pp_purify_state_init(&context->pure, input, options);
+
+   memset(&context->process_state, 0, sizeof(context->process_state));
+
+   return context;
+}
+
+void
+sl_pp_context_destroy(struct sl_pp_context *context)
+{
+   if (context) {
+      free(context->cstr_pool);
+      sl_pp_macro_free(context->macro);
+      free(context->getc_buf);
+      sl_pp_token_buffer_destroy(&context->tokens);
+      free(context->process_state.out);
+      free(context);
+   }
+}
+
+const char *
+sl_pp_context_error_message(const struct sl_pp_context *context)
+{
+   return context->error_msg;
+}
+
+void
+sl_pp_context_error_position(const struct sl_pp_context *context,
+                             unsigned int *file,
+                             unsigned int *line)
+{
+   if (file) {
+      *file = 0;
+   }
+   if (line) {
+      *line = context->error_line;
+   }
+}
+
+int
+sl_pp_context_add_predefined(struct sl_pp_context *context,
+                             const char *name,
+                             const char *value)
+{
+   struct sl_pp_predefined pre;
+
+   if (context->num_predefined == SL_PP_MAX_PREDEFINED) {
+      return -1;
+   }
+
+   pre.name = sl_pp_context_add_unique_str(context, name);
+   if (pre.name == -1) {
+      return -1;
+   }
+
+   pre.value = sl_pp_context_add_unique_str(context, value);
+   if (pre.value == -1) {
+      return -1;
+   }
+
+   context->predefined[context->num_predefined++] = pre;
+   return 0;
+}
+
+int
+sl_pp_context_add_unique_str(struct sl_pp_context *context,
+                             const char *str)
+{
+   unsigned int size;
+   unsigned int offset = 0;
+
+   size = strlen(str) + 1;
+
+   /* Find out if this is a unique string. */
+   while (offset < context->cstr_pool_len) {
+      const char *str2;
+      unsigned int size2;
+
+      str2 = &context->cstr_pool[offset];
+      size2 = strlen(str2) + 1;
+      if (size == size2 && !memcmp(str, str2, size - 1)) {
+         return offset;
+      }
+
+      offset += size2;
+   }
+
+   if (context->cstr_pool_len + size > context->cstr_pool_max) {
+      context->cstr_pool_max = (context->cstr_pool_len + size + 0xffff) & ~0xffff;
+      context->cstr_pool = realloc(context->cstr_pool, context->cstr_pool_max);
+   }
+
+   if (!context->cstr_pool) {
+      strcpy(context->error_msg, "out of memory");
+      return -1;
+   }
+
+   offset = context->cstr_pool_len;
+   memcpy(&context->cstr_pool[offset], str, size);
+   context->cstr_pool_len += size;
+
+   return offset;
+}
+
+const char *
+sl_pp_context_cstr(const struct sl_pp_context *context,
+                   int offset)
+{
+   if (offset == -1) {
+      return NULL;
+   }
+   return &context->cstr_pool[offset];
+}
diff --git a/src/glsl/pp/sl_pp_context.h b/src/glsl/pp/sl_pp_context.h
new file mode 100644 (file)
index 0000000..3eada38
--- /dev/null
@@ -0,0 +1,92 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef SL_PP_CONTEXT_H
+#define SL_PP_CONTEXT_H
+
+#include "sl_pp_dict.h"
+#include "sl_pp_macro.h"
+#include "sl_pp_process.h"
+#include "sl_pp_purify.h"
+#include "sl_pp_token_util.h"
+
+
+#define SL_PP_MAX_IF_NESTING  64
+
+#define SL_PP_MAX_ERROR_MSG   1024
+
+#define SL_PP_MAX_EXTENSIONS  16
+
+#define SL_PP_MAX_PREDEFINED  16
+
+struct sl_pp_extension {
+   int name;         /*< VENDOR_extension_name */
+   int name_string;  /*< GL_VENDOR_extension_name */
+};
+
+struct sl_pp_predefined {
+   int name;
+   int value;
+};
+
+struct sl_pp_context {
+   char *cstr_pool;
+   unsigned int cstr_pool_max;
+   unsigned int cstr_pool_len;
+   struct sl_pp_dict dict;
+
+   struct sl_pp_macro *macro;
+   struct sl_pp_macro **macro_tail;
+
+   struct sl_pp_extension extensions[SL_PP_MAX_EXTENSIONS];
+   unsigned int num_extensions;
+
+   struct sl_pp_predefined predefined[SL_PP_MAX_PREDEFINED];
+   unsigned int num_predefined;
+
+   unsigned int if_stack[SL_PP_MAX_IF_NESTING];
+   unsigned int if_ptr;
+   unsigned int if_value;
+
+   char error_msg[SL_PP_MAX_ERROR_MSG];
+   unsigned int error_line;
+
+   unsigned int line;
+   unsigned int file;
+
+   struct sl_pp_purify_state pure;
+
+   char *getc_buf;
+   unsigned int getc_buf_size;
+   unsigned int getc_buf_capacity;
+
+   struct sl_pp_token_buffer tokens;
+
+   struct sl_pp_process_state process_state;
+};
+
+#endif /* SL_PP_CONTEXT_H */
diff --git a/src/glsl/pp/sl_pp_define.c b/src/glsl/pp/sl_pp_define.c
new file mode 100644 (file)
index 0000000..808a6a0
--- /dev/null
@@ -0,0 +1,238 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_context.h"
+#include "sl_pp_process.h"
+#include "sl_pp_public.h"
+
+
+static void
+skip_whitespace(const struct sl_pp_token_info *input,
+                unsigned int *first,
+                unsigned int last)
+{
+   while (*first < last && input[*first].token == SL_PP_WHITESPACE) {
+      (*first)++;
+   }
+}
+
+
+static int
+_parse_formal_args(struct sl_pp_context *context,
+                   const struct sl_pp_token_info *input,
+                   unsigned int *first,
+                   unsigned int last,
+                   struct sl_pp_macro *macro)
+{
+   struct sl_pp_macro_formal_arg **arg;
+
+   macro->num_args = 0;
+
+   skip_whitespace(input, first, last);
+   if (*first < last) {
+      if (input[*first].token == SL_PP_RPAREN) {
+         (*first)++;
+         return 0;
+      }
+   } else {
+      strcpy(context->error_msg, "expected either macro formal argument or `)'");
+      return -1;
+   }
+
+   arg = &macro->arg;
+
+   for (;;) {
+      if (*first < last && input[*first].token != SL_PP_IDENTIFIER) {
+         strcpy(context->error_msg, "expected macro formal argument");
+         return -1;
+      }
+
+      *arg = malloc(sizeof(struct sl_pp_macro_formal_arg));
+      if (!*arg) {
+         strcpy(context->error_msg, "out of memory");
+         return -1;
+      }
+
+      (**arg).name = input[*first].data.identifier;
+      (*first)++;
+
+      (**arg).next = NULL;
+      arg = &(**arg).next;
+
+      macro->num_args++;
+
+      skip_whitespace(input, first, last);
+      if (*first < last) {
+         if (input[*first].token == SL_PP_COMMA) {
+            (*first)++;
+            skip_whitespace(input, first, last);
+         } else if (input[*first].token == SL_PP_RPAREN) {
+            (*first)++;
+            return 0;
+         } else {
+            strcpy(context->error_msg, "expected either `,' or `)'");
+            return -1;
+         }
+      } else {
+         strcpy(context->error_msg, "expected either `,' or `)'");
+         return -1;
+      }
+   }
+
+   /* Should not gete here. */
+}
+
+
+int
+sl_pp_process_define(struct sl_pp_context *context,
+                     const struct sl_pp_token_info *input,
+                     unsigned int first,
+                     unsigned int last)
+{
+   int macro_name = -1;
+   struct sl_pp_macro *macro;
+   unsigned int i;
+   unsigned int body_len;
+   unsigned int j;
+
+   if (first < last && input[first].token == SL_PP_IDENTIFIER) {
+      macro_name = input[first].data.identifier;
+      first++;
+   }
+   if (macro_name == -1) {
+      strcpy(context->error_msg, "expected macro name");
+      return -1;
+   }
+
+   /* Check for reserved macro names */
+   {
+      const char *name = sl_pp_context_cstr(context, macro_name);
+
+      if (strstr(name, "__")) {
+         strcpy(context->error_msg, "macro names containing `__' are reserved");
+         return 1;
+      }
+      if (name[0] == 'G' && name[1] == 'L' && name[2] == '_') {
+         strcpy(context->error_msg, "macro names prefixed with `GL_' are reserved");
+         return 1;
+      }
+   }
+
+   for (macro = context->macro; macro; macro = macro->next) {
+      if (macro->name == macro_name) {
+         break;
+      }
+   }
+
+   if (!macro) {
+      macro = sl_pp_macro_new();
+      if (!macro) {
+         strcpy(context->error_msg, "out of memory");
+         return -1;
+      }
+
+      *context->macro_tail = macro;
+      context->macro_tail = &macro->next;
+   } else {
+      sl_pp_macro_reset(macro);
+   }
+
+   macro->name = macro_name;
+
+   /*
+    * If there is no whitespace between macro name and left paren, a macro
+    * formal argument list follows. This is the only place where the presence
+    * of a whitespace matters and it's the only reason why we are dealing
+    * with whitespace at this level.
+    */
+   if (first < last && input[first].token == SL_PP_LPAREN) {
+      first++;
+      if (_parse_formal_args(context, input, &first, last, macro)) {
+         return -1;
+      }
+   }
+
+   /* Calculate body size, trim out whitespace, make room for EOF. */
+   body_len = 1;
+   for (i = first; i < last; i++) {
+      if (input[i].token != SL_PP_WHITESPACE) {
+         body_len++;
+      }
+   }
+
+   macro->body = malloc(sizeof(struct sl_pp_token_info) * body_len);
+   if (!macro->body) {
+      strcpy(context->error_msg, "out of memory");
+      return -1;
+   }
+
+   for (j = 0, i = first; i < last; i++) {
+      if (input[i].token != SL_PP_WHITESPACE) {
+         macro->body[j++] = input[i];
+      }
+   }
+   macro->body[j++].token = SL_PP_EOF;
+
+   return 0;
+}
+
+
+int
+sl_pp_process_undef(struct sl_pp_context *context,
+                    const struct sl_pp_token_info *input,
+                    unsigned int first,
+                    unsigned int last)
+{
+   int macro_name = -1;
+   struct sl_pp_macro **pmacro;
+   struct sl_pp_macro *macro;
+
+   if (first < last && input[first].token == SL_PP_IDENTIFIER) {
+      macro_name = input[first].data.identifier;
+   }
+   if (macro_name == -1) {
+      return 0;
+   }
+
+   for (pmacro = &context->macro; *pmacro; pmacro = &(**pmacro).next) {
+      if ((**pmacro).name == macro_name) {
+         break;
+      }
+   }
+   if (!*pmacro) {
+      return 0;
+   }
+
+   macro = *pmacro;
+   *pmacro = macro->next;
+   macro->next = NULL;
+   sl_pp_macro_free(macro);
+
+   return 0;
+}
diff --git a/src/glsl/pp/sl_pp_dict.c b/src/glsl/pp/sl_pp_dict.c
new file mode 100644 (file)
index 0000000..062139e
--- /dev/null
@@ -0,0 +1,85 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include "sl_pp_public.h"
+#include "sl_pp_context.h"
+#include "sl_pp_dict.h"
+
+
+#define ADD_NAME_STR(CTX, NAME, STR)\
+   do {\
+      (CTX)->dict.NAME = sl_pp_context_add_unique_str((CTX), (STR));\
+      if ((CTX)->dict.NAME == -1) {\
+         return -1;\
+      }\
+   } while (0)
+
+#define ADD_NAME(CTX, NAME) ADD_NAME_STR(CTX, NAME, #NAME)
+
+
+int
+sl_pp_dict_init(struct sl_pp_context *context)
+{
+   ADD_NAME(context, all);
+
+   ADD_NAME(context, require);
+   ADD_NAME(context, enable);
+   ADD_NAME(context, warn);
+   ADD_NAME(context, disable);
+
+   ADD_NAME(context, defined);
+
+   ADD_NAME_STR(context, ___LINE__, "__LINE__");
+   ADD_NAME_STR(context, ___FILE__, "__FILE__");
+   ADD_NAME_STR(context, ___VERSION__, "__VERSION__");
+
+   ADD_NAME(context, optimize);
+   ADD_NAME(context, debug);
+
+   ADD_NAME(context, off);
+   ADD_NAME(context, on);
+
+   ADD_NAME(context, define);
+   ADD_NAME(context, elif);
+   ADD_NAME_STR(context, _else, "else");
+   ADD_NAME(context, endif);
+   ADD_NAME(context, error);
+   ADD_NAME(context, extension);
+   ADD_NAME_STR(context, _if, "if");
+   ADD_NAME(context, ifdef);
+   ADD_NAME(context, ifndef);
+   ADD_NAME(context, line);
+   ADD_NAME(context, pragma);
+   ADD_NAME(context, undef);
+
+   ADD_NAME(context, version);
+
+   ADD_NAME_STR(context, _0, "0");
+   ADD_NAME_STR(context, _1, "1");
+
+   return 0;
+}
diff --git a/src/glsl/pp/sl_pp_dict.h b/src/glsl/pp/sl_pp_dict.h
new file mode 100644 (file)
index 0000000..875217b
--- /dev/null
@@ -0,0 +1,77 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef SL_PP_DICT_H
+#define SL_PP_DICT_H
+
+
+struct sl_pp_context;
+
+struct sl_pp_dict {
+   int all;
+
+   int require;
+   int enable;
+   int warn;
+   int disable;
+
+   int defined;
+
+   int ___LINE__;
+   int ___FILE__;
+   int ___VERSION__;
+
+   int optimize;
+   int debug;
+
+   int off;
+   int on;
+
+   int define;
+   int elif;
+   int _else;
+   int endif;
+   int error;
+   int extension;
+   int _if;
+   int ifdef;
+   int ifndef;
+   int line;
+   int pragma;
+   int undef;
+
+   int version;
+
+   int _0;
+   int _1;
+};
+
+
+int
+sl_pp_dict_init(struct sl_pp_context *context);
+
+#endif /* SL_PP_DICT_H */
diff --git a/src/glsl/pp/sl_pp_error.c b/src/glsl/pp/sl_pp_error.c
new file mode 100644 (file)
index 0000000..b628e37
--- /dev/null
@@ -0,0 +1,270 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_context.h"
+#include "sl_pp_process.h"
+#include "sl_pp_public.h"
+
+
+void
+sl_pp_process_error(struct sl_pp_context *context,
+                    const struct sl_pp_token_info *input,
+                    unsigned int first,
+                    unsigned int last)
+{
+   unsigned int out_len = 0;
+   unsigned int i;
+
+   for (i = first; i < last; i++) {
+      const char *s = NULL;
+      char buf[2];
+
+      switch (input[i].token) {
+      case SL_PP_WHITESPACE:
+         s = " ";
+         break;
+
+      case SL_PP_NEWLINE:
+         s = "\n";
+         break;
+
+      case SL_PP_HASH:
+         s = "#";
+         break;
+
+      case SL_PP_COMMA:
+         s = ",";
+         break;
+
+      case SL_PP_SEMICOLON:
+         s = ";";
+         break;
+
+      case SL_PP_LBRACE:
+         s = "{";
+         break;
+
+      case SL_PP_RBRACE:
+         s = "}";
+         break;
+
+      case SL_PP_LPAREN:
+         s = "(";
+         break;
+
+      case SL_PP_RPAREN:
+         s = ")";
+         break;
+
+      case SL_PP_LBRACKET:
+         s = "[";
+         break;
+
+      case SL_PP_RBRACKET:
+         s = "]";
+         break;
+
+      case SL_PP_DOT:
+         s = ".";
+         break;
+
+      case SL_PP_INCREMENT:
+         s = "++";
+         break;
+
+      case SL_PP_ADDASSIGN:
+         s = "+=";
+         break;
+
+      case SL_PP_PLUS:
+         s = "+";
+         break;
+
+      case SL_PP_DECREMENT:
+         s = "--";
+         break;
+
+      case SL_PP_SUBASSIGN:
+         s = "-=";
+         break;
+
+      case SL_PP_MINUS:
+         s = "-";
+         break;
+
+      case SL_PP_BITNOT:
+         s = "~";
+         break;
+
+      case SL_PP_NOTEQUAL:
+         s = "!=";
+         break;
+
+      case SL_PP_NOT:
+         s = "!";
+         break;
+
+      case SL_PP_MULASSIGN:
+         s = "*=";
+         break;
+
+      case SL_PP_STAR:
+         s = "*";
+         break;
+
+      case SL_PP_DIVASSIGN:
+         s = "/=";
+         break;
+
+      case SL_PP_SLASH:
+         s = "/";
+         break;
+
+      case SL_PP_MODASSIGN:
+         s = "%=";
+         break;
+
+      case SL_PP_MODULO:
+         s = "%";
+         break;
+
+      case SL_PP_LSHIFTASSIGN:
+         s = "<<=";
+         break;
+
+      case SL_PP_LSHIFT:
+         s = "<<";
+         break;
+
+      case SL_PP_LESSEQUAL:
+         s = "<=";
+         break;
+
+      case SL_PP_LESS:
+         s = "<";
+         break;
+
+      case SL_PP_RSHIFTASSIGN:
+         s = ">>=";
+         break;
+
+      case SL_PP_RSHIFT:
+         s = ">>";
+         break;
+
+      case SL_PP_GREATEREQUAL:
+         s = ">=";
+         break;
+
+      case SL_PP_GREATER:
+         s = ">";
+         break;
+
+      case SL_PP_EQUAL:
+         s = "==";
+         break;
+
+      case SL_PP_ASSIGN:
+         s = "=";
+         break;
+
+      case SL_PP_AND:
+         s = "&&";
+         break;
+
+      case SL_PP_BITANDASSIGN:
+         s = "&=";
+         break;
+
+      case SL_PP_BITAND:
+         s = "&";
+         break;
+
+      case SL_PP_XOR:
+         s = "^^";
+         break;
+
+      case SL_PP_BITXORASSIGN:
+         s = "^=";
+         break;
+
+      case SL_PP_BITXOR:
+         s = "^";
+         break;
+
+      case SL_PP_OR:
+         s = "||";
+         break;
+
+      case SL_PP_BITORASSIGN:
+         s = "|=";
+         break;
+
+      case SL_PP_BITOR:
+         s = "|";
+         break;
+
+      case SL_PP_QUESTION:
+         s = "?";
+         break;
+
+      case SL_PP_COLON:
+         s = ":";
+         break;
+
+      case SL_PP_IDENTIFIER:
+         s = sl_pp_context_cstr(context, input[i].data.identifier);
+         break;
+
+      case SL_PP_UINT:
+         s = sl_pp_context_cstr(context, input[i].data._uint);
+         break;
+
+      case SL_PP_FLOAT:
+         s = sl_pp_context_cstr(context, input[i].data._float);
+         break;
+
+      case SL_PP_OTHER:
+         buf[0] = input[i].data.other;
+         buf[1] = '\0';
+         s = buf;
+         break;
+
+      default:
+         strcpy(context->error_msg, "internal error");
+         return;
+      }
+
+      while (*s != '\0' && out_len < sizeof(context->error_msg) - 1) {
+         context->error_msg[out_len++] = *s++;
+      }
+   }
+
+   context->error_msg[out_len] = '\0';
+}
diff --git a/src/glsl/pp/sl_pp_expression.c b/src/glsl/pp/sl_pp_expression.c
new file mode 100644 (file)
index 0000000..ec90478
--- /dev/null
@@ -0,0 +1,411 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_expression.h"
+#include "sl_pp_public.h"
+
+
+struct parse_context {
+   struct sl_pp_context *context;
+   const struct sl_pp_token_info *input;
+};
+
+static int
+_parse_or(struct parse_context *ctx,
+          int *result);
+
+static int
+_parse_primary(struct parse_context *ctx,
+               int *result)
+{
+   if (ctx->input->token == SL_PP_UINT) {
+      *result = atoi(sl_pp_context_cstr(ctx->context, ctx->input->data._uint));
+      ctx->input++;
+   } else {
+      if (ctx->input->token != SL_PP_LPAREN) {
+         strcpy(ctx->context->error_msg, "expected `('");
+         return -1;
+      }
+      ctx->input++;
+      if (_parse_or(ctx, result)) {
+         return -1;
+      }
+      if (ctx->input->token != SL_PP_RPAREN) {
+         strcpy(ctx->context->error_msg, "expected `)'");
+         return -1;
+      }
+      ctx->input++;
+   }
+   return 0;
+}
+
+static int
+_parse_unary(struct parse_context *ctx,
+             int *result)
+{
+   if (!_parse_primary(ctx, result)) {
+      return 0;
+   }
+
+   switch (ctx->input->token) {
+   case SL_PP_PLUS:
+      ctx->input++;
+      if (_parse_unary(ctx, result)) {
+         return -1;
+      }
+      *result = +*result;
+      break;
+
+   case SL_PP_MINUS:
+      ctx->input++;
+      if (_parse_unary(ctx, result)) {
+         return -1;
+      }
+      *result = -*result;
+      break;
+
+   case SL_PP_NOT:
+      ctx->input++;
+      if (_parse_unary(ctx, result)) {
+         return -1;
+      }
+      *result = !*result;
+      break;
+
+   case SL_PP_BITNOT:
+      ctx->input++;
+      if (_parse_unary(ctx, result)) {
+         return -1;
+      }
+      *result = ~*result;
+      break;
+
+   default:
+      return -1;
+   }
+
+   return 0;
+}
+
+static int
+_parse_multiplicative(struct parse_context *ctx,
+                      int *result)
+{
+   if (_parse_unary(ctx, result)) {
+      return -1;
+   }
+   for (;;) {
+      int right;
+
+      switch (ctx->input->token) {
+      case SL_PP_STAR:
+         ctx->input++;
+         if (_parse_unary(ctx, &right)) {
+            return -1;
+         }
+         *result = *result * right;
+         break;
+
+      case SL_PP_SLASH:
+         ctx->input++;
+         if (_parse_unary(ctx, &right)) {
+            return -1;
+         }
+         *result = *result / right;
+         break;
+
+      case SL_PP_MODULO:
+         ctx->input++;
+         if (_parse_unary(ctx, &right)) {
+            return -1;
+         }
+         *result = *result % right;
+         break;
+
+      default:
+         return 0;
+      }
+   }
+}
+
+static int
+_parse_additive(struct parse_context *ctx,
+                int *result)
+{
+   if (_parse_multiplicative(ctx, result)) {
+      return -1;
+   }
+   for (;;) {
+      int right;
+
+      switch (ctx->input->token) {
+      case SL_PP_PLUS:
+         ctx->input++;
+         if (_parse_multiplicative(ctx, &right)) {
+            return -1;
+         }
+         *result = *result + right;
+         break;
+
+      case SL_PP_MINUS:
+         ctx->input++;
+         if (_parse_multiplicative(ctx, &right)) {
+            return -1;
+         }
+         *result = *result - right;
+         break;
+
+      default:
+         return 0;
+      }
+   }
+}
+
+static int
+_parse_shift(struct parse_context *ctx,
+             int *result)
+{
+   if (_parse_additive(ctx, result)) {
+      return -1;
+   }
+   for (;;) {
+      int right;
+
+      switch (ctx->input->token) {
+      case SL_PP_LSHIFT:
+         ctx->input++;
+         if (_parse_additive(ctx, &right)) {
+            return -1;
+         }
+         *result = *result << right;
+         break;
+
+      case SL_PP_RSHIFT:
+         ctx->input++;
+         if (_parse_additive(ctx, &right)) {
+            return -1;
+         }
+         *result = *result >> right;
+         break;
+
+      default:
+         return 0;
+      }
+   }
+}
+
+static int
+_parse_relational(struct parse_context *ctx,
+                  int *result)
+{
+   if (_parse_shift(ctx, result)) {
+      return -1;
+   }
+   for (;;) {
+      int right;
+
+      switch (ctx->input->token) {
+      case SL_PP_LESSEQUAL:
+         ctx->input++;
+         if (_parse_shift(ctx, &right)) {
+            return -1;
+         }
+         *result = *result <= right;
+         break;
+
+      case SL_PP_GREATEREQUAL:
+         ctx->input++;
+         if (_parse_shift(ctx, &right)) {
+            return -1;
+         }
+         *result = *result >= right;
+         break;
+
+      case SL_PP_LESS:
+         ctx->input++;
+         if (_parse_shift(ctx, &right)) {
+            return -1;
+         }
+         *result = *result < right;
+         break;
+
+      case SL_PP_GREATER:
+         ctx->input++;
+         if (_parse_shift(ctx, &right)) {
+            return -1;
+         }
+         *result = *result > right;
+         break;
+
+      default:
+         return 0;
+      }
+   }
+}
+
+static int
+_parse_equality(struct parse_context *ctx,
+                int *result)
+{
+   if (_parse_relational(ctx, result)) {
+      return -1;
+   }
+   for (;;) {
+      int right;
+
+      switch (ctx->input->token) {
+      case SL_PP_EQUAL:
+         ctx->input++;
+         if (_parse_relational(ctx, &right)) {
+            return -1;
+         }
+         *result = *result == right;
+         break;
+
+      case SL_PP_NOTEQUAL:
+         ctx->input++;
+         if (_parse_relational(ctx, &right)) {
+            return -1;
+         }
+         *result = *result != right;
+         break;
+
+      default:
+         return 0;
+      }
+   }
+}
+
+static int
+_parse_bitand(struct parse_context *ctx,
+              int *result)
+{
+   if (_parse_equality(ctx, result)) {
+      return -1;
+   }
+   while (ctx->input->token == SL_PP_BITAND) {
+      int right;
+
+      ctx->input++;
+      if (_parse_equality(ctx, &right)) {
+         return -1;
+      }
+      *result = *result & right;
+   }
+   return 0;
+}
+
+static int
+_parse_xor(struct parse_context *ctx,
+           int *result)
+{
+   if (_parse_bitand(ctx, result)) {
+      return -1;
+   }
+   while (ctx->input->token == SL_PP_XOR) {
+      int right;
+
+      ctx->input++;
+      if (_parse_bitand(ctx, &right)) {
+         return -1;
+      }
+      *result = *result ^ right;
+   }
+   return 0;
+}
+
+static int
+_parse_bitor(struct parse_context *ctx,
+             int *result)
+{
+   if (_parse_xor(ctx, result)) {
+      return -1;
+   }
+   while (ctx->input->token == SL_PP_BITOR) {
+      int right;
+
+      ctx->input++;
+      if (_parse_xor(ctx, &right)) {
+         return -1;
+      }
+      *result = *result | right;
+   }
+   return 0;
+}
+
+static int
+_parse_and(struct parse_context *ctx,
+           int *result)
+{
+   if (_parse_bitor(ctx, result)) {
+      return -1;
+   }
+   while (ctx->input->token == SL_PP_AND) {
+      int right;
+
+      ctx->input++;
+      if (_parse_bitor(ctx, &right)) {
+         return -1;
+      }
+      *result = *result && right;
+   }
+   return 0;
+}
+
+static int
+_parse_or(struct parse_context *ctx,
+          int *result)
+{
+   if (_parse_and(ctx, result)) {
+      return -1;
+   }
+   while (ctx->input->token == SL_PP_OR) {
+      int right;
+
+      ctx->input++;
+      if (_parse_and(ctx, &right)) {
+         return -1;
+      }
+      *result = *result || right;
+   }
+   return 0;
+}
+
+int
+sl_pp_execute_expression(struct sl_pp_context *context,
+                         const struct sl_pp_token_info *input,
+                         int *result)
+{
+   struct parse_context ctx;
+
+   ctx.context = context;
+   ctx.input = input;
+
+   return _parse_or(&ctx, result);
+}
diff --git a/src/glsl/pp/sl_pp_expression.h b/src/glsl/pp/sl_pp_expression.h
new file mode 100644 (file)
index 0000000..377d5b4
--- /dev/null
@@ -0,0 +1,40 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef SL_PP_EXPRESSION_H
+#define SL_PP_EXPRESSION_H
+
+#include "sl_pp_context.h"
+#include "sl_pp_token.h"
+
+
+int
+sl_pp_execute_expression(struct sl_pp_context *context,
+                         const struct sl_pp_token_info *input,
+                         int *result);
+
+#endif /* SL_PP_EXPRESSION_H */
diff --git a/src/glsl/pp/sl_pp_extension.c b/src/glsl/pp/sl_pp_extension.c
new file mode 100644 (file)
index 0000000..8af5731
--- /dev/null
@@ -0,0 +1,171 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_context.h"
+#include "sl_pp_process.h"
+#include "sl_pp_public.h"
+
+
+int
+sl_pp_context_add_extension(struct sl_pp_context *context,
+                            const char *name,
+                            const char *name_string)
+{
+   struct sl_pp_extension ext;
+
+   if (context->num_extensions == SL_PP_MAX_EXTENSIONS) {
+      return -1;
+   }
+
+   ext.name = sl_pp_context_add_unique_str(context, name);
+   if (ext.name == -1) {
+      return -1;
+   }
+
+   ext.name_string = sl_pp_context_add_unique_str(context, name_string);
+   if (ext.name_string == -1) {
+      return -1;
+   }
+
+   context->extensions[context->num_extensions++] = ext;
+   return 0;
+}
+
+int
+sl_pp_process_extension(struct sl_pp_context *context,
+                        const struct sl_pp_token_info *input,
+                        unsigned int first,
+                        unsigned int last,
+                        struct sl_pp_process_state *state)
+{
+   int extension_name = -1;
+   int behavior = -1;
+   struct sl_pp_token_info out;
+
+   /* Grab the extension name. */
+   if (first < last && input[first].token == SL_PP_IDENTIFIER) {
+      extension_name = input[first].data.identifier;
+      first++;
+   }
+   if (extension_name == -1) {
+      strcpy(context->error_msg, "expected identifier after `#extension'");
+      return -1;
+   }
+
+   /* Make sure the extension is supported. */
+   if (extension_name == context->dict.all) {
+      out.data.extension = extension_name;
+   } else {
+      unsigned int i;
+
+      out.data.extension = -1;
+      for (i = 0; i < context->num_extensions; i++) {
+         if (extension_name == context->extensions[i].name_string) {
+            out.data.extension = extension_name;
+            break;
+         }
+      }
+   }
+
+   /* Grab the colon separating the extension name and behavior. */
+   while (first < last && input[first].token == SL_PP_WHITESPACE) {
+      first++;
+   }
+   if (first < last && input[first].token == SL_PP_COLON) {
+      first++;
+   } else {
+      strcpy(context->error_msg, "expected `:' after extension name");
+      return -1;
+   }
+   while (first < last && input[first].token == SL_PP_WHITESPACE) {
+      first++;
+   }
+
+   /* Grab the behavior name. */
+   if (first < last && input[first].token == SL_PP_IDENTIFIER) {
+      behavior = input[first].data.identifier;
+      first++;
+   }
+   if (behavior == -1) {
+      strcpy(context->error_msg, "expected identifier after `:'");
+      return -1;
+   }
+
+   if (behavior == context->dict.require) {
+      if (out.data.extension == -1) {
+         strcpy(context->error_msg, "the required extension is not supported");
+         return -1;
+      }
+      if (out.data.extension == context->dict.all) {
+         strcpy(context->error_msg, "invalid behavior for `all' extension: `require'");
+         return -1;
+      }
+      out.token = SL_PP_EXTENSION_REQUIRE;
+   } else if (behavior == context->dict.enable) {
+      if (out.data.extension == -1) {
+         /* Warning: the extension cannot be enabled. */
+         return 0;
+      }
+      if (out.data.extension == context->dict.all) {
+         strcpy(context->error_msg, "invalid behavior for `all' extension: `enable'");
+         return -1;
+      }
+      out.token = SL_PP_EXTENSION_ENABLE;
+   } else if (behavior == context->dict.warn) {
+      if (out.data.extension == -1) {
+         /* Warning: the extension is not supported. */
+         return 0;
+      }
+      out.token = SL_PP_EXTENSION_WARN;
+   } else if (behavior == context->dict.disable) {
+      if (out.data.extension == -1) {
+         /* Warning: the extension is not supported. */
+         return 0;
+      }
+      out.token = SL_PP_EXTENSION_DISABLE;
+   } else {
+      strcpy(context->error_msg, "unrecognised behavior name");
+      return -1;
+   }
+
+   /* Grab the end of line. */
+   while (first < last && input[first].token == SL_PP_WHITESPACE) {
+      first++;
+   }
+   if (first < last) {
+      strcpy(context->error_msg, "expected end of line after behavior name");
+      return -1;
+   }
+
+   if (sl_pp_process_out(state, &out)) {
+      return -1;
+   }
+
+   return 0;
+}
diff --git a/src/glsl/pp/sl_pp_if.c b/src/glsl/pp/sl_pp_if.c
new file mode 100644 (file)
index 0000000..f12f0f1
--- /dev/null
@@ -0,0 +1,344 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_expression.h"
+#include "sl_pp_process.h"
+
+
+static int
+_parse_defined(struct sl_pp_context *context,
+               struct sl_pp_token_buffer *buffer,
+               struct sl_pp_process_state *state)
+{
+   struct sl_pp_token_info input;
+   int parens = 0;
+   int macro_name;
+   struct sl_pp_macro *macro;
+   int defined = 0;
+   struct sl_pp_token_info result;
+
+   if (sl_pp_token_buffer_skip_white(buffer, &input)) {
+      return -1;
+   }
+
+   if (input.token == SL_PP_LPAREN) {
+      if (sl_pp_token_buffer_skip_white(buffer, &input)) {
+         return -1;
+      }
+      parens = 1;
+   }
+
+   if (input.token != SL_PP_IDENTIFIER) {
+      strcpy(context->error_msg, "expected an identifier");
+      return -1;
+   }
+
+   macro_name = input.data.identifier;
+   for (macro = context->macro; macro; macro = macro->next) {
+      if (macro->name == macro_name) {
+         defined = 1;
+         break;
+      }
+   }
+
+   if (parens) {
+      if (sl_pp_token_buffer_skip_white(buffer, &input)) {
+         return -1;
+      }
+      if (input.token != SL_PP_RPAREN) {
+         strcpy(context->error_msg, "expected `)'");
+         return -1;
+      }
+   }
+
+   result.token = SL_PP_UINT;
+   result.data._uint = (defined ? context->dict._1 : context->dict._0);
+
+   if (sl_pp_process_out(state, &result)) {
+      strcpy(context->error_msg, "out of memory");
+      return -1;
+   }
+
+   return 0;
+}
+
+static unsigned int
+_evaluate_if_stack(struct sl_pp_context *context)
+{
+   unsigned int i;
+
+   for (i = context->if_ptr; i < SL_PP_MAX_IF_NESTING; i++) {
+      if (!(context->if_stack[i] & 1)) {
+         return 0;
+      }
+   }
+   return 1;
+}
+
+static int
+_parse_if(struct sl_pp_context *context,
+          struct sl_pp_token_buffer *buffer)
+{
+   struct sl_pp_process_state state;
+   int found_end = 0;
+   struct sl_pp_token_info eof;
+   int result;
+
+   if (!context->if_ptr) {
+      strcpy(context->error_msg, "`#if' nesting too deep");
+      return -1;
+   }
+
+   memset(&state, 0, sizeof(state));
+   while (!found_end) {
+      struct sl_pp_token_info input;
+
+      sl_pp_token_buffer_get(buffer, &input);
+      switch (input.token) {
+      case SL_PP_WHITESPACE:
+         break;
+
+      case SL_PP_IDENTIFIER:
+         if (input.data.identifier == context->dict.defined) {
+            if (_parse_defined(context, buffer, &state)) {
+               free(state.out);
+               return -1;
+            }
+         } else {
+            sl_pp_token_buffer_unget(buffer, &input);
+            if (sl_pp_macro_expand(context, buffer, NULL, &state, sl_pp_macro_expand_unknown_to_0)) {
+               free(state.out);
+               return -1;
+            }
+         }
+         break;
+
+      case SL_PP_NEWLINE:
+      case SL_PP_EOF:
+         found_end = 1;
+         break;
+
+      default:
+         if (sl_pp_process_out(&state, &input)) {
+            strcpy(context->error_msg, "out of memory");
+            free(state.out);
+            return -1;
+         }
+      }
+   }
+
+   eof.token = SL_PP_EOF;
+   if (sl_pp_process_out(&state, &eof)) {
+      strcpy(context->error_msg, "out of memory");
+      free(state.out);
+      return -1;
+   }
+
+   if (sl_pp_execute_expression(context, state.out, &result)) {
+      free(state.out);
+      return -1;
+   }
+
+   free(state.out);
+
+   context->if_ptr--;
+   context->if_stack[context->if_ptr] = result ? 1 : 0;
+   context->if_value = _evaluate_if_stack(context);
+
+   return 0;
+}
+
+static int
+_parse_else(struct sl_pp_context *context)
+{
+   if (context->if_ptr == SL_PP_MAX_IF_NESTING) {
+      strcpy(context->error_msg, "no matching `#if'");
+      return -1;
+   }
+
+   /* Bit b1 indicates we already went through #else. */
+   if (context->if_stack[context->if_ptr] & 2) {
+      strcpy(context->error_msg, "no matching `#if'");
+      return -1;
+   }
+
+   /* Invert current condition value and mark that we are in the #else block. */
+   context->if_stack[context->if_ptr] = (1 - (context->if_stack[context->if_ptr] & 1)) | 2;
+   context->if_value = _evaluate_if_stack(context);
+
+   return 0;
+}
+
+int
+sl_pp_process_if(struct sl_pp_context *context,
+                 struct sl_pp_token_buffer *buffer)
+{
+   return _parse_if(context, buffer);
+}
+
+int
+sl_pp_process_ifdef(struct sl_pp_context *context,
+                    const struct sl_pp_token_info *input,
+                    unsigned int first,
+                    unsigned int last)
+{
+   unsigned int i;
+
+   if (!context->if_ptr) {
+      strcpy(context->error_msg, "`#if' nesting too deep");
+      return -1;
+   }
+
+   for (i = first; i < last; i++) {
+      switch (input[i].token) {
+      case SL_PP_IDENTIFIER:
+         {
+            struct sl_pp_macro *macro;
+            int macro_name = input[i].data.identifier;
+            int defined = 0;
+
+            for (macro = context->macro; macro; macro = macro->next) {
+               if (macro->name == macro_name) {
+                  defined = 1;
+                  break;
+               }
+            }
+
+            context->if_ptr--;
+            context->if_stack[context->if_ptr] = defined ? 1 : 0;
+            context->if_value = _evaluate_if_stack(context);
+         }
+         return 0;
+
+      case SL_PP_WHITESPACE:
+         break;
+
+      default:
+         strcpy(context->error_msg, "expected an identifier");
+         return -1;
+      }
+   }
+
+   strcpy(context->error_msg, "expected an identifier");
+   return -1;
+}
+
+int
+sl_pp_process_ifndef(struct sl_pp_context *context,
+                     const struct sl_pp_token_info *input,
+                     unsigned int first,
+                     unsigned int last)
+{
+   unsigned int i;
+
+   if (!context->if_ptr) {
+      strcpy(context->error_msg, "`#if' nesting too deep");
+      return -1;
+   }
+
+   for (i = first; i < last; i++) {
+      switch (input[i].token) {
+      case SL_PP_IDENTIFIER:
+         {
+            struct sl_pp_macro *macro;
+            int macro_name = input[i].data.identifier;
+            int defined = 0;
+
+            for (macro = context->macro; macro; macro = macro->next) {
+               if (macro->name == macro_name) {
+                  defined = 1;
+                  break;
+               }
+            }
+
+            context->if_ptr--;
+            context->if_stack[context->if_ptr] = defined ? 0 : 1;
+            context->if_value = _evaluate_if_stack(context);
+         }
+         return 0;
+
+      case SL_PP_WHITESPACE:
+         break;
+
+      default:
+         strcpy(context->error_msg, "expected an identifier");
+         return -1;
+      }
+   }
+
+   strcpy(context->error_msg, "expected an identifier");
+   return -1;
+}
+
+int
+sl_pp_process_elif(struct sl_pp_context *context,
+                   struct sl_pp_token_buffer *buffer)
+{
+   if (_parse_else(context)) {
+      return -1;
+   }
+
+   if (context->if_stack[context->if_ptr] & 1) {
+      context->if_ptr++;
+      if (_parse_if(context, buffer)) {
+         return -1;
+      }
+   }
+
+   /* We are still in the #if block. */
+   context->if_stack[context->if_ptr] = context->if_stack[context->if_ptr] & ~2;
+
+   return 0;
+}
+
+int
+sl_pp_process_else(struct sl_pp_context *context,
+                   const struct sl_pp_token_info *input,
+                   unsigned int first,
+                   unsigned int last)
+{
+   return _parse_else(context);
+}
+
+int
+sl_pp_process_endif(struct sl_pp_context *context,
+                    const struct sl_pp_token_info *input,
+                    unsigned int first,
+                    unsigned int last)
+{
+   if (context->if_ptr == SL_PP_MAX_IF_NESTING) {
+      strcpy(context->error_msg, "no matching `#if'");
+      return -1;
+   }
+
+   context->if_ptr++;
+   context->if_value = _evaluate_if_stack(context);
+
+   return 0;
+}
diff --git a/src/glsl/pp/sl_pp_line.c b/src/glsl/pp/sl_pp_line.c
new file mode 100644 (file)
index 0000000..6f7e9eb
--- /dev/null
@@ -0,0 +1,127 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_context.h"
+#include "sl_pp_public.h"
+#include "sl_pp_process.h"
+
+
+int
+sl_pp_process_line(struct sl_pp_context *context,
+                   struct sl_pp_token_buffer *buffer,
+                   struct sl_pp_process_state *pstate)
+{
+   struct sl_pp_process_state state;
+   int found_end = 0;
+   int line_number = -1;
+   int file_number = -1;
+   unsigned int line;
+   unsigned int file;
+
+   memset(&state, 0, sizeof(state));
+   while (!found_end) {
+      struct sl_pp_token_info input;
+
+      sl_pp_token_buffer_get(buffer, &input);
+      switch (input.token) {
+      case SL_PP_WHITESPACE:
+         break;
+
+      case SL_PP_IDENTIFIER:
+         sl_pp_token_buffer_unget(buffer, &input);
+         if (sl_pp_macro_expand(context, buffer, NULL, &state, sl_pp_macro_expand_normal)) {
+            free(state.out);
+            return -1;
+         }
+         break;
+
+      case SL_PP_NEWLINE:
+      case SL_PP_EOF:
+         found_end = 1;
+         break;
+
+      default:
+         if (sl_pp_process_out(&state, &input)) {
+            strcpy(context->error_msg, "out of memory");
+            free(state.out);
+            return -1;
+         }
+      }
+   }
+
+   if (state.out_len > 0 && state.out[0].token == SL_PP_UINT) {
+      line_number = state.out[0].data._uint;
+   } else {
+      strcpy(context->error_msg, "expected a number after `#line'");
+      free(state.out);
+      return -1;
+   }
+
+   if (state.out_len > 1) {
+      if (state.out[1].token == SL_PP_UINT) {
+         file_number = state.out[1].data._uint;
+      } else {
+         strcpy(context->error_msg, "expected a number after line number");
+         free(state.out);
+         return -1;
+      }
+
+      if (state.out_len > 2) {
+         strcpy(context->error_msg, "expected an end of line after file number");
+         free(state.out);
+         return -1;
+      }
+   }
+
+   free(state.out);
+
+   line = atoi(sl_pp_context_cstr(context, line_number));
+   if (file_number != -1) {
+      file = atoi(sl_pp_context_cstr(context, file_number));
+   } else {
+      file = context->file;
+   }
+
+   if (context->line != line || context->file != file) {
+      struct sl_pp_token_info ti;
+
+      ti.token = SL_PP_LINE;
+      ti.data.line.lineno = line;
+      ti.data.line.fileno = file;
+      if (sl_pp_process_out(pstate, &ti)) {
+         strcpy(context->error_msg, "out of memory");
+         return -1;
+      }
+
+      context->line = line;
+      context->file = file;
+   }
+
+   return 0;
+}
diff --git a/src/glsl/pp/sl_pp_macro.c b/src/glsl/pp/sl_pp_macro.c
new file mode 100644 (file)
index 0000000..9f520b8
--- /dev/null
@@ -0,0 +1,414 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <stdio.h>
+#include <string.h>
+#include "sl_pp_context.h"
+#include "sl_pp_public.h"
+#include "sl_pp_macro.h"
+#include "sl_pp_process.h"
+
+
+static void
+_macro_init(struct sl_pp_macro *macro)
+{
+   macro->name = -1;
+   macro->num_args = -1;
+   macro->arg = NULL;
+   macro->body = NULL;
+}
+
+struct sl_pp_macro *
+sl_pp_macro_new(void)
+{
+   struct sl_pp_macro *macro;
+
+   macro = calloc(1, sizeof(struct sl_pp_macro));
+   if (macro) {
+      _macro_init(macro);
+   }
+   return macro;
+}
+
+static void
+_macro_destroy(struct sl_pp_macro *macro)
+{
+   struct sl_pp_macro_formal_arg *arg = macro->arg;
+
+   while (arg) {
+      struct sl_pp_macro_formal_arg *next_arg = arg->next;
+
+      free(arg);
+      arg = next_arg;
+   }
+
+   free(macro->body);
+}
+
+void
+sl_pp_macro_free(struct sl_pp_macro *macro)
+{
+   while (macro) {
+      struct sl_pp_macro *next_macro = macro->next;
+
+      _macro_destroy(macro);
+      free(macro);
+      macro = next_macro;
+   }
+}
+
+void
+sl_pp_macro_reset(struct sl_pp_macro *macro)
+{
+   _macro_destroy(macro);
+   _macro_init(macro);
+}
+
+static int
+_out_number(struct sl_pp_context *context,
+            struct sl_pp_process_state *state,
+            unsigned int number)
+{
+   char buf[32];
+   struct sl_pp_token_info ti;
+
+   sprintf(buf, "%u", number);
+
+   ti.token = SL_PP_UINT;
+   ti.data._uint = sl_pp_context_add_unique_str(context, buf);
+   if (sl_pp_process_out(state, &ti)) {
+      strcpy(context->error_msg, "out of memory");
+      return -1;
+   }
+
+   return 0;
+}
+
+int
+sl_pp_macro_expand(struct sl_pp_context *context,
+                   struct sl_pp_token_buffer *tokens,
+                   struct sl_pp_macro *local,
+                   struct sl_pp_process_state *state,
+                   enum sl_pp_macro_expand_behaviour behaviour)
+{
+   int mute = (behaviour == sl_pp_macro_expand_mute);
+   struct sl_pp_token_info input;
+   int macro_name;
+   struct sl_pp_macro *macro = NULL;
+   struct sl_pp_macro *actual_arg = NULL;
+   unsigned int j;
+
+   if (sl_pp_token_buffer_get(tokens, &input)) {
+      return -1;
+   }
+
+   if (input.token != SL_PP_IDENTIFIER) {
+      strcpy(context->error_msg, "expected an identifier");
+      return -1;
+   }
+
+   macro_name = input.data.identifier;
+
+   /* First look for predefined macros.
+    */
+
+   if (macro_name == context->dict.___LINE__) {
+      if (!mute && _out_number(context, state, context->line)) {
+         return -1;
+      }
+      return 0;
+   }
+   if (macro_name == context->dict.___FILE__) {
+      if (!mute && _out_number(context, state, context->file)) {
+         return -1;
+      }
+      return 0;
+   }
+   if (macro_name == context->dict.___VERSION__) {
+      if (!mute && _out_number(context, state, 110)) {
+         return -1;
+      }
+      return 0;
+   }
+
+   for (j = 0; j < context->num_predefined; j++) {
+      if (macro_name == context->predefined[j].name) {
+         if (!mute) {
+            struct sl_pp_token_info ti;
+
+            ti.token = SL_PP_UINT;
+            ti.data._uint = context->predefined[j].value;
+            if (sl_pp_process_out(state, &ti)) {
+               strcpy(context->error_msg, "out of memory");
+               return -1;
+            }
+         }
+         return 0;
+      }
+   }
+
+   /* Replace extension names with 1.
+    */
+   for (j = 0; j < context->num_extensions; j++) {
+      if (macro_name == context->extensions[j].name) {
+         if (!mute && _out_number(context, state, 1)) {
+            return -1;
+         }
+         return 0;
+      }
+   }
+
+   if (local) {
+      for (macro = local; macro; macro = macro->next) {
+         if (macro->name == macro_name) {
+            break;
+         }
+      }
+   }
+
+   if (!macro) {
+      for (macro = context->macro; macro; macro = macro->next) {
+         if (macro->name == macro_name) {
+            break;
+         }
+      }
+   }
+
+   if (!macro) {
+      if (behaviour == sl_pp_macro_expand_unknown_to_0) {
+         if (_out_number(context, state, 0)) {
+            strcpy(context->error_msg, "out of memory");
+            return -1;
+         }
+      } else if (!mute) {
+         if (sl_pp_process_out(state, &input)) {
+            strcpy(context->error_msg, "out of memory");
+            return -1;
+         }
+      }
+      return 0;
+   }
+
+   if (macro->num_args >= 0) {
+      if (sl_pp_token_buffer_skip_white(tokens, &input)) {
+         return -1;
+      }
+      if (input.token != SL_PP_LPAREN) {
+         strcpy(context->error_msg, "expected `('");
+         return -1;
+      }
+      if (sl_pp_token_buffer_skip_white(tokens, &input)) {
+         return -1;
+      }
+      sl_pp_token_buffer_unget(tokens, &input);
+   }
+
+   if (macro->num_args > 0) {
+      struct sl_pp_macro_formal_arg *formal_arg = macro->arg;
+      struct sl_pp_macro **pmacro = &actual_arg;
+
+      for (j = 0; j < (unsigned int)macro->num_args; j++) {
+         struct sl_pp_process_state arg_state;
+         int done = 0;
+         unsigned int paren_nesting = 0;
+         struct sl_pp_token_info eof;
+
+         memset(&arg_state, 0, sizeof(arg_state));
+
+         while (!done) {
+            if (sl_pp_token_buffer_get(tokens, &input)) {
+               goto fail_arg;
+            }
+            switch (input.token) {
+            case SL_PP_WHITESPACE:
+               break;
+
+            case SL_PP_COMMA:
+               if (!paren_nesting) {
+                  if (j < (unsigned int)macro->num_args - 1) {
+                     done = 1;
+                  } else {
+                     strcpy(context->error_msg, "too many actual macro arguments");
+                     goto fail_arg;
+                  }
+               } else {
+                  if (sl_pp_process_out(&arg_state, &input)) {
+                     strcpy(context->error_msg, "out of memory");
+                     goto fail_arg;
+                  }
+               }
+               break;
+
+            case SL_PP_LPAREN:
+               paren_nesting++;
+               if (sl_pp_process_out(&arg_state, &input)) {
+                  goto oom_arg;
+               }
+               break;
+
+            case SL_PP_RPAREN:
+               if (!paren_nesting) {
+                  if (j == (unsigned int)macro->num_args - 1) {
+                     done = 1;
+                  } else {
+                     strcpy(context->error_msg, "too few actual macro arguments");
+                     goto fail_arg;
+                  }
+               } else {
+                  paren_nesting--;
+                  if (sl_pp_process_out(&arg_state, &input)) {
+                     goto oom_arg;
+                  }
+               }
+               break;
+
+            case SL_PP_IDENTIFIER:
+               sl_pp_token_buffer_unget(tokens, &input);
+               if (sl_pp_macro_expand(context, tokens, local, &arg_state, sl_pp_macro_expand_normal)) {
+                  goto fail_arg;
+               }
+               break;
+
+            case SL_PP_EOF:
+               strcpy(context->error_msg, "too few actual macro arguments");
+               goto fail_arg;
+
+            default:
+               if (sl_pp_process_out(&arg_state, &input)) {
+                  goto oom_arg;
+               }
+            }
+         }
+
+         eof.token = SL_PP_EOF;
+         if (sl_pp_process_out(&arg_state, &eof)) {
+            goto oom_arg;
+         }
+
+         *pmacro = sl_pp_macro_new();
+         if (!*pmacro) {
+            goto oom_arg;
+         }
+
+         (**pmacro).name = formal_arg->name;
+         (**pmacro).body = arg_state.out;
+
+         formal_arg = formal_arg->next;
+         pmacro = &(**pmacro).next;
+
+         continue;
+
+oom_arg:
+         strcpy(context->error_msg, "out of memory");
+fail_arg:
+         free(arg_state.out);
+         goto fail;
+      }
+   }
+
+   /* Right paren for non-empty argument list has already been eaten. */
+   if (macro->num_args == 0) {
+      if (sl_pp_token_buffer_skip_white(tokens, &input)) {
+         goto fail;
+      }
+      if (input.token != SL_PP_RPAREN) {
+         strcpy(context->error_msg, "expected `)'");
+         goto fail;
+      }
+   }
+
+   /* XXX: This is all wrong, we should be ungetting all tokens
+    *      back to the main token buffer.
+    */
+   {
+      struct sl_pp_token_buffer buffer;
+
+      /* Seek to the end.
+       */
+      for (j = 0; macro->body[j].token != SL_PP_EOF; j++) {
+      }
+      j++;
+
+      /* Create a context-less token buffer since we are not going to underrun
+       * its internal buffer.
+       */
+      if (sl_pp_token_buffer_init(&buffer, NULL)) {
+         strcpy(context->error_msg, "out of memory");
+         goto fail;
+      }
+
+      /* Unget the tokens in reverse order so later they will be fetched correctly.
+       */
+      for (; j > 0; j--) {
+         sl_pp_token_buffer_unget(&buffer, &macro->body[j - 1]);
+      }
+
+      /* Expand.
+       */
+      for (;;) {
+         struct sl_pp_token_info input;
+
+         sl_pp_token_buffer_get(&buffer, &input);
+         switch (input.token) {
+         case SL_PP_NEWLINE:
+            if (sl_pp_process_out(state, &input)) {
+               strcpy(context->error_msg, "out of memory");
+               sl_pp_token_buffer_destroy(&buffer);
+               goto fail;
+            }
+            break;
+
+         case SL_PP_IDENTIFIER:
+            sl_pp_token_buffer_unget(&buffer, &input);
+            if (sl_pp_macro_expand(context, &buffer, actual_arg, state, behaviour)) {
+               sl_pp_token_buffer_destroy(&buffer);
+               goto fail;
+            }
+            break;
+
+         case SL_PP_EOF:
+            sl_pp_token_buffer_destroy(&buffer);
+            sl_pp_macro_free(actual_arg);
+            return 0;
+
+         default:
+            if (!mute) {
+               if (sl_pp_process_out(state, &input)) {
+                  strcpy(context->error_msg, "out of memory");
+                  sl_pp_token_buffer_destroy(&buffer);
+                  goto fail;
+               }
+            }
+         }
+      }
+   }
+
+fail:
+   sl_pp_macro_free(actual_arg);
+   return -1;
+}
diff --git a/src/glsl/pp/sl_pp_macro.h b/src/glsl/pp/sl_pp_macro.h
new file mode 100644 (file)
index 0000000..1d21068
--- /dev/null
@@ -0,0 +1,73 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef SL_PP_MACRO_H
+#define SL_PP_MACRO_H
+
+#include "sl_pp_token.h"
+
+
+struct sl_pp_context;
+struct sl_pp_process_state;
+struct sl_pp_token_buffer;
+
+struct sl_pp_macro_formal_arg {
+   int name;
+   struct sl_pp_macro_formal_arg *next;
+};
+
+struct sl_pp_macro {
+   int name;
+   int num_args;                       /* -1 means no args, 0 means `()' */
+   struct sl_pp_macro_formal_arg *arg;
+   struct sl_pp_token_info *body;
+   struct sl_pp_macro *next;
+};
+
+struct sl_pp_macro *
+sl_pp_macro_new(void);
+
+void
+sl_pp_macro_free(struct sl_pp_macro *macro);
+
+void
+sl_pp_macro_reset(struct sl_pp_macro *macro);
+
+enum sl_pp_macro_expand_behaviour {
+   sl_pp_macro_expand_normal,
+   sl_pp_macro_expand_mute,
+   sl_pp_macro_expand_unknown_to_0
+};
+
+int
+sl_pp_macro_expand(struct sl_pp_context *context,
+                   struct sl_pp_token_buffer *tokens,
+                   struct sl_pp_macro *local,
+                   struct sl_pp_process_state *state,
+                   enum sl_pp_macro_expand_behaviour behaviour);
+
+#endif /* SL_PP_MACRO_H */
diff --git a/src/glsl/pp/sl_pp_pragma.c b/src/glsl/pp/sl_pp_pragma.c
new file mode 100644 (file)
index 0000000..caf4c63
--- /dev/null
@@ -0,0 +1,109 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_context.h"
+#include "sl_pp_process.h"
+
+
+int
+sl_pp_process_pragma(struct sl_pp_context *context,
+                     const struct sl_pp_token_info *input,
+                     unsigned int first,
+                     unsigned int last,
+                     struct sl_pp_process_state *state)
+{
+   int pragma_name = -1;
+   struct sl_pp_token_info out;
+   int arg_name = -1;
+
+   if (first < last && input[first].token == SL_PP_IDENTIFIER) {
+      pragma_name = input[first].data.identifier;
+      first++;
+   }
+   if (pragma_name == -1) {
+      return 0;
+   }
+
+   if (pragma_name == context->dict.optimize) {
+      out.token = SL_PP_PRAGMA_OPTIMIZE;
+   } else if (pragma_name == context->dict.debug) {
+      out.token = SL_PP_PRAGMA_DEBUG;
+   } else {
+      return 0;
+   }
+
+   while (first < last && input[first].token == SL_PP_WHITESPACE) {
+      first++;
+   }
+
+   if (first < last && input[first].token == SL_PP_LPAREN) {
+      first++;
+   } else {
+      return 0;
+   }
+
+   while (first < last && input[first].token == SL_PP_WHITESPACE) {
+      first++;
+   }
+
+   if (first < last && input[first].token == SL_PP_IDENTIFIER) {
+      arg_name = input[first].data.identifier;
+      first++;
+   }
+   if (arg_name == -1) {
+      return 0;
+   }
+
+   if (arg_name == context->dict.off) {
+      out.data.pragma = 0;
+   } else if (arg_name == context->dict.on) {
+      out.data.pragma = 1;
+   } else {
+      return 0;
+   }
+
+   while (first < last && input[first].token == SL_PP_WHITESPACE) {
+      first++;
+   }
+
+   if (first < last && input[first].token == SL_PP_RPAREN) {
+      first++;
+   } else {
+      return 0;
+   }
+
+   /* Ignore the tokens that follow. */
+
+   if (sl_pp_process_out(state, &out)) {
+      strcpy(context->error_msg, "out of memory");
+      return -1;
+   }
+
+   return 0;
+}
diff --git a/src/glsl/pp/sl_pp_process.c b/src/glsl/pp/sl_pp_process.c
new file mode 100644 (file)
index 0000000..f89986d
--- /dev/null
@@ -0,0 +1,327 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_context.h"
+#include "sl_pp_process.h"
+#include "sl_pp_public.h"
+
+
+int
+sl_pp_process_out(struct sl_pp_process_state *state,
+                  const struct sl_pp_token_info *token)
+{
+   if (state->out_len >= state->out_max) {
+      unsigned int new_max = state->out_max;
+
+      if (new_max < 0x100) {
+         new_max = 0x100;
+      } else if (new_max < 0x10000) {
+         new_max *= 2;
+      } else {
+         new_max += 0x10000;
+      }
+
+      state->out = realloc(state->out, new_max * sizeof(struct sl_pp_token_info));
+      if (!state->out) {
+         return -1;
+      }
+      state->out_max = new_max;
+   }
+
+   state->out[state->out_len++] = *token;
+   return 0;
+}
+
+int
+sl_pp_process_get(struct sl_pp_context *context,
+                  struct sl_pp_token_info *output)
+{
+   if (!context->process_state.out) {
+      if (context->line > 1) {
+         struct sl_pp_token_info ti;
+
+         ti.token = SL_PP_LINE;
+         ti.data.line.lineno = context->line - 1;
+         ti.data.line.fileno = context->file;
+         if (sl_pp_process_out(&context->process_state, &ti)) {
+            strcpy(context->error_msg, "out of memory");
+            return -1;
+         }
+
+         ti.token = SL_PP_NEWLINE;
+         if (sl_pp_process_out(&context->process_state, &ti)) {
+            strcpy(context->error_msg, "out of memory");
+            return -1;
+         }
+      }
+   }
+
+   for (;;) {
+      struct sl_pp_token_info input;
+      int found_eof = 0;
+
+      if (context->process_state.out_len) {
+         *output = context->process_state.out[0];
+
+         if (context->process_state.out_len > 1) {
+            unsigned int i;
+
+            for (i = 1; i < context->process_state.out_len; i++) {
+               context->process_state.out[i - 1] = context->process_state.out[i];
+            }
+         }
+         context->process_state.out_len--;
+
+         return 0;
+      }
+
+      if (sl_pp_token_buffer_skip_white(&context->tokens, &input)) {
+         return -1;
+      }
+      if (input.token == SL_PP_HASH) {
+         if (sl_pp_token_buffer_skip_white(&context->tokens, &input)) {
+            return -1;
+         }
+         switch (input.token) {
+         case SL_PP_IDENTIFIER:
+            {
+               int name;
+               int found_eol = 0;
+               struct sl_pp_token_info endof;
+               struct sl_pp_token_peek peek;
+               int result = 0;
+
+               /* Directive name. */
+               name = input.data.identifier;
+
+               if (sl_pp_token_buffer_skip_white(&context->tokens, &input)) {
+                  return -1;
+               }
+               sl_pp_token_buffer_unget(&context->tokens, &input);
+
+               if (sl_pp_token_peek_init(&peek, &context->tokens)) {
+                  return -1;
+               }
+
+               while (!found_eol) {
+                  if (sl_pp_token_peek_get(&peek, &input)) {
+                     sl_pp_token_peek_destroy(&peek);
+                     return -1;
+                  }
+                  switch (input.token) {
+                  case SL_PP_NEWLINE:
+                     /* Preserve newline just for the sake of line numbering. */
+                     endof = input;
+                     found_eol = 1;
+                     break;
+
+                  case SL_PP_EOF:
+                     endof = input;
+                     found_eof = 1;
+                     found_eol = 1;
+                     break;
+
+                  default:
+                     break;
+                  }
+               }
+
+               if (name == context->dict._if) {
+                  struct sl_pp_token_buffer buffer;
+
+                  result = sl_pp_token_peek_to_buffer(&peek, &buffer);
+                  if (result == 0) {
+                     result = sl_pp_process_if(context, &buffer);
+                     sl_pp_token_buffer_destroy(&buffer);
+                  }
+               } else if (name == context->dict.ifdef) {
+                  result = sl_pp_process_ifdef(context, peek.tokens, 0, peek.size - 1);
+               } else if (name == context->dict.ifndef) {
+                  result = sl_pp_process_ifndef(context, peek.tokens, 0, peek.size - 1);
+               } else if (name == context->dict.elif) {
+                  struct sl_pp_token_buffer buffer;
+
+                  result = sl_pp_token_peek_to_buffer(&peek, &buffer);
+                  if (result == 0) {
+                     result = sl_pp_process_elif(context, &buffer);
+                     sl_pp_token_buffer_destroy(&buffer);
+                  }
+               } else if (name == context->dict._else) {
+                  result = sl_pp_process_else(context, peek.tokens, 0, peek.size - 1);
+               } else if (name == context->dict.endif) {
+                  result = sl_pp_process_endif(context, peek.tokens, 0, peek.size - 1);
+               } else if (context->if_value) {
+                  if (name == context->dict.define) {
+                     result = sl_pp_process_define(context, peek.tokens, 0, peek.size - 1);
+                  } else if (name == context->dict.error) {
+                     sl_pp_process_error(context, peek.tokens, 0, peek.size - 1);
+                     result = -1;
+                  } else if (name == context->dict.extension) {
+                     result = sl_pp_process_extension(context, peek.tokens, 0, peek.size - 1, &context->process_state);
+                  } else if (name == context->dict.line) {
+                     struct sl_pp_token_buffer buffer;
+
+                     result = sl_pp_token_peek_to_buffer(&peek, &buffer);
+                     if (result == 0) {
+                        result = sl_pp_process_line(context, &buffer, &context->process_state);
+                        sl_pp_token_buffer_destroy(&buffer);
+                     }
+                  } else if (name == context->dict.pragma) {
+                     result = sl_pp_process_pragma(context, peek.tokens, 0, peek.size - 1, &context->process_state);
+                  } else if (name == context->dict.undef) {
+                     result = sl_pp_process_undef(context, peek.tokens, 0, peek.size - 1);
+                  } else {
+                     strcpy(context->error_msg, "unrecognised directive name");
+                     result = -1;
+                  }
+               }
+
+               sl_pp_token_peek_commit(&peek);
+               sl_pp_token_peek_destroy(&peek);
+
+               if (result) {
+                  return result;
+               }
+
+               if (sl_pp_process_out(&context->process_state, &endof)) {
+                  strcpy(context->error_msg, "out of memory");
+                  return -1;
+               }
+               context->line++;
+            }
+            break;
+
+         case SL_PP_NEWLINE:
+            /* Empty directive. */
+            if (sl_pp_process_out(&context->process_state, &input)) {
+               strcpy(context->error_msg, "out of memory");
+               return -1;
+            }
+            context->line++;
+            break;
+
+         case SL_PP_EOF:
+            /* Empty directive. */
+            if (sl_pp_process_out(&context->process_state, &input)) {
+               strcpy(context->error_msg, "out of memory");
+               return -1;
+            }
+            found_eof = 1;
+            break;
+
+         default:
+            strcpy(context->error_msg, "expected a directive name");
+            return -1;
+         }
+      } else {
+         int found_eol = 0;
+
+         sl_pp_token_buffer_unget(&context->tokens, &input);
+
+         while (!found_eol) {
+            if (sl_pp_token_buffer_get(&context->tokens, &input)) {
+               return -1;
+            }
+
+            switch (input.token) {
+            case SL_PP_WHITESPACE:
+               /* Drop whitespace all together at this point. */
+               break;
+
+            case SL_PP_NEWLINE:
+               /* Preserve newline just for the sake of line numbering. */
+               if (sl_pp_process_out(&context->process_state, &input)) {
+                  strcpy(context->error_msg, "out of memory");
+                  return -1;
+               }
+               context->line++;
+               found_eol = 1;
+               break;
+
+            case SL_PP_EOF:
+               if (sl_pp_process_out(&context->process_state, &input)) {
+                  strcpy(context->error_msg, "out of memory");
+                  return -1;
+               }
+               found_eof = 1;
+               found_eol = 1;
+               break;
+
+            case SL_PP_IDENTIFIER:
+               sl_pp_token_buffer_unget(&context->tokens, &input);
+               if (sl_pp_macro_expand(context, &context->tokens, NULL, &context->process_state,
+                                      context->if_value ? sl_pp_macro_expand_normal : sl_pp_macro_expand_mute)) {
+                  return -1;
+               }
+               break;
+
+            default:
+               if (context->if_value) {
+                  if (sl_pp_process_out(&context->process_state, &input)) {
+                     strcpy(context->error_msg, "out of memory");
+                     return -1;
+                  }
+               }
+            }
+         }
+      }
+
+      if (found_eof) {
+         if (context->if_ptr != SL_PP_MAX_IF_NESTING) {
+            strcpy(context->error_msg, "expected `#endif' directive");
+            return -1;
+         }
+      }
+   }
+}
+
+int
+sl_pp_process(struct sl_pp_context *context,
+              struct sl_pp_token_info **output)
+{
+   struct sl_pp_process_state state;
+
+   memset(&state, 0, sizeof(state));
+   for (;;) {
+      struct sl_pp_token_info input;
+
+      if (sl_pp_process_get(context, &input)) {
+         free(state.out);
+         return -1;
+      }
+      if (sl_pp_process_out(&state, &input)) {
+         free(state.out);
+         return -1;
+      }
+      if (input.token == SL_PP_EOF) {
+         *output = state.out;
+         return 0;
+      }
+   }
+}
diff --git a/src/glsl/pp/sl_pp_process.h b/src/glsl/pp/sl_pp_process.h
new file mode 100644 (file)
index 0000000..fe6ff0d
--- /dev/null
@@ -0,0 +1,116 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef SL_PP_PROCESS_H
+#define SL_PP_PROCESS_H
+
+#include "sl_pp_macro.h"
+#include "sl_pp_token.h"
+
+
+struct sl_pp_context;
+
+struct sl_pp_process_state {
+   struct sl_pp_token_info *out;
+   unsigned int out_len;
+   unsigned int out_max;
+};
+
+int
+sl_pp_process_define(struct sl_pp_context *context,
+                     const struct sl_pp_token_info *input,
+                     unsigned int first,
+                     unsigned int last);
+
+int
+sl_pp_process_undef(struct sl_pp_context *context,
+                    const struct sl_pp_token_info *input,
+                    unsigned int first,
+                    unsigned int last);
+
+int
+sl_pp_process_if(struct sl_pp_context *context,
+                 struct sl_pp_token_buffer *input);
+
+int
+sl_pp_process_ifdef(struct sl_pp_context *context,
+                    const struct sl_pp_token_info *input,
+                    unsigned int first,
+                    unsigned int last);
+
+int
+sl_pp_process_ifndef(struct sl_pp_context *context,
+                     const struct sl_pp_token_info *input,
+                     unsigned int first,
+                     unsigned int last);
+
+int
+sl_pp_process_elif(struct sl_pp_context *context,
+                   struct sl_pp_token_buffer *buffer);
+
+int
+sl_pp_process_else(struct sl_pp_context *context,
+                   const struct sl_pp_token_info *input,
+                   unsigned int first,
+                   unsigned int last);
+
+int
+sl_pp_process_endif(struct sl_pp_context *context,
+                    const struct sl_pp_token_info *input,
+                    unsigned int first,
+                    unsigned int last);
+
+void
+sl_pp_process_error(struct sl_pp_context *context,
+                    const struct sl_pp_token_info *input,
+                    unsigned int first,
+                    unsigned int last);
+
+int
+sl_pp_process_pragma(struct sl_pp_context *context,
+                     const struct sl_pp_token_info *input,
+                     unsigned int first,
+                     unsigned int last,
+                     struct sl_pp_process_state *state);
+
+int
+sl_pp_process_extension(struct sl_pp_context *context,
+                        const struct sl_pp_token_info *input,
+                        unsigned int first,
+                        unsigned int last,
+                        struct sl_pp_process_state *state);
+
+int
+sl_pp_process_line(struct sl_pp_context *context,
+                   struct sl_pp_token_buffer *buffer,
+                   struct sl_pp_process_state *state);
+
+int
+sl_pp_process_out(struct sl_pp_process_state *state,
+                  const struct sl_pp_token_info *token);
+
+#endif /* SL_PP_PROCESS_H */
diff --git a/src/glsl/pp/sl_pp_public.h b/src/glsl/pp/sl_pp_public.h
new file mode 100644 (file)
index 0000000..12528d6
--- /dev/null
@@ -0,0 +1,84 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef SL_PP_PUBLIC_H
+#define SL_PP_PUBLIC_H
+
+
+struct sl_pp_context;
+
+
+#include "sl_pp_purify.h"
+#include "sl_pp_token.h"
+
+
+struct sl_pp_context *
+sl_pp_context_create(const char *input,
+                     const struct sl_pp_purify_options *options);
+
+void
+sl_pp_context_destroy(struct sl_pp_context *context);
+
+const char *
+sl_pp_context_error_message(const struct sl_pp_context *context);
+
+void
+sl_pp_context_error_position(const struct sl_pp_context *context,
+                             unsigned int *file,
+                             unsigned int *line);
+
+int
+sl_pp_context_add_extension(struct sl_pp_context *context,
+                            const char *name,
+                            const char *name_string);
+
+int
+sl_pp_context_add_predefined(struct sl_pp_context *context,
+                             const char *name,
+                             const char *value);
+
+int
+sl_pp_context_add_unique_str(struct sl_pp_context *context,
+                             const char *str);
+
+const char *
+sl_pp_context_cstr(const struct sl_pp_context *context,
+                   int offset);
+
+int
+sl_pp_version(struct sl_pp_context *context,
+              unsigned int *version);
+
+int
+sl_pp_process_get(struct sl_pp_context *context,
+                  struct sl_pp_token_info *output);
+
+int
+sl_pp_process(struct sl_pp_context *context,
+              struct sl_pp_token_info **output);
+
+#endif /* SL_PP_PUBLIC_H */
diff --git a/src/glsl/pp/sl_pp_purify.c b/src/glsl/pp/sl_pp_purify.c
new file mode 100644 (file)
index 0000000..b50f819
--- /dev/null
@@ -0,0 +1,302 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <stdarg.h>
+#include <stdio.h>
+#include "sl_pp_purify.h"
+
+
+/*
+ * Preprocessor purifier performs the following tasks.
+ * - Convert all variants of newlines into a Unix newline.
+ * - Merge continued lines into a single long line.
+ * - Remove line comments and replace block comments with whitespace.
+ */
+
+
+static unsigned int
+_purify_newline(const char *input,
+                char *out,
+                unsigned int *current_line)
+{
+   if (input[0] == '\n') {
+      *out = '\n';
+      (*current_line)++;
+      if (input[1] == '\r') {
+         /*
+          * The GLSL spec is not explicit about whether this
+          * combination is a valid newline or not.
+          * Let's assume it is acceptable.
+          */
+         return 2;
+      }
+      return 1;
+   }
+   if (input[0] == '\r') {
+      *out = '\n';
+      (*current_line)++;
+      if (input[1] == '\n') {
+         return 2;
+      }
+      return 1;
+   }
+   *out = input[0];
+   return 1;
+}
+
+
+static unsigned int
+_purify_backslash(const char *input,
+                  char *out,
+                  unsigned int *current_line)
+{
+   unsigned int eaten = 0;
+
+   for (;;) {
+      if (input[0] == '\\') {
+         char next;
+         unsigned int next_eaten;
+         unsigned int next_line = *current_line;
+
+         eaten++;
+         input++;
+
+         next_eaten = _purify_newline(input, &next, &next_line);
+         if (next == '\n') {
+            /*
+             * If this is really a line continuation sequence, eat
+             * it and do not exit the loop.
+             */
+            eaten += next_eaten;
+            input += next_eaten;
+            *current_line = next_line;
+         } else {
+            /*
+             * It is an error to put anything between a backslash
+             * and a newline and still expect it to behave like a line
+             * continuation sequence.
+             * Even if it is an innocent whitespace.
+             */
+            *out = '\\';
+            break;
+         }
+      } else {
+         eaten += _purify_newline(input, out, current_line);
+         break;
+      }
+   }
+   return eaten;
+}
+
+
+static void
+_report_error(char *buf,
+              unsigned int cbbuf,
+              const char *msg,
+              ...)
+{
+   va_list args;
+
+   va_start(args, msg);
+   vsnprintf(buf, cbbuf, msg, args);
+   va_end(args);
+}
+
+
+void
+sl_pp_purify_state_init(struct sl_pp_purify_state *state,
+                        const char *input,
+                        const struct sl_pp_purify_options *options)
+{
+   state->options = *options;
+   state->input = input;
+   state->current_line = 1;
+   state->inside_c_comment = 0;
+}
+
+
+unsigned int
+_purify_comment(struct sl_pp_purify_state *state,
+                char *output,
+                unsigned int *current_line,
+                char *errormsg,
+                unsigned int cberrormsg)
+{
+   for (;;) {
+      unsigned int eaten;
+      char next;
+
+      eaten = _purify_backslash(state->input, &next, current_line);
+      state->input += eaten;
+      while (next == '*') {
+         eaten = _purify_backslash(state->input, &next, current_line);
+         state->input += eaten;
+         if (next == '/') {
+            *output = ' ';
+            state->inside_c_comment = 0;
+            return 1;
+         }
+      }
+      if (next == '\n') {
+         *output = '\n';
+         state->inside_c_comment = 1;
+         return 1;
+      }
+      if (next == '\0') {
+         _report_error(errormsg, cberrormsg, "expected `*/' but end of translation unit found");
+         return 0;
+      }
+   }
+}
+
+
+unsigned int
+sl_pp_purify_getc(struct sl_pp_purify_state *state,
+                  char *output,
+                  unsigned int *current_line,
+                  char *errormsg,
+                  unsigned int cberrormsg)
+{
+   unsigned int eaten;
+
+   if (state->inside_c_comment) {
+      return _purify_comment(state, output, current_line, errormsg, cberrormsg);
+   }
+
+   eaten = _purify_backslash(state->input, output, current_line);
+   state->input += eaten;
+   if (*output == '/') {
+      char next;
+      unsigned int next_line = *current_line;
+
+      eaten = _purify_backslash(state->input, &next, &next_line);
+      if (next == '/') {
+         state->input += eaten;
+         *current_line = next_line;
+
+         /* Replace a line comment with either a newline or nil. */
+         for (;;) {
+            eaten = _purify_backslash(state->input, &next, current_line);
+            state->input += eaten;
+            if (next == '\n' || next == '\0') {
+               *output = next;
+               return eaten;
+            }
+         }
+      } else if (next == '*') {
+         state->input += eaten;
+         *current_line = next_line;
+
+         return _purify_comment(state, output, current_line, errormsg, cberrormsg);
+      }
+   }
+   return eaten;
+}
+
+
+struct out_buf {
+   char *out;
+   unsigned int len;
+   unsigned int capacity;
+   unsigned int current_line;
+   char *errormsg;
+   unsigned int cberrormsg;
+};
+
+
+static int
+_out_buf_putc(struct out_buf *obuf,
+              char c)
+{
+   if (obuf->len >= obuf->capacity) {
+      unsigned int new_max = obuf->capacity;
+
+      if (new_max < 0x100) {
+         new_max = 0x100;
+      } else if (new_max < 0x10000) {
+         new_max *= 2;
+      } else {
+         new_max += 0x10000;
+      }
+
+      obuf->out = realloc(obuf->out, new_max);
+      if (!obuf->out) {
+         _report_error(obuf->errormsg, obuf->cberrormsg, "out of memory");
+         return -1;
+      }
+      obuf->capacity = new_max;
+   }
+
+   obuf->out[obuf->len++] = c;
+
+   return 0;
+}
+
+
+int
+sl_pp_purify(const char *input,
+             const struct sl_pp_purify_options *options,
+             char **output,
+             char *errormsg,
+             unsigned int cberrormsg,
+             unsigned int *errorline)
+{
+   struct out_buf obuf;
+   struct sl_pp_purify_state state;
+
+   obuf.out = NULL;
+   obuf.len = 0;
+   obuf.capacity = 0;
+   obuf.current_line = 1;
+   obuf.errormsg = errormsg;
+   obuf.cberrormsg = cberrormsg;
+
+   sl_pp_purify_state_init(&state, input, options);
+
+   for (;;) {
+      unsigned int eaten;
+      char c;
+
+      eaten = sl_pp_purify_getc(&state, &c, &obuf.current_line, errormsg, cberrormsg);
+      if (!eaten) {
+         *errorline = obuf.current_line;
+         return -1;
+      }
+      if (_out_buf_putc(&obuf, c)) {
+         *errorline = obuf.current_line;
+         return -1;
+      }
+
+      if (c == '\0') {
+         break;
+      }
+   }
+
+   *output = obuf.out;
+   return 0;
+}
diff --git a/src/glsl/pp/sl_pp_purify.h b/src/glsl/pp/sl_pp_purify.h
new file mode 100644 (file)
index 0000000..c0f55cb
--- /dev/null
@@ -0,0 +1,63 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef SL_PP_PURIFY_H
+#define SL_PP_PURIFY_H
+
+struct sl_pp_purify_options {
+   unsigned int preserve_columns:1;
+   unsigned int tab_width:4;
+};
+
+int
+sl_pp_purify(const char *input,
+             const struct sl_pp_purify_options *options,
+             char **output,
+             char *errormsg,
+             unsigned int cberrormsg,
+             unsigned int *errorline);
+
+struct sl_pp_purify_state {
+   struct sl_pp_purify_options options;
+   const char *input;
+   unsigned int current_line;
+   unsigned int inside_c_comment:1;
+};
+
+void
+sl_pp_purify_state_init(struct sl_pp_purify_state *state,
+                        const char *input,
+                        const struct sl_pp_purify_options *options);
+
+unsigned int
+sl_pp_purify_getc(struct sl_pp_purify_state *state,
+                  char *output,
+                  unsigned int *current_line,
+                  char *errormsg,
+                  unsigned int cberrormsg);
+
+#endif /* SL_PP_PURIFY_H */
diff --git a/src/glsl/pp/sl_pp_token.c b/src/glsl/pp/sl_pp_token.c
new file mode 100644 (file)
index 0000000..a708978
--- /dev/null
@@ -0,0 +1,854 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <assert.h>
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_public.h"
+#include "sl_pp_context.h"
+#include "sl_pp_token.h"
+
+
+#define PURE_ERROR 256
+
+static int
+_pure_getc(struct sl_pp_context *context)
+{
+   char c;
+
+   if (context->getc_buf_size) {
+      return context->getc_buf[--context->getc_buf_size];
+   }
+
+   if (sl_pp_purify_getc(&context->pure, &c, &context->error_line, context->error_msg, sizeof(context->error_msg)) == 0) {
+      return PURE_ERROR;
+   }
+   return c;
+}
+
+
+static void
+_pure_ungetc(struct sl_pp_context *context,
+             int c)
+{
+   assert(c != PURE_ERROR);
+
+   if (context->getc_buf_size == context->getc_buf_capacity) {
+      context->getc_buf_capacity += 64;
+      context->getc_buf = realloc(context->getc_buf, context->getc_buf_capacity * sizeof(char));
+      assert(context->getc_buf);                            
+   }
+
+   context->getc_buf[context->getc_buf_size++] = (char)c;
+}
+
+
+struct lookahead_state {
+   char buf[256];
+   unsigned int pos;
+   struct sl_pp_context *context;
+};
+
+
+static void
+_lookahead_init(struct lookahead_state *lookahead,
+                struct sl_pp_context *context)
+{
+   lookahead->pos = 0;
+   lookahead->context = context;
+}
+
+
+static unsigned int
+_lookahead_tell(const struct lookahead_state *lookahead)
+{
+   return lookahead->pos;
+}
+
+
+static const void *
+_lookahead_buf(const struct lookahead_state *lookahead)
+{
+   return lookahead->buf;
+}
+
+
+static void
+_lookahead_revert(struct lookahead_state *lookahead,
+                  unsigned int pos)
+{
+   assert(pos <= lookahead->pos);
+
+   while (lookahead->pos > pos) {
+      _pure_ungetc(lookahead->context, lookahead->buf[--lookahead->pos]);
+   }
+}
+
+
+static int
+_lookahead_getc(struct lookahead_state *lookahead)
+{
+   int c;
+
+   assert(lookahead->pos < sizeof(lookahead->buf) / sizeof(lookahead->buf[0]));
+
+   c = _pure_getc(lookahead->context);
+   if (c != PURE_ERROR) {
+      lookahead->buf[lookahead->pos++] = (char)c;
+   }
+   return c;
+}
+
+
+static int
+_is_identifier_char(char c)
+{
+   return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || (c >= '0' && c <= '9') || c == '_';
+}
+
+
+static int
+_tokenise_identifier(struct sl_pp_context *context,
+                     struct sl_pp_token_info *out)
+{
+   int c;
+   char identifier[256];   /* XXX: Remove this artifical limit. */
+   unsigned int i = 0;
+
+   out->token = SL_PP_IDENTIFIER;
+   out->data.identifier = -1;
+
+   c = _pure_getc(context);
+   if (c == PURE_ERROR) {
+      return -1;
+   }
+   identifier[i++] = (char)c;
+   for (;;) {
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+
+      if (_is_identifier_char((char)c)) {
+         if (i >= sizeof(identifier) / sizeof(char) - 1) {
+            strcpy(context->error_msg, "out of memory");
+            _pure_ungetc(context, c);
+            while (i) {
+               _pure_ungetc(context, identifier[--i]);
+            }
+            return -1;
+         }
+         identifier[i++] = (char)c;
+      } else {
+         _pure_ungetc(context, c);
+         break;
+      }
+   }
+   identifier[i] = '\0';
+
+   out->data.identifier = sl_pp_context_add_unique_str(context, identifier);
+   if (out->data.identifier == -1) {
+      while (i) {
+         _pure_ungetc(context, identifier[--i]);
+      }
+      return -1;
+   }
+
+   return 0;
+}
+
+
+/*
+ * Return the number of consecutive decimal digits in the input stream.
+ */
+static unsigned int
+_parse_float_digits(struct lookahead_state *lookahead)
+{
+   unsigned int eaten;
+
+   for (eaten = 0;; eaten++) {
+      unsigned int pos = _lookahead_tell(lookahead);
+      char c = _lookahead_getc(lookahead);
+
+      if (c < '0' || c > '9') {
+         _lookahead_revert(lookahead, pos);
+         break;
+      }
+   }
+   return eaten;
+}
+
+
+/*
+ * Try to match one of the following patterns for the fractional part
+ * of a floating point number.
+ *
+ * digits . [digits]
+ * . digits
+ *
+ * Return 0 if the pattern could not be matched, otherwise the number
+ * of eaten characters from the input stream.
+ */
+static unsigned int
+_parse_float_frac(struct lookahead_state *lookahead)
+{
+   unsigned int pos;
+   int c;
+   unsigned int eaten;
+
+   pos = _lookahead_tell(lookahead);
+   c = _lookahead_getc(lookahead);
+   if (c == '.') {
+      eaten = _parse_float_digits(lookahead);
+      if (eaten) {
+         return eaten + 1;
+      }
+      _lookahead_revert(lookahead, pos);
+      return 0;
+   }
+
+   _lookahead_revert(lookahead, pos);
+   eaten = _parse_float_digits(lookahead);
+   if (eaten) {
+      c = _lookahead_getc(lookahead);
+      if (c == '.') {
+         return eaten + 1 + _parse_float_digits(lookahead);
+      }
+   }
+
+   _lookahead_revert(lookahead, pos);
+   return 0;
+}
+
+
+/*
+ * Try to match the following pattern for the exponential part
+ * of a floating point number.
+ *
+ * (e|E) [(+|-)] digits
+ *
+ * Return 0 if the pattern could not be matched, otherwise the number
+ * of eaten characters from the input stream.
+ */
+static unsigned int
+_parse_float_exp(struct lookahead_state *lookahead)
+{
+   unsigned int pos, pos2;
+   int c;
+   unsigned int eaten, digits;
+
+   pos = _lookahead_tell(lookahead);
+   c = _lookahead_getc(lookahead);
+   if (c != 'e' && c != 'E') {
+      _lookahead_revert(lookahead, pos);
+      return 0;
+   }
+
+   pos2 = _lookahead_tell(lookahead);
+   c = _lookahead_getc(lookahead);
+   if (c == '-' || c == '+') {
+      eaten = 2;
+   } else {
+      _lookahead_revert(lookahead, pos2);
+      eaten = 1;
+   }
+
+   digits = _parse_float_digits(lookahead);
+   if (!digits) {
+      _lookahead_revert(lookahead, pos);
+      return 0;
+   }
+
+   return eaten + digits;
+}
+
+
+/*
+ * Try to match one of the following patterns for a floating point number.
+ *
+ * fract [exp] [(f|F)]
+ * digits exp [(f|F)]
+ *
+ * Return 0 if the pattern could not be matched, otherwise the number
+ * of eaten characters from the input stream.
+ */
+static unsigned int
+_parse_float(struct lookahead_state *lookahead)
+{
+   unsigned int eaten;
+
+   eaten = _parse_float_frac(lookahead);
+   if (eaten) {
+      unsigned int pos;
+      int c;
+
+      eaten += _parse_float_exp(lookahead);
+
+      pos = _lookahead_tell(lookahead);
+      c = _lookahead_getc(lookahead);
+      if (c == 'f' || c == 'F') {
+         eaten++;
+      } else {
+         _lookahead_revert(lookahead, pos);
+      }
+
+      return eaten;
+   }
+
+   eaten = _parse_float_digits(lookahead);
+   if (eaten) {
+      unsigned int exponent;
+
+      exponent = _parse_float_exp(lookahead);
+      if (exponent) {
+         unsigned int pos;
+         int c;
+
+         eaten += exponent;
+
+         pos = _lookahead_tell(lookahead);
+         c = _lookahead_getc(lookahead);
+         if (c == 'f' || c == 'F') {
+            eaten++;
+         } else {
+            _lookahead_revert(lookahead, pos);
+         }
+
+         return eaten;
+      }
+   }
+
+   _lookahead_revert(lookahead, 0);
+   return 0;
+}
+
+
+static unsigned int
+_parse_hex(struct lookahead_state *lookahead)
+{
+   int c;
+   unsigned int n;
+
+   c = _lookahead_getc(lookahead);
+   if (c != '0') {
+      _lookahead_revert(lookahead, 0);
+      return 0;
+   }
+
+   c = _lookahead_getc(lookahead);
+   if (c != 'x' && c != 'X') {
+      _lookahead_revert(lookahead, 0);
+      return 0;
+   }
+
+   for (n = 2;;) {
+      unsigned int pos = _lookahead_tell(lookahead);
+
+      c = _lookahead_getc(lookahead);
+      if ((c >= '0' && c <= '9') || (c >= 'a' && c <= 'f') || (c >= 'A' && c <= 'F')) {
+         n++;
+      } else {
+         _lookahead_revert(lookahead, pos);
+         break;
+      }
+   }
+
+   if (n > 2) {
+      return n;
+   }
+
+   _lookahead_revert(lookahead, 0);
+   return 0;
+}
+
+
+static unsigned int
+_parse_oct(struct lookahead_state *lookahead)
+{
+   int c;
+   unsigned int n;
+
+   c = _lookahead_getc(lookahead);
+   if (c != '0') {
+      _lookahead_revert(lookahead, 0);
+      return 0;
+   }
+
+   for (n = 1;;) {
+      unsigned int pos = _lookahead_tell(lookahead);
+
+      c = _lookahead_getc(lookahead);
+      if ((c >= '0' && c <= '7')) {
+         n++;
+      } else {
+         _lookahead_revert(lookahead, pos);
+         break;
+      }
+   }
+
+   return n;
+}
+
+
+static unsigned int
+_parse_dec(struct lookahead_state *lookahead)
+{
+   unsigned int n = 0;
+
+   for (;;) {
+      unsigned int pos = _lookahead_tell(lookahead);
+      int c = _lookahead_getc(lookahead);
+
+      if ((c >= '0' && c <= '9')) {
+         n++;
+      } else {
+         _lookahead_revert(lookahead, pos);
+         break;
+      }
+   }
+
+   return n;
+}
+
+
+static int
+_tokenise_number(struct sl_pp_context *context,
+                 struct sl_pp_token_info *out)
+{
+   struct lookahead_state lookahead;
+   unsigned int eaten;
+   unsigned int is_float = 0;
+   unsigned int pos;
+   int c;
+   char number[256];   /* XXX: Remove this artifical limit. */
+
+   _lookahead_init(&lookahead, context);
+
+   eaten = _parse_float(&lookahead);
+   if (!eaten) {
+      eaten = _parse_hex(&lookahead);
+      if (!eaten) {
+         eaten = _parse_oct(&lookahead);
+         if (!eaten) {
+            eaten = _parse_dec(&lookahead);
+         }
+      }
+   } else {
+      is_float = 1;
+   }
+
+   if (!eaten) {
+      strcpy(context->error_msg, "expected a number");
+      return -1;
+   }
+
+   pos = _lookahead_tell(&lookahead);
+   c = _lookahead_getc(&lookahead);
+   _lookahead_revert(&lookahead, pos);
+
+   if (_is_identifier_char(c)) {
+      strcpy(context->error_msg, "expected a number");
+      _lookahead_revert(&lookahead, 0);
+      return -1;
+   }
+
+   if (eaten > sizeof(number) - 1) {
+      strcpy(context->error_msg, "out of memory");
+      _lookahead_revert(&lookahead, 0);
+      return -1;
+   }
+
+   assert(_lookahead_tell(&lookahead) == eaten);
+
+   memcpy(number, _lookahead_buf(&lookahead), eaten);
+   number[eaten] = '\0';
+
+   if (is_float) {
+      out->token = SL_PP_FLOAT;
+      out->data._float = sl_pp_context_add_unique_str(context, number);
+      if (out->data._float == -1) {
+         _lookahead_revert(&lookahead, 0);
+         return -1;
+      }
+   } else {
+      out->token = SL_PP_UINT;
+      out->data._uint = sl_pp_context_add_unique_str(context, number);
+      if (out->data._uint == -1) {
+         _lookahead_revert(&lookahead, 0);
+         return -1;
+      }
+   }
+
+   return 0;
+}
+
+
+int
+sl_pp_token_get(struct sl_pp_context *context,
+                struct sl_pp_token_info *out)
+{
+   int c = _pure_getc(context);
+
+   switch (c) {
+   case ' ':
+   case '\t':
+      out->token = SL_PP_WHITESPACE;
+      break;
+
+   case '\n':
+      out->token = SL_PP_NEWLINE;
+      break;
+
+   case '#':
+      out->token = SL_PP_HASH;
+      break;
+
+   case ',':
+      out->token = SL_PP_COMMA;
+      break;
+
+   case ';':
+      out->token = SL_PP_SEMICOLON;
+      break;
+
+   case '{':
+      out->token = SL_PP_LBRACE;
+      break;
+
+   case '}':
+      out->token = SL_PP_RBRACE;
+      break;
+
+   case '(':
+      out->token = SL_PP_LPAREN;
+      break;
+
+   case ')':
+      out->token = SL_PP_RPAREN;
+      break;
+
+   case '[':
+      out->token = SL_PP_LBRACKET;
+      break;
+
+   case ']':
+      out->token = SL_PP_RBRACKET;
+      break;
+
+   case '.':
+      {
+         int c2 = _pure_getc(context);
+
+         if (c2 == PURE_ERROR) {
+            return -1;
+         }
+         if (c2 >= '0' && c2 <= '9') {
+            _pure_ungetc(context, c2);
+            _pure_ungetc(context, c);
+            if (_tokenise_number(context, out)) {
+               return -1;
+            }
+         } else {
+            _pure_ungetc(context, c2);
+            out->token = SL_PP_DOT;
+         }
+      }
+      break;
+
+   case '+':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '+') {
+         out->token = SL_PP_INCREMENT;
+      } else if (c == '=') {
+         out->token = SL_PP_ADDASSIGN;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_PLUS;
+      }
+      break;
+
+   case '-':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '-') {
+         out->token = SL_PP_DECREMENT;
+      } else if (c == '=') {
+         out->token = SL_PP_SUBASSIGN;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_MINUS;
+      }
+      break;
+
+   case '~':
+      out->token = SL_PP_BITNOT;
+      break;
+
+   case '!':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '=') {
+         out->token = SL_PP_NOTEQUAL;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_NOT;
+      }
+      break;
+
+   case '*':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '=') {
+         out->token = SL_PP_MULASSIGN;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_STAR;
+      }
+      break;
+
+   case '/':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '=') {
+         out->token = SL_PP_DIVASSIGN;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_SLASH;
+      }
+      break;
+
+   case '%':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '=') {
+         out->token = SL_PP_MODASSIGN;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_MODULO;
+      }
+      break;
+
+   case '<':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '<') {
+         c = _pure_getc(context);
+         if (c == PURE_ERROR) {
+            return -1;
+         }
+         if (c == '=') {
+            out->token = SL_PP_LSHIFTASSIGN;
+         } else {
+            _pure_ungetc(context, c);
+            out->token = SL_PP_LSHIFT;
+         }
+      } else if (c == '=') {
+         out->token = SL_PP_LESSEQUAL;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_LESS;
+      }
+      break;
+
+   case '>':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '>') {
+         c = _pure_getc(context);
+         if (c == PURE_ERROR) {
+            return -1;
+         }
+         if (c == '=') {
+            out->token = SL_PP_RSHIFTASSIGN;
+         } else {
+            _pure_ungetc(context, c);
+            out->token = SL_PP_RSHIFT;
+         }
+      } else if (c == '=') {
+         out->token = SL_PP_GREATEREQUAL;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_GREATER;
+      }
+      break;
+
+   case '=':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '=') {
+         out->token = SL_PP_EQUAL;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_ASSIGN;
+      }
+      break;
+
+   case '&':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '&') {
+         out->token = SL_PP_AND;
+      } else if (c == '=') {
+         out->token = SL_PP_BITANDASSIGN;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_BITAND;
+      }
+      break;
+
+   case '^':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '^') {
+         out->token = SL_PP_XOR;
+      } else if (c == '=') {
+         out->token = SL_PP_BITXORASSIGN;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_BITXOR;
+      }
+      break;
+
+   case '|':
+      c = _pure_getc(context);
+      if (c == PURE_ERROR) {
+         return -1;
+      }
+      if (c == '|') {
+         out->token = SL_PP_OR;
+      } else if (c == '=') {
+         out->token = SL_PP_BITORASSIGN;
+      } else {
+         _pure_ungetc(context, c);
+         out->token = SL_PP_BITOR;
+      }
+      break;
+
+   case '?':
+      out->token = SL_PP_QUESTION;
+      break;
+
+   case ':':
+      out->token = SL_PP_COLON;
+      break;
+
+   case '\0':
+      out->token = SL_PP_EOF;
+      break;
+
+   case PURE_ERROR:
+      return -1;
+
+   default:
+      if ((c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || c == '_') {
+         _pure_ungetc(context, c);
+         if (_tokenise_identifier(context, out)) {
+            return -1;
+         }
+      } else if (c >= '0' && c <= '9') {
+         _pure_ungetc(context, c);
+         if (_tokenise_number(context, out)) {
+            return -1;
+         }
+      } else {
+         out->data.other = c;
+         out->token = SL_PP_OTHER;
+      }
+   }
+
+   return 0;
+}
+
+
+int
+sl_pp_tokenise(struct sl_pp_context *context,
+               struct sl_pp_token_info **output)
+{
+   struct sl_pp_token_info *out = NULL;
+   unsigned int out_len = 0;
+   unsigned int out_max = 0;
+
+   for (;;) {
+      struct sl_pp_token_info info;
+
+      if (sl_pp_token_buffer_get(&context->tokens, &info)) {
+         free(out);
+         return -1;
+      }
+
+      if (out_len >= out_max) {
+         unsigned int new_max = out_max;
+
+         if (new_max < 0x100) {
+            new_max = 0x100;
+         } else if (new_max < 0x10000) {
+            new_max *= 2;
+         } else {
+            new_max += 0x10000;
+         }
+
+         out = realloc(out, new_max * sizeof(struct sl_pp_token_info));
+         if (!out) {
+            strcpy(context->error_msg, "out of memory");
+            return -1;
+         }
+         out_max = new_max;
+      }
+
+      out[out_len++] = info;
+
+      if (info.token == SL_PP_EOF) {
+         break;
+      }
+   }
+
+   *output = out;
+   return 0;
+}
diff --git a/src/glsl/pp/sl_pp_token.h b/src/glsl/pp/sl_pp_token.h
new file mode 100644 (file)
index 0000000..a12b193
--- /dev/null
@@ -0,0 +1,133 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef SL_PP_TOKEN_H
+#define SL_PP_TOKEN_H
+
+
+struct sl_pp_context;
+
+enum sl_pp_token {
+   SL_PP_WHITESPACE,
+   SL_PP_NEWLINE,
+   SL_PP_HASH,             /* #   */
+
+   SL_PP_COMMA,            /* ,   */
+   SL_PP_SEMICOLON,        /* ;   */
+   SL_PP_LBRACE,           /* {   */
+   SL_PP_RBRACE,           /* }   */
+   SL_PP_LPAREN,           /* (   */
+   SL_PP_RPAREN,           /* )   */
+   SL_PP_LBRACKET,         /* [   */
+   SL_PP_RBRACKET,         /* ]   */
+   SL_PP_DOT,              /* .   */
+   SL_PP_INCREMENT,        /* ++  */
+   SL_PP_ADDASSIGN,        /* +=  */
+   SL_PP_PLUS,             /* +   */
+   SL_PP_DECREMENT,        /* --  */
+   SL_PP_SUBASSIGN,        /* -=  */
+   SL_PP_MINUS,            /* -   */
+   SL_PP_BITNOT,           /* ~   */
+   SL_PP_NOTEQUAL,         /* !=  */
+   SL_PP_NOT,              /* !   */
+   SL_PP_MULASSIGN,        /* *=  */
+   SL_PP_STAR,             /* *   */
+   SL_PP_DIVASSIGN,        /* /=  */
+   SL_PP_SLASH,            /* /   */
+   SL_PP_MODASSIGN,        /* %=  */
+   SL_PP_MODULO,           /* %   */
+   SL_PP_LSHIFTASSIGN,     /* <<= */
+   SL_PP_LSHIFT,           /* <<  */
+   SL_PP_LESSEQUAL,        /* <=  */
+   SL_PP_LESS,             /* <   */
+   SL_PP_RSHIFTASSIGN,     /* >>= */
+   SL_PP_RSHIFT,           /* >>  */
+   SL_PP_GREATEREQUAL,     /* >=  */
+   SL_PP_GREATER,          /* >   */
+   SL_PP_EQUAL,            /* ==  */
+   SL_PP_ASSIGN,           /* =   */
+   SL_PP_AND,              /* &&  */
+   SL_PP_BITANDASSIGN,     /* &=  */
+   SL_PP_BITAND,           /* &   */
+   SL_PP_XOR,              /* ^^  */
+   SL_PP_BITXORASSIGN,     /* ^=  */
+   SL_PP_BITXOR,           /* ^   */
+   SL_PP_OR,               /* ||  */
+   SL_PP_BITORASSIGN,      /* |=  */
+   SL_PP_BITOR,            /* |   */
+   SL_PP_QUESTION,         /* ?   */
+   SL_PP_COLON,            /* :   */
+
+   SL_PP_IDENTIFIER,
+
+   SL_PP_UINT,
+   SL_PP_FLOAT,
+
+   SL_PP_OTHER,
+
+   SL_PP_PRAGMA_OPTIMIZE,
+   SL_PP_PRAGMA_DEBUG,
+
+   SL_PP_EXTENSION_REQUIRE,
+   SL_PP_EXTENSION_ENABLE,
+   SL_PP_EXTENSION_WARN,
+   SL_PP_EXTENSION_DISABLE,
+
+   SL_PP_LINE,
+
+   SL_PP_EOF
+};
+
+union sl_pp_token_data {
+   int identifier;
+   int _uint;
+   int _float;
+   char other;
+   int pragma;
+   int extension;
+   struct {
+      unsigned int lineno: 24;
+      unsigned int fileno: 8;
+   } line;
+};
+
+struct sl_pp_token_info {
+   enum sl_pp_token token;
+   union sl_pp_token_data data;
+};
+
+struct sl_pp_purify_options;
+
+int
+sl_pp_token_get(struct sl_pp_context *context,
+                struct sl_pp_token_info *out);
+
+int
+sl_pp_tokenise(struct sl_pp_context *context,
+               struct sl_pp_token_info **output);
+
+#endif /* SL_PP_TOKEN_H */
diff --git a/src/glsl/pp/sl_pp_token_util.c b/src/glsl/pp/sl_pp_token_util.c
new file mode 100644 (file)
index 0000000..c85263d
--- /dev/null
@@ -0,0 +1,182 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <assert.h>
+#include <stdlib.h>
+#include "sl_pp_token_util.h"
+
+
+int
+sl_pp_token_buffer_init(struct sl_pp_token_buffer *buffer,
+                        struct sl_pp_context *context)
+{
+   buffer->context = context;
+   buffer->size = 0;
+   buffer->capacity = 64;
+   buffer->tokens = malloc(buffer->capacity * sizeof(struct sl_pp_token_info));
+   if (!buffer->tokens) {
+      return -1;
+   }
+   return 0;
+}
+
+void
+sl_pp_token_buffer_destroy(struct sl_pp_token_buffer *buffer)
+{
+   free(buffer->tokens);
+}
+
+int
+sl_pp_token_buffer_get(struct sl_pp_token_buffer *buffer,
+                       struct sl_pp_token_info *out)
+{
+   /* Pop from stack first if not empty. */
+   if (buffer->size) {
+      *out = buffer->tokens[--buffer->size];
+      return 0;
+   }
+
+   assert(buffer->context);
+   return sl_pp_token_get(buffer->context, out);
+}
+
+void
+sl_pp_token_buffer_unget(struct sl_pp_token_buffer *buffer,
+                         const struct sl_pp_token_info *in)
+{
+   /* Resize if needed. */
+   if (buffer->size == buffer->capacity) {
+      buffer->capacity += 64;
+      buffer->tokens = realloc(buffer->tokens,
+                               buffer->capacity * sizeof(struct sl_pp_token_info));
+      assert(buffer->tokens);
+   }
+
+   /* Push token on stack. */
+   buffer->tokens[buffer->size++] = *in;
+}
+
+int
+sl_pp_token_buffer_skip_white(struct sl_pp_token_buffer *buffer,
+                              struct sl_pp_token_info *out)
+{
+   if (sl_pp_token_buffer_get(buffer, out)) {
+      return -1;
+   }
+
+   while (out->token == SL_PP_WHITESPACE) {
+      if (sl_pp_token_buffer_get(buffer, out)) {
+         return -1;
+      }
+   }
+
+   return 0;
+}
+
+
+
+int
+sl_pp_token_peek_init(struct sl_pp_token_peek *peek,
+                      struct sl_pp_token_buffer *buffer)
+{
+   peek->buffer = buffer;
+   peek->size = 0;
+   peek->capacity = 64;
+   peek->tokens = malloc(peek->capacity * sizeof(struct sl_pp_token_info));
+   if (!peek->tokens) {
+      return -1;
+   }
+   return 0;
+}
+
+void
+sl_pp_token_peek_destroy(struct sl_pp_token_peek *peek)
+{
+   /* Abort. */
+   while (peek->size) {
+      sl_pp_token_buffer_unget(peek->buffer, &peek->tokens[--peek->size]);
+   }
+   free(peek->tokens);
+}
+
+int
+sl_pp_token_peek_get(struct sl_pp_token_peek *peek,
+                     struct sl_pp_token_info *out)
+{
+   /* Get token from buffer. */
+   if (sl_pp_token_buffer_get(peek->buffer, out)) {
+      return -1;
+   }
+
+   /* Save it. */
+   if (peek->size == peek->capacity) {
+      peek->capacity += 64;
+      peek->tokens = realloc(peek->tokens,
+                             peek->capacity * sizeof(struct sl_pp_token_info));
+      assert(peek->tokens);
+   }
+   peek->tokens[peek->size++] = *out;
+   return 0;
+}
+
+void
+sl_pp_token_peek_commit(struct sl_pp_token_peek *peek)
+{
+   peek->size = 0;
+}
+
+int
+sl_pp_token_peek_to_buffer(const struct sl_pp_token_peek *peek,
+                           struct sl_pp_token_buffer *buffer)
+{
+   unsigned int i;
+
+   if (sl_pp_token_buffer_init(buffer, NULL)) {
+      return -1;
+   }
+   for (i = peek->size; i > 0; i--) {
+      sl_pp_token_buffer_unget(buffer, &peek->tokens[i - 1]);
+   }
+   return 0;
+}
+
+int
+sl_pp_token_peek_skip_white(struct sl_pp_token_peek *peek,
+                            struct sl_pp_token_info *out)
+{
+   if (sl_pp_token_peek_get(peek, out)) {
+      return -1;
+   }
+
+   while (out->token == SL_PP_WHITESPACE) {
+      if (sl_pp_token_peek_get(peek, out)) {
+         return -1;
+      }
+   }
+
+   return 0;
+}
diff --git a/src/glsl/pp/sl_pp_token_util.h b/src/glsl/pp/sl_pp_token_util.h
new file mode 100644 (file)
index 0000000..2a668ad
--- /dev/null
@@ -0,0 +1,103 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#ifndef SL_PP_TOKEN_UTIL_H
+#define SL_PP_TOKEN_UTIL_H
+
+#include <assert.h>
+#include <stdlib.h>
+#include "sl_pp_token.h"
+
+
+struct sl_pp_context;
+
+/*
+ * A token buffer allows one to get and unget a token
+ * from a preprocessor context.
+ */
+struct sl_pp_token_buffer {
+   struct sl_pp_context *context;
+   unsigned int size;
+   unsigned int capacity;
+   struct sl_pp_token_info *tokens;
+};
+
+int
+sl_pp_token_buffer_init(struct sl_pp_token_buffer *buffer,
+                        struct sl_pp_context *context);
+
+void
+sl_pp_token_buffer_destroy(struct sl_pp_token_buffer *buffer);
+
+int
+sl_pp_token_buffer_get(struct sl_pp_token_buffer *buffer,
+                       struct sl_pp_token_info *out);
+
+void
+sl_pp_token_buffer_unget(struct sl_pp_token_buffer *buffer,
+                         const struct sl_pp_token_info *in);
+
+int
+sl_pp_token_buffer_skip_white(struct sl_pp_token_buffer *buffer,
+                              struct sl_pp_token_info *out);
+
+
+/*
+ * A token peek allows one to get a number of tokens from a buffer
+ * and then either commit the operation or abort it,
+ * effectively ungetting the peeked tokens.
+ */
+struct sl_pp_token_peek {
+   struct sl_pp_token_buffer *buffer;
+   unsigned int size;
+   unsigned int capacity;
+   struct sl_pp_token_info *tokens;
+};
+
+int
+sl_pp_token_peek_init(struct sl_pp_token_peek *peek,
+                      struct sl_pp_token_buffer *buffer);
+
+void
+sl_pp_token_peek_destroy(struct sl_pp_token_peek *peek);
+
+int
+sl_pp_token_peek_get(struct sl_pp_token_peek *peek,
+                     struct sl_pp_token_info *out);
+
+void
+sl_pp_token_peek_commit(struct sl_pp_token_peek *peek);
+
+int
+sl_pp_token_peek_to_buffer(const struct sl_pp_token_peek *peek,
+                           struct sl_pp_token_buffer *buffer);
+
+int
+sl_pp_token_peek_skip_white(struct sl_pp_token_peek *peek,
+                            struct sl_pp_token_info *out);
+
+#endif /* SL_PP_TOKEN_UTIL_H */
diff --git a/src/glsl/pp/sl_pp_version.c b/src/glsl/pp/sl_pp_version.c
new file mode 100644 (file)
index 0000000..3c995b7
--- /dev/null
@@ -0,0 +1,161 @@
+/**************************************************************************
+ * 
+ * Copyright 2009 VMware, Inc.
+ * All Rights Reserved.
+ * 
+ * Permission is hereby granted, free of charge, to any person obtaining a
+ * copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sub license, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ * 
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial portions
+ * of the Software.
+ * 
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
+ * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NON-INFRINGEMENT.
+ * IN NO EVENT SHALL TUNGSTEN GRAPHICS AND/OR ITS SUPPLIERS BE LIABLE FOR
+ * ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT,
+ * TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE
+ * SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ * 
+ **************************************************************************/
+
+#include <stdlib.h>
+#include <string.h>
+#include "sl_pp_public.h"
+#include "sl_pp_context.h"
+
+
+int
+sl_pp_version(struct sl_pp_context *context,
+              unsigned int *version)
+{
+   struct sl_pp_token_peek peek;
+   unsigned int line = context->line;
+
+   /* Default values if `#version' is not present. */
+   *version = 110;
+
+   if (sl_pp_token_peek_init(&peek, &context->tokens)) {
+      return -1;
+   }
+
+   /* There can be multiple `#version' directives present.
+    * Accept the value of the last one.
+    */
+   for (;;) {
+      struct sl_pp_token_info input;
+      int found_hash = 0;
+      int found_version = 0;
+      int found_number = 0;
+      int found_end = 0;
+
+      /* Skip whitespace and newlines and seek for hash. */
+      while (!found_hash) {
+         if (sl_pp_token_peek_get(&peek, &input)) {
+            sl_pp_token_peek_destroy(&peek);
+            return -1;
+         }
+
+         switch (input.token) {
+         case SL_PP_NEWLINE:
+            line++;
+            break;
+
+         case SL_PP_WHITESPACE:
+            break;
+
+         case SL_PP_HASH:
+            found_hash = 1;
+            break;
+
+         default:
+            sl_pp_token_peek_destroy(&peek);
+            return 0;
+         }
+      }
+
+      /* Skip whitespace and seek for `version'. */
+      while (!found_version) {
+         if (sl_pp_token_peek_get(&peek, &input)) {
+            sl_pp_token_peek_destroy(&peek);
+            return -1;
+         }
+
+         switch (input.token) {
+         case SL_PP_WHITESPACE:
+            break;
+
+         case SL_PP_IDENTIFIER:
+            if (input.data.identifier != context->dict.version) {
+               sl_pp_token_peek_destroy(&peek);
+               return 0;
+            }
+            found_version = 1;
+            break;
+
+         default:
+            sl_pp_token_peek_destroy(&peek);
+            return 0;
+         }
+      }
+
+      sl_pp_token_peek_commit(&peek);
+
+      /* Skip whitespace and seek for version number. */
+      while (!found_number) {
+         if (sl_pp_token_buffer_get(&context->tokens, &input)) {
+            sl_pp_token_peek_destroy(&peek);
+            return -1;
+         }
+
+         switch (input.token) {
+         case SL_PP_WHITESPACE:
+            break;
+
+         case SL_PP_UINT:
+            *version = atoi(sl_pp_context_cstr(context, input.data._uint));
+            found_number = 1;
+            break;
+
+         default:
+            strcpy(context->error_msg, "expected version number after `#version'");
+            sl_pp_token_peek_destroy(&peek);
+            return -1;
+         }
+      }
+
+      /* Skip whitespace and seek for either newline or eof. */
+      while (!found_end) {
+         if (sl_pp_token_buffer_get(&context->tokens, &input)) {
+            sl_pp_token_peek_destroy(&peek);
+            return -1;
+         }
+
+         switch (input.token) {
+         case SL_PP_WHITESPACE:
+            break;
+
+         case SL_PP_NEWLINE:
+            line++;
+            /* pass thru */
+         case SL_PP_EOF:
+            context->line = line;
+            found_end = 1;
+            break;
+
+         default:
+            strcpy(context->error_msg, "expected end of line after version number");
+            sl_pp_token_peek_destroy(&peek);
+            return -1;
+         }
+      }
+   }
+
+   /* Should not get here. */
+}
index 6f58ad61617de2f6afcf152f4375af28d53d195f..a815f46b4a1551514d0f57460522b4b19142539d 100644 (file)
@@ -18,11 +18,10 @@ include sources.mak
 
 
 
-
-# Default: build dependencies, then asm_subdirs, then convenience
-# libs (.a) and finally the device drivers:
-default: depend asm_subdirs libmesa.a libmesagallium.a libglapi.a \
-       driver_subdirs
+# Default: build dependencies, then asm_subdirs, GLSL built-in lib,
+# then convenience libs (.a) and finally the device drivers:
+default: depend asm_subdirs glsl_builtin libmesa.a libmesagallium.a \
+       libglapi.a driver_subdirs
 
 
 
@@ -30,12 +29,12 @@ default: depend asm_subdirs libmesa.a libmesagallium.a libglapi.a \
 # Helper libraries used by many drivers:
 
 # Make archive of core mesa object files
-libmesa.a: $(MESA_OBJECTS)
-       @ $(MKLIB) -o mesa -static $(MESA_OBJECTS)
+libmesa.a: $(MESA_OBJECTS) $(GLSL_LIBS)
+       @ $(MKLIB) -o mesa -static $(MESA_OBJECTS) $(GLSL_LIBS)
 
 # Make archive of subset of core mesa object files for gallium
-libmesagallium.a: $(MESA_GALLIUM_OBJECTS)
-       @ $(MKLIB) -o mesagallium -static $(MESA_GALLIUM_OBJECTS)
+libmesagallium.a: $(MESA_GALLIUM_OBJECTS) $(GLSL_LIBS)
+       @ $(MKLIB) -o mesagallium -static $(MESA_GALLIUM_OBJECTS) $(GLSL_LIBS)
 
 # Make archive of gl* API dispatcher functions only
 libglapi.a: $(GLAPI_OBJECTS)
@@ -59,6 +58,12 @@ asm_subdirs:
        fi
 
 
+######################################################################
+# GLSL built-in library
+glsl_builtin:
+       (cd shader/slang/library && $(MAKE)) || exit 1 ;
+
+
 ######################################################################
 # Dependency generation
 
@@ -152,6 +157,7 @@ clean:
        -rm -f depend depend.bak libmesa.a libglapi.a
        -rm -f drivers/*/*.o
        -rm -f *.pc
+       -rm -f shader/slang/library/*_gc.h
        -@cd drivers/dri && $(MAKE) clean
        -@cd drivers/x11 && $(MAKE) clean
        -@cd drivers/osmesa && $(MAKE) clean
index 097c390a83e2625e402cc415fcfb2cb021b836ad..e894c6277d19a220df3df28daf7068770e557d18 100644 (file)
@@ -218,7 +218,6 @@ clean:
        -$(call UNLINK,vbo/*.o)
        -$(call UNLINK,shader/*.o)
        -$(call UNLINK,shader/slang/*.o)
-       -$(call UNLINK,shader/grammar/*.o)
        -$(call UNLINK,sparc/*.o)
        -$(call UNLINK,ppc/*.o)
        -$(call UNLINK,swrast/*.o)
index ca4a9afce5cdee581de2920e8995fcc9f560e33d..f4e0b98570da482f2a1b6d82ef94053fb60d9448 100644 (file)
@@ -189,7 +189,6 @@ if env['platform'] != 'winddk':
                'shader/arbprogparse.c',
                'shader/arbprogram.c',
                'shader/atifragshader.c',
-               'shader/grammar/grammar_mesa.c',
                'shader/hash_table.c',
                'shader/lex.yy.c',
                'shader/nvfragparse.c',
@@ -227,7 +226,6 @@ if env['platform'] != 'winddk':
                'shader/slang/slang_link.c',
                'shader/slang/slang_log.c',
                'shader/slang/slang_mem.c',
-               'shader/slang/slang_preprocess.c',
                'shader/slang/slang_print.c',
                'shader/slang/slang_simplify.c',
                'shader/slang/slang_storage.c',
@@ -338,7 +336,9 @@ if env['platform'] != 'winddk':
                # Add the dir containing the generated header (somewhere inside  the
                # build dir) to the include path  
                env.Append(CPPPATH = [matypes[0].dir])
-       
+
+       SConscript('shader/slang/library/SConscript')
+
        #
        # Libraries
        # 
index 4ca0e7bcc37712cc3341353b0f0d69ea70447e8e..9b271f85e92ca14e18bf73bcde25385975076f8e 100644 (file)
@@ -124,7 +124,6 @@ _mesa_init_driver_functions(struct dd_function_table *driver)
    driver->UnmapTexture = NULL;
    driver->TextureMemCpy = _mesa_memcpy; 
    driver->IsTextureResident = NULL;
-   driver->ActiveTexture = NULL;
    driver->UpdateTexturePalette = NULL;
 
    /* imaging */
index 0495d0e2767d900dd86d5aca31c6549fae9cfe76..c548ef3ad58fc35b709ce61c439ed38bea5251d4 100644 (file)
@@ -38,11 +38,11 @@ static void TAG(emit)(GLcontext *ctx, GLuint start, GLuint end)
 #endif
 
 #if (IND & (FFB_VB_RGBA_BIT))
-       col0 = VB->ColorPtr[0]->data;
-       col0_stride = VB->ColorPtr[0]->stride;
+       col0 = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+       col0_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
 #if (IND & (FFB_VB_TWOSIDE_BIT))
-       col1 = VB->ColorPtr[1]->data;
-       col1_stride = VB->ColorPtr[1]->stride;
+       col1 = VB->BackfaceColorPtr->data;
+       col1_stride = VB->BackfaceColorPtr->stride;
 #endif
 #endif
 
index 1b9fd169f46270349ef195df89ed53aaaa6f5c19..a03a93d132cf1c1099abefc3ab40100428d2e888 100644 (file)
@@ -53,13 +53,13 @@ static void gamma_emit( GLcontext *ctx, GLuint start, GLuint end)
    GLfloat (*tc0)[4] = 0;
    GLuint tc0_size = 0;
 
-   col = VB->ColorPtr[0]->data;
-   col_stride = VB->ColorPtr[0]->stride;
+   col = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+   col_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
 
    if (ctx->Texture.Unit[0]._ReallyEnabled) {
-      tc0_stride = VB->TexCoordPtr[0]->stride;
-      tc0 = VB->TexCoordPtr[0]->data;
-      tc0_size = VB->TexCoordPtr[0]->size;
+      tc0_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0]->stride;
+      tc0 = VB->AttribPtr[_TNL_ATTRIB_TEX0]->data;
+      tc0_size = VB->AttribPtr[_TNL_ATTRIB_TEX0]->size;
       coord = VB->ClipPtr->data;
       coord_stride = VB->ClipPtr->stride;
    } else {
index 645ebe305774410b90153bf7a31ceca0fa9ae888..acda7e70de0dc50ca4bf708700778825fa3203f9 100644 (file)
@@ -620,7 +620,7 @@ i830LineWidth(GLcontext * ctx, GLfloat widthf)
    DBG("%s\n", __FUNCTION__);
    
    width = (int) (widthf * 2);
-   CLAMP_SELF(width, 1, 15);
+   width = CLAMP(width, 1, 15);
 
    state5 = i830->state.Ctx[I830_CTXREG_STATE5] & ~FIXED_LINE_WIDTH_MASK;
    state5 |= (ENABLE_FIXED_LINE_WIDTH | FIXED_LINE_WIDTH(width));
@@ -639,7 +639,7 @@ i830PointSize(GLcontext * ctx, GLfloat size)
 
    DBG("%s\n", __FUNCTION__);
    
-   CLAMP_SELF(point_size, 1, 256);
+   point_size = CLAMP(point_size, 1, 256);
    I830_STATECHANGE(i830, I830_UPLOAD_CTX);
    i830->state.Ctx[I830_CTXREG_STATE5] &= ~FIXED_POINT_WIDTH_MASK;
    i830->state.Ctx[I830_CTXREG_STATE5] |= (ENABLE_FIXED_POINT_WIDTH |
index ce409b3a60c62a16d4b5d1da5ee563dbdcfd5668..27c5aa1e08590258b895d4b660596f5763c4d4f7 100644 (file)
@@ -27,6 +27,7 @@
 
 #include "main/mtypes.h"
 #include "main/enums.h"
+#include "main/colormac.h"
 
 #include "intel_mipmap_tree.h"
 #include "intel_tex.h"
@@ -308,11 +309,10 @@ i830_update_tex_unit(struct intel_context *intel, GLuint unit, GLuint ss3)
    CLAMPED_FLOAT_TO_UBYTE(border[2], tObj->BorderColor[2]);
    CLAMPED_FLOAT_TO_UBYTE(border[3], tObj->BorderColor[3]);
 
-   state[I830_TEXREG_TM0S4] = INTEL_PACKCOLOR8888(border[0],
-                                                  border[1],
-                                                  border[2],
-                                                  border[3]);
-
+   state[I830_TEXREG_TM0S4] = PACK_COLOR_8888(border[3],
+                                             border[0],
+                                             border[1],
+                                             border[2]);
 
    I830_ACTIVESTATE(i830, I830_UPLOAD_TEX(unit), GL_TRUE);
    /* memcmp was already disabled, but definitely won't work as the
index e8c8d5a0486925c2076d42f171654ba391860c9c..1e3c8301d8d183270e8be0857fe3b6d435492947 100644 (file)
@@ -126,7 +126,7 @@ i830_render_start(struct intel_context *intel)
 
       for (i = 0; i < I830_TEX_UNITS; i++) {
          if (RENDERINPUTS_TEST(index_bitset, _TNL_ATTRIB_TEX(i))) {
-            GLuint sz = VB->TexCoordPtr[i]->size;
+            GLuint sz = VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]->size;
             GLuint emit;
             GLuint mcs = (i830->state.Tex[i][I830_TEXREG_MCS] &
                           ~TEXCOORDTYPE_MASK);
@@ -714,9 +714,6 @@ i830_new_batch(struct intel_context *intel)
 {
    struct i830_context *i830 = i830_context(&intel->ctx);
    i830->state.emitted = 0;
-
-   /* Check that we didn't just wrap our batchbuffer at a bad time. */
-   assert(!intel->no_batch_wrap);
 }
 
 static void 
index 7d4c7cfbabbdd1a1d6e346f7f826c3242a5c30b7..0485be2cc1feafd502ed0cf2ec0778de453e1da9 100644 (file)
@@ -143,6 +143,9 @@ i915CreateContext(const __GLcontextModes * mesaVis,
    ctx->Const.MaxTextureImageUnits = I915_TEX_UNITS;
    ctx->Const.MaxTextureCoordUnits = I915_TEX_UNITS;
    ctx->Const.MaxVarying = I915_TEX_UNITS;
+   ctx->Const.MaxCombinedTextureImageUnits =
+      ctx->Const.MaxVertexTextureImageUnits +
+      ctx->Const.MaxTextureImageUnits;
 
    /* Advertise the full hardware capabilities.  The new memory
     * manager should cope much better with overload situations:
index d9c61446f52d8e8bcbafceafe5019a60c0111011..a273bd28ea1fe8d8e97a7898de0d3fda3c63b839 100644 (file)
@@ -663,7 +663,7 @@ upload_program(struct i915_fragment_program *p)
                         A0_MOV,
                         get_result_vector(p, inst),
                         get_result_flags(inst), 0,
-                        swizzle(src0, ZERO, ZERO, ZERO, ZERO), 0, 0);
+                        swizzle(tmp, ZERO, ZERO, ZERO, ZERO), 0, 0);
 
       case OPCODE_POW:
          src0 = src_vector(p, &inst->SrcReg[0], program);
@@ -1301,7 +1301,7 @@ i915ValidateFragmentProgram(struct i915_context *i915)
 
    for (i = 0; i < p->ctx->Const.MaxTextureCoordUnits; i++) {
       if (inputsRead & FRAG_BIT_TEX(i)) {
-         int sz = VB->TexCoordPtr[i]->size;
+         int sz = VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]->size;
 
          s2 &= ~S2_TEXCOORD_FMT(i, S2_TEXCOORD_FMT0_MASK);
          s2 |= S2_TEXCOORD_FMT(i, SZ_TO_HW(sz));
index cc98d125dbcffee2ab6402d3e010eb5c80d5edd1..9d7a9e1dfeaa72822d19f727a6db401019b8a393 100644 (file)
@@ -571,7 +571,7 @@ i915LineWidth(GLcontext * ctx, GLfloat widthf)
    DBG("%s\n", __FUNCTION__);
    
    width = (int) (widthf * 2);
-   CLAMP_SELF(width, 1, 0xf);
+   width = CLAMP(width, 1, 0xf);
    lis4 |= width << S4_LINE_WIDTH_SHIFT;
 
    if (lis4 != i915->state.Ctx[I915_CTXREG_LIS4]) {
@@ -589,7 +589,7 @@ i915PointSize(GLcontext * ctx, GLfloat size)
 
    DBG("%s\n", __FUNCTION__);
    
-   CLAMP_SELF(point_size, 1, 255);
+   point_size = CLAMP(point_size, 1, 255);
    lis4 |= point_size << S4_POINT_WIDTH_SHIFT;
 
    if (lis4 != i915->state.Ctx[I915_CTXREG_LIS4]) {
index de258484d6cbb12c26290b034b153452fe919d4c..221bf033327ca19a001c139506e22415581ad0f6 100644 (file)
@@ -28,6 +28,7 @@
 #include "main/mtypes.h"
 #include "main/enums.h"
 #include "main/macros.h"
+#include "main/colormac.h"
 
 #include "intel_mipmap_tree.h"
 #include "intel_tex.h"
@@ -356,15 +357,15 @@ i915_update_tex_unit(struct intel_context *intel, GLuint unit, GLuint ss3)
        * R channel, while the hardware uses A.  Spam R into all the channels
        * for safety.
        */
-      state[I915_TEXREG_SS4] = INTEL_PACKCOLOR8888(border[0],
-                                                  border[0],
-                                                  border[0],
-                                                  border[0]);
+      state[I915_TEXREG_SS4] = PACK_COLOR_8888(border[0],
+                                              border[0],
+                                              border[0],
+                                              border[0]);
    } else {
-      state[I915_TEXREG_SS4] = INTEL_PACKCOLOR8888(border[0],
-                                                  border[1],
-                                                  border[2],
-                                                  border[3]);
+      state[I915_TEXREG_SS4] = PACK_COLOR_8888(border[3],
+                                              border[0],
+                                              border[1],
+                                              border[2]);
    }
 
 
index ff97e5a944f18d2d9299ec497b6e9f611f84a2ac..9f7635a9538452f4222faa52e33e4639a1469ef8 100644 (file)
@@ -611,7 +611,7 @@ i915_state_draw_region(struct intel_context *intel,
     * the value of this bit, the pipeline needs to be MI_FLUSHed.  And it
     * can only be set when a depth buffer is already defined.
     */
-   if (IS_945(intel->intelScreen->deviceID) && intel->use_early_z &&
+   if (intel->is_945 && intel->use_early_z &&
        depth_region->tiling != I915_TILING_NONE)
       value |= CLASSIC_EARLY_DEPTH;
 
@@ -667,9 +667,6 @@ i915_new_batch(struct intel_context *intel)
     * difficulties associated with them (physical address requirements).
     */
    i915->state.emitted = 0;
-
-   /* Check that we didn't just wrap our batchbuffer at a bad time. */
-   assert(!intel->no_batch_wrap);
 }
 
 static void 
index bc527aae47a8c165860e63b080ae917b96e706cb..63c5ae96dc7a62c7a087c17b62798d79a934ad08 100644 (file)
@@ -221,7 +221,7 @@ void intel_flush_prim(struct intel_context *intel)
    intel->prim.count = 0;
    offset = intel->prim.start_offset;
    intel->prim.start_offset = intel->prim.current_offset;
-   if (!IS_9XX(intel->intelScreen->deviceID))
+   if (!intel->gen >= 3)
       intel->prim.start_offset = ALIGN(intel->prim.start_offset, 128);
    intel->prim.flush = NULL;
 
@@ -251,7 +251,7 @@ void intel_flush_prim(struct intel_context *intel)
          intel->vertex_size * 4);
 #endif
 
-   if (IS_9XX(intel->intelScreen->deviceID)) {
+   if (intel->gen >= 3) {
       BEGIN_BATCH(5, LOOP_CLIPRECTS);
       OUT_BATCH(_3DSTATE_LOAD_STATE_IMMEDIATE_1 |
                I1_LOAD_S(0) | I1_LOAD_S(1) | 1);
@@ -1250,81 +1250,6 @@ union fi
    GLint i;
 };
 
-
-/**********************************************************************/
-/*             Used only with the metaops callbacks.                  */
-/**********************************************************************/
-static void
-intel_meta_draw_poly(struct intel_context *intel,
-                     GLuint n,
-                     GLfloat xy[][2],
-                     GLfloat z, GLuint color, GLfloat tex[][2])
-{
-   union fi *vb;
-   GLint i;
-   unsigned int saved_vertex_size = intel->vertex_size;
-
-   LOCK_HARDWARE(intel);
-
-   intel->vertex_size = 6;
-
-   /* All 3d primitives should be emitted with LOOP_CLIPRECTS,
-    * otherwise the drawing origin (DR4) might not be set correctly.
-    */
-   intel_set_prim(intel, PRIM3D_TRIFAN);
-   vb = (union fi *) intel_get_prim_space(intel, n);
-
-   for (i = 0; i < n; i++) {
-      vb[0].f = xy[i][0];
-      vb[1].f = xy[i][1];
-      vb[2].f = z;
-      vb[3].i = color;
-      vb[4].f = tex[i][0];
-      vb[5].f = tex[i][1];
-      vb += 6;
-   }
-
-   INTEL_FIREVERTICES(intel);
-
-   intel->vertex_size = saved_vertex_size;
-
-   UNLOCK_HARDWARE(intel);
-}
-
-static void
-intel_meta_draw_quad(struct intel_context *intel,
-                     GLfloat x0, GLfloat x1,
-                     GLfloat y0, GLfloat y1,
-                     GLfloat z,
-                     GLuint color,
-                     GLfloat s0, GLfloat s1, GLfloat t0, GLfloat t1)
-{
-   GLfloat xy[4][2];
-   GLfloat tex[4][2];
-
-   xy[0][0] = x0;
-   xy[0][1] = y0;
-   xy[1][0] = x1;
-   xy[1][1] = y0;
-   xy[2][0] = x1;
-   xy[2][1] = y1;
-   xy[3][0] = x0;
-   xy[3][1] = y1;
-
-   tex[0][0] = s0;
-   tex[0][1] = t0;
-   tex[1][0] = s1;
-   tex[1][1] = t0;
-   tex[2][0] = s1;
-   tex[2][1] = t1;
-   tex[3][0] = s0;
-   tex[3][1] = t1;
-
-   intel_meta_draw_poly(intel, 4, xy, z, color, tex);
-}
-
-
-
 /**********************************************************************/
 /*                            Initialization.                         */
 /**********************************************************************/
@@ -1333,7 +1258,6 @@ intel_meta_draw_quad(struct intel_context *intel,
 void
 intelInitTriFuncs(GLcontext * ctx)
 {
-   struct intel_context *intel = intel_context(ctx);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    static int firsttime = 1;
 
@@ -1350,6 +1274,4 @@ intelInitTriFuncs(GLcontext * ctx)
    tnl->Driver.Render.BuildVertices = _tnl_build_vertices;
    tnl->Driver.Render.CopyPV = _tnl_copy_pv;
    tnl->Driver.Render.Interp = _tnl_interp;
-
-   intel->vtbl.meta_draw_quad = intel_meta_draw_quad;
 }
index dbd10a5297480ab8817e7113f589d8f631e96270..af1d975de95cdf477d3c82f896216a6c719fae86 100644 (file)
@@ -50,6 +50,7 @@
 static void compile_clip_prog( struct brw_context *brw,
                             struct brw_clip_prog_key *key )
 {
+   struct intel_context *intel = &brw->intel;
    struct brw_clip_compile c;
    const GLuint *program;
    GLuint program_size;
@@ -65,14 +66,13 @@ static void compile_clip_prog( struct brw_context *brw,
    c.func.single_program_flow = 1;
 
    c.key = *key;
-   c.need_ff_sync = BRW_IS_IGDNG(brw);
 
    /* Need to locate the two positions present in vertex + header.
     * These are currently hardcoded:
     */
    c.header_position_offset = ATTR_SIZE;
 
-   if (BRW_IS_IGDNG(brw))
+   if (intel->is_ironlake)
        delta = 3 * REG_SIZE;
    else
        delta = REG_SIZE;
@@ -85,7 +85,7 @@ static void compile_clip_prog( struct brw_context *brw,
 
    c.nr_attrs = brw_count_bits(c.key.attrs);
    
-   if (BRW_IS_IGDNG(brw))
+   if (intel->is_ironlake)
        c.nr_regs = (c.nr_attrs + 1) / 2 + 3;  /* are vertices packed, or reg-aligned? */
    else
        c.nr_regs = (c.nr_attrs + 1) / 2 + 1;  /* are vertices packed, or reg-aligned? */
@@ -143,7 +143,8 @@ static void compile_clip_prog( struct brw_context *brw,
  */
 static void upload_clip_prog(struct brw_context *brw)
 {
-   GLcontext *ctx = &brw->intel.ctx;
+   struct intel_context *intel = &brw->intel;
+   GLcontext *ctx = &intel->ctx;
    struct brw_clip_prog_key key;
 
    memset(&key, 0, sizeof(key));
@@ -160,7 +161,7 @@ static void upload_clip_prog(struct brw_context *brw)
    /* _NEW_TRANSFORM */
    key.nr_userclip = brw_count_bits(ctx->Transform.ClipPlanesEnabled);
 
-   if (BRW_IS_IGDNG(brw))
+   if (intel->is_ironlake)
        key.clip_mode = BRW_CLIPMODE_KERNEL_CLIP;
    else
        key.clip_mode = BRW_CLIPMODE_NORMAL;
index 1c6825510a2995e13b2d0c6faa370b4f24755cdc..d71bac7f617512fceb6bdc37a55aea96ae55f9e9 100644 (file)
@@ -118,7 +118,6 @@ struct brw_clip_compile {
 
    GLuint header_position_offset;
    GLuint offset[VERT_ATTRIB_MAX];
-   GLboolean need_ff_sync;
 };
 
 #define ATTR_SIZE  (4*4)
index fa9648f50f312dadfbf2dd4c4cd9a9d3d03f0d69..afc0b11049406575fc09944c85d939171ab886db 100644 (file)
@@ -46,6 +46,7 @@
 
 static void brw_clip_line_alloc_regs( struct brw_clip_compile *c )
 {
+   struct intel_context *intel = &c->func.brw->intel;
    GLuint i = 0,j;
 
    /* Register usage is static, precompute here:
@@ -85,7 +86,7 @@ static void brw_clip_line_alloc_regs( struct brw_clip_compile *c )
       i++;
    }
 
-   if (c->need_ff_sync) {
+   if (intel->needs_ff_sync) {
       c->reg.ff_sync = retype(brw_vec1_grf(i, 0), BRW_REGISTER_TYPE_UD);
       i++;
    }
@@ -126,6 +127,7 @@ static void brw_clip_line_alloc_regs( struct brw_clip_compile *c )
 static void clip_and_emit_line( struct brw_clip_compile *c )
 {
    struct brw_compile *p = &c->func;
+   struct brw_context *brw = p->brw;
    struct brw_indirect vtx0     = brw_indirect(0, 0);
    struct brw_indirect vtx1      = brw_indirect(1, 0);
    struct brw_indirect newvtx0   = brw_indirect(2, 0);
@@ -152,7 +154,7 @@ static void clip_and_emit_line( struct brw_clip_compile *c )
    brw_clip_init_clipmask(c);
 
    /* -ve rhw workaround */
-   if (BRW_IS_965(p->brw)) {
+   if (brw->has_negative_rhw_bug) {
       brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
       brw_AND(p, brw_null_reg(), get_element_ud(c->reg.R0, 2),
               brw_imm_ud(1<<20));
@@ -189,7 +191,7 @@ static void clip_and_emit_line( struct brw_clip_compile *c )
               * Both can be negative on GM965/G965 due to RHW workaround
               * if so, this object should be rejected.
               */
-             if (BRW_IS_965(p->brw)) {
+             if (brw->has_negative_rhw_bug) {
                  brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_LE, c->reg.dp0, brw_imm_f(0.0));
                  is_neg2 = brw_IF(p, BRW_EXECUTE_1);
                  {
@@ -214,7 +216,7 @@ static void clip_and_emit_line( struct brw_clip_compile *c )
 
              /* If both are positive, do nothing */
              /* Only on GM965/G965 */
-             if (BRW_IS_965(p->brw)) {
+             if (brw->has_negative_rhw_bug) {
                  brw_CMP(p, vec1(brw_null_reg()), BRW_CONDITIONAL_L, c->reg.dp0, brw_imm_f(0.0));
                  is_neg2 = brw_IF(p, BRW_EXECUTE_1);
              }
@@ -229,7 +231,7 @@ static void clip_and_emit_line( struct brw_clip_compile *c )
                  brw_set_predicate_control(p, BRW_PREDICATE_NONE);
              }
 
-             if (BRW_IS_965(p->brw)) {
+             if (brw->has_negative_rhw_bug) {
                  brw_ENDIF(p, is_neg2);
              }
          }
index 234b3744bfc6717e986b8c41d2db72ec1992b8f3..c8f24a94e4d5feb574ccfd0fbb6ec7f48ef97557 100644 (file)
@@ -74,6 +74,7 @@ static dri_bo *
 clip_unit_create_from_key(struct brw_context *brw,
                          struct brw_clip_unit_key *key)
 {
+   struct intel_context *intel = &brw->intel;
    struct brw_clip_unit_state clip;
    dri_bo *bo;
 
@@ -105,7 +106,7 @@ clip_unit_create_from_key(struct brw_context *brw,
       /* Although up to 16 concurrent Clip threads are allowed on IGDNG, 
        * only 2 threads can output VUEs at a time.
        */
-      if (BRW_IS_IGDNG(brw))
+      if (intel->is_ironlake)
          clip.thread4.max_threads = 16 - 1;        
       else
          clip.thread4.max_threads = 2 - 1;
@@ -130,7 +131,7 @@ clip_unit_create_from_key(struct brw_context *brw,
    clip.clip5.api_mode = BRW_CLIP_API_OGL;
    clip.clip5.clip_mode = key->clip_mode;
 
-   if (BRW_IS_G4X(brw))
+   if (intel->is_g4x)
       clip.clip5.negative_w_clip_test = 1;
 
    clip.clip6.clipper_viewport_state_ptr = 0;
index cf79224be4385ae31bbedbc15cf82fd857271ce5..cfbb8f2686aee6421471c5f2d23b439ba14df8cb 100644 (file)
@@ -51,6 +51,7 @@ static void release_tmps( struct brw_clip_compile *c )
 void brw_clip_tri_alloc_regs( struct brw_clip_compile *c, 
                              GLuint nr_verts )
 {
+   struct intel_context *intel = &c->func.brw->intel;
    GLuint i = 0,j;
 
    /* Register usage is static, precompute here:
@@ -78,7 +79,7 @@ void brw_clip_tri_alloc_regs( struct brw_clip_compile *c,
       for (j = 0; j < 3; j++) {
         GLuint delta = c->nr_attrs*16 + 32;
 
-         if (BRW_IS_IGDNG(c->func.brw))
+         if (intel->is_ironlake)
              delta = c->nr_attrs * 16 + 32 * 3;
 
         brw_MOV(&c->func, byte_offset(c->reg.vertex[j], delta), brw_imm_f(0));
@@ -119,7 +120,7 @@ void brw_clip_tri_alloc_regs( struct brw_clip_compile *c,
       i++;
    }
 
-   if (c->need_ff_sync) {
+   if (intel->needs_ff_sync) {
       c->reg.ff_sync = retype(brw_vec1_grf(i, 0), BRW_REGISTER_TYPE_UD);
       i++;
    }
@@ -571,6 +572,7 @@ void brw_emit_tri_clip( struct brw_clip_compile *c )
 {
    struct brw_instruction *neg_rhw;
    struct brw_compile *p = &c->func;
+   struct brw_context *brw = p->brw;
    brw_clip_tri_alloc_regs(c, 3 + c->key.nr_userclip + 6);
    brw_clip_tri_init_vertices(c);
    brw_clip_init_clipmask(c);
@@ -578,7 +580,7 @@ void brw_emit_tri_clip( struct brw_clip_compile *c )
 
    /* if -ve rhw workaround bit is set, 
       do cliptest */
-   if (BRW_IS_965(p->brw)) {
+   if (brw->has_negative_rhw_bug) {
       brw_set_conditionalmod(p, BRW_CONDITIONAL_NZ);
       brw_AND(p, brw_null_reg(), get_element_ud(c->reg.R0, 2), 
               brw_imm_ud(1<<20));
index 5a73abdfee9d994eaed52b2d5d2d0ef34e1efb29..86fed59fa4b89dfbc90b7747a3407d72d7a38b95 100644 (file)
@@ -135,6 +135,7 @@ void brw_clip_interp_vertex( struct brw_clip_compile *c,
                             GLboolean force_edgeflag)
 {
    struct brw_compile *p = &c->func;
+   struct intel_context *intel = &p->brw->intel;
    struct brw_reg tmp = get_tmp(c);
    GLuint i;
 
@@ -142,7 +143,7 @@ void brw_clip_interp_vertex( struct brw_clip_compile *c,
     */
    /*
     * After CLIP stage, only first 256 bits of the VUE are read
-    * back on IGDNG, so needn't change it
+    * back on Ironlake, so needn't change it
     */
    brw_copy_indirect_to_indirect(p, dest_ptr, v0_ptr, 1);
       
@@ -151,7 +152,7 @@ void brw_clip_interp_vertex( struct brw_clip_compile *c,
    for (i = 0; i < c->nr_attrs; i++) {
       GLuint delta = i*16 + 32;
 
-      if (BRW_IS_IGDNG(p->brw))
+      if (intel->is_ironlake)
           delta = i * 16 + 32 * 3;
 
       if (delta == c->offset[VERT_RESULT_EDGE]) {
@@ -185,7 +186,7 @@ void brw_clip_interp_vertex( struct brw_clip_compile *c,
    if (i & 1) {
       GLuint delta = i*16 + 32;
 
-      if (BRW_IS_IGDNG(p->brw))
+      if (intel->is_ironlake)
           delta = i * 16 + 32 * 3;
 
       brw_MOV(p, deref_4f(dest_ptr, delta), brw_imm_f(0));
@@ -359,7 +360,9 @@ void brw_clip_init_clipmask( struct brw_clip_compile *c )
 
 void brw_clip_ff_sync(struct brw_clip_compile *c)
 {
-    if (c->need_ff_sync) {
+    struct intel_context *intel = &c->func.brw->intel;
+
+    if (intel->needs_ff_sync) {
         struct brw_compile *p = &c->func;
         struct brw_instruction *need_ff_sync;
 
@@ -388,7 +391,9 @@ void brw_clip_ff_sync(struct brw_clip_compile *c)
 
 void brw_clip_init_ff_sync(struct brw_clip_compile *c)
 {
-    if (c->need_ff_sync) {
+    struct intel_context *intel = &c->func.brw->intel;
+
+    if (intel->needs_ff_sync) {
        struct brw_compile *p = &c->func;
         
         brw_MOV(p, c->reg.ff_sync, brw_imm_ud(0));
index 48685c087b82b95a01ad0f786ff8fe78cd3956e5..d8af2c512ba8c63fb014c87bc085d804dfc8c9d0 100644 (file)
@@ -111,6 +111,9 @@ GLboolean brwCreateContext( const __GLcontextModes *mesaVis,
    ctx->Const.MaxTextureUnits = MIN2(ctx->Const.MaxTextureCoordUnits,
                                      ctx->Const.MaxTextureImageUnits);
    ctx->Const.MaxVertexTextureImageUnits = 0; /* no vertex shader textures */
+   ctx->Const.MaxCombinedTextureImageUnits =
+      ctx->Const.MaxVertexTextureImageUnits +
+      ctx->Const.MaxTextureImageUnits;
 
    /* Mesa limits textures to 4kx4k; it would be nice to fix that someday
     */
@@ -155,6 +158,38 @@ GLboolean brwCreateContext( const __GLcontextModes *mesaVis,
       MIN2(ctx->Const.FragmentProgram.MaxNativeParameters,
           ctx->Const.FragmentProgram.MaxEnvParams);
 
+   if (intel->is_ironlake || intel->is_g4x) {
+      brw->CMD_VF_STATISTICS = CMD_VF_STATISTICS_GM45;
+      brw->CMD_PIPELINE_SELECT = CMD_PIPELINE_SELECT_GM45;
+      brw->has_surface_tile_offset = GL_TRUE;
+      brw->has_compr4 = GL_TRUE;
+      brw->has_aa_line_parameters = GL_TRUE;
+  } else {
+      brw->CMD_VF_STATISTICS = CMD_VF_STATISTICS_965;
+      brw->CMD_PIPELINE_SELECT = CMD_PIPELINE_SELECT_965;
+   }
+
+   /* WM maximum threads is number of EUs times number of threads per EU. */
+   if (intel->is_ironlake) {
+      brw->urb.size = 1024;
+      brw->vs_max_threads = 72;
+      brw->wm_max_threads = 12 * 6;
+   } else if (intel->is_g4x) {
+      brw->urb.size = 384;
+      brw->vs_max_threads = 32;
+      brw->wm_max_threads = 10 * 5;
+   } else {
+      brw->urb.size = 256;
+      brw->vs_max_threads = 16;
+      brw->wm_max_threads = 8 * 4;
+      brw->has_negative_rhw_bug = GL_TRUE;
+   }
+
+   if (INTEL_DEBUG & DEBUG_SINGLE_THREAD) {
+      brw->vs_max_threads = 1;
+      brw->wm_max_threads = 1;
+   }
+
    brw_init_state( brw );
 
    brw->state.dirty.mesa = ~0;
index fded47aa2ff7478129cd570f9e1d0c6792e2b9d3..ea5503e2fe14a30c065a763fcee29f0aef1b7a6a 100644 (file)
@@ -172,8 +172,8 @@ struct brw_fragment_program {
    GLuint id;  /**< serial no. to identify frag progs, never re-used */
    GLboolean isGLSL;  /**< really, any IF/LOOP/CONT/BREAK instructions */
 
-   dri_bo *const_buffer;    /** Program constant buffer/surface */
    GLboolean use_const_buffer;
+   dri_bo *const_buffer;    /** Program constant buffer/surface */
 
    /** for debugging, which texture units are referenced */
    GLbitfield tex_units_used;
@@ -438,8 +438,11 @@ struct brw_context
    GLuint primitive;
 
    GLboolean emit_state_always;
-   GLboolean no_batch_wrap;
-
+   GLboolean has_surface_tile_offset;
+   GLboolean has_compr4;
+   GLboolean has_negative_rhw_bug;
+   GLboolean has_aa_line_parameters;
+;
    struct {
       struct brw_state_flags dirty;
 
@@ -515,6 +518,12 @@ struct brw_context
     */
    GLuint next_free_page;
 
+   /* hw-dependent 3DSTATE_VF_STATISTICS opcode */
+   uint32_t CMD_VF_STATISTICS;
+   /* hw-dependent 3DSTATE_PIPELINE_SELECT opcode */
+   uint32_t CMD_PIPELINE_SELECT;
+   int vs_max_threads;
+   int wm_max_threads;
 
    /* BRW_NEW_URB_ALLOCATIONS:
     */
@@ -542,6 +551,7 @@ struct brw_context
       GLuint clip_start;
       GLuint sf_start;
       GLuint cs_start;
+      GLuint size; /* Hardware URB size, in KB. */
    } urb;
 
    
index c19510bbd4206a72065fb1b1261c1cf4a490a7eb..ea0d7e05d4184254b66087962c98fed05bb5f18b 100644 (file)
 
 #include "intel_chipset.h"
 
-#define BRW_IS_G4X(brw)         (IS_G4X((brw)->intel.intelScreen->deviceID))
-#define BRW_IS_IGDNG(brw)         (IS_IGDNG((brw)->intel.intelScreen->deviceID))
-#define BRW_IS_965(brw)         (!(BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw)))
-#define CMD_PIPELINE_SELECT(brw)        ((BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw)) ? CMD_PIPELINE_SELECT_GM45 : CMD_PIPELINE_SELECT_965)
-#define CMD_VF_STATISTICS(brw)          ((BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw)) ? CMD_VF_STATISTICS_GM45 : CMD_VF_STATISTICS_965)
-#define URB_SIZES(brw)                  (BRW_IS_IGDNG(brw) ? 1024 : \
-                                         (BRW_IS_G4X(brw) ? 384 : 256))  /* 512 bit units */
-
 #endif
index 9fef230507fa5970c4513616e595a948fb8af411..47629dfa3d65ba0e1378f8fb4444eabe9c2b611c 100644 (file)
@@ -239,7 +239,7 @@ char *imm_encoding[8] = {
     [2] = "UW",
     [3] = "W",
     [5] = "VF",
-    [5] = "V",
+    [6] = "V",
     [7] = "F"
 };
 
index 8bcb6083f71400306d3fbd12bcb966d47d51e0b7..7ad860898fc20bf45d1ec654634482acbc8027c8 100644 (file)
@@ -145,7 +145,7 @@ static void brw_emit_prim(struct brw_context *brw,
    prim_packet.base_vert_location = prim->basevertex;
 
    /* Can't wrap here, since we rely on the validated state. */
-   brw->no_batch_wrap = GL_TRUE;
+   intel->no_batch_wrap = GL_TRUE;
 
    /* If we're set to always flush, do it before and after the primitive emit.
     * We want to catch both missed flushes that hurt instruction/state cache
@@ -163,7 +163,7 @@ static void brw_emit_prim(struct brw_context *brw,
       intel_batchbuffer_emit_mi_flush(intel->batch);
    }
 
-   brw->no_batch_wrap = GL_FALSE;
+   intel->no_batch_wrap = GL_FALSE;
 }
 
 static void brw_merge_inputs( struct brw_context *brw,
index 271a88dae0b6206581c658c0d4b4a3c6f0202704..2c9902c90f0ee10be31c6f048c9e4c48ccf6d46b 100644 (file)
@@ -243,14 +243,6 @@ static void wrap_buffers( struct brw_context *brw,
       dri_bo_unreference(brw->vb.upload.bo);
    brw->vb.upload.bo = dri_bo_alloc(brw->intel.bufmgr, "temporary VBO",
                                    size, 1);
-
-   /* Set the internal VBO\ to no-backing-store.  We only use them as a
-    * temporary within a brw_try_draw_prims while the lock is held.
-    */
-   /* DON'T DO THIS AS IF WE HAVE TO RE-ORG MEMORY WE NEED SOMEWHERE WITH
-      FAKE TO PUSH THIS STUFF */
-//   if (!brw->intel.ttm)
-//      dri_bo_fake_disable_backing_store(brw->vb.upload.bo, NULL, NULL);
 }
 
 static void get_space( struct brw_context *brw,
@@ -535,7 +527,7 @@ static void brw_emit_vertices(struct brw_context *brw)
       OUT_RELOC(input->bo,
                I915_GEM_DOMAIN_VERTEX, 0,
                input->offset);
-      if (BRW_IS_IGDNG(brw)) {
+      if (intel->is_ironlake) {
         OUT_RELOC(input->bo,
                   I915_GEM_DOMAIN_VERTEX, 0,
                   input->bo->size - 1);
@@ -571,7 +563,7 @@ static void brw_emit_vertices(struct brw_context *brw)
                (format << BRW_VE0_FORMAT_SHIFT) |
                (0 << BRW_VE0_SRC_OFFSET_SHIFT));
 
-      if (BRW_IS_IGDNG(brw))
+      if (intel->is_ironlake)
           OUT_BATCH((comp0 << BRW_VE1_COMPONENT_0_SHIFT) |
                     (comp1 << BRW_VE1_COMPONENT_1_SHIFT) |
                     (comp2 << BRW_VE1_COMPONENT_2_SHIFT) |
index 7ceabba288a771509d69e5a16af27cec824725b4..3413c969281e6fbef6acf2bd5eca5a6013be3d27 100644 (file)
@@ -252,9 +252,10 @@ static void brw_set_math_message( struct brw_context *brw,
                                  GLboolean saturate,
                                  GLuint dataType )
 {
+   struct intel_context *intel = &brw->intel;
    brw_set_src1(insn, brw_imm_d(0));
 
-   if (BRW_IS_IGDNG(brw)) {
+   if (intel->is_ironlake) {
        insn->bits3.math_igdng.function = function;
        insn->bits3.math_igdng.int_type = integer_type;
        insn->bits3.math_igdng.precision = low_precision;
@@ -319,9 +320,10 @@ static void brw_set_urb_message( struct brw_context *brw,
                                 GLuint offset,
                                 GLuint swizzle_control )
 {
+    struct intel_context *intel = &brw->intel;
     brw_set_src1(insn, brw_imm_d(0));
 
-    if (BRW_IS_IGDNG(brw)) {
+    if (intel->is_ironlake) {
         insn->bits3.urb_igdng.opcode = 0;      /* ? */
         insn->bits3.urb_igdng.offset = offset;
         insn->bits3.urb_igdng.swizzle_control = swizzle_control;
@@ -358,9 +360,10 @@ static void brw_set_dp_write_message( struct brw_context *brw,
                                      GLuint response_length,
                                      GLuint end_of_thread )
 {
+   struct intel_context *intel = &brw->intel;
    brw_set_src1(insn, brw_imm_d(0));
 
-   if (BRW_IS_IGDNG(brw)) {
+   if (intel->is_ironlake) {
        insn->bits3.dp_write_igdng.binding_table_index = binding_table_index;
        insn->bits3.dp_write_igdng.msg_control = msg_control;
        insn->bits3.dp_write_igdng.pixel_scoreboard_clear = pixel_scoreboard_clear;
@@ -395,9 +398,10 @@ static void brw_set_dp_read_message( struct brw_context *brw,
                                      GLuint response_length,
                                      GLuint end_of_thread )
 {
+   struct intel_context *intel = &brw->intel;
    brw_set_src1(insn, brw_imm_d(0));
 
-   if (BRW_IS_IGDNG(brw)) {
+   if (intel->is_ironlake) {
        insn->bits3.dp_read_igdng.binding_table_index = binding_table_index;
        insn->bits3.dp_read_igdng.msg_control = msg_control;
        insn->bits3.dp_read_igdng.msg_type = msg_type;
@@ -433,10 +437,11 @@ static void brw_set_sampler_message(struct brw_context *brw,
                                     GLuint header_present,
                                     GLuint simd_mode)
 {
+   struct intel_context *intel = &brw->intel;
    assert(eot == 0);
    brw_set_src1(insn, brw_imm_d(0));
 
-   if (BRW_IS_IGDNG(brw)) {
+   if (intel->is_ironlake) {
       insn->bits3.sampler_igdng.binding_table_index = binding_table_index;
       insn->bits3.sampler_igdng.sampler = sampler;
       insn->bits3.sampler_igdng.msg_type = msg_type;
@@ -447,7 +452,7 @@ static void brw_set_sampler_message(struct brw_context *brw,
       insn->bits3.sampler_igdng.end_of_thread = eot;
       insn->bits2.send_igdng.sfid = BRW_MESSAGE_TARGET_SAMPLER;
       insn->bits2.send_igdng.end_of_thread = eot;
-   } else if (BRW_IS_G4X(brw)) {
+   } else if (intel->is_g4x) {
       insn->bits3.sampler_g4x.binding_table_index = binding_table_index;
       insn->bits3.sampler_g4x.sampler = sampler;
       insn->bits3.sampler_g4x.msg_type = msg_type;
@@ -648,10 +653,11 @@ struct brw_instruction *brw_IF(struct brw_compile *p, GLuint execute_size)
 struct brw_instruction *brw_ELSE(struct brw_compile *p, 
                                 struct brw_instruction *if_insn)
 {
+   struct intel_context *intel = &p->brw->intel;
    struct brw_instruction *insn;
    GLuint br = 1;
 
-   if (BRW_IS_IGDNG(p->brw))
+   if (intel->is_ironlake)
       br = 2;
 
    if (p->single_program_flow) {
@@ -690,9 +696,10 @@ struct brw_instruction *brw_ELSE(struct brw_compile *p,
 void brw_ENDIF(struct brw_compile *p, 
               struct brw_instruction *patch_insn)
 {
+   struct intel_context *intel = &p->brw->intel;
    GLuint br = 1;
 
-   if (BRW_IS_IGDNG(p->brw))
+   if (intel->is_ironlake)
       br = 2; 
  
    if (p->single_program_flow) {
@@ -803,10 +810,11 @@ struct brw_instruction *brw_DO(struct brw_compile *p, GLuint execute_size)
 struct brw_instruction *brw_WHILE(struct brw_compile *p, 
                                   struct brw_instruction *do_insn)
 {
+   struct intel_context *intel = &p->brw->intel;
    struct brw_instruction *insn;
    GLuint br = 1;
 
-   if (BRW_IS_IGDNG(p->brw))
+   if (intel->is_ironlake)
       br = 2;
 
    if (p->single_program_flow)
@@ -846,10 +854,11 @@ struct brw_instruction *brw_WHILE(struct brw_compile *p,
 void brw_land_fwd_jump(struct brw_compile *p, 
                       struct brw_instruction *jmp_insn)
 {
+   struct intel_context *intel = &p->brw->intel;
    struct brw_instruction *landing = &p->store[p->nr_insn];
    GLuint jmpi = 1;
 
-   if (BRW_IS_IGDNG(p->brw))
+   if (intel->is_ironlake)
        jmpi = 2;
 
    assert(jmp_insn->header.opcode == BRW_OPCODE_JMPI);
index 562a17844b43ed4793d3682bdbdc0d5a094dcb6a..fe5c1ae279100d730ef05eb530e44e9441e42467 100644 (file)
@@ -47,7 +47,6 @@
 
 static GLboolean do_check_fallback(struct brw_context *brw)
 {
-   struct intel_context *intel = &brw->intel;
    GLcontext *ctx = &brw->intel.ctx;
    GLuint i;
 
@@ -86,8 +85,7 @@ static GLboolean do_check_fallback(struct brw_context *brw)
    }
 
    /* _NEW_BUFFERS */
-   if (IS_965(intel->intelScreen->deviceID) &&
-       !IS_G4X(intel->intelScreen->deviceID)) {
+   if (!brw->has_surface_tile_offset) {
       for (i = 0; i < ctx->DrawBuffer->_NumColorDrawBuffers; i++) {
         struct gl_renderbuffer *rb = ctx->DrawBuffer->_ColorDrawBuffers[i];
         struct intel_renderbuffer *irb = intel_renderbuffer(rb);
index 610b6c35e2a15ce672b8283f965975d11b9c612a..1bc3eccf49d9ad05869b5d29f23db42ed3b6fc34 100644 (file)
@@ -47,6 +47,7 @@
 static void compile_gs_prog( struct brw_context *brw,
                             struct brw_gs_prog_key *key )
 {
+   struct intel_context *intel = &brw->intel;
    struct brw_gs_compile c;
    const GLuint *program;
    GLuint program_size;
@@ -54,13 +55,12 @@ static void compile_gs_prog( struct brw_context *brw,
    memset(&c, 0, sizeof(c));
    
    c.key = *key;
-   c.need_ff_sync = BRW_IS_IGDNG(brw);
    /* Need to locate the two positions present in vertex + header.
     * These are currently hardcoded:
     */
    c.nr_attrs = brw_count_bits(c.key.attrs);
 
-   if (BRW_IS_IGDNG(brw))
+   if (intel->is_ironlake)
        c.nr_regs = (c.nr_attrs + 1) / 2 + 3;  /* are vertices packed, or reg-aligned? */
    else
        c.nr_regs = (c.nr_attrs + 1) / 2 + 1;  /* are vertices packed, or reg-aligned? */
index 010c1c235219ca91d91887f200423d5f3eca9b6e..813b8d447aebe49fc7878825255a4ee6a9a3631f 100644 (file)
@@ -63,7 +63,6 @@ struct brw_gs_compile {
    GLuint nr_attrs;
    GLuint nr_regs;
    GLuint nr_bytes;
-   GLboolean need_ff_sync;
 };
 
 #define ATTR_SIZE  (4*4)
index 0fc5b02c619590fa61dc5268a9abd594367e0aa5..a81b972ef4b12a574c2cc39b8a45ee161d09ada0 100644 (file)
@@ -122,12 +122,14 @@ static void brw_gs_ff_sync(struct brw_gs_compile *c, int num_prim)
 
 void brw_gs_quads( struct brw_gs_compile *c, struct brw_gs_prog_key *key )
 {
+   struct intel_context *intel = &c->func.brw->intel;
+
    brw_gs_alloc_regs(c, 4);
    
    /* Use polygons for correct edgeflag behaviour. Note that vertex 3
     * is the PV for quads, but vertex 0 for polygons:
     */
-   if (c->need_ff_sync)
+   if (intel->needs_ff_sync)
           brw_gs_ff_sync(c, 1);
    if (key->pv_first) {
       brw_gs_emit_vue(c, c->reg.vertex[0], 0, ((_3DPRIM_POLYGON << 2) | R02_PRIM_START));
@@ -145,9 +147,11 @@ void brw_gs_quads( struct brw_gs_compile *c, struct brw_gs_prog_key *key )
 
 void brw_gs_quad_strip( struct brw_gs_compile *c, struct brw_gs_prog_key *key )
 {
+   struct intel_context *intel = &c->func.brw->intel;
+
    brw_gs_alloc_regs(c, 4);
    
-   if (c->need_ff_sync)
+   if (intel->needs_ff_sync)
           brw_gs_ff_sync(c, 1);      
    if (key->pv_first) {
       brw_gs_emit_vue(c, c->reg.vertex[0], 0, ((_3DPRIM_POLYGON << 2) | R02_PRIM_START));
@@ -165,9 +169,11 @@ void brw_gs_quad_strip( struct brw_gs_compile *c, struct brw_gs_prog_key *key )
 
 void brw_gs_tris( struct brw_gs_compile *c )
 {
+   struct intel_context *intel = &c->func.brw->intel;
+
    brw_gs_alloc_regs(c, 3);
 
-   if (c->need_ff_sync)
+   if (intel->needs_ff_sync)
           brw_gs_ff_sync(c, 1);      
    brw_gs_emit_vue(c, c->reg.vertex[0], 0, ((_3DPRIM_TRILIST << 2) | R02_PRIM_START));
    brw_gs_emit_vue(c, c->reg.vertex[1], 0, (_3DPRIM_TRILIST << 2));
@@ -176,9 +182,11 @@ void brw_gs_tris( struct brw_gs_compile *c )
 
 void brw_gs_lines( struct brw_gs_compile *c )
 {
+   struct intel_context *intel = &c->func.brw->intel;
+
    brw_gs_alloc_regs(c, 2);
 
-   if (c->need_ff_sync)
+   if (intel->needs_ff_sync)
           brw_gs_ff_sync(c, 1);      
    brw_gs_emit_vue(c, c->reg.vertex[0], 0, ((_3DPRIM_LINESTRIP << 2) | R02_PRIM_START));
    brw_gs_emit_vue(c, c->reg.vertex[1], 1, ((_3DPRIM_LINESTRIP << 2) | R02_PRIM_END));
@@ -186,9 +194,11 @@ void brw_gs_lines( struct brw_gs_compile *c )
 
 void brw_gs_points( struct brw_gs_compile *c )
 {
+   struct intel_context *intel = &c->func.brw->intel;
+
    brw_gs_alloc_regs(c, 1);
 
-   if (c->need_ff_sync)
+   if (intel->needs_ff_sync)
           brw_gs_ff_sync(c, 1);      
    brw_gs_emit_vue(c, c->reg.vertex[0], 1, ((_3DPRIM_POINTLIST << 2) | R02_PRIM_START | R02_PRIM_END));
 }
index ed9d2ffe605e49ee89f9ee007cd058b0afd9bb9b..1af5790a676445ed17d03dccb937100602de8ccd 100644 (file)
@@ -72,6 +72,7 @@ gs_unit_populate_key(struct brw_context *brw, struct brw_gs_unit_key *key)
 static dri_bo *
 gs_unit_create_from_key(struct brw_context *brw, struct brw_gs_unit_key *key)
 {
+   struct intel_context *intel = &brw->intel;
    struct brw_gs_unit_state gs;
    dri_bo *bo;
 
@@ -98,7 +99,7 @@ gs_unit_create_from_key(struct brw_context *brw, struct brw_gs_unit_key *key)
    else
       gs.thread4.max_threads = 0;
 
-   if (BRW_IS_IGDNG(brw))
+   if (intel->is_ironlake)
       gs.thread4.rendering_enable = 1;
 
    if (INTEL_DEBUG & DEBUG_STATS)
index 4b0d5983360f221852ee9cef1ffc894ec721c8d4..d437b1e03095692fc037eb8c05c4ff9b1e9665ed 100644 (file)
@@ -212,7 +212,7 @@ static void emit_depthbuffer(struct brw_context *brw)
 {
    struct intel_context *intel = &brw->intel;
    struct intel_region *region = brw->state.depth_region;
-   unsigned int len = (BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw)) ? 6 : 5;
+   unsigned int len = (intel->is_g4x || intel->is_ironlake) ? 6 : 5;
 
    if (region == NULL) {
       BEGIN_BATCH(len, IGNORE_CLIPRECTS);
@@ -223,7 +223,7 @@ static void emit_depthbuffer(struct brw_context *brw)
       OUT_BATCH(0);
       OUT_BATCH(0);
 
-      if (BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw))
+      if (intel->is_g4x || intel->is_ironlake)
          OUT_BATCH(0);
 
       ADVANCE_BATCH();
@@ -262,7 +262,7 @@ static void emit_depthbuffer(struct brw_context *brw)
                ((region->height - 1) << 19));
       OUT_BATCH(0);
 
-      if (BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw))
+      if (intel->is_g4x || intel->is_ironlake)
          OUT_BATCH(0);
 
       ADVANCE_BATCH();
@@ -374,8 +374,8 @@ const struct brw_tracked_state brw_polygon_stipple_offset = {
 static void upload_aa_line_parameters(struct brw_context *brw)
 {
    struct brw_aa_line_parameters balp;
-   
-   if (BRW_IS_965(brw))
+
+   if (!brw->has_aa_line_parameters)
       return;
 
    /* use legacy aa line coverage computation */
@@ -444,7 +444,7 @@ static void upload_invarient_state( struct brw_context *brw )
       struct brw_pipeline_select ps;
 
       memset(&ps, 0, sizeof(ps));
-      ps.header.opcode = CMD_PIPELINE_SELECT(brw);
+      ps.header.opcode = brw->CMD_PIPELINE_SELECT;
       ps.header.pipeline_select = 0;
       BRW_BATCH_STRUCT(brw, &ps);
    }
@@ -480,7 +480,7 @@ static void upload_invarient_state( struct brw_context *brw )
       struct brw_vf_statistics vfs;
       memset(&vfs, 0, sizeof(vfs));
 
-      vfs.opcode = CMD_VF_STATISTICS(brw);
+      vfs.opcode = brw->CMD_VF_STATISTICS;
       if (INTEL_DEBUG & DEBUG_STATS)
         vfs.statistics_enable = 1; 
 
@@ -512,7 +512,7 @@ static void upload_state_base_address( struct brw_context *brw )
    /* Output the structure (brw_state_base_address) directly to the
     * batchbuffer, so we can emit relocations inline.
     */
-   if (BRW_IS_IGDNG(brw)) {
+   if (intel->is_ironlake) {
        BEGIN_BATCH(8, IGNORE_CLIPRECTS);
        OUT_BATCH(CMD_STATE_BASE_ADDRESS << 16 | (8 - 2));
        OUT_BATCH(1); /* General state base address */
index 3eae41ee7424743163a8d1620104923b83c6ac74..bb08055e3bb6db23587a92aa858cb88aed9d7dd6 100644 (file)
@@ -149,6 +149,7 @@ static void copy_colors( struct brw_sf_compile *c,
 static void do_flatshade_triangle( struct brw_sf_compile *c )
 {
    struct brw_compile *p = &c->func;
+   struct intel_context *intel = &p->brw->intel;
    struct brw_reg ip = brw_ip_reg();
    GLuint nr = brw_count_bits(c->key.attrs & VERT_RESULT_COLOR_BITS);
    GLuint jmpi = 1;
@@ -161,7 +162,7 @@ static void do_flatshade_triangle( struct brw_sf_compile *c )
    if (c->key.primitive == SF_UNFILLED_TRIS)
       return;
 
-   if (BRW_IS_IGDNG(p->brw))
+   if (intel->is_ironlake)
        jmpi = 2;
 
    brw_push_insn_state(p);
@@ -187,6 +188,7 @@ static void do_flatshade_triangle( struct brw_sf_compile *c )
 static void do_flatshade_line( struct brw_sf_compile *c )
 {
    struct brw_compile *p = &c->func;
+   struct intel_context *intel = &p->brw->intel;
    struct brw_reg ip = brw_ip_reg();
    GLuint nr = brw_count_bits(c->key.attrs & VERT_RESULT_COLOR_BITS);
    GLuint jmpi = 1;
@@ -199,7 +201,7 @@ static void do_flatshade_line( struct brw_sf_compile *c )
    if (c->key.primitive == SF_UNFILLED_TRIS)
       return;
 
-   if (BRW_IS_IGDNG(p->brw))
+   if (intel->is_ironlake)
        jmpi = 2;
 
    brw_push_insn_state(p);
index bb69435ec0a845cc3b73aad41425d14b7715f18d..09223b7cfb1d95b01f9be4f39e8567b942ab65fb 100644 (file)
@@ -165,6 +165,7 @@ static dri_bo *
 sf_unit_create_from_key(struct brw_context *brw, struct brw_sf_unit_key *key,
                        dri_bo **reloc_bufs)
 {
+   struct intel_context *intel = &brw->intel;
    struct brw_sf_unit_state sf;
    dri_bo *bo;
    int chipset_max_threads;
@@ -177,7 +178,7 @@ sf_unit_create_from_key(struct brw_context *brw, struct brw_sf_unit_key *key,
 
    sf.thread3.dispatch_grf_start_reg = 3;
 
-   if (BRW_IS_IGDNG(brw))
+   if (intel->is_ironlake)
        sf.thread3.urb_entry_read_offset = 3;
    else
        sf.thread3.urb_entry_read_offset = 1;
@@ -187,10 +188,10 @@ sf_unit_create_from_key(struct brw_context *brw, struct brw_sf_unit_key *key,
    sf.thread4.nr_urb_entries = key->nr_urb_entries;
    sf.thread4.urb_entry_allocation_size = key->sfsize - 1;
 
-   /* Each SF thread produces 1 PUE, and there can be up to 24(Pre-IGDNG) or 
-    * 48(IGDNG) threads 
+   /* Each SF thread produces 1 PUE, and there can be up to 24 (Pre-Ironlake) or
+    * 48 (Ironlake) threads.
     */
-   if (BRW_IS_IGDNG(brw))
+   if (intel->is_ironlake)
       chipset_max_threads = 48;
    else
       chipset_max_threads = 24;
index e59e52ed8612da84da15c1d99b3e34198d8531c1..64a9535282bd43f8a0cf17fbf502bc559c859cfc 100644 (file)
@@ -49,7 +49,7 @@ GLboolean brw_miptree_layout(struct intel_context *intel,
 
    switch (mt->target) {
    case GL_TEXTURE_CUBE_MAP:
-      if (IS_IGDNG(intel->intelScreen->deviceID)) {
+      if (intel->is_ironlake) {
           GLuint align_h = 2, align_w = 4;
           GLuint level;
           GLuint x = 0;
index 8c6f4355a6eccd1bede10b814e48a4886df1694c..f2cdb203b85235ffddf27d1f14e7dbc9524dd85a 100644 (file)
@@ -105,7 +105,8 @@ static GLboolean check_urb_layout( struct brw_context *brw )
    brw->urb.sf_start = brw->urb.clip_start + brw->urb.nr_clip_entries * brw->urb.vsize;
    brw->urb.cs_start = brw->urb.sf_start + brw->urb.nr_sf_entries * brw->urb.sfsize;
 
-   return brw->urb.cs_start + brw->urb.nr_cs_entries * brw->urb.csize <= URB_SIZES(brw);
+   return brw->urb.cs_start + brw->urb.nr_cs_entries *
+      brw->urb.csize <= brw->urb.size;
 }
 
 /* Most minimal update, forces re-emit of URB fence packet after GS
@@ -113,6 +114,7 @@ static GLboolean check_urb_layout( struct brw_context *brw )
  */
 static void recalculate_urb_fence( struct brw_context *brw )
 {
+   struct intel_context *intel = &brw->intel;
    GLuint csize = brw->curbe.total_size;
    GLuint vsize = brw->vs.prog_data->urb_entry_size;
    GLuint sfsize = brw->sf.prog_data->urb_entry_size;
@@ -146,7 +148,7 @@ static void recalculate_urb_fence( struct brw_context *brw )
 
       brw->urb.constrained = 0;
 
-      if (BRW_IS_IGDNG(brw)) {
+      if (intel->is_ironlake) {
          brw->urb.nr_vs_entries = 128;
          brw->urb.nr_sf_entries = 48;
          if (check_urb_layout(brw)) {
@@ -156,7 +158,7 @@ static void recalculate_urb_fence( struct brw_context *brw )
             brw->urb.nr_vs_entries = limits[VS].preferred_nr_entries;
             brw->urb.nr_sf_entries = limits[SF].preferred_nr_entries;
          }
-      } else if (BRW_IS_G4X(brw)) {
+      } else if (intel->is_g4x) {
         brw->urb.nr_vs_entries = 64;
         if (check_urb_layout(brw)) {
            goto done;
@@ -200,7 +202,7 @@ done:
                      brw->urb.clip_start,
                      brw->urb.sf_start,
                      brw->urb.cs_start, 
-                     URB_SIZES(brw));
+                     brw->urb.size);
       
       brw->state.dirty.brw |= BRW_NEW_URB_FENCE;
    }
@@ -244,7 +246,7 @@ void brw_upload_urb_fence(struct brw_context *brw)
    uf.bits0.gs_fence  = brw->urb.clip_start; 
    uf.bits0.clp_fence = brw->urb.sf_start; 
    uf.bits1.sf_fence  = brw->urb.cs_start; 
-   uf.bits1.cs_fence  = URB_SIZES(brw);
+   uf.bits1.cs_fence  = brw->urb.size;
 
    BRW_BATCH_STRUCT(brw, &uf);
 }
index 27aac8b54a95da8b1279d4cc073c45be4163c10d..1b84dd505f60be1e057bb7c6d6519d983b365041 100644 (file)
@@ -67,6 +67,7 @@ static void release_tmps( struct brw_vs_compile *c )
  */
 static void brw_vs_alloc_regs( struct brw_vs_compile *c )
 {
+   struct intel_context *intel = &c->func.brw->intel;
    GLuint i, reg = 0, mrf;
    int attributes_in_vue;
 
@@ -141,7 +142,7 @@ static void brw_vs_alloc_regs( struct brw_vs_compile *c )
    c->first_output = reg;
    c->first_overflow_output = 0;
 
-   if (BRW_IS_IGDNG(c->func.brw))
+   if (intel->is_ironlake)
        mrf = 8;
    else
        mrf = 4;
@@ -238,7 +239,7 @@ static void brw_vs_alloc_regs( struct brw_vs_compile *c )
     */
    attributes_in_vue = MAX2(c->nr_outputs, c->nr_inputs);
 
-   if (BRW_IS_IGDNG(c->func.brw))
+   if (intel->is_ironlake)
        c->prog_data.urb_entry_size = (attributes_in_vue + 6 + 3) / 4;
    else
        c->prog_data.urb_entry_size = (attributes_in_vue + 2 + 3) / 4;
@@ -1113,6 +1114,8 @@ static void emit_swz( struct brw_vs_compile *c,
 static void emit_vertex_write( struct brw_vs_compile *c)
 {
    struct brw_compile *p = &c->func;
+   struct brw_context *brw = p->brw;
+   struct intel_context *intel = &brw->intel;
    struct brw_reg m0 = brw_message_reg(0);
    struct brw_reg pos = c->regs[PROGRAM_OUTPUT][VERT_RESULT_HPOS];
    struct brw_reg ndc;
@@ -1136,7 +1139,7 @@ static void emit_vertex_write( struct brw_vs_compile *c)
     * workaround.
     */
    if ((c->prog_data.outputs_written & BITFIELD64_BIT(VERT_RESULT_PSIZ)) ||
-       c->key.nr_userclip || BRW_IS_965(p->brw))
+       c->key.nr_userclip || brw->has_negative_rhw_bug)
    {
       struct brw_reg header1 = retype(get_tmp(c), BRW_REGISTER_TYPE_UD);
       GLuint i;
@@ -1167,7 +1170,7 @@ static void emit_vertex_write( struct brw_vs_compile *c)
        * Later, clipping will detect ucp[6] and ensure the primitive is
        * clipped against all fixed planes.
        */
-      if (BRW_IS_965(p->brw)) {
+      if (brw->has_negative_rhw_bug) {
         brw_CMP(p,
                 vec8(brw_null_reg()),
                 BRW_CONDITIONAL_L,
@@ -1195,8 +1198,8 @@ static void emit_vertex_write( struct brw_vs_compile *c)
    brw_set_access_mode(p, BRW_ALIGN_1);
    brw_MOV(p, offset(m0, 2), ndc);
 
-   if (BRW_IS_IGDNG(p->brw)) {
-       /* There are 20 DWs (D0-D19) in VUE vertex header on IGDNG */
+   if (intel->is_ironlake) {
+       /* There are 20 DWs (D0-D19) in VUE vertex header on Ironlake */
        brw_MOV(p, offset(m0, 3), pos); /* a portion of vertex header */
        /* m4, m5 contain the distances from vertex to the user clip planeXXX. 
         * Seems it is useless for us.
@@ -1359,6 +1362,7 @@ void brw_vs_emit(struct brw_vs_compile *c )
 #define MAX_LOOP_DEPTH 32
    struct brw_compile *p = &c->func;
    struct brw_context *brw = p->brw;
+   struct intel_context *intel = &brw->intel;
    const GLuint nr_insns = c->vp->program.Base.NumInstructions;
    GLuint insn, if_depth = 0, loop_depth = 0;
    GLuint end_offset = 0;
@@ -1592,7 +1596,7 @@ void brw_vs_emit(struct brw_vs_compile *c )
 
             loop_depth--;
 
-           if (BRW_IS_IGDNG(brw))
+           if (intel->is_ironlake)
               br = 2;
 
             inst0 = inst1 = brw_WHILE(p, loop_inst[loop_depth]);
index 7285466645b66a1aa7457c44fbccfaaf535f25ae..345ffa7ee11bc1184ff88353c8ad164f6d62c20f 100644 (file)
@@ -82,9 +82,9 @@ vs_unit_populate_key(struct brw_context *brw, struct brw_vs_unit_key *key)
 static dri_bo *
 vs_unit_create_from_key(struct brw_context *brw, struct brw_vs_unit_key *key)
 {
+   struct intel_context *intel = &brw->intel;
    struct brw_vs_unit_state vs;
    dri_bo *bo;
-   int chipset_max_threads;
 
    memset(&vs, 0, sizeof(vs));
 
@@ -98,7 +98,7 @@ vs_unit_create_from_key(struct brw_context *brw, struct brw_vs_unit_key *key)
     */
    vs.thread1.single_program_flow = 0;
 
-   if (BRW_IS_IGDNG(brw))
+   if (intel->is_ironlake)
       vs.thread1.binding_table_entry_count = 0; /* hardware requirement */
    else
       vs.thread1.binding_table_entry_count = key->nr_surfaces;
@@ -109,7 +109,7 @@ vs_unit_create_from_key(struct brw_context *brw, struct brw_vs_unit_key *key)
    vs.thread3.urb_entry_read_offset = 0;
    vs.thread3.const_urb_entry_read_offset = key->curbe_offset * 2;
 
-   if (BRW_IS_IGDNG(brw)) {
+   if (intel->is_ironlake) {
       switch (key->nr_urb_entries) {
       case 8:
       case 12:
@@ -135,7 +135,7 @@ vs_unit_create_from_key(struct brw_context *brw, struct brw_vs_unit_key *key)
       case 32:
         break;
       case 64:
-        assert(BRW_IS_G4X(brw));
+        assert(intel->is_g4x);
         break;
       default:
         assert(0);
@@ -145,17 +145,8 @@ vs_unit_create_from_key(struct brw_context *brw, struct brw_vs_unit_key *key)
 
    vs.thread4.urb_entry_allocation_size = key->urb_size - 1;
 
-   if (BRW_IS_IGDNG(brw))
-      chipset_max_threads = 72;
-   else if (BRW_IS_G4X(brw))
-      chipset_max_threads = 32;
-   else
-      chipset_max_threads = 16;
    vs.thread4.max_threads = CLAMP(key->nr_urb_entries / 2,
-                                 1, chipset_max_threads) - 1;
-
-   if (INTEL_DEBUG & DEBUG_SINGLE_THREAD)
-      vs.thread4.max_threads = 0;
+                                 1, brw->vs_max_threads) - 1;
 
    /* No samplers for ARB_vp programs:
     */
index 34aaea373606dc209d6fb7a6426d7058ebd28644..72749b3859d3a72b8c186caa256339ea7f618597 100644 (file)
@@ -150,9 +150,6 @@ static void brw_new_batch( struct intel_context *intel )
 {
    struct brw_context *brw = brw_context(&intel->ctx);
 
-   /* Check that we didn't just wrap our batchbuffer at a bad time. */
-   assert(!brw->no_batch_wrap);
-
    brw->curbe.need_new_bo = GL_TRUE;
 
    /* Mark all context state as needing to be re-emitted.
index 9dcb6e14bbe4d5c9689cf568be1248ee62ff605e..b9b987ea70a987f47da8bc061d832bd59348b7be 100644 (file)
@@ -76,10 +76,10 @@ struct brw_wm_prog_key {
 
    GLushort tex_swizzles[BRW_MAX_TEX_UNIT];
 
-   GLuint program_string_id:32;
    GLushort origin_x, origin_y;
    GLushort drawable_height;
    GLbitfield64 vp_outputs_written;
+   GLuint program_string_id:32;
 };
 
 
index 5390fd25849d16127d7e470f4f794cd25522ac65..7811689d6ad83d9b26a92bd0bd753ab550e53a53 100644 (file)
@@ -830,6 +830,7 @@ void emit_tex(struct brw_wm_compile *c,
              GLboolean shadow)
 {
    struct brw_compile *p = &c->func;
+   struct intel_context *intel = &p->brw->intel;
    struct brw_reg dst_retyped;
    GLuint cur_mrf = 2, response_length;
    GLuint i, nr_texcoords;
@@ -873,7 +874,7 @@ void emit_tex(struct brw_wm_compile *c,
    }
 
    /* Pre-Ironlake, the 8-wide sampler always took u,v,r. */
-   if (!BRW_IS_IGDNG(p->brw) && c->dispatch_width == 8)
+   if (!intel->is_ironlake && c->dispatch_width == 8)
       nr_texcoords = 3;
 
    /* For shadow comparisons, we have to supply u,v,r. */
@@ -891,7 +892,7 @@ void emit_tex(struct brw_wm_compile *c,
 
    /* Fill in the shadow comparison reference value. */
    if (shadow) {
-      if (BRW_IS_IGDNG(p->brw)) {
+      if (intel->is_ironlake) {
         /* Fill in the cube map array index value. */
         brw_MOV(p, brw_message_reg(cur_mrf), brw_imm_f(0));
         cur_mrf += mrf_per_channel;
@@ -904,7 +905,7 @@ void emit_tex(struct brw_wm_compile *c,
       cur_mrf += mrf_per_channel;
    }
 
-   if (BRW_IS_IGDNG(p->brw)) {
+   if (intel->is_ironlake) {
       if (shadow)
         msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_COMPARE_IGDNG;
       else
@@ -944,6 +945,7 @@ void emit_txb(struct brw_wm_compile *c,
              GLuint sampler)
 {
    struct brw_compile *p = &c->func;
+   struct intel_context *intel = &p->brw->intel;
    GLuint msgLength;
    GLuint msg_type;
    GLuint mrf_per_channel;
@@ -955,8 +957,8 @@ void emit_txb(struct brw_wm_compile *c,
     * undefined, and trust the execution mask to keep the undefined pixels
     * from mattering.
     */
-   if (c->dispatch_width == 16 || !BRW_IS_IGDNG(p->brw)) {
-      if (BRW_IS_IGDNG(p->brw))
+   if (c->dispatch_width == 16 || !intel->is_ironlake) {
+      if (intel->is_ironlake)
         msg_type = BRW_SAMPLER_MESSAGE_SAMPLE_BIAS_IGDNG;
       else
         msg_type = BRW_SAMPLER_MESSAGE_SIMD16_SAMPLE_BIAS;
@@ -1174,7 +1176,7 @@ void emit_fb_write(struct brw_wm_compile *c,
    brw_push_insn_state(p);
 
    for (channel = 0; channel < 4; channel++) {
-      if (c->dispatch_width == 16 && (BRW_IS_G4X(brw) || BRW_IS_IGDNG(brw))) {
+      if (c->dispatch_width == 16 && brw->has_compr4) {
         /* By setting the high bit of the MRF register number, we indicate
          * that we want COMPR4 mode - instead of doing the usual destination
          * + 1 for the second half we get destination + 4.
index 7d031795883a1d35e7a68caf577c8f8aa0acb518..3737faf26fb33390cda38f12704c4399a12296a7 100644 (file)
@@ -138,7 +138,6 @@ static struct prog_dst_register dst_reg(GLuint file, GLuint idx)
    reg.CondMask = COND_TR;
    reg.CondSwizzle = 0;
    reg.CondSrc = 0;
-   reg.pad = 0;
    return reg;
 }
 
index e8c2cb66ec3217c02affbfb5e8de87fff76b7f2b..eccbfd5fa3ae8d0325de344c9f6ba810e93ceeb9 100644 (file)
@@ -1826,6 +1826,7 @@ get_argument_regs(struct brw_wm_compile *c,
 
 static void brw_wm_emit_glsl(struct brw_context *brw, struct brw_wm_compile *c)
 {
+   struct intel_context *intel = &brw->intel;
 #define MAX_IF_DEPTH 32
 #define MAX_LOOP_DEPTH 32
     struct brw_instruction *if_inst[MAX_IF_DEPTH], *loop_inst[MAX_LOOP_DEPTH];
@@ -1876,10 +1877,6 @@ static void brw_wm_emit_glsl(struct brw_context *brw, struct brw_wm_compile *c)
        else
            brw_set_conditionalmod(p, BRW_CONDITIONAL_NONE);
 
-       dst_flags = inst->DstReg.WriteMask;
-       if (inst->SaturateMode == SATURATE_ZERO_ONE)
-           dst_flags |= SATURATE;
-
        switch (inst->Opcode) {
            case WM_PIXELXY:
                emit_pixel_xy(c, dst, dst_flags);
@@ -2043,6 +2040,7 @@ static void brw_wm_emit_glsl(struct brw_context *brw, struct brw_wm_compile *c)
                if_inst[if_depth++] = brw_IF(p, BRW_EXECUTE_8);
                break;
            case OPCODE_ELSE:
+               assert(if_depth > 0);
                if_inst[if_depth-1]  = brw_ELSE(p, if_inst[if_depth-1]);
                break;
            case OPCODE_ENDIF:
@@ -2096,9 +2094,10 @@ static void brw_wm_emit_glsl(struct brw_context *brw, struct brw_wm_compile *c)
                   struct brw_instruction *inst0, *inst1;
                   GLuint br = 1;
 
-                  if (BRW_IS_IGDNG(brw))
+                  if (intel->is_ironlake)
                      br = 2;
+
+                 assert(loop_depth > 0);
                   loop_depth--;
                   inst0 = inst1 = brw_WHILE(p, loop_inst[loop_depth]);
                   /* patch all the BREAK/CONT instructions from last BGNLOOP */
index f89ed9bce706fb30de5d73a70214cc15a36af9dd..d3373ea79e80fb9b333adc3ad8da01b8f992ef39 100644 (file)
@@ -49,8 +49,6 @@ struct brw_wm_unit_key {
    unsigned int curbe_offset;
    unsigned int urb_size;
 
-   unsigned int max_threads;
-
    unsigned int nr_surfaces, sampler_count;
    GLboolean uses_depth, computes_depth, uses_kill, is_glsl;
    GLboolean polygon_stipple, stats_wm, line_stipple, offset_enable;
@@ -67,18 +65,6 @@ wm_unit_populate_key(struct brw_context *brw, struct brw_wm_unit_key *key)
 
    memset(key, 0, sizeof(*key));
 
-   if (INTEL_DEBUG & DEBUG_SINGLE_THREAD)
-      key->max_threads = 1;
-   else {
-      /* WM maximum threads is number of EUs times number of threads per EU. */
-      if (BRW_IS_IGDNG(brw))
-         key->max_threads = 12 * 6;
-      else if (BRW_IS_G4X(brw))
-        key->max_threads = 10 * 5;
-      else
-        key->max_threads = 8 * 4;
-   }
-
    /* CACHE_NEW_WM_PROG */
    key->total_grf = brw->wm.prog_data->total_grf;
    key->urb_entry_read_length = brw->wm.prog_data->urb_read_length;
@@ -140,6 +126,7 @@ static dri_bo *
 wm_unit_create_from_key(struct brw_context *brw, struct brw_wm_unit_key *key,
                        dri_bo **reloc_bufs)
 {
+   struct intel_context *intel = &brw->intel;
    struct brw_wm_unit_state wm;
    dri_bo *bo;
 
@@ -150,7 +137,7 @@ wm_unit_create_from_key(struct brw_context *brw, struct brw_wm_unit_key *key,
    wm.thread1.depth_coef_urb_read_offset = 1;
    wm.thread1.floating_point_mode = BRW_FLOATING_POINT_NON_IEEE_754;
 
-   if (BRW_IS_IGDNG(brw))
+   if (intel->is_ironlake)
       wm.thread1.binding_table_entry_count = 0; /* hardware requirement */
    else
       wm.thread1.binding_table_entry_count = key->nr_surfaces;
@@ -170,7 +157,7 @@ wm_unit_create_from_key(struct brw_context *brw, struct brw_wm_unit_key *key,
    wm.thread3.const_urb_entry_read_length = key->curb_entry_read_length;
    wm.thread3.const_urb_entry_read_offset = key->curbe_offset * 2;
 
-   if (BRW_IS_IGDNG(brw)) 
+   if (intel->is_ironlake)
       wm.wm4.sampler_count = 0; /* hardware requirement */
    else
       wm.wm4.sampler_count = (key->sampler_count + 1) / 4;
@@ -191,7 +178,7 @@ wm_unit_create_from_key(struct brw_context *brw, struct brw_wm_unit_key *key,
    else
       wm.wm5.enable_16_pix = 1;
 
-   wm.wm5.max_threads = key->max_threads - 1;
+   wm.wm5.max_threads = brw->wm_max_threads - 1;
    wm.wm5.thread_dispatch_enable = 1;  /* AKA: color_write */
    wm.wm5.legacy_line_rast = 0;
    wm.wm5.legacy_global_depth_bias = 0;
@@ -268,7 +255,7 @@ static void upload_wm_unit( struct brw_context *brw )
     */
    assert(key.total_scratch <= 12 * 1024);
    if (key.total_scratch) {
-      GLuint total = key.total_scratch * key.max_threads;
+      GLuint total = key.total_scratch * brw->wm_max_threads;
 
       if (brw->wm.scratch_bo && total > brw->wm.scratch_bo->size) {
         dri_bo_unreference(brw->wm.scratch_bo);
index 8335e5a650ebcbba8aadc4dcf277ce6c3f85b7e3..8810f2a3808925b4a8c4bc1ad5de16d4e392da94 100644 (file)
@@ -608,8 +608,7 @@ brw_update_renderbuffer_surface(struct brw_context *brw,
 
         surf.ss1.base_addr = key.draw_offset - tile_offset;
 
-        assert(BRW_IS_G4X(brw) || tile_offset == 0);
-        if (BRW_IS_G4X(brw)) {
+        if (brw->has_surface_tile_offset) {
            if (key.tiling == I915_TILING_X) {
               /* Note that the low bits of these fields are missing, so
                * there's the possibility of getting in trouble.
@@ -620,6 +619,8 @@ brw_update_renderbuffer_surface(struct brw_context *brw,
               surf.ss5.x_offset = (tile_offset % 128) / key.cpp / 4;
               surf.ss5.y_offset = tile_offset / 128 / 2;
            }
+        } else {
+           assert(tile_offset == 0);
         }
       }
       if (region_bo != NULL)
index ca6e2fa5b11b1b5db0f72458c212e978ddb9d5c5..2eae9b66d87ba58ef955d13b2df3c722bd712496 100644 (file)
@@ -80,7 +80,7 @@ intel_batchbuffer_reset(struct intel_batchbuffer *batch)
       batch->buf = NULL;
    }
 
-   if (!batch->buffer && intel->ttm == GL_TRUE)
+   if (!batch->buffer)
       batch->buffer = malloc (intel->maxBatchSize);
 
    batch->buf = dri_bo_alloc(intel->bufmgr, "batchbuffer",
@@ -212,7 +212,7 @@ _intel_batchbuffer_flush(struct intel_batchbuffer *batch, const char *file,
 
    batch->reserved_space = 0;
    /* Emit a flush if the bufmgr doesn't do it for us. */
-   if (intel->always_flush_cache || !intel->ttm) {
+   if (intel->always_flush_cache) {
       intel_batchbuffer_emit_mi_flush(batch);
       used = batch->ptr - batch->map;
    }
@@ -244,6 +244,9 @@ _intel_batchbuffer_flush(struct intel_batchbuffer *batch, const char *file,
    if (intel->vtbl.finish_batch)
       intel->vtbl.finish_batch(intel);
 
+   /* Check that we didn't just wrap our batchbuffer at a bad time. */
+   assert(!intel->no_batch_wrap);
+
    batch->reserved_space = BATCH_RESERVED;
 
    /* TODO: Just pass the relocation list and dma buffer up to the
index 9f638b0ef985c41a116fa0851db4e7151dbba057..da9beba030d5597374821991cea18fd96da4b2eb 100644 (file)
@@ -52,7 +52,6 @@ intelCopyBuffer(const __DRIdrawablePrivate * dPriv,
 {
 
    struct intel_context *intel;
-   const intelScreenPrivate *intelScreen;
 
    DBG("%s\n", __FUNCTION__);
 
@@ -62,8 +61,6 @@ intelCopyBuffer(const __DRIdrawablePrivate * dPriv,
    if (!intel)
       return;
 
-   intelScreen = intel->intelScreen;
-
    /* The LOCK_HARDWARE is required for the cliprects.  Buffer offsets
     * should work regardless.
     */
@@ -499,10 +496,11 @@ intelClearWithBlit(GLcontext *ctx, GLbitfield mask)
                  switch (irb->Base.Format) {
                  case MESA_FORMAT_ARGB8888:
                  case MESA_FORMAT_XRGB8888:
-                    clearVal = intel->ClearColor8888;
+                    clearVal = PACK_COLOR_8888(clear[3], clear[0],
+                                               clear[1], clear[2]);
                     break;
                  case MESA_FORMAT_RGB565:
-                    clearVal = intel->ClearColor565;
+                    clearVal = PACK_COLOR_565(clear[0], clear[1], clear[2]);
                     break;
                  case MESA_FORMAT_ARGB4444:
                     clearVal = PACK_COLOR_4444(clear[3], clear[0],
index 26ff11ba61f5e0a57949e0113ebc778878c06782..02e0cc7b332de1e38e3a2fa4863db308ff625762 100644 (file)
@@ -68,7 +68,7 @@ int INTEL_DEBUG = (0);
 #endif
 
 
-#define DRIVER_DATE                     "20091221 2009Q4"
+#define DRIVER_DATE                     "20091221 DEVELOPMENT"
 #define DRIVER_DATE_GEM                 "GEM " DRIVER_DATE
 
 
@@ -176,9 +176,7 @@ intelGetString(GLcontext * ctx, GLenum name)
          break;
       }
 
-      (void) driGetRendererString(buffer, chipset, 
-                                 (intel->ttm) ? DRIVER_DATE_GEM : DRIVER_DATE,
-                                 0);
+      (void) driGetRendererString(buffer, chipset, DRIVER_DATE_GEM, 0);
       return (GLubyte *) buffer;
 
    default:
@@ -601,6 +599,7 @@ intelInitContext(struct intel_context *intel,
    __DRIscreenPrivate *sPriv = driContextPriv->driScreenPriv;
    intelScreenPrivate *intelScreen = (intelScreenPrivate *) sPriv->private;
    int fthrottle_mode;
+   int bo_reuse_mode;
 
    if (!_mesa_initialize_context(&intel->ctx, mesaVis, shareCtx,
                                  functions, (void *) intel)) {
@@ -614,12 +613,25 @@ intelInitContext(struct intel_context *intel,
    intel->sarea = intelScreen->sarea;
    intel->driContext = driContextPriv;
 
-   if (IS_965(intel->intelScreen->deviceID))
+   if (IS_965(intel->intelScreen->deviceID)) {
       intel->gen = 4;
-   else if (IS_9XX(intel->intelScreen->deviceID))
+   } else if (IS_9XX(intel->intelScreen->deviceID)) {
       intel->gen = 3;
-   else
+      if (IS_945(intel->intelScreen->deviceID)) {
+        intel->is_945 = GL_TRUE;
+      }
+   } else {
       intel->gen = 2;
+   }
+
+   if (IS_IGDNG(intel->intelScreen->deviceID)) {
+      intel->is_ironlake = GL_TRUE;
+      intel->needs_ff_sync = GL_TRUE;
+      intel->has_luminance_srgb = GL_TRUE;
+   } else if (IS_G4X(intel->intelScreen->deviceID)) {
+      intel->has_luminance_srgb = GL_TRUE;
+      intel->is_g4x = GL_TRUE;
+   }
 
    /* Dri stuff */
    intel->hHWContext = driContextPriv->hHWContext;
@@ -635,18 +647,14 @@ intelInitContext(struct intel_context *intel,
       intel->maxBatchSize = BATCH_SZ;
 
    intel->bufmgr = intelScreen->bufmgr;
-   intel->ttm = intelScreen->ttm;
-   if (intel->ttm) {
-      int bo_reuse_mode;
 
-      bo_reuse_mode = driQueryOptioni(&intel->optionCache, "bo_reuse");
-      switch (bo_reuse_mode) {
-      case DRI_CONF_BO_REUSE_DISABLED:
-        break;
-      case DRI_CONF_BO_REUSE_ALL:
-        intel_bufmgr_gem_enable_reuse(intel->bufmgr);
-        break;
-      }
+   bo_reuse_mode = driQueryOptioni(&intel->optionCache, "bo_reuse");
+   switch (bo_reuse_mode) {
+   case DRI_CONF_BO_REUSE_DISABLED:
+      break;
+   case DRI_CONF_BO_REUSE_ALL:
+      intel_bufmgr_gem_enable_reuse(intel->bufmgr);
+      break;
    }
 
    /* This doesn't yet catch all non-conformant rendering, but it's a
@@ -733,12 +741,6 @@ intelInitContext(struct intel_context *intel,
    intel->RenderIndex = ~0;
 
    fthrottle_mode = driQueryOptioni(&intel->optionCache, "fthrottle_mode");
-   intel->irqsEmitted = 0;
-
-   intel->do_irqs = (intel->intelScreen->irq_active &&
-                     fthrottle_mode == DRI_CONF_FTHROTTLE_IRQS);
-
-   intel->do_usleeps = (fthrottle_mode == DRI_CONF_FTHROTTLE_USLEEPS);
 
    if (intel->gen >= 4 && !intel->intelScreen->irq_active) {
       _mesa_printf("IRQs not active.  Exiting\n");
@@ -1058,21 +1060,6 @@ intelContendedLock(struct intel_context *intel, GLuint flags)
       sarea->ctxOwner = me;
    }
 
-   /* If the last consumer of the texture memory wasn't us, notify the fake
-    * bufmgr and record the new owner.  We should have the memory shared
-    * between contexts of a single fake bufmgr, but this will at least make
-    * things correct for now.
-    */
-   if (!intel->ttm && sarea->texAge != intel->hHWContext) {
-      sarea->texAge = intel->hHWContext;
-      intel_bufmgr_fake_contended_lock_take(intel->bufmgr);
-      if (INTEL_DEBUG & DEBUG_BATCH)
-        intel_decode_context_reset();
-      if (INTEL_DEBUG & DEBUG_BUFMGR)
-        fprintf(stderr, "Lost Textures: sarea->texAge %x hw context %x\n",
-                sarea->ctxOwner, intel->hHWContext);
-   }
-
    /* Drawable changed?
     */
    if (dPriv && intel->lastStamp != dPriv->lastStamp) {
index eb7be7ddd0fb904cd0698e48232f089a267e889f..e85886db82fb4de7ab92fa240d3f016a447d971e 100644 (file)
@@ -135,14 +135,6 @@ struct intel_context
                                 struct intel_region * draw_region,
                                 struct intel_region * depth_region);
 
-      void (*meta_draw_quad)(struct intel_context *intel,
-                            GLfloat x0, GLfloat x1,
-                            GLfloat y0, GLfloat y1,
-                            GLfloat z,
-                            GLuint color, /* ARGB32 */
-                            GLfloat s0, GLfloat s1,
-                            GLfloat t0, GLfloat t1);
-
       void (*meta_color_mask) (struct intel_context * intel, GLboolean);
 
       void (*meta_stencil_replace) (struct intel_context * intel,
@@ -184,17 +176,18 @@ struct intel_context
     * Generation number of the hardware: 2 is 8xx, 3 is 9xx pre-965, 4 is 965.
     */
    int gen;
+   GLboolean needs_ff_sync;
+   GLboolean is_ironlake;
+   GLboolean is_g4x;
+   GLboolean is_945;
+   GLboolean has_luminance_srgb;
+
+   int urb_size;
 
    struct intel_region *front_region;
    struct intel_region *back_region;
    struct intel_region *depth_region;
 
-   /**
-    * This value indicates that the kernel memory manager is being used
-    * instead of the fake client-side memory manager.
-    */
-   GLboolean ttm;
-
    struct intel_batchbuffer *batch;
    drm_intel_bo *first_post_swapbuffers_batch;
    GLboolean no_batch_wrap;
@@ -217,10 +210,6 @@ struct intel_context
    char *prevLockFile;
    int prevLockLine;
 
-   GLuint ClearColor565;
-   GLuint ClearColor8888;
-
-
    /* Offsets of fields within the current vertex:
     */
    GLuint coloroffset;
@@ -237,6 +226,7 @@ struct intel_context
    GLboolean hw_stipple;
    GLboolean depth_buffer_is_float;
    GLboolean no_rast;
+   GLboolean no_hw;
    GLboolean always_flush_batch;
    GLboolean always_flush_cache;
 
@@ -302,13 +292,6 @@ struct intel_context
    GLboolean use_early_z;
    drm_clip_rect_t fboRect;     /**< cliprect for FBO rendering */
 
-   int perf_boxes;
-
-   GLuint do_usleeps;
-   int do_irqs;
-   GLuint irqsEmitted;
-
-   GLboolean scissor;
    drm_clip_rect_t draw_rect;
    drm_clip_rect_t scissor_rect;
 
@@ -325,8 +308,6 @@ struct intel_context
 
    GLuint lastStamp;
 
-   GLboolean no_hw;
-
    /**
     * Configuration cache
     */
@@ -373,29 +354,6 @@ do {                                               \
       (intel)->prim.flush(intel);              \
 } while (0)
 
-/* ================================================================
- * Color packing:
- */
-
-#define INTEL_PACKCOLOR4444(r,g,b,a) \
-  ((((a) & 0xf0) << 8) | (((r) & 0xf0) << 4) | ((g) & 0xf0) | ((b) >> 4))
-
-#define INTEL_PACKCOLOR1555(r,g,b,a) \
-  ((((r) & 0xf8) << 7) | (((g) & 0xf8) << 2) | (((b) & 0xf8) >> 3) | \
-    ((a) ? 0x8000 : 0))
-
-#define INTEL_PACKCOLOR565(r,g,b) \
-  ((((r) & 0xf8) << 8) | (((g) & 0xfc) << 3) | (((b) & 0xf8) >> 3))
-
-#define INTEL_PACKCOLOR8888(r,g,b,a) \
-  ((a<<24) | (r<<16) | (g<<8) | b)
-
-#define INTEL_PACKCOLOR(format, r,  g,  b, a)          \
-(format == DV_PF_555 ? INTEL_PACKCOLOR1555(r,g,b,a) :  \
- (format == DV_PF_565 ? INTEL_PACKCOLOR565(r,g,b) :    \
-  (format == DV_PF_8888 ? INTEL_PACKCOLOR8888(r,g,b,a) :       \
-   0)))
-
 /* ================================================================
  * From linux kernel i386 header files, copes with odd sizes better
  * than COPY_DWORDS would:
index 48cdae509e37cdfe6bd08ff906f9d89684a6f99d..5ac5ce10afbfd2dfb8b1c90f1b55aa348e33f212 100644 (file)
@@ -79,6 +79,7 @@ static const struct dri_extension card_extensions[] = {
    { "GL_ARB_half_float_pixel",           NULL },
    { "GL_ARB_map_buffer_range",           GL_ARB_map_buffer_range_functions },
    { "GL_ARB_multitexture",               NULL },
+   { "GL_ARB_pixel_buffer_object",      NULL },
    { "GL_ARB_point_parameters",           GL_ARB_point_parameters_functions },
    { "GL_ARB_point_sprite",               NULL },
    { "GL_ARB_shader_objects",             GL_ARB_shader_objects_functions },
@@ -104,6 +105,8 @@ static const struct dri_extension card_extensions[] = {
    { "GL_EXT_blend_logic_op",             NULL },
    { "GL_EXT_blend_subtract",             NULL },
    { "GL_EXT_cull_vertex",                GL_EXT_cull_vertex_functions },
+   { "GL_EXT_framebuffer_blit",         GL_EXT_framebuffer_blit_functions },
+   { "GL_EXT_framebuffer_object",       GL_EXT_framebuffer_object_functions },
    { "GL_EXT_fog_coord",                  GL_EXT_fog_coord_functions },
    { "GL_EXT_gpu_program_parameters",     GL_EXT_gpu_program_parameters_functions },
    { "GL_EXT_packed_depth_stencil",       NULL },
@@ -175,14 +178,6 @@ static const struct dri_extension arb_oq_extensions[] = {
    { NULL, NULL }
 };
 
-
-static const struct dri_extension ttm_extensions[] = {
-   { "GL_ARB_pixel_buffer_object",      NULL },
-   { "GL_EXT_framebuffer_blit",         GL_EXT_framebuffer_blit_functions },
-   { "GL_EXT_framebuffer_object",       GL_EXT_framebuffer_object_functions },
-   { NULL, NULL }
-};
-
 static const struct dri_extension fragment_shader_extensions[] = {
    { "GL_ARB_fragment_shader",            NULL },
    { NULL, NULL }
@@ -201,14 +196,10 @@ intelInitExtensions(GLcontext *ctx)
     */
    driInitExtensions(ctx, card_extensions, GL_FALSE);
 
-   if (intel->ttm)
-      driInitExtensions(ctx, ttm_extensions, GL_FALSE);
-
-   if (IS_965(intel->intelScreen->deviceID))
+   if (intel->gen >= 4)
       driInitExtensions(ctx, brw_extensions, GL_FALSE);
 
-   if (IS_915(intel->intelScreen->deviceID)
-       || IS_945(intel->intelScreen->deviceID)) {
+   if (intel->gen == 3) {
       driInitExtensions(ctx, i915_extensions, GL_FALSE);
 
       if (driQueryOptionb(&intel->optionCache, "fragment_shader"))
index 608f75b82406e93029db5f82e2a3af7432dae6e7..fc502a87d37fe5557b93b4188f0c61c47fc9c10f 100644 (file)
@@ -398,8 +398,6 @@ static GLboolean
 intel_update_wrapper(GLcontext *ctx, struct intel_renderbuffer *irb, 
                     struct gl_texture_image *texImage)
 {
-   gl_format texFormat;
-
    if (texImage->TexFormat == MESA_FORMAT_ARGB8888) {
       irb->Base.DataType = GL_UNSIGNED_BYTE;
       DBG("Render to RGBA8 texture OK\n");
@@ -435,8 +433,6 @@ intel_update_wrapper(GLcontext *ctx, struct intel_renderbuffer *irb,
 
    irb->Base.Format = texImage->TexFormat;
 
-   texFormat = texImage->TexFormat;
-
    irb->Base.InternalFormat = texImage->InternalFormat;
    irb->Base._BaseFormat = _mesa_base_fbo_format(ctx, irb->Base.InternalFormat);
    irb->Base.Width = texImage->Width;
index abb3024bfbb54b28bd0b1aba2dfe6f99a0fb8b19..82e4150c6a9dc016082985ee774edabf400fb86f 100644 (file)
@@ -87,7 +87,7 @@ intel_miptree_create_internal(struct intel_context *intel,
    mt->pitch = 0;
 
 #ifdef I915
-   if (IS_945(intel->intelScreen->deviceID))
+   if (intel->is_945)
       ok = i945_miptree_layout(intel, mt, tiling);
    else
       ok = i915_miptree_layout(intel, mt, tiling);
@@ -224,16 +224,12 @@ int intel_miptree_pitch_align (struct intel_context *intel,
    if (!mt->compressed) {
       int pitch_align;
 
-      if (intel->ttm) {
-        /* XXX: Align pitch to multiple of 64 bytes for now to allow
-         * render-to-texture to work in all cases. This should probably be
-         * replaced at some point by some scheme to only do this when really
-         * necessary.
-         */
-        pitch_align = 64;
-      } else {
-        pitch_align = 4;
-      }
+      /* XXX: Align pitch to multiple of 64 bytes for now to allow
+       * render-to-texture to work in all cases. This should probably be
+       * replaced at some point by some scheme to only do this when really
+       * necessary.
+       */
+      pitch_align = 64;
 
       if (tiling == I915_TILING_X)
         pitch_align = 512;
index 204a233173778b7cc5990c5b04f74b108552eea1..d0bcc2a360291d081e3c1a30bbceded055a1657b 100644 (file)
@@ -228,10 +228,9 @@ do_blit_bitmap( GLcontext *ctx,
    UNCLAMPED_FLOAT_TO_UBYTE(ubcolor[3], tmpColor[3]);
 
    if (dst->cpp == 2)
-      color = INTEL_PACKCOLOR565(ubcolor[0], ubcolor[1], ubcolor[2]);
+      color = PACK_COLOR_565(ubcolor[0], ubcolor[1], ubcolor[2]);
    else
-      color = INTEL_PACKCOLOR8888(ubcolor[0], ubcolor[1],
-                                 ubcolor[2], ubcolor[3]);
+      color = PACK_COLOR_8888(ubcolor[3], ubcolor[0], ubcolor[1], ubcolor[2]);
 
    if (!intel_check_blit_fragment_ops(ctx, tmpColor[3] == 1.0F))
       return GL_FALSE;
@@ -504,6 +503,7 @@ intel_texture_bitmap(GLcontext * ctx,
    meta_restore_fragment_program(&intel->meta);
    meta_restore_vertex_program(&intel->meta);
 
+   _mesa_ActiveTextureARB(GL_TEXTURE0_ARB + old_active_texture);
    _mesa_PopClientAttrib();
    _mesa_PopAttrib();
 
index 9b382e36225c4e98a1f50b209c9d8ec951f0b417..b870e9315e0ce75f71f18165c1b031eeffad010c 100644 (file)
@@ -69,7 +69,6 @@ intel_stencil_drawpixels(GLcontext * ctx,
    GLfloat vertices[4][2];
    struct intel_renderbuffer *irb;
    struct intel_renderbuffer *depth_irb;
-   struct gl_renderbuffer *rb;
    struct gl_pixelstore_attrib old_unpack;
    GLstencil *stencil_pixels;
    int row, y1, y2;
@@ -170,7 +169,6 @@ intel_stencil_drawpixels(GLcontext * ctx,
     */
    depth_irb = intel_get_renderbuffer(ctx->DrawBuffer, BUFFER_DEPTH);
    irb = intel_create_renderbuffer(MESA_FORMAT_ARGB8888);
-   rb = &irb->Base;
    irb->Base.Width = depth_irb->Base.Width;
    irb->Base.Height = depth_irb->Base.Height;
    intel_renderbuffer_set_region(irb, depth_irb->region);
index 80975163d479451a7715a43e611a56fa4bb18913..d6b9dc444667534d65bfae57e3c4105c79acd9c5 100644 (file)
@@ -542,55 +542,18 @@ intel_recreate_static(struct intel_context *intel,
       region->buffer = NULL;
    }
 
-   if (intel->ttm) {
-      assert(region_desc->bo_handle != -1);
-      region->buffer = intel_bo_gem_create_from_name(intel->bufmgr,
-                                                    name,
-                                                    region_desc->bo_handle);
-
-      ret = dri_bo_get_tiling(region->buffer, &region->tiling,
-                             &region->bit_6_swizzle);
-      if (ret != 0) {
-        fprintf(stderr, "Couldn't get tiling of buffer %d (%s): %s\n",
-                region_desc->bo_handle, name, strerror(-ret));
-        intel_region_release(&region);
-        return NULL;
-      }
-   } else {
-      if (region->classic_map != NULL) {
-        drmUnmap(region->classic_map,
-                 region->pitch * region->cpp * region->height);
-        region->classic_map = NULL;
-      }
-      ret = drmMap(intel->driFd, region_desc->handle,
-                  region->pitch * region->cpp * region->height,
-                  &region->classic_map);
-      if (ret != 0) {
-        fprintf(stderr, "Failed to drmMap %s buffer\n", name);
-        free(region);
-        return NULL;
-      }
-
-      region->buffer = intel_bo_fake_alloc_static(intel->bufmgr,
+   assert(region_desc->bo_handle != -1);
+   region->buffer = intel_bo_gem_create_from_name(intel->bufmgr,
                                                  name,
-                                                 region_desc->offset,
-                                                 region->pitch * region->cpp *
-                                                 region->height,
-                                                 region->classic_map);
-
-      /* The sarea just gives us a boolean for whether it's tiled or not,
-       * instead of which tiling mode it is.  Guess.
-       */
-      if (region_desc->tiled) {
-        if (intel->gen >= 4 && region_desc == &intelScreen->depth)
-           region->tiling = I915_TILING_Y;
-        else
-           region->tiling = I915_TILING_X;
-      } else {
-        region->tiling = I915_TILING_NONE;
-      }
+                                                 region_desc->bo_handle);
 
-      region->bit_6_swizzle = I915_BIT_6_SWIZZLE_NONE;
+   ret = dri_bo_get_tiling(region->buffer, &region->tiling,
+                          &region->bit_6_swizzle);
+   if (ret != 0) {
+      fprintf(stderr, "Couldn't get tiling of buffer %d (%s): %s\n",
+             region_desc->bo_handle, name, strerror(-ret));
+      intel_region_release(&region);
+      return NULL;
    }
 
    assert(region->buffer != NULL);
index 789135b49f98314aa7860766e92d067a96d8a34b..68e8db1de677f07d95f902f3e9027b76cddc7e84 100644 (file)
@@ -605,9 +605,7 @@ intelFillInModes(__DRIscreenPrivate *psp,
 static GLboolean
 intel_init_bufmgr(intelScreenPrivate *intelScreen)
 {
-   GLboolean gem_disable = getenv("INTEL_NO_GEM") != NULL;
    int gem_kernel = 0;
-   GLboolean gem_supported;
    struct drm_i915_getparam gp;
    __DRIscreenPrivate *spriv = intelScreen->driScrnPriv;
    int num_fences = 0;
@@ -622,43 +620,19 @@ intel_init_bufmgr(intelScreenPrivate *intelScreen)
    /* If we've got a new enough DDX that's initializing GEM and giving us
     * object handles for the shared buffers, use that.
     */
-   intelScreen->ttm = GL_FALSE;
-   if (intelScreen->driScrnPriv->dri2.enabled)
-       gem_supported = GL_TRUE;
-   else if (intelScreen->driScrnPriv->ddx_version.minor >= 9 &&
-           gem_kernel &&
-           intelScreen->front.bo_handle != -1)
-       gem_supported = GL_TRUE;
-   else
-       gem_supported = GL_FALSE;
-
-   if (!gem_disable && gem_supported) {
-      intelScreen->bufmgr = intel_bufmgr_gem_init(spriv->fd, BATCH_SZ);
-      if (intelScreen->bufmgr != NULL)
-        intelScreen->ttm = GL_TRUE;
+   if (!intelScreen->driScrnPriv->dri2.enabled &&
+       intelScreen->driScrnPriv->ddx_version.minor < 9) {
+      fprintf(stderr, "[%s:%u] Error initializing GEM.\n",
+             __func__, __LINE__);
+      return GL_FALSE;
    }
+
+   intelScreen->bufmgr = intel_bufmgr_gem_init(spriv->fd, BATCH_SZ);
    /* Otherwise, use the classic buffer manager. */
    if (intelScreen->bufmgr == NULL) {
-      if (gem_disable) {
-        _mesa_warning(NULL, "GEM disabled.  Using classic.");
-      } else {
-        _mesa_warning(NULL,
-                       "Failed to initialize GEM.  Falling back to classic.");
-      }
-
-      if (intelScreen->tex.size == 0) {
-        fprintf(stderr, "[%s:%u] Error initializing buffer manager.\n",
-                __func__, __LINE__);
-        return GL_FALSE;
-      }
-
-      intelScreen->bufmgr =
-        intel_bufmgr_fake_init(spriv->fd,
-                               intelScreen->tex.offset,
-                               intelScreen->tex.map,
-                               intelScreen->tex.size,
-                               (unsigned int * volatile)
-                               &intelScreen->sarea->last_dispatch);
+      fprintf(stderr, "[%s:%u] Error initializing buffer manager.\n",
+             __func__, __LINE__);
+      return GL_FALSE;
    }
 
    if (intel_get_param(spriv, I915_PARAM_NUM_FENCES_AVAIL, &num_fences))
index a9b9e109a6aeec05838e0900471dd248c466f192..14ca0903b6c332bdbf9c3538e70a19ca9d306aba 100644 (file)
@@ -77,7 +77,6 @@ typedef struct
    GLboolean no_hw;
 
    GLboolean no_vbo;
-   int ttm;
    dri_bufmgr *bufmgr;
    GLboolean kernel_exec_fencing;
 
index 34c3d9df74c99db401c4dae8adbe712b32924f9f..d1681e9088a36aa0a613e8258f836820355dabdc 100644 (file)
@@ -613,15 +613,7 @@ intel_set_span_functions(struct intel_context *intel,
                         struct gl_renderbuffer *rb)
 {
    struct intel_renderbuffer *irb = (struct intel_renderbuffer *) rb;
-   uint32_t tiling;
-
-   /* If in GEM mode, we need to do the tile address swizzling ourselves,
-    * instead of the fence registers handling it.
-    */
-   if (intel->ttm)
-      tiling = irb->region->tiling;
-   else
-      tiling = I915_TILING_NONE;
+   uint32_t tiling = irb->region->tiling;
 
    if (intel->intelScreen->kernel_exec_fencing) {
       switch (irb->Base.Format) {
@@ -667,6 +659,9 @@ intel_set_span_functions(struct intel_context *intel,
       return;
    }
 
+   /* If in GEM mode, we need to do the tile address swizzling ourselves,
+    * instead of the fence registers handling it.
+    */
    switch (irb->Base.Format) {
    case MESA_FORMAT_RGB565:
       switch (tiling) {
index 4ee742377d696e563756afcb3c3ebd2f746123b4..aefae53eb20750fe965b6ceed8db290189726347 100644 (file)
@@ -196,25 +196,6 @@ intel_translate_logic_op(GLenum opcode)
    }
 }
 
-
-static void
-intelClearColor(GLcontext *ctx, const GLfloat color[4])
-{
-   struct intel_context *intel = intel_context(ctx);
-   GLubyte clear[4];
-
-   CLAMPED_FLOAT_TO_UBYTE(clear[0], color[0]);
-   CLAMPED_FLOAT_TO_UBYTE(clear[1], color[1]);
-   CLAMPED_FLOAT_TO_UBYTE(clear[2], color[2]);
-   CLAMPED_FLOAT_TO_UBYTE(clear[3], color[3]);
-
-   /* compute both 32 and 16-bit clear values */
-   intel->ClearColor8888 = INTEL_PACKCOLOR8888(clear[0], clear[1],
-                                               clear[2], clear[3]);
-   intel->ClearColor565 = INTEL_PACKCOLOR565(clear[0], clear[1], clear[2]);
-}
-
-
 /* Fallback to swrast for select and feedback.
  */
 static void
@@ -229,5 +210,4 @@ void
 intelInitStateFuncs(struct dd_function_table *functions)
 {
    functions->RenderMode = intelRenderMode;
-   functions->ClearColor = intelClearColor;
 }
index 7d035b9f6e41aff5f55a229eaafa70fc21c5b323..5ae124071871fd7d7405f6ba32a09e108041a3f3 100644 (file)
@@ -56,7 +56,7 @@ intelFixupVblank(struct intel_context *intel, __DRIdrawablePrivate *dPriv)
                                     .y2 = sarea->planeB_y + sarea->planeB_h };
       GLint areaA = driIntersectArea( drw_rect, planeA_rect );
       GLint areaB = driIntersectArea( drw_rect, planeB_rect );
-      GLuint flags = dPriv->vblFlags;
+      GLuint flags;
 
       /* Update vblank info
        */
index 87efb72cc519221e25b509872d42005f44b1ab4e..a7c6c45ffe41da25a3a25b8ee324a719de0352df 100644 (file)
@@ -173,13 +173,13 @@ intelChooseTextureFormat(GLcontext * ctx, GLint internalFormat,
       return MESA_FORMAT_SARGB8;
    case GL_SLUMINANCE_EXT:
    case GL_SLUMINANCE8_EXT:
-      if (IS_G4X(intel->intelScreen->deviceID))
+      if (intel->has_luminance_srgb)
          return MESA_FORMAT_SL8;
       else
          return MESA_FORMAT_SARGB8;
    case GL_SLUMINANCE_ALPHA_EXT:
    case GL_SLUMINANCE8_ALPHA8_EXT:
-      if (IS_G4X(intel->intelScreen->deviceID))
+      if (intel->has_luminance_srgb)
          return MESA_FORMAT_SLA8;
       else
          return MESA_FORMAT_SARGB8;
index 99f1a14e171c25ab644df7b099a6389cd4e1e96b..816682ec5f13d3b0f6f2d2fbe3a0d16866cd51d7 100644 (file)
@@ -207,19 +207,19 @@ INTERP_QUALIFIER void TAG(interp_extras)( GLcontext *ctx,
    LOCALVARS
    struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
-   if (VB->ColorPtr[1]) {
-      assert(VB->ColorPtr[1]->stride == 4 * sizeof(GLfloat));
+   if (VB->BackfaceColorPtr) {
+      assert(VB->BackfaceColorPtr->stride == 4 * sizeof(GLfloat));
       
       INTERP_4F( t,
-                   GET_COLOR(VB->ColorPtr[1], dst),
-                   GET_COLOR(VB->ColorPtr[1], out),
-                   GET_COLOR(VB->ColorPtr[1], in) );
+                GET_COLOR(VB->BackfaceColorPtr, dst),
+                GET_COLOR(VB->BackfaceColorPtr, out),
+                GET_COLOR(VB->BackfaceColorPtr, in) );
 
-      if (VB->SecondaryColorPtr[1]) {
+      if (VB->BackfaceSecondaryColorPtr) {
         INTERP_3F( t,
-                      GET_COLOR(VB->SecondaryColorPtr[1], dst),
-                      GET_COLOR(VB->SecondaryColorPtr[1], out),
-                      GET_COLOR(VB->SecondaryColorPtr[1], in) );
+                   GET_COLOR(VB->BackfaceSecondaryColorPtr, dst),
+                   GET_COLOR(VB->BackfaceSecondaryColorPtr, out),
+                   GET_COLOR(VB->BackfaceSecondaryColorPtr, in) );
       }
    }
 
@@ -236,13 +236,13 @@ INTERP_QUALIFIER void TAG(copy_pv_extras)( GLcontext *ctx,
    LOCALVARS
       struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
-   if (VB->ColorPtr[1]) {
-      COPY_4FV( GET_COLOR(VB->ColorPtr[1], dst), 
-               GET_COLOR(VB->ColorPtr[1], src) );
+   if (VB->BackfaceColorPtr) {
+      COPY_4FV( GET_COLOR(VB->BackfaceColorPtr, dst),
+               GET_COLOR(VB->BackfaceColorPtr, src) );
 
-      if (VB->SecondaryColorPtr[1]) {
-        COPY_4FV( GET_COLOR(VB->SecondaryColorPtr[1], dst), 
-                  GET_COLOR(VB->SecondaryColorPtr[1], src) );
+      if (VB->BackfaceSecondaryColorPtr) {
+        COPY_4FV( GET_COLOR(VB->BackfaceSecondaryColorPtr, dst),
+                  GET_COLOR(VB->BackfaceSecondaryColorPtr, src) );
       }
    }
 
index 684f2acc890849ac4fba0e0ad05927879bc5b7e0..6e5fa3520e111f0eb5b06feb0198117bc570c7ac 100644 (file)
@@ -103,10 +103,10 @@ static void TAG(emit)( GLcontext *ctx,
 #if DO_TEX1
    {
       const GLuint t1 = GET_TEXSOURCE(1);
-      tc1 = VB->TexCoordPtr[t1]->data;
-      tc1_stride = VB->TexCoordPtr[t1]->stride;
+      tc1 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->data;
+      tc1_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->stride;
 #if DO_PTEX
-      tc1_size = VB->TexCoordPtr[t1]->size;
+      tc1_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->size;
 #endif
    }
 #endif
@@ -114,18 +114,18 @@ static void TAG(emit)( GLcontext *ctx,
 #if DO_TEX0
    {
       const GLuint t0 = GET_TEXSOURCE(0);
-      tc0 = VB->TexCoordPtr[t0]->data;
-      tc0_stride = VB->TexCoordPtr[t0]->stride;
+      tc0 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->data;
+      tc0_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->stride;
 #if DO_PTEX
-      tc0_size = VB->TexCoordPtr[t0]->size;
+      tc0_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->size;
 #endif
    }
 #endif
 
 #if DO_SPEC
-   if (VB->SecondaryColorPtr[0]) {
-      spec = VB->SecondaryColorPtr[0]->data;
-      spec_stride = VB->SecondaryColorPtr[0]->stride;
+   if (VB->AttribPtr[_TNL_ATTRIB_COLOR1]) {
+      spec = VB->AttribPtr[_TNL_ATTRIB_COLOR1]->data;
+      spec_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR1]->stride;
    } else {
       spec = (GLfloat (*)[4])ctx->Current.Attrib[VERT_ATTRIB_COLOR1];
       spec_stride = 0;
@@ -133,9 +133,9 @@ static void TAG(emit)( GLcontext *ctx,
 #endif
 
 #if DO_FOG
-   if (VB->FogCoordPtr) {
-      fog = VB->FogCoordPtr->data;
-      fog_stride = VB->FogCoordPtr->stride;
+   if (VB->AttribPtr[_TNL_ATTRIB_FOG]) {
+      fog = VB->AttribPtr[_TNL_ATTRIB_FOG]->data;
+      fog_stride = VB->AttribPtr[_TNL_ATTRIB_FOG]->stride;
    } else {
       static GLfloat tmp[4] = {0, 0, 0, 0};
       fog = &tmp;
@@ -144,8 +144,8 @@ static void TAG(emit)( GLcontext *ctx,
 #endif
 
 #if DO_RGBA
-   col = VB->ColorPtr[0]->data;
-   col_stride = VB->ColorPtr[0]->stride;
+   col = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+   col_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
 #endif
 
    coord = VB->NdcPtr->data;
@@ -319,8 +319,8 @@ static GLboolean TAG(check_tex_sizes)( GLcontext *ctx )
 
    /* Force 'missing' texcoords to something valid.
     */
-   if (DO_TEX1 && VB->TexCoordPtr[0] == 0)
-      VB->TexCoordPtr[0] = VB->TexCoordPtr[1];
+   if (DO_TEX1 && VB->AttribPtr[_TNL_ATTRIB_TEX0] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX0] = VB->AttribPtr[_TNL_ATTRIB_TEX1];
 
    if (DO_PTEX)
       return GL_TRUE;
@@ -328,12 +328,12 @@ static GLboolean TAG(check_tex_sizes)( GLcontext *ctx )
    /* No hardware support for projective texture.  Can fake it for
     * TEX0 only.
     */
-   if ((DO_TEX1 && VB->TexCoordPtr[GET_TEXSOURCE(1)]->size == 4)) {
+   if ((DO_TEX1 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(1)]->size == 4)) {
       PTEX_FALLBACK();
       return GL_FALSE;
    }
 
-   if (DO_TEX0 && VB->TexCoordPtr[GET_TEXSOURCE(0)]->size == 4) {
+   if (DO_TEX0 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(0)]->size == 4) {
       if (DO_TEX1) {
         PTEX_FALLBACK();
       }
index a757362b11d82cf75d55863441675f675abce781..72917ee13bfa76180df67c6ea80edc0da289fc1f 100644 (file)
@@ -565,7 +565,6 @@ void mach64InitTextureFuncs( struct dd_function_table *functions )
    functions->IsTextureResident                = driIsTextureResident;
 
    functions->UpdateTexturePalette     = NULL;
-   functions->ActiveTexture            = NULL;
 
    driInitTextureFormats();
 }
index 938804af9ecb97f6a851c7c683f8d2102458443a..60bfab8f6dc7f78c60bc7e810a2f06626a434a9b 100644 (file)
@@ -156,53 +156,53 @@ static void TAG(emit)( GLcontext *ctx,
 
    if (DO_TEX3) {
       const GLuint t3 = GET_TEXSOURCE(3);
-      tc3 = VB->TexCoordPtr[t3]->data;
-      tc3_stride = VB->TexCoordPtr[t3]->stride;
+      tc3 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t3]->data;
+      tc3_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t3]->stride;
       if (DO_PTEX)
-        tc3_size = VB->TexCoordPtr[t3]->size;
+        tc3_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t3]->size;
    }
 
    if (DO_TEX2) {
       const GLuint t2 = GET_TEXSOURCE(2);
-      tc2 = VB->TexCoordPtr[t2]->data;
-      tc2_stride = VB->TexCoordPtr[t2]->stride;
+      tc2 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t2]->data;
+      tc2_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t2]->stride;
       if (DO_PTEX)
-        tc2_size = VB->TexCoordPtr[t2]->size;
+        tc2_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t2]->size;
    }
 
    if (DO_TEX1) {
       const GLuint t1 = GET_TEXSOURCE(1);
-      tc1 = VB->TexCoordPtr[t1]->data;
-      tc1_stride = VB->TexCoordPtr[t1]->stride;
+      tc1 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->data;
+      tc1_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->stride;
       if (DO_PTEX)
-        tc1_size = VB->TexCoordPtr[t1]->size;
+        tc1_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->size;
    }
 
    if (DO_TEX0) {
       const GLuint t0 = GET_TEXSOURCE(0);
-      tc0_stride = VB->TexCoordPtr[t0]->stride;
-      tc0 = VB->TexCoordPtr[t0]->data;
+      tc0_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->stride;
+      tc0 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->data;
       if (DO_PTEX) 
-        tc0_size = VB->TexCoordPtr[t0]->size;
+        tc0_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->size;
    }
 
    if (DO_RGBA) {
-      col = VB->ColorPtr[0]->data;
-      col_stride = VB->ColorPtr[0]->stride;
+      col = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+      col_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
    }
 
    if (DO_SPEC) {
-      spec = VB->SecondaryColorPtr[0]->data;
-      spec_stride = VB->SecondaryColorPtr[0]->stride;
+      spec = VB->AttribPtr[_TNL_ATTRIB_COLOR1]->data;
+      spec_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR1]->stride;
    } else {
       spec = (GLfloat (*)[4])ctx->Current.Attrib[VERT_ATTRIB_COLOR1];
       spec_stride = 0;
    }
 
    if (DO_FOG) {
-      if (VB->FogCoordPtr) {
-        fog = VB->FogCoordPtr->data;
-        fog_stride = VB->FogCoordPtr->stride;
+      if (VB->AttribPtr[_TNL_ATTRIB_FOG]) {
+        fog = VB->AttribPtr[_TNL_ATTRIB_FOG]->data;
+        fog_stride = VB->AttribPtr[_TNL_ATTRIB_FOG]->stride;
       } else {
         static GLfloat tmp[4] = {0, 0, 0, 0};
         fog = &tmp;
@@ -384,8 +384,8 @@ static void TAG(emit)( GLcontext *ctx, GLuint start, GLuint end,
 
    ASSERT(stride == 4);
 
-   col = VB->ColorPtr[0]->data;
-   col_stride = VB->ColorPtr[0]->stride;
+   col = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+   col_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
 
    /* Pack what's left into a 4-dword vertex.  Color is in a different
     * place, and there is no 'w' coordinate.
@@ -432,8 +432,8 @@ static void TAG(emit)( GLcontext *ctx, GLuint start, GLuint end,
    GLfloat *v = (GLfloat *)dest;
    int i;
 
-   col = VB->ColorPtr[0]->data;
-   col_stride = VB->ColorPtr[0]->stride;
+   col = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+   col_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
 
    if (start)
       STRIDE_4F(col, col_stride * start);
@@ -473,22 +473,22 @@ static GLboolean TAG(check_tex_sizes)( GLcontext *ctx )
 
    /* Force 'missing' texcoords to something valid.
     */
-   if (DO_TEX3 && VB->TexCoordPtr[2] == 0)
-      VB->TexCoordPtr[2] = VB->TexCoordPtr[3];
+   if (DO_TEX3 && VB->AttribPtr[_TNL_ATTRIB_TEX2] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX2] = VB->AttribPtr[_TNL_ATTRIB_TEX3];
 
-   if (DO_TEX2 && VB->TexCoordPtr[1] == 0)
-      VB->TexCoordPtr[1] = VB->TexCoordPtr[2];
+   if (DO_TEX2 && VB->AttribPtr[_TNL_ATTRIB_TEX1] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX1] = VB->AttribPtr[_TNL_ATTRIB_TEX2];
 
-   if (DO_TEX1 && VB->TexCoordPtr[0] == 0)
-      VB->TexCoordPtr[0] = VB->TexCoordPtr[1];
+   if (DO_TEX1 && VB->AttribPtr[_TNL_ATTRIB_TEX0] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX0] = VB->AttribPtr[_TNL_ATTRIB_TEX1];
 
    if (DO_PTEX)
       return GL_TRUE;
    
-   if ((DO_TEX3 && VB->TexCoordPtr[GET_TEXSOURCE(3)]->size == 4) ||
-       (DO_TEX2 && VB->TexCoordPtr[GET_TEXSOURCE(2)]->size == 4) ||
-       (DO_TEX1 && VB->TexCoordPtr[GET_TEXSOURCE(1)]->size == 4) ||
-       (DO_TEX0 && VB->TexCoordPtr[GET_TEXSOURCE(0)]->size == 4))
+   if ((DO_TEX3 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(3)]->size == 4) ||
+       (DO_TEX2 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(2)]->size == 4) ||
+       (DO_TEX1 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(1)]->size == 4) ||
+       (DO_TEX0 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(0)]->size == 4))
       return GL_FALSE;
 
    return GL_TRUE;
@@ -501,14 +501,14 @@ static GLboolean TAG(check_tex_sizes)( GLcontext *ctx )
 
    /* Force 'missing' texcoords to something valid.
     */
-   if (DO_TEX3 && VB->TexCoordPtr[2] == 0)
-      VB->TexCoordPtr[2] = VB->TexCoordPtr[3];
+   if (DO_TEX3 && VB->AttribPtr[_TNL_ATTRIB_TEX2] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX2] = VB->AttribPtr[_TNL_ATTRIB_TEX3];
 
-   if (DO_TEX2 && VB->TexCoordPtr[1] == 0)
-      VB->TexCoordPtr[1] = VB->TexCoordPtr[2];
+   if (DO_TEX2 && VB->AttribPtr[_TNL_ATTRIB_TEX1] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX1] = VB->AttribPtr[_TNL_ATTRIB_TEX2];
 
-   if (DO_TEX1 && VB->TexCoordPtr[0] == 0)
-      VB->TexCoordPtr[0] = VB->TexCoordPtr[1];
+   if (DO_TEX1 && VB->AttribPtr[_TNL_ATTRIB_TEX0] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX0] = VB->AttribPtr[_TNL_ATTRIB_TEX1];
 
    if (DO_PTEX)
       return GL_TRUE;
@@ -516,14 +516,14 @@ static GLboolean TAG(check_tex_sizes)( GLcontext *ctx )
    /* No hardware support for projective texture.  Can fake it for
     * TEX0 only.
     */
-   if ((DO_TEX3 && VB->TexCoordPtr[GET_TEXSOURCE(3)]->size == 4) ||
-       (DO_TEX2 && VB->TexCoordPtr[GET_TEXSOURCE(2)]->size == 4) ||
-       (DO_TEX1 && VB->TexCoordPtr[GET_TEXSOURCE(1)]->size == 4)) {
+   if ((DO_TEX3 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(3)]->size == 4) ||
+       (DO_TEX2 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(2)]->size == 4) ||
+       (DO_TEX1 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(1)]->size == 4)) {
       PTEX_FALLBACK();
       return GL_FALSE;
    }
 
-   if (DO_TEX0 && VB->TexCoordPtr[GET_TEXSOURCE(0)]->size == 4) {
+   if (DO_TEX0 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(0)]->size == 4) {
       if (DO_TEX1 || DO_TEX2 || DO_TEX3) {
         PTEX_FALLBACK();
       }
index 82d825b8d282e3e78d5a3b8863f6ac85f5014c74..86d4717b050fa94015877f4e2653518a0ff4ee5e 100644 (file)
@@ -651,12 +651,12 @@ static void r128RenderStart( GLcontext *ctx )
    }
 
    if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(rmesa->tmu_source[0]) )) {
-      if ( VB->TexCoordPtr[rmesa->tmu_source[0]]->size > 2 )
+      if ( VB->AttribPtr[_TNL_ATTRIB_TEX0 + rmesa->tmu_source[0]]->size > 2 )
         fallback_projtex = GL_TRUE;
       EMIT_ATTR( _TNL_ATTRIB_TEX0, EMIT_2F, R128_CCE_VC_FRMT_S_T, 8 );
    }
    if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(rmesa->tmu_source[1]) )) {
-      if ( VB->TexCoordPtr[rmesa->tmu_source[1]]->size > 2 )
+      if ( VB->AttribPtr[_TNL_ATTRIB_TEX0 + rmesa->tmu_source[1]]->size > 2 )
         fallback_projtex = GL_TRUE;
       EMIT_ATTR( _TNL_ATTRIB_TEX1, EMIT_2F, R128_CCE_VC_FRMT_S2_T2, 8 );
    }
index 383a0c4b0d3ea6c67c3f8ea1d973a77fcec32f4c..249c0bbc11d2ef7236f5d08d786c010ed7a1e8bc 100644 (file)
@@ -90,12 +90,14 @@ static void r200_emit_vecfog(GLcontext *ctx, struct radeon_aos *aos,
        aos->components = size;
        aos->count = count;
 
+       radeon_bo_map(aos->bo, 1);
        out = (uint32_t*)((char*)aos->bo->ptr + aos->offset);
        for (i = 0; i < count; i++) {
          out[0] = r200ComputeFogBlendFactor( ctx, *(GLfloat *)data );
          out++;
          data += stride;
        }
+       radeon_bo_unmap(aos->bo);
 }
 
 /* Emit any changed arrays to new GART memory, re-emit a packet to
index 240fb450780b6107542df24a4a28e30a54d2859c..4596912ddc75ed33bf0676b0b9a8134cba026af8 100644 (file)
@@ -168,7 +168,7 @@ static void r200SetVertexFormat( GLcontext *ctx )
 
       for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
         if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(i) )) {
-           GLuint sz = VB->TexCoordPtr[i]->size;
+           GLuint sz = VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]->size;
 
            fmt_1 |= sz << (3 * i);
            EMIT_ATTR( _TNL_ATTRIB_TEX0+i, EMIT_1F + sz - 1, 0 );
@@ -297,7 +297,7 @@ void r200_swtcl_flush(GLcontext *ctx, uint32_t current_offset)
    radeonEmitState(&rmesa->radeon);
    r200EmitVertexAOS( rmesa,
                      rmesa->radeon.swtcl.vertex_size,
-                     first_elem(&rmesa->radeon.dma.reserved)->bo,
+                     rmesa->radeon.swtcl.bo,
                      current_offset);
 
 
index 7782404a7944711d3ea6f9b42ae5830589b96159..e2f9cf0ea86f9a03e3f1764682dfb5e59f9cdfe0 100644 (file)
@@ -797,24 +797,13 @@ void r200SetTexBuffer2(__DRIcontext *pDRICtx, GLint target, GLint glx_texture_fo
            return;
        }
 
-       radeon_update_renderbuffers(pDRICtx, dPriv);
-       /* back & depth buffer are useless free them right away */
-       rb = (void*)rfb->base.Attachment[BUFFER_DEPTH].Renderbuffer;
-       if (rb && rb->bo) {
-               radeon_bo_unref(rb->bo);
-        rb->bo = NULL;
-       }
-       rb = (void*)rfb->base.Attachment[BUFFER_BACK_LEFT].Renderbuffer;
-       if (rb && rb->bo) {
-               radeon_bo_unref(rb->bo);
-               rb->bo = NULL;
-       }
+       radeon_update_renderbuffers(pDRICtx, dPriv, GL_TRUE);
        rb = rfb->color_rb[0];
        if (rb->bo == NULL) {
                /* Failed to BO for the buffer */
                return;
        }
-       
+
        _mesa_lock_texture(radeon->glCtx, texObj);
        if (t->bo) {
                radeon_bo_unref(t->bo);
index f87a6dd2ddcffc16ae0572498efc99c9cc5d34a8..be005bd1641f3edeac04c52d9f06641f036e9d1b 100644 (file)
@@ -43,13 +43,14 @@ RADEON_COMMON_SOURCES = \
 
 DRIVER_SOURCES = \
                 radeon_screen.c \
+                r300_blit.c \
                 r300_context.c \
                 r300_draw.c \
-                r300_ioctl.c \
                 r300_cmdbuf.c \
                 r300_state.c \
                 r300_render.c \
                 r300_tex.c \
+                r300_texcopy.c \
                 r300_texstate.c \
                 r300_vertprog.c \
                 r300_fragprog_common.c \
diff --git a/src/mesa/drivers/dri/r300/r300_blit.c b/src/mesa/drivers/dri/r300/r300_blit.c
new file mode 100644 (file)
index 0000000..ea626d9
--- /dev/null
@@ -0,0 +1,607 @@
+/*
+ * Copyright (C) 2009 Maciej Cencora <m.cencora@gmail.com>
+ *
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial
+ * portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "radeon_common.h"
+#include "r300_context.h"
+
+#include "r300_blit.h"
+#include "r300_cmdbuf.h"
+#include "r300_emit.h"
+#include "r300_tex.h"
+#include "compiler/radeon_compiler.h"
+#include "compiler/radeon_opcodes.h"
+
+static void vp_ins_outs(struct r300_vertex_program_compiler *c)
+{
+    c->code->inputs[VERT_ATTRIB_POS] = 0;
+    c->code->inputs[VERT_ATTRIB_TEX0] = 1;
+    c->code->outputs[VERT_RESULT_HPOS] = 0;
+    c->code->outputs[VERT_RESULT_TEX0] = 1;
+}
+
+static void fp_allocate_hw_inputs(
+    struct r300_fragment_program_compiler * c,
+    void (*allocate)(void * data, unsigned input, unsigned hwreg),
+    void * mydata)
+{
+    allocate(mydata, FRAG_ATTRIB_TEX0, 0);
+}
+
+static void create_vertex_program(struct r300_context *r300)
+{
+    struct r300_vertex_program_compiler compiler;
+    struct rc_instruction *inst;
+
+    rc_init(&compiler.Base);
+
+    inst = rc_insert_new_instruction(&compiler.Base, compiler.Base.Program.Instructions.Prev);
+    inst->U.I.Opcode = RC_OPCODE_MOV;
+    inst->U.I.DstReg.File = RC_FILE_OUTPUT;
+    inst->U.I.DstReg.Index = VERT_RESULT_HPOS;
+    inst->U.I.DstReg.RelAddr = 0;
+    inst->U.I.DstReg.WriteMask = RC_MASK_XYZW;
+    inst->U.I.SrcReg[0].Abs = 0;
+    inst->U.I.SrcReg[0].File = RC_FILE_INPUT;
+    inst->U.I.SrcReg[0].Index = VERT_ATTRIB_POS;
+    inst->U.I.SrcReg[0].Negate = 0;
+    inst->U.I.SrcReg[0].RelAddr = 0;
+    inst->U.I.SrcReg[0].Swizzle = RC_SWIZZLE_XYZW;
+
+    inst = rc_insert_new_instruction(&compiler.Base, compiler.Base.Program.Instructions.Prev);
+    inst->U.I.Opcode = RC_OPCODE_MOV;
+    inst->U.I.DstReg.File = RC_FILE_OUTPUT;
+    inst->U.I.DstReg.Index = VERT_RESULT_TEX0;
+    inst->U.I.DstReg.RelAddr = 0;
+    inst->U.I.DstReg.WriteMask = RC_MASK_XYZW;
+    inst->U.I.SrcReg[0].Abs = 0;
+    inst->U.I.SrcReg[0].File = RC_FILE_INPUT;
+    inst->U.I.SrcReg[0].Index = VERT_ATTRIB_TEX0;
+    inst->U.I.SrcReg[0].Negate = 0;
+    inst->U.I.SrcReg[0].RelAddr = 0;
+    inst->U.I.SrcReg[0].Swizzle = RC_SWIZZLE_XYZW;
+
+    compiler.Base.Program.InputsRead = (1 << VERT_ATTRIB_POS) | (1 << VERT_ATTRIB_TEX0);
+    compiler.RequiredOutputs = compiler.Base.Program.OutputsWritten = (1 << VERT_RESULT_HPOS) | (1 << VERT_RESULT_TEX0);
+    compiler.SetHwInputOutput = vp_ins_outs;
+    compiler.code = &r300->blit.vp_code;
+
+    r3xx_compile_vertex_program(&compiler);
+}
+
+static void create_fragment_program(struct r300_context *r300)
+{
+    struct r300_fragment_program_compiler compiler;
+    struct rc_instruction *inst;
+
+    rc_init(&compiler.Base);
+
+    inst = rc_insert_new_instruction(&compiler.Base, compiler.Base.Program.Instructions.Prev);
+    inst->U.I.Opcode = RC_OPCODE_TEX;
+    inst->U.I.TexSrcTarget = RC_TEXTURE_2D;
+    inst->U.I.TexSrcUnit = 0;
+    inst->U.I.DstReg.File = RC_FILE_OUTPUT;
+    inst->U.I.DstReg.Index = FRAG_RESULT_COLOR;
+    inst->U.I.DstReg.WriteMask = RC_MASK_XYZW;
+    inst->U.I.SrcReg[0].Abs = 0;
+    inst->U.I.SrcReg[0].File = RC_FILE_INPUT;
+    inst->U.I.SrcReg[0].Index = FRAG_ATTRIB_TEX0;
+    inst->U.I.SrcReg[0].Negate = 0;
+    inst->U.I.SrcReg[0].RelAddr = 0;
+    inst->U.I.SrcReg[0].Swizzle = RC_SWIZZLE_XYZW;
+
+    compiler.Base.Program.InputsRead = (1 << FRAG_ATTRIB_TEX0);
+    compiler.OutputColor = FRAG_RESULT_COLOR;
+    compiler.OutputDepth = FRAG_RESULT_DEPTH;
+    compiler.is_r500 = (r300->radeon.radeonScreen->chip_family >= CHIP_FAMILY_RV515);
+    compiler.code = &r300->blit.fp_code;
+    compiler.AllocateHwInputs = fp_allocate_hw_inputs;
+
+    r3xx_compile_fragment_program(&compiler);
+}
+
+void r300_blit_init(struct r300_context *r300)
+{
+    create_vertex_program(r300);
+    create_fragment_program(r300);
+}
+
+static void r300_emit_tx_setup(struct r300_context *r300,
+                               gl_format mesa_format,
+                               struct radeon_bo *bo,
+                               intptr_t offset,
+                               unsigned width,
+                               unsigned height,
+                               unsigned pitch)
+{
+    BATCH_LOCALS(&r300->radeon);
+
+    assert(width <= 2048);
+    assert(height <= 2048);
+    assert(r300TranslateTexFormat(mesa_format) >= 0);
+    assert(offset % 32 == 0);
+
+    BEGIN_BATCH(17);
+    OUT_BATCH_REGVAL(R300_TX_FILTER0_0,
+                     (R300_TX_CLAMP_TO_EDGE  << R300_TX_WRAP_S_SHIFT) |
+                     (R300_TX_CLAMP_TO_EDGE  << R300_TX_WRAP_T_SHIFT) |
+                     (R300_TX_CLAMP_TO_EDGE  << R300_TX_WRAP_R_SHIFT) |
+                     R300_TX_MIN_FILTER_MIP_NONE |
+                     R300_TX_MIN_FILTER_LINEAR |
+                     R300_TX_MAG_FILTER_LINEAR |
+                     (0 << 28));
+    OUT_BATCH_REGVAL(R300_TX_FILTER1_0, 0);
+    OUT_BATCH_REGVAL(R300_TX_SIZE_0,
+                     ((width-1) << R300_TX_WIDTHMASK_SHIFT) |
+                     ((height-1) << R300_TX_HEIGHTMASK_SHIFT) |
+                     (0 << R300_TX_DEPTHMASK_SHIFT) |
+                     (0 << R300_TX_MAX_MIP_LEVEL_SHIFT) |
+                     R300_TX_SIZE_TXPITCH_EN);
+
+    OUT_BATCH_REGVAL(R300_TX_FORMAT_0, r300TranslateTexFormat(mesa_format));
+    OUT_BATCH_REGVAL(R300_TX_FORMAT2_0, pitch - 1);
+    OUT_BATCH_REGSEQ(R300_TX_OFFSET_0, 1);
+    OUT_BATCH_RELOC(0, bo, offset, RADEON_GEM_DOMAIN_GTT|RADEON_GEM_DOMAIN_VRAM, 0, 0);
+
+    OUT_BATCH_REGSEQ(R300_TX_INVALTAGS, 2);
+    OUT_BATCH(0);
+    OUT_BATCH(1);
+
+    END_BATCH();
+}
+
+#define EASY_US_FORMAT(FMT, C0, C1, C2, C3, SIGN) \
+    (FMT  | R500_C0_SEL_##C0 | R500_C1_SEL_##C1 | \
+    R500_C2_SEL_##C2 | R500_C3_SEL_##C3 | R500_OUT_SIGN(SIGN))
+
+static uint32_t mesa_format_to_us_format(gl_format mesa_format)
+{
+    switch(mesa_format)
+    {
+        case MESA_FORMAT_S8_Z24:
+        case MESA_FORMAT_X8_Z24:
+        case MESA_FORMAT_RGBA8888: // x
+            return EASY_US_FORMAT(R500_OUT_FMT_C4_8, A, B, G, R, 0);
+        case MESA_FORMAT_RGB565: // x
+        case MESA_FORMAT_ARGB1555: // x
+        case MESA_FORMAT_RGBA8888_REV: // x
+            return EASY_US_FORMAT(R500_OUT_FMT_C4_8, R, G, B, A, 0);
+        case MESA_FORMAT_ARGB8888: // x
+            return EASY_US_FORMAT(R500_OUT_FMT_C4_8, B, G, R, A, 0);
+        case MESA_FORMAT_ARGB8888_REV:
+            return EASY_US_FORMAT(R500_OUT_FMT_C4_8, A, R, G, B, 0);
+        case MESA_FORMAT_XRGB8888:
+            return EASY_US_FORMAT(R500_OUT_FMT_C4_8, A, R, G, B, 0);
+
+        case MESA_FORMAT_RGB332:
+            return EASY_US_FORMAT(R500_OUT_FMT_C_3_3_2, A, R, G, B, 0);
+
+        case MESA_FORMAT_RGBA_FLOAT32:
+            return EASY_US_FORMAT(R500_OUT_FMT_C4_32_FP, R, G, B, A, 0);
+        case MESA_FORMAT_RGBA_FLOAT16:
+            return EASY_US_FORMAT(R500_OUT_FMT_C4_16_FP, R, G, B, A, 0);
+        case MESA_FORMAT_ALPHA_FLOAT32:
+            return EASY_US_FORMAT(R500_OUT_FMT_C_32_FP, A, A, A, A, 0);
+        case MESA_FORMAT_ALPHA_FLOAT16:
+            return EASY_US_FORMAT(R500_OUT_FMT_C_16_FP, A, A, A, A, 0);
+
+        case MESA_FORMAT_SIGNED_RGBA8888:
+            return EASY_US_FORMAT(R500_OUT_FMT_C4_8, R, G, B, A, 0xf);
+        case MESA_FORMAT_SIGNED_RGBA8888_REV:
+            return EASY_US_FORMAT(R500_OUT_FMT_C4_8, A, B, G, R, 0xf);
+        case MESA_FORMAT_SIGNED_RGBA_16:
+            return EASY_US_FORMAT(R500_OUT_FMT_C4_16, R, G, B, A, 0xf);
+
+        default:
+            assert(!"Invalid format for US output\n");
+            return 0;
+    }
+}
+#undef EASY_US_FORMAT
+
+static void r500_emit_fp_setup(struct r300_context *r300,
+                               struct r500_fragment_program_code *fp,
+                               gl_format dst_format)
+{
+    r500_emit_fp(r300, (uint32_t *)fp->inst, (fp->inst_end + 1) * 6, 0, 0, 0);
+    BATCH_LOCALS(&r300->radeon);
+
+    BEGIN_BATCH(10);
+    OUT_BATCH_REGSEQ(R500_US_CODE_ADDR, 3);
+    OUT_BATCH(R500_US_CODE_START_ADDR(0) | R500_US_CODE_END_ADDR(fp->inst_end));
+    OUT_BATCH(R500_US_CODE_RANGE_ADDR(0) | R500_US_CODE_RANGE_SIZE(fp->inst_end));
+    OUT_BATCH(0);
+    OUT_BATCH_REGVAL(R500_US_CONFIG, 0);
+    OUT_BATCH_REGVAL(R500_US_OUT_FMT_0, mesa_format_to_us_format(dst_format));
+    OUT_BATCH_REGVAL(R500_US_PIXSIZE, fp->max_temp_idx);
+    END_BATCH();
+}
+
+static void r500_emit_rs_setup(struct r300_context *r300)
+{
+    BATCH_LOCALS(&r300->radeon);
+
+    BEGIN_BATCH(7);
+    OUT_BATCH_REGSEQ(R300_RS_COUNT, 2);
+    OUT_BATCH((4 << R300_IT_COUNT_SHIFT) | R300_HIRES_EN);
+    OUT_BATCH(0);
+    OUT_BATCH_REGVAL(R500_RS_INST_0,
+                     (0 << R500_RS_INST_TEX_ID_SHIFT) |
+                     (0 << R500_RS_INST_TEX_ADDR_SHIFT) |
+                     R500_RS_INST_TEX_CN_WRITE |
+                     R500_RS_INST_COL_CN_NO_WRITE);
+    OUT_BATCH_REGVAL(R500_RS_IP_0,
+                     (0 << R500_RS_IP_TEX_PTR_S_SHIFT) |
+                     (1 << R500_RS_IP_TEX_PTR_T_SHIFT) |
+                     (2 << R500_RS_IP_TEX_PTR_R_SHIFT) |
+                     (3 << R500_RS_IP_TEX_PTR_Q_SHIFT));
+    END_BATCH();
+}
+
+static void r300_emit_fp_setup(struct r300_context *r300,
+                               struct r300_fragment_program_code *code,
+                               gl_format dst_format)
+{
+    unsigned i;
+    BATCH_LOCALS(&r300->radeon);
+
+    BEGIN_BATCH((code->alu.length + 1) * 4 + code->tex.length + 1 + 11);
+
+    OUT_BATCH_REGSEQ(R300_US_ALU_RGB_INST_0, code->alu.length);
+    for (i = 0; i < code->alu.length; i++) {
+        OUT_BATCH(code->alu.inst[i].rgb_inst);
+    }
+    OUT_BATCH_REGSEQ(R300_US_ALU_RGB_ADDR_0, code->alu.length);
+    for (i = 0; i < code->alu.length; i++) {
+        OUT_BATCH(code->alu.inst[i].rgb_addr);
+    }
+    OUT_BATCH_REGSEQ(R300_US_ALU_ALPHA_INST_0, code->alu.length);
+    for (i = 0; i < code->alu.length; i++) {
+        OUT_BATCH(code->alu.inst[i].alpha_inst);
+    }
+    OUT_BATCH_REGSEQ(R300_US_ALU_ALPHA_ADDR_0, code->alu.length);
+    for (i = 0; i < code->alu.length; i++) {
+        OUT_BATCH(code->alu.inst[i].alpha_addr);
+    }
+
+    OUT_BATCH_REGSEQ(R300_US_TEX_INST_0, code->tex.length);
+    OUT_BATCH_TABLE(code->tex.inst, code->tex.length);
+
+    OUT_BATCH_REGSEQ(R300_US_CONFIG, 3);
+    OUT_BATCH(R300_PFS_CNTL_FIRST_NODE_HAS_TEX);
+    OUT_BATCH(code->pixsize);
+    OUT_BATCH(code->code_offset);
+    OUT_BATCH_REGSEQ(R300_US_CODE_ADDR_0, 4);
+    OUT_BATCH_TABLE(code->code_addr, 4);
+    OUT_BATCH_REGVAL(R500_US_OUT_FMT_0, mesa_format_to_us_format(dst_format));
+    END_BATCH();
+}
+
+static void r300_emit_rs_setup(struct r300_context *r300)
+{
+    BATCH_LOCALS(&r300->radeon);
+
+    BEGIN_BATCH(7);
+    OUT_BATCH_REGSEQ(R300_RS_COUNT, 2);
+    OUT_BATCH((4 << R300_IT_COUNT_SHIFT) | R300_HIRES_EN);
+    OUT_BATCH(0);
+    OUT_BATCH_REGVAL(R300_RS_INST_0,
+                     R300_RS_INST_TEX_ID(0) |
+                     R300_RS_INST_TEX_ADDR(0) |
+                     R300_RS_INST_TEX_CN_WRITE);
+    OUT_BATCH_REGVAL(R300_RS_IP_0,
+                     R300_RS_TEX_PTR(0) |
+                     R300_RS_SEL_S(R300_RS_SEL_C0) |
+                     R300_RS_SEL_T(R300_RS_SEL_C1) |
+                     R300_RS_SEL_R(R300_RS_SEL_K0) |
+                     R300_RS_SEL_Q(R300_RS_SEL_K1));
+    END_BATCH();
+}
+
+static void emit_pvs_setup(struct r300_context *r300,
+                           uint32_t *vp_code,
+                           unsigned vp_len)
+{
+    BATCH_LOCALS(&r300->radeon);
+
+    r300_emit_vpu(r300, vp_code, vp_len * 4, R300_PVS_CODE_START);
+
+    BEGIN_BATCH(4);
+    OUT_BATCH_REGSEQ(R300_VAP_PVS_CODE_CNTL_0, 3);
+    OUT_BATCH((0 << R300_PVS_FIRST_INST_SHIFT) |
+              ((vp_len - 1)  << R300_PVS_XYZW_VALID_INST_SHIFT) |
+              ((vp_len - 1)<< R300_PVS_LAST_INST_SHIFT));
+    OUT_BATCH(0);
+    OUT_BATCH((vp_len - 1) << R300_PVS_LAST_VTX_SRC_INST_SHIFT);
+    END_BATCH();
+}
+
+static void emit_vap_setup(struct r300_context *r300)
+{
+    BATCH_LOCALS(&r300->radeon);
+
+    BEGIN_BATCH(12);
+    OUT_BATCH_REGSEQ(R300_SE_VTE_CNTL, 2);
+    OUT_BATCH(R300_VTX_XY_FMT | R300_VTX_Z_FMT);
+    OUT_BATCH(4);
+
+    OUT_BATCH_REGVAL(R300_VAP_PSC_SGN_NORM_CNTL, 0xaaaaaaaa);
+    OUT_BATCH_REGVAL(R300_VAP_PROG_STREAM_CNTL_0,
+                     ((R300_DATA_TYPE_FLOAT_2 | (0 << R300_DST_VEC_LOC_SHIFT)) << 0) |
+                     (((1 << R300_DST_VEC_LOC_SHIFT) | R300_DATA_TYPE_FLOAT_2 | R300_LAST_VEC) << 16));
+    OUT_BATCH_REGVAL(R300_VAP_PROG_STREAM_CNTL_EXT_0,
+                    ((((R300_SWIZZLE_SELECT_X << R300_SWIZZLE_SELECT_X_SHIFT) |
+                       (R300_SWIZZLE_SELECT_Y << R300_SWIZZLE_SELECT_Y_SHIFT) |
+                       (R300_SWIZZLE_SELECT_FP_ZERO << R300_SWIZZLE_SELECT_Z_SHIFT) |
+                       (R300_SWIZZLE_SELECT_FP_ONE << R300_SWIZZLE_SELECT_W_SHIFT) | 
+                       (0xf << R300_WRITE_ENA_SHIFT) ) << 0) |
+                     (((R300_SWIZZLE_SELECT_X << R300_SWIZZLE_SELECT_X_SHIFT) |
+                       (R300_SWIZZLE_SELECT_Y << R300_SWIZZLE_SELECT_Y_SHIFT) |
+                       (R300_SWIZZLE_SELECT_FP_ZERO << R300_SWIZZLE_SELECT_Z_SHIFT) |
+                       (R300_SWIZZLE_SELECT_FP_ONE << R300_SWIZZLE_SELECT_W_SHIFT) |
+                       (0xf << R300_WRITE_ENA_SHIFT) ) << 16) ) );
+    OUT_BATCH_REGSEQ(R300_VAP_OUTPUT_VTX_FMT_0, 2);
+    OUT_BATCH(R300_VAP_OUTPUT_VTX_FMT_0__POS_PRESENT);
+    OUT_BATCH(R300_VAP_OUTPUT_VTX_FMT_1__4_COMPONENTS);
+    END_BATCH();
+}
+
+static GLboolean validate_buffers(struct r300_context *r300,
+                                  struct radeon_bo *src_bo,
+                                  struct radeon_bo *dst_bo)
+{
+    int ret;
+    radeon_cs_space_add_persistent_bo(r300->radeon.cmdbuf.cs,
+                                      src_bo, RADEON_GEM_DOMAIN_VRAM, 0);
+
+    radeon_cs_space_add_persistent_bo(r300->radeon.cmdbuf.cs,
+                                      dst_bo, 0, RADEON_GEM_DOMAIN_VRAM);
+
+    ret = radeon_cs_space_check_with_bo(r300->radeon.cmdbuf.cs,
+                                        first_elem(&r300->radeon.dma.reserved)->bo,
+                                        RADEON_GEM_DOMAIN_GTT, 0);
+    if (ret)
+        return GL_FALSE;
+
+    return GL_TRUE;
+}
+
+/**
+ * Calculate texcoords for given image region.
+ * Output values are [minx, maxx, miny, maxy]
+ */
+static void calc_tex_coords(float img_width, float img_height,
+                            float x, float y,
+                            float reg_width, float reg_height,
+                            unsigned flip_y, float *buf)
+{
+    buf[0] = x / img_width;
+    buf[1] = buf[0] + reg_width / img_width;
+    buf[2] = y / img_height;
+    buf[3] = buf[2] + reg_height / img_height;
+    if (flip_y)
+    {
+        float tmp = buf[2];
+        buf[2] = 1.0 - buf[3];
+        buf[3] = 1.0 - tmp;
+    }
+}
+
+static void emit_draw_packet(struct r300_context *r300,
+                             unsigned src_width, unsigned src_height,
+                             unsigned src_x_offset, unsigned src_y_offset,
+                             unsigned dst_x_offset, unsigned dst_y_offset,
+                             unsigned reg_width, unsigned reg_height,
+                             unsigned flip_y)
+{
+    float texcoords[4];
+
+    calc_tex_coords(src_width, src_height,
+                    src_x_offset, src_y_offset,
+                    reg_width, reg_height,
+                    flip_y, texcoords);
+
+    float verts[] = { dst_x_offset, dst_y_offset,
+                      texcoords[0], texcoords[3],
+                      dst_x_offset, dst_y_offset + reg_height,
+                      texcoords[0], texcoords[2],
+                      dst_x_offset + reg_width, dst_y_offset + reg_height,
+                      texcoords[1], texcoords[2],
+                      dst_x_offset + reg_width, dst_y_offset,
+                      texcoords[1], texcoords[3] };
+
+    BATCH_LOCALS(&r300->radeon);
+
+    BEGIN_BATCH(19);
+    OUT_BATCH_PACKET3(R300_PACKET3_3D_DRAW_IMMD_2, 16);
+    OUT_BATCH(R300_VAP_VF_CNTL__PRIM_WALK_VERTEX_EMBEDDED |
+              (4 << 16) | R300_VAP_VF_CNTL__PRIM_QUADS);
+    OUT_BATCH_TABLE(verts, 16);
+    END_BATCH();
+}
+
+static void other_stuff(struct r300_context *r300)
+{
+    BATCH_LOCALS(&r300->radeon);
+
+    BEGIN_BATCH(15);
+    OUT_BATCH_REGVAL(R300_GA_POLY_MODE,
+                     R300_GA_POLY_MODE_FRONT_PTYPE_TRI | R300_GA_POLY_MODE_BACK_PTYPE_TRI);
+    OUT_BATCH_REGVAL(R300_SU_CULL_MODE, R300_FRONT_FACE_CCW);
+    OUT_BATCH_REGVAL(R300_FG_FOG_BLEND, 0);
+    OUT_BATCH_REGVAL(R300_FG_ALPHA_FUNC, 0);
+    OUT_BATCH_REGSEQ(R300_RB3D_CBLEND, 2);
+    OUT_BATCH(0x0);
+    OUT_BATCH(0x0);
+    OUT_BATCH_REGVAL(R300_VAP_CLIP_CNTL, R300_CLIP_DISABLE);
+    OUT_BATCH_REGVAL(R300_ZB_CNTL, 0);
+    END_BATCH();
+}
+
+static void emit_cb_setup(struct r300_context *r300,
+                          struct radeon_bo *bo,
+                          intptr_t offset,
+                          gl_format mesa_format,
+                          unsigned pitch,
+                          unsigned width,
+                          unsigned height)
+{
+    BATCH_LOCALS(&r300->radeon);
+
+    unsigned x1, y1, x2, y2;
+    x1 = 0;
+    y1 = 0;
+    x2 = width - 1;
+    y2 = height - 1;
+
+    if (r300->radeon.radeonScreen->chip_family < CHIP_FAMILY_RV515) {
+        x1 += R300_SCISSORS_OFFSET;
+        y1 += R300_SCISSORS_OFFSET;
+        x2 += R300_SCISSORS_OFFSET;
+        y2 += R300_SCISSORS_OFFSET;
+    }
+
+    r300_emit_cb_setup(r300, bo, offset, mesa_format,
+                       _mesa_get_format_bytes(mesa_format),
+                       _mesa_format_row_stride(mesa_format, pitch));
+
+    BEGIN_BATCH_NO_AUTOSTATE(5);
+    OUT_BATCH_REGSEQ(R300_SC_SCISSORS_TL, 2);
+    OUT_BATCH((x1 << R300_SCISSORS_X_SHIFT)|(y1 << R300_SCISSORS_Y_SHIFT));
+    OUT_BATCH((x2 << R300_SCISSORS_X_SHIFT)|(y2 << R300_SCISSORS_Y_SHIFT));
+    OUT_BATCH_REGVAL(R300_RB3D_CCTL, 0);
+    END_BATCH();
+}
+
+/**
+ * Copy a region of [@a width x @a height] pixels from source buffer
+ * to destination buffer.
+ * @param[in] r300 r300 context
+ * @param[in] src_bo source radeon buffer object
+ * @param[in] src_offset offset of the source image in the @a src_bo
+ * @param[in] src_mesaformat source image format
+ * @param[in] src_pitch aligned source image width
+ * @param[in] src_width source image width
+ * @param[in] src_height source image height
+ * @param[in] src_x_offset x offset in the source image
+ * @param[in] src_y_offset y offset in the source image
+ * @param[in] dst_bo destination radeon buffer object
+ * @param[in] dst_offset offset of the destination image in the @a dst_bo
+ * @param[in] dst_mesaformat destination image format
+ * @param[in] dst_pitch aligned destination image width
+ * @param[in] dst_width destination image width
+ * @param[in] dst_height destination image height
+ * @param[in] dst_x_offset x offset in the destination image
+ * @param[in] dst_y_offset y offset in the destination image
+ * @param[in] width region width
+ * @param[in] height region height
+ * @param[in] flip_y set if y coords of the source image need to be flipped
+ */
+GLboolean r300_blit(struct r300_context *r300,
+                    struct radeon_bo *src_bo,
+                    intptr_t src_offset,
+                    gl_format src_mesaformat,
+                    unsigned src_pitch,
+                    unsigned src_width,
+                    unsigned src_height,
+                    unsigned src_x_offset,
+                    unsigned src_y_offset,
+                    struct radeon_bo *dst_bo,
+                    intptr_t dst_offset,
+                    gl_format dst_mesaformat,
+                    unsigned dst_pitch,
+                    unsigned dst_width,
+                    unsigned dst_height,
+                    unsigned dst_x_offset,
+                    unsigned dst_y_offset,
+                    unsigned reg_width,
+                    unsigned reg_height,
+                    unsigned flip_y)
+{
+    /* Need to clamp the region size to make sure
+     * we don't read outside of the source buffer
+     * or write outside of the destination buffer.
+     */
+    if (reg_width + src_x_offset > src_width)
+        reg_width = src_width - src_x_offset;
+    if (reg_height + src_y_offset > src_height)
+        reg_height = src_height - src_y_offset;
+    if (reg_width + dst_x_offset > dst_width)
+        reg_width = dst_width - dst_x_offset;
+    if (reg_height + dst_y_offset > dst_height)
+        reg_height = dst_height - dst_y_offset;
+
+    if (src_bo == dst_bo) {
+        return GL_FALSE;
+    }
+
+    if (0) {
+        fprintf(stderr, "src: size [%d x %d], pitch %d, "
+                "offset [%d x %d], format %s, bo %p\n",
+                src_width, src_height, src_pitch,
+                src_offset, src_y_offset,
+                _mesa_get_format_name(src_mesaformat),
+                src_bo);
+        fprintf(stderr, "dst: pitch %d, offset[%d x %d], format %s, bo %p\n",
+                dst_pitch, dst_x_offset, dst_y_offset,
+                _mesa_get_format_name(dst_mesaformat), dst_bo);
+        fprintf(stderr, "region: %d x %d\n", reg_width, reg_height);
+    }
+
+    if (!validate_buffers(r300, src_bo, dst_bo))
+        return GL_FALSE;
+
+    rcommonEnsureCmdBufSpace(&r300->radeon, 200, __FUNCTION__);
+
+    other_stuff(r300);
+
+    r300_emit_tx_setup(r300, src_mesaformat, src_bo, src_offset, src_width, src_height, src_pitch);
+
+    if (r300->radeon.radeonScreen->chip_family >= CHIP_FAMILY_RV515) {
+        r500_emit_fp_setup(r300, &r300->blit.fp_code.code.r500, dst_mesaformat);
+        r500_emit_rs_setup(r300);
+    } else {
+        r300_emit_fp_setup(r300, &r300->blit.fp_code.code.r300, dst_mesaformat);
+        r300_emit_rs_setup(r300);
+    }
+
+    emit_pvs_setup(r300, r300->blit.vp_code.body.d, 2);
+    emit_vap_setup(r300);
+
+    emit_cb_setup(r300, dst_bo, dst_offset, dst_mesaformat, dst_pitch, dst_width, dst_height);
+
+    emit_draw_packet(r300, src_width, src_height,
+                     src_x_offset, src_y_offset,
+                     dst_x_offset, dst_y_offset,
+                     reg_width, reg_height,
+                     flip_y);
+
+    r300EmitCacheFlush(r300);
+
+    radeonFlush(r300->radeon.glCtx);
+
+    return GL_TRUE;
+}
\ No newline at end of file
diff --git a/src/mesa/drivers/dri/r300/r300_blit.h b/src/mesa/drivers/dri/r300/r300_blit.h
new file mode 100644 (file)
index 0000000..dc21e88
--- /dev/null
@@ -0,0 +1,54 @@
+/*
+ * Copyright (C) 2009 Maciej Cencora <m.cencora@gmail.com>
+ *
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial
+ * portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#ifndef R300_BLIT_H
+#define R300_BLIT_H
+
+void r300_blit_init(struct r300_context *r300);
+
+GLboolean r300_blit(struct r300_context *r300,
+                    struct radeon_bo *src_bo,
+                    intptr_t src_offset,
+                    gl_format src_mesaformat,
+                    unsigned src_pitch,
+                    unsigned src_width,
+                    unsigned src_height,
+                    unsigned src_x_offset,
+                    unsigned src_y_offset,
+                    struct radeon_bo *dst_bo,
+                    intptr_t dst_offset,
+                    gl_format dst_mesaformat,
+                    unsigned dst_pitch,
+                    unsigned dst_width,
+                    unsigned dst_height,
+                    unsigned dst_x_offset,
+                    unsigned dst_y_offset,
+                    unsigned width,
+                    unsigned height,
+                    unsigned flip_y);
+
+#endif // R300_BLIT_H
\ No newline at end of file
index ad8db6e68e0c37177ad7c0d70e2d3c44f32fbbee..e1c33bbb2cfba446c8186f3b5f380d4904840d23 100644 (file)
@@ -45,7 +45,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "radeon_drm.h"
 
 #include "r300_context.h"
-#include "r300_ioctl.h"
 #include "r300_reg.h"
 #include "r300_cmdbuf.h"
 #include "r300_emit.h"
@@ -72,7 +71,7 @@ static unsigned packet0_count(r300ContextPtr r300, uint32_t *pkt)
 #define vpu_count(ptr) (((drm_r300_cmd_header_t*)(ptr))->vpu.count)
 #define r500fp_count(ptr) (((drm_r300_cmd_header_t*)(ptr))->r500fp.count)
 
-int check_vpu(GLcontext *ctx, struct radeon_state_atom *atom)
+static int check_vpu(GLcontext *ctx, struct radeon_state_atom *atom)
 {
        r300ContextPtr r300 = R300_CONTEXT(ctx);
        int cnt;
@@ -86,54 +85,73 @@ int check_vpu(GLcontext *ctx, struct radeon_state_atom *atom)
        return cnt ? (cnt * 4) + extra : 0;
 }
 
+void r300_emit_vpu(struct r300_context *r300,
+                   uint32_t *data,
+                   unsigned len,
+                   uint32_t addr)
+{
+    BATCH_LOCALS(&r300->radeon);
+
+    BEGIN_BATCH_NO_AUTOSTATE(5 + len);
+    OUT_BATCH_REGVAL(R300_VAP_PVS_STATE_FLUSH_REG, 0);
+    OUT_BATCH_REGVAL(R300_VAP_PVS_VECTOR_INDX_REG, addr);
+    OUT_BATCH(CP_PACKET0(R300_VAP_PVS_UPLOAD_DATA, len-1) | RADEON_ONE_REG_WR);
+    OUT_BATCH_TABLE(data, len);
+    END_BATCH();
+}
 
-void emit_vpu(GLcontext *ctx, struct radeon_state_atom * atom)
+static void emit_vpu_state(GLcontext *ctx, struct radeon_state_atom * atom)
 {
-       r300ContextPtr r300 = R300_CONTEXT(ctx);
-       BATCH_LOCALS(&r300->radeon);
-       drm_r300_cmd_header_t cmd;
-       uint32_t addr, ndw;
+    r300ContextPtr r300 = R300_CONTEXT(ctx);
+    drm_r300_cmd_header_t cmd;
+    uint32_t addr, ndw;
 
-       cmd.u = atom->cmd[0];
-       addr = (cmd.vpu.adrhi << 8) | cmd.vpu.adrlo;
-       ndw = atom->check(ctx, atom);
+    cmd.u = atom->cmd[0];
+    addr = (cmd.vpu.adrhi << 8) | cmd.vpu.adrlo;
+    ndw = atom->check(ctx, atom);
 
-       BEGIN_BATCH_NO_AUTOSTATE(ndw);
+    r300_emit_vpu(r300, &atom->cmd[1], vpu_count(atom->cmd) * 4, addr);
+}
 
-       ndw -= 5;
-       OUT_BATCH_REGVAL(R300_VAP_PVS_VECTOR_INDX_REG, addr);
-       OUT_BATCH(CP_PACKET0(R300_VAP_PVS_UPLOAD_DATA, ndw-1) | RADEON_ONE_REG_WR);
-       OUT_BATCH_TABLE(&atom->cmd[1], ndw);
-       OUT_BATCH_REGVAL(R300_VAP_PVS_STATE_FLUSH_REG, 0);
-       END_BATCH();
+void r500_emit_fp(struct r300_context *r300,
+                  uint32_t *data,
+                  unsigned len,
+                  uint32_t addr,
+                  unsigned type,
+                  unsigned clamp)
+{
+    BATCH_LOCALS(&r300->radeon);
+
+    addr |= (type << 16);
+    addr |= (clamp << 17);
+
+    BEGIN_BATCH_NO_AUTOSTATE(len + 3);
+    OUT_BATCH(CP_PACKET0(R500_GA_US_VECTOR_INDEX, 0));
+    OUT_BATCH(addr);
+    OUT_BATCH(CP_PACKET0(R500_GA_US_VECTOR_DATA, len-1) | RADEON_ONE_REG_WR);
+    OUT_BATCH_TABLE(data, len);
+    END_BATCH();
 }
 
-void emit_r500fp(GLcontext *ctx, struct radeon_state_atom * atom)
+static void emit_r500fp_atom(GLcontext *ctx, struct radeon_state_atom * atom)
 {
-       r300ContextPtr r300 = R300_CONTEXT(ctx);
-       BATCH_LOCALS(&r300->radeon);
-       drm_r300_cmd_header_t cmd;
-       uint32_t addr, ndw, sz;
-       int type, clamp;
-
-       ndw = atom->check(ctx, atom);
-
-       cmd.u = atom->cmd[0];
-       sz = cmd.r500fp.count;
-       addr = ((cmd.r500fp.adrhi_flags & 1) << 8) | cmd.r500fp.adrlo;
-       type = !!(cmd.r500fp.adrhi_flags & R500FP_CONSTANT_TYPE);
-       clamp = !!(cmd.r500fp.adrhi_flags & R500FP_CONSTANT_CLAMP);
-
-       addr |= (type << 16);
-       addr |= (clamp << 17);
-
-       BEGIN_BATCH_NO_AUTOSTATE(ndw);
-       OUT_BATCH(CP_PACKET0(R500_GA_US_VECTOR_INDEX, 0));
-       OUT_BATCH(addr);
-       ndw-=3;
-       OUT_BATCH(CP_PACKET0(R500_GA_US_VECTOR_DATA, ndw-1) | RADEON_ONE_REG_WR);
-       OUT_BATCH_TABLE(&atom->cmd[1], ndw);
-       END_BATCH();
+    r300ContextPtr r300 = R300_CONTEXT(ctx);
+    drm_r300_cmd_header_t cmd;
+    uint32_t addr, count;
+    int type, clamp;
+
+    cmd.u = atom->cmd[0];
+    addr = ((cmd.r500fp.adrhi_flags & 1) << 8) | cmd.r500fp.adrlo;
+    type = !!(cmd.r500fp.adrhi_flags & R500FP_CONSTANT_TYPE);
+    clamp = !!(cmd.r500fp.adrhi_flags & R500FP_CONSTANT_CLAMP);
+
+    if (type) {
+        count = r500fp_count(atom->cmd) * 4;
+    } else {
+        count = r500fp_count(atom->cmd) * 6;
+    }
+
+    r500_emit_fp(r300, &atom->cmd[1], count, addr, type, clamp);
 }
 
 static int check_tex_offsets(GLcontext *ctx, struct radeon_state_atom * atom)
@@ -256,110 +274,136 @@ static int check_cb_offset(GLcontext *ctx, struct radeon_state_atom * atom)
        return dw;
 }
 
-static void emit_cb_offset(GLcontext *ctx, struct radeon_state_atom * atom)
+static void emit_scissor(struct r300_context *r300,
+                         unsigned width,
+                         unsigned height)
 {
-       r300ContextPtr r300 = R300_CONTEXT(ctx);
-       BATCH_LOCALS(&r300->radeon);
-       struct radeon_renderbuffer *rrb;
-       uint32_t cbpitch;
-       uint32_t offset = r300->radeon.state.color.draw_offset;
-       uint32_t dw = 6;
-       int i;
+    int i;
+    BATCH_LOCALS(&r300->radeon);
+    if (r300->radeon.radeonScreen->chip_family >= CHIP_FAMILY_RV515) {
+        BEGIN_BATCH_NO_AUTOSTATE(3);
+        OUT_BATCH_REGSEQ(R300_SC_SCISSORS_TL, 2);
+        OUT_BATCH(0);
+        OUT_BATCH(((width - 1) << R300_SCISSORS_X_SHIFT) |
+                ((height - 1) << R300_SCISSORS_Y_SHIFT));
+        END_BATCH();
+        BEGIN_BATCH_NO_AUTOSTATE(16);
+        for (i = 0; i < 4; i++) {
+            OUT_BATCH_REGSEQ(R300_SC_CLIPRECT_TL_0 + (i * 8), 2);
+            OUT_BATCH((0 << R300_CLIPRECT_X_SHIFT) | (0 << R300_CLIPRECT_Y_SHIFT));
+            OUT_BATCH(((width - 1) << R300_CLIPRECT_X_SHIFT) | ((height - 1) << R300_CLIPRECT_Y_SHIFT));
+        }
+        OUT_BATCH_REGSEQ(R300_SC_CLIP_RULE, 1);
+        OUT_BATCH(0xAAAA);
+        OUT_BATCH_REGSEQ(R300_SC_SCREENDOOR, 1);
+        OUT_BATCH(0xffffff);
+        END_BATCH();
+    } else {
+        BEGIN_BATCH_NO_AUTOSTATE(3);
+        OUT_BATCH_REGSEQ(R300_SC_SCISSORS_TL, 2);
+        OUT_BATCH((R300_SCISSORS_OFFSET << R300_SCISSORS_X_SHIFT) |
+                (R300_SCISSORS_OFFSET << R300_SCISSORS_Y_SHIFT));
+        OUT_BATCH(((width + R300_SCISSORS_OFFSET - 1) << R300_SCISSORS_X_SHIFT) |
+                ((height + R300_SCISSORS_OFFSET - 1) << R300_SCISSORS_Y_SHIFT));
+        END_BATCH();
+        BEGIN_BATCH_NO_AUTOSTATE(16);
+        for (i = 0; i < 4; i++) {
+            OUT_BATCH_REGSEQ(R300_SC_CLIPRECT_TL_0 + (i * 8), 2);
+            OUT_BATCH((R300_SCISSORS_OFFSET << R300_CLIPRECT_X_SHIFT) | (R300_SCISSORS_OFFSET << R300_CLIPRECT_Y_SHIFT));
+            OUT_BATCH(((R300_SCISSORS_OFFSET + width - 1) << R300_CLIPRECT_X_SHIFT) |
+                        ((R300_SCISSORS_OFFSET + height - 1) << R300_CLIPRECT_Y_SHIFT));
+        }
+        OUT_BATCH_REGSEQ(R300_SC_CLIP_RULE, 1);
+        OUT_BATCH(0xAAAA);
+        OUT_BATCH_REGSEQ(R300_SC_SCREENDOOR, 1);
+        OUT_BATCH(0xffffff);
+        END_BATCH();
+    }
+}
 
-       rrb = radeon_get_colorbuffer(&r300->radeon);
-       if (!rrb || !rrb->bo) {
-               fprintf(stderr, "no rrb\n");
-               return;
-       }
+void r300_emit_cb_setup(struct r300_context *r300,
+                        struct radeon_bo *bo,
+                        uint32_t offset,
+                        GLuint format,
+                        unsigned cpp,
+                        unsigned pitch)
+{
+    BATCH_LOCALS(&r300->radeon);
+    uint32_t cbpitch = pitch / cpp;
+    uint32_t dw = 6;
 
-        if (RADEON_DEBUG & RADEON_STATE)
-           fprintf(stderr,"rrb is %p %d %dx%d\n", rrb, offset, rrb->base.Width, rrb->base.Height);
-       cbpitch = (rrb->pitch / rrb->cpp);
-       if (rrb->cpp == 4)
-               cbpitch |= R300_COLOR_FORMAT_ARGB8888;
-       else switch (rrb->base.Format) {
+    assert(offset % 32 == 0);
+
+    switch (format) {
         case MESA_FORMAT_RGB565:
-               assert(_mesa_little_endian());
-               cbpitch |= R300_COLOR_FORMAT_RGB565;
-               break;
+            assert(_mesa_little_endian());
+            cbpitch |= R300_COLOR_FORMAT_RGB565;
+            break;
         case MESA_FORMAT_RGB565_REV:
-               assert(!_mesa_little_endian());
-               cbpitch |= R300_COLOR_FORMAT_RGB565;
-               break;
+            assert(!_mesa_little_endian());
+            cbpitch |= R300_COLOR_FORMAT_RGB565;
+            break;
         case MESA_FORMAT_ARGB4444:
-               assert(_mesa_little_endian());
-               cbpitch |= R300_COLOR_FORMAT_ARGB4444;
-               break;
+            assert(_mesa_little_endian());
+            cbpitch |= R300_COLOR_FORMAT_ARGB4444;
+            break;
         case MESA_FORMAT_ARGB4444_REV:
-               assert(!_mesa_little_endian());
-               cbpitch |= R300_COLOR_FORMAT_ARGB4444;
-               break;
-       case MESA_FORMAT_ARGB1555:
-               assert(_mesa_little_endian());
-               cbpitch |= R300_COLOR_FORMAT_ARGB1555;
-               break;
-       case MESA_FORMAT_ARGB1555_REV:
-               assert(!_mesa_little_endian());
-               cbpitch |= R300_COLOR_FORMAT_ARGB1555;
-               break;
-       default:
-               _mesa_problem(ctx, "unexpected format in emit_cb_offset()");
-       }
+            assert(!_mesa_little_endian());
+            cbpitch |= R300_COLOR_FORMAT_ARGB4444;
+            break;
+        case MESA_FORMAT_ARGB1555:
+            assert(_mesa_little_endian());
+            cbpitch |= R300_COLOR_FORMAT_ARGB1555;
+            break;
+        case MESA_FORMAT_ARGB1555_REV:
+            assert(!_mesa_little_endian());
+            cbpitch |= R300_COLOR_FORMAT_ARGB1555;
+            break;
+        default:
+            if (cpp == 4) {
+                cbpitch |= R300_COLOR_FORMAT_ARGB8888;
+            } else {
+                _mesa_problem(r300->radeon.glCtx, "unexpected format in emit_cb_offset()");;
+            }
+            break;
+    }
 
-       if (rrb->bo->flags & RADEON_BO_FLAGS_MACRO_TILE)
-               cbpitch |= R300_COLOR_TILE_ENABLE;
+    if (bo->flags & RADEON_BO_FLAGS_MACRO_TILE)
+        cbpitch |= R300_COLOR_TILE_ENABLE;
+
+    if (r300->radeon.radeonScreen->kernel_mm)
+        dw += 2;
+
+    BEGIN_BATCH_NO_AUTOSTATE(dw);
+    OUT_BATCH_REGSEQ(R300_RB3D_COLOROFFSET0, 1);
+    OUT_BATCH_RELOC(offset, bo, offset, 0, RADEON_GEM_DOMAIN_VRAM, 0);
+    OUT_BATCH_REGSEQ(R300_RB3D_COLORPITCH0, 1);
+    if (!r300->radeon.radeonScreen->kernel_mm)
+        OUT_BATCH(cbpitch);
+    else
+        OUT_BATCH_RELOC(cbpitch, bo, cbpitch, 0, RADEON_GEM_DOMAIN_VRAM, 0);
+    END_BATCH();
+}
+
+static void emit_cb_offset_atom(GLcontext *ctx, struct radeon_state_atom * atom)
+{
+    r300ContextPtr r300 = R300_CONTEXT(ctx);
+    struct radeon_renderbuffer *rrb;
+    uint32_t offset = r300->radeon.state.color.draw_offset;
+
+    rrb = radeon_get_colorbuffer(&r300->radeon);
+    if (!rrb || !rrb->bo) {
+        fprintf(stderr, "no rrb\n");
+        return;
+    }
+
+    if (RADEON_DEBUG & RADEON_STATE)
+        fprintf(stderr,"rrb is %p %d %dx%d\n", rrb, offset, rrb->base.Width, rrb->base.Height);
+
+    r300_emit_cb_setup(r300, rrb->bo, offset, rrb->base.Format, rrb->cpp, rrb->pitch);
 
-       if (r300->radeon.radeonScreen->kernel_mm)
-               dw += 2;
-       BEGIN_BATCH_NO_AUTOSTATE(dw);
-       OUT_BATCH_REGSEQ(R300_RB3D_COLOROFFSET0, 1);
-       OUT_BATCH_RELOC(offset, rrb->bo, offset, 0, RADEON_GEM_DOMAIN_VRAM, 0);
-       OUT_BATCH_REGSEQ(R300_RB3D_COLORPITCH0, 1);
-       if (!r300->radeon.radeonScreen->kernel_mm)
-               OUT_BATCH(cbpitch);
-       else
-               OUT_BATCH_RELOC(cbpitch, rrb->bo, cbpitch, 0, RADEON_GEM_DOMAIN_VRAM, 0);
-       END_BATCH();
     if (r300->radeon.radeonScreen->driScreen->dri2.enabled) {
-        if (r300->radeon.radeonScreen->chip_family >= CHIP_FAMILY_RV515) {
-            BEGIN_BATCH_NO_AUTOSTATE(3);
-            OUT_BATCH_REGSEQ(R300_SC_SCISSORS_TL, 2);
-            OUT_BATCH(0);
-            OUT_BATCH(((rrb->base.Width - 1) << R300_SCISSORS_X_SHIFT) |
-                    ((rrb->base.Height - 1) << R300_SCISSORS_Y_SHIFT));
-            END_BATCH();
-            BEGIN_BATCH_NO_AUTOSTATE(16);
-            for (i = 0; i < 4; i++) {
-                OUT_BATCH_REGSEQ(R300_SC_CLIPRECT_TL_0 + (i * 8), 2);
-                OUT_BATCH((0 << R300_CLIPRECT_X_SHIFT) | (0 << R300_CLIPRECT_Y_SHIFT));
-                OUT_BATCH(((rrb->base.Width - 1) << R300_CLIPRECT_X_SHIFT) | ((rrb->base.Height - 1) << R300_CLIPRECT_Y_SHIFT));
-            }
-            OUT_BATCH_REGSEQ(R300_SC_CLIP_RULE, 1);
-            OUT_BATCH(0xAAAA);
-            OUT_BATCH_REGSEQ(R300_SC_SCREENDOOR, 1);
-            OUT_BATCH(0xffffff);
-            END_BATCH();
-        } else {
-            BEGIN_BATCH_NO_AUTOSTATE(3);
-            OUT_BATCH_REGSEQ(R300_SC_SCISSORS_TL, 2);
-            OUT_BATCH((R300_SCISSORS_OFFSET << R300_SCISSORS_X_SHIFT) |
-                    (R300_SCISSORS_OFFSET << R300_SCISSORS_Y_SHIFT));
-            OUT_BATCH(((rrb->base.Width + R300_SCISSORS_OFFSET - 1) << R300_SCISSORS_X_SHIFT) |
-                    ((rrb->base.Height + R300_SCISSORS_OFFSET - 1) << R300_SCISSORS_Y_SHIFT));
-            END_BATCH();
-            BEGIN_BATCH_NO_AUTOSTATE(16);
-            for (i = 0; i < 4; i++) {
-                OUT_BATCH_REGSEQ(R300_SC_CLIPRECT_TL_0 + (i * 8), 2);
-                OUT_BATCH((R300_SCISSORS_OFFSET << R300_CLIPRECT_X_SHIFT) | (R300_SCISSORS_OFFSET << R300_CLIPRECT_Y_SHIFT));
-                OUT_BATCH(((R300_SCISSORS_OFFSET + rrb->base.Width - 1) << R300_CLIPRECT_X_SHIFT) |
-                          ((R300_SCISSORS_OFFSET + rrb->base.Height - 1) << R300_CLIPRECT_Y_SHIFT));
-            }
-            OUT_BATCH_REGSEQ(R300_SC_CLIP_RULE, 1);
-            OUT_BATCH(0xAAAA);
-            OUT_BATCH_REGSEQ(R300_SC_SCREENDOOR, 1);
-            OUT_BATCH(0xffffff);
-            END_BATCH();
-        }
+        emit_scissor(r300, rrb->base.Width, rrb->base.Height);
     }
 }
 
@@ -455,7 +499,7 @@ static int check_variable(GLcontext *ctx, struct radeon_state_atom *atom)
        return cnt ? cnt + 1 : 0;
 }
 
-int check_r500fp(GLcontext *ctx, struct radeon_state_atom *atom)
+static int check_r500fp(GLcontext *ctx, struct radeon_state_atom *atom)
 {
        int cnt;
        r300ContextPtr r300 = R300_CONTEXT(ctx);
@@ -467,7 +511,7 @@ int check_r500fp(GLcontext *ctx, struct radeon_state_atom *atom)
        return cnt ? (cnt * 6) + extra : 0;
 }
 
-int check_r500fp_const(GLcontext *ctx, struct radeon_state_atom *atom)
+static int check_r500fp_const(GLcontext *ctx, struct radeon_state_atom *atom)
 {
        int cnt;
        r300ContextPtr r300 = R300_CONTEXT(ctx);
@@ -644,13 +688,13 @@ void r300InitCmdBuf(r300ContextPtr r300)
                r300->hw.r500fp.cmd[R300_FPI_CMD_0] =
                        cmdr500fp(r300->radeon.radeonScreen, 0, 0, 0, 0);
                if (r300->radeon.radeonScreen->kernel_mm)
-                       r300->hw.r500fp.emit = emit_r500fp;
+                       r300->hw.r500fp.emit = emit_r500fp_atom;
 
                ALLOC_STATE(r500fp_const, r500fp_const, R500_FPP_CMDSIZE, 0);
                r300->hw.r500fp_const.cmd[R300_FPI_CMD_0] =
                        cmdr500fp(r300->radeon.radeonScreen, 0, 0, 1, 0);
                if (r300->radeon.radeonScreen->kernel_mm)
-                       r300->hw.r500fp_const.emit = emit_r500fp;
+                       r300->hw.r500fp_const.emit = emit_r500fp_atom;
        } else {
                ALLOC_STATE(fp, always, R300_FP_CMDSIZE, 0);
                r300->hw.fp.cmd[R300_FP_CMD_0] = cmdpacket0(r300->radeon.radeonScreen, R300_US_CONFIG, 3);
@@ -694,7 +738,7 @@ void r300InitCmdBuf(r300ContextPtr r300)
        ALLOC_STATE(rop, always, 2, 0);
        r300->hw.rop.cmd[0] = cmdpacket0(r300->radeon.radeonScreen, R300_RB3D_ROPCNTL, 1);
        ALLOC_STATE(cb, cb_offset, R300_CB_CMDSIZE, 0);
-       r300->hw.cb.emit = &emit_cb_offset;
+       r300->hw.cb.emit = &emit_cb_offset_atom;
        ALLOC_STATE(rb3d_dither_ctl, always, 10, 0);
        r300->hw.rb3d_dither_ctl.cmd[0] = cmdpacket0(r300->radeon.radeonScreen, R300_RB3D_DITHER_CTL, 9);
        ALLOC_STATE(rb3d_aaresolve_ctl, always, 2, 0);
@@ -758,20 +802,20 @@ void r300InitCmdBuf(r300ContextPtr r300)
                r300->hw.vpi.cmd[0] =
                        cmdvpu(r300->radeon.radeonScreen, R300_PVS_CODE_START, 0);
                if (r300->radeon.radeonScreen->kernel_mm)
-                       r300->hw.vpi.emit = emit_vpu;
+                       r300->hw.vpi.emit = emit_vpu_state;
 
                if (is_r500) {
                        ALLOC_STATE(vpp, vpu, R300_VPP_CMDSIZE, 0);
                        r300->hw.vpp.cmd[0] =
                                cmdvpu(r300->radeon.radeonScreen, R500_PVS_CONST_START, 0);
                        if (r300->radeon.radeonScreen->kernel_mm)
-                               r300->hw.vpp.emit = emit_vpu;
+                               r300->hw.vpp.emit = emit_vpu_state;
 
                        ALLOC_STATE(vps, vpu, R300_VPS_CMDSIZE, 0);
                        r300->hw.vps.cmd[0] =
                                cmdvpu(r300->radeon.radeonScreen, R500_POINT_VPORT_SCALE_OFFSET, 1);
                        if (r300->radeon.radeonScreen->kernel_mm)
-                               r300->hw.vps.emit = emit_vpu;
+                               r300->hw.vps.emit = emit_vpu_state;
 
                        for (i = 0; i < 6; i++) {
                                ALLOC_STATE(vpucp[i], vpu, R300_VPUCP_CMDSIZE, 0);
@@ -779,20 +823,20 @@ void r300InitCmdBuf(r300ContextPtr r300)
                                        cmdvpu(r300->radeon.radeonScreen,
                                                        R500_PVS_UCP_START + i, 1);
                                if (r300->radeon.radeonScreen->kernel_mm)
-                                       r300->hw.vpucp[i].emit = emit_vpu;
+                                       r300->hw.vpucp[i].emit = emit_vpu_state;
                        }
                } else {
                        ALLOC_STATE(vpp, vpu, R300_VPP_CMDSIZE, 0);
                        r300->hw.vpp.cmd[0] =
                                cmdvpu(r300->radeon.radeonScreen, R300_PVS_CONST_START, 0);
                        if (r300->radeon.radeonScreen->kernel_mm)
-                               r300->hw.vpp.emit = emit_vpu;
+                               r300->hw.vpp.emit = emit_vpu_state;
 
                        ALLOC_STATE(vps, vpu, R300_VPS_CMDSIZE, 0);
                        r300->hw.vps.cmd[0] =
                                cmdvpu(r300->radeon.radeonScreen, R300_POINT_VPORT_SCALE_OFFSET, 1);
                        if (r300->radeon.radeonScreen->kernel_mm)
-                               r300->hw.vps.emit = emit_vpu;
+                               r300->hw.vps.emit = emit_vpu_state;
 
                        for (i = 0; i < 6; i++) {
                                ALLOC_STATE(vpucp[i], vpu, R300_VPUCP_CMDSIZE, 0);
@@ -800,7 +844,7 @@ void r300InitCmdBuf(r300ContextPtr r300)
                                        cmdvpu(r300->radeon.radeonScreen,
                                                        R300_PVS_UCP_START + i, 1);
                                if (r300->radeon.radeonScreen->kernel_mm)
-                                       r300->hw.vpucp[i].emit = emit_vpu;
+                                       r300->hw.vpucp[i].emit = emit_vpu_state;
                        }
                }
        }
index 1b703e518a0c403196553c27de42e7bf7d5c8201..0e68da928ed51122ddb81487633b3e02b375fb73 100644 (file)
@@ -44,14 +44,26 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #define FIREAOS_BUFSZ          (3)
 #define SCISSORS_BUFSZ         (3)
 
-extern void r300InitCmdBuf(r300ContextPtr r300);
+void r300InitCmdBuf(r300ContextPtr r300);
 void r300_emit_scissor(GLcontext *ctx);
 
-void emit_vpu(GLcontext *ctx, struct radeon_state_atom * atom);
-int check_vpu(GLcontext *ctx, struct radeon_state_atom *atom);
+void r300_emit_vpu(struct r300_context *ctx,
+                   uint32_t *data,
+                   unsigned len,
+                   uint32_t addr);
 
-void emit_r500fp(GLcontext *ctx, struct radeon_state_atom * atom);
-int check_r500fp(GLcontext *ctx, struct radeon_state_atom *atom);
-int check_r500fp_const(GLcontext *ctx, struct radeon_state_atom *atom);
+void r500_emit_fp(struct r300_context *r300,
+                  uint32_t *data,
+                  unsigned len,
+                  uint32_t addr,
+                  unsigned type,
+                  unsigned clamp);
 
-#endif                         /* __R300_CMDBUF_H__ */
+void r300_emit_cb_setup(struct r300_context *r300,
+                        struct radeon_bo *bo,
+                        uint32_t offset,
+                        GLuint format,
+                        unsigned cpp,
+                        unsigned pitch);
+
+#endif /* __R300_CMDBUF_H__ */
index 5f07b95634956acdb870b097dadcde007ae368d2..3c6ec2a34a8fc9f85fa338198e25904da1266021 100644 (file)
@@ -55,13 +55,14 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "tnl/t_vp_build.h"
 
 #include "drivers/common/driverfuncs.h"
+#include "drivers/common/meta.h"
 
 #include "r300_context.h"
 #include "radeon_context.h"
 #include "radeon_span.h"
+#include "r300_blit.h"
 #include "r300_cmdbuf.h"
 #include "r300_state.h"
-#include "r300_ioctl.h"
 #include "r300_tex.h"
 #include "r300_emit.h"
 #include "r300_swtcl.h"
@@ -92,6 +93,7 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "main/remap_helper.h"
 
+void r300_init_texcopy_functions(struct dd_function_table *table);
 
 static const struct dri_extension card_extensions[] = {
   /* *INDENT-OFF* */
@@ -451,6 +453,13 @@ static void r300InitGLExtensions(GLcontext *ctx)
        }
 }
 
+static void r300InitIoctlFuncs(struct dd_function_table *functions)
+{
+       functions->Clear = _mesa_meta_Clear;
+       functions->Finish = radeonFinish;
+       functions->Flush = radeonFlush;
+}
+
 /* Create the device specific rendering context.
  */
 GLboolean r300CreateContext(const __GLcontextModes * glVisual,
@@ -484,6 +493,10 @@ GLboolean r300CreateContext(const __GLcontextModes * glVisual,
        radeonInitQueryObjFunctions(&functions);
        radeonInitBufferObjectFuncs(&functions);
 
+       if (r300->radeon.radeonScreen->kernel_mm) {
+               r300_init_texcopy_functions(&functions);
+       }
+
        if (!radeonInitContext(&r300->radeon, &functions,
                               glVisual, driContextPriv,
                               sharedContextPrivate)) {
@@ -530,6 +543,7 @@ GLboolean r300CreateContext(const __GLcontextModes * glVisual,
                r300InitSwtcl(ctx);
        }
 
+       r300_blit_init(r300);
        radeon_fbo_init(&r300->radeon);
        radeonInitSpanFuncs( ctx );
        r300InitCmdBuf(r300);
index 518d5cdbf4f5908a93ecc61091d9e81e411f70a0..54a92a2e44727b7f14b839fa1d639a51a2a257d7 100644 (file)
@@ -533,6 +533,11 @@ struct r300_context {
 
        uint32_t fallback;
 
+       struct {
+               struct r300_vertex_program_code vp_code;
+               struct rX00_fragment_program_code fp_code;
+       } blit;
+
        DECLARE_RENDERINPUTS(render_inputs_bitset);
 };
 
@@ -549,6 +554,8 @@ extern void r300InitShaderFunctions(r300ContextPtr r300);
 
 extern void r300InitDraw(GLcontext *ctx);
 
+extern void r300_init_texcopy_functions(struct dd_function_table *table);
+
 #define r300PackFloat32 radeonPackFloat32
 #define r300PackFloat24 radeonPackFloat24
 
index e9968f9ffeffb8d958ac9f1245e8cd41e196c4b8..3dcd986e2267f3ae09abc6342e0efaa30dbe4456 100644 (file)
@@ -100,7 +100,7 @@ static void r300FixupIndexBuffer(GLcontext *ctx, const struct _mesa_index_buffer
                GLubyte *in = (GLubyte *)src_ptr;
 
                radeonAllocDmaRegion(&r300->radeon, &r300->ind_buf.bo, &r300->ind_buf.bo_offset, size, 4);
-
+               radeon_bo_map(r300->ind_buf.bo, 1);
                assert(r300->ind_buf.bo->ptr != NULL);
                out = (GLuint *)ADD_POINTERS(r300->ind_buf.bo->ptr, r300->ind_buf.bo_offset);
 
@@ -111,7 +111,7 @@ static void r300FixupIndexBuffer(GLcontext *ctx, const struct _mesa_index_buffer
                if (i < mesa_ind_buf->count) {
                        *out++ = in[i];
                }
-
+               radeon_bo_unmap(r300->ind_buf.bo);
 #if MESA_BIG_ENDIAN
        } else { /* if (mesa_ind_buf->type == GL_UNSIGNED_SHORT) */
                GLushort *in = (GLushort *)src_ptr;
@@ -120,6 +120,7 @@ static void r300FixupIndexBuffer(GLcontext *ctx, const struct _mesa_index_buffer
                radeonAllocDmaRegion(&r300->radeon, &r300->ind_buf.bo,
                                     &r300->ind_buf.bo_offset, size, 4);
 
+               radeon_bo_map(r300->ind_buf.bo, 1);
                assert(r300->ind_buf.bo->ptr != NULL);
                out = (GLuint *)ADD_POINTERS(r300->ind_buf.bo->ptr, r300->ind_buf.bo_offset);
 
@@ -130,6 +131,7 @@ static void r300FixupIndexBuffer(GLcontext *ctx, const struct _mesa_index_buffer
                if (i < mesa_ind_buf->count) {
                        *out++ = in[i];
                }
+               radeon_bo_unmap(r300->ind_buf.bo);
 #endif
        }
 
@@ -173,10 +175,12 @@ static void r300SetupIndexBuffer(GLcontext *ctx, const struct _mesa_index_buffer
 
                radeonAllocDmaRegion(&r300->radeon, &r300->ind_buf.bo, &r300->ind_buf.bo_offset, size, 4);
 
+               radeon_bo_map(r300->ind_buf.bo, 1);
                assert(r300->ind_buf.bo->ptr != NULL);
                dst_ptr = ADD_POINTERS(r300->ind_buf.bo->ptr, r300->ind_buf.bo_offset);
                _mesa_memcpy(dst_ptr, src_ptr, size);
 
+               radeon_bo_unmap(r300->ind_buf.bo);
                r300->ind_buf.is_32bit = (mesa_ind_buf->type == GL_UNSIGNED_INT);
                r300->ind_buf.count = mesa_ind_buf->count;
 
@@ -242,6 +246,7 @@ static void r300ConvertAttrib(GLcontext *ctx, int count, const struct gl_client_
        }
 
        radeonAllocDmaRegion(&r300->radeon, &attr->bo, &attr->bo_offset, sizeof(GLfloat) * input->Size * count, 32);
+       radeon_bo_map(attr->bo, 1);
        dst_ptr = (GLfloat *)ADD_POINTERS(attr->bo->ptr, attr->bo_offset);
 
        radeon_print(RADEON_FALLBACKS, RADEON_IMPORTANT,
@@ -280,6 +285,7 @@ static void r300ConvertAttrib(GLcontext *ctx, int count, const struct gl_client_
                        break;
        }
 
+       radeon_bo_unmap(attr->bo);
        if (mapped_named_bo) {
                ctx->Driver.UnmapBuffer(ctx, GL_ARRAY_BUFFER, input->BufferObj);
        }
@@ -294,6 +300,8 @@ static void r300AlignDataToDword(GLcontext *ctx, const struct gl_client_array *i
 
        radeonAllocDmaRegion(&r300->radeon, &attr->bo, &attr->bo_offset, size, 32);
 
+       radeon_bo_map(attr->bo, 1);
+
        if (!input->BufferObj->Pointer) {
                ctx->Driver.MapBuffer(ctx, GL_ARRAY_BUFFER, GL_READ_ONLY_ARB, input->BufferObj);
                mapped_named_bo = GL_TRUE;
@@ -317,6 +325,7 @@ static void r300AlignDataToDword(GLcontext *ctx, const struct gl_client_array *i
                ctx->Driver.UnmapBuffer(ctx, GL_ARRAY_BUFFER, input->BufferObj);
        }
 
+       radeon_bo_unmap(attr->bo);
        attr->stride = dst_stride;
 }
 
@@ -527,6 +536,7 @@ static void r300AllocDmaRegions(GLcontext *ctx, const struct gl_client_array *in
                                }
 
                                radeonAllocDmaRegion(&r300->radeon, &vbuf->attribs[index].bo, &vbuf->attribs[index].bo_offset, size, 32);
+                               radeon_bo_map(vbuf->attribs[index].bo, 1);
                                assert(vbuf->attribs[index].bo->ptr != NULL);
                                dst = (uint32_t *)ADD_POINTERS(vbuf->attribs[index].bo->ptr, vbuf->attribs[index].bo_offset);
                                switch (vbuf->attribs[index].dwords) {
@@ -536,6 +546,7 @@ static void r300AllocDmaRegions(GLcontext *ctx, const struct gl_client_array *in
                                        case 4: radeonEmitVec16(dst, input[i]->Ptr, input[i]->StrideB, local_count); break;
                                        default: assert(0); break;
                                }
+                               radeon_bo_unmap(vbuf->attribs[index].bo);
 
                        }
                }
index 07e62230874cd462727764f769aaca20512e8e8f..15aeaf051480f40554de4a1958e7309c5eac3b7b 100644 (file)
@@ -49,7 +49,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "r300_context.h"
 #include "r300_state.h"
 #include "r300_emit.h"
-#include "r300_ioctl.h"
 #include "r300_render.h"
 #include "r300_swtcl.h"
 
@@ -118,7 +117,7 @@ GLuint r300VAPOutputCntl1(GLcontext * ctx, GLuint vp_writes)
 
        if (first_free_texcoord > 8) {
                fprintf(stderr, "\tout of free texcoords\n");
-               _mesa_exit(-1);
+               exit(-1);
        }
 
        return ret;
diff --git a/src/mesa/drivers/dri/r300/r300_ioctl.c b/src/mesa/drivers/dri/r300/r300_ioctl.c
deleted file mode 100644 (file)
index 5cb04e2..0000000
+++ /dev/null
@@ -1,782 +0,0 @@
-/*
-Copyright (C) The Weather Channel, Inc.  2002.
-Copyright (C) 2004 Nicolai Haehnle.
-All Rights Reserved.
-
-The Weather Channel (TM) funded Tungsten Graphics to develop the
-initial release of the Radeon 8500 driver under the XFree86 license.
-This notice must be preserved.
-
-Permission is hereby granted, free of charge, to any person obtaining
-a copy of this software and associated documentation files (the
-"Software"), to deal in the Software without restriction, including
-without limitation the rights to use, copy, modify, merge, publish,
-distribute, sublicense, and/or sell copies of the Software, and to
-permit persons to whom the Software is furnished to do so, subject to
-the following conditions:
-
-The above copyright notice and this permission notice (including the
-next paragraph) shall be included in all copies or substantial
-portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
-EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
-MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
-IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
-LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
-OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
-WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
-
-**************************************************************************/
-
-/**
- * \file
- *
- * \author Keith Whitwell <keith@tungstengraphics.com>
- *
- * \author Nicolai Haehnle <prefect_@gmx.net>
- */
-
-#include <sched.h>
-#include <errno.h>
-
-#include "main/glheader.h"
-#include "main/imports.h"
-#include "main/macros.h"
-#include "main/context.h"
-#include "main/simple_list.h"
-#include "swrast/swrast.h"
-
-#include "radeon_common.h"
-#include "radeon_lock.h"
-#include "r300_context.h"
-#include "r300_ioctl.h"
-#include "r300_cmdbuf.h"
-#include "r300_state.h"
-#include "r300_vertprog.h"
-#include "radeon_reg.h"
-#include "r300_emit.h"
-#include "r300_context.h"
-
-#include "vblank.h"
-
-#define R200_3D_DRAW_IMMD_2      0xC0003500
-
-#define CLEARBUFFER_COLOR      0x1
-#define CLEARBUFFER_DEPTH      0x2
-#define CLEARBUFFER_STENCIL    0x4
-
-#if 1
-
-/**
- * Fragment program helper macros
- */
-
-/* Produce unshifted source selectors */
-#define FP_TMP(idx) (idx)
-#define FP_CONST(idx) ((idx) | (1 << 5))
-
-/* Produce source/dest selector dword */
-#define FP_SELC_MASK_NO                0
-#define FP_SELC_MASK_X         1
-#define FP_SELC_MASK_Y         2
-#define FP_SELC_MASK_XY                3
-#define FP_SELC_MASK_Z         4
-#define FP_SELC_MASK_XZ                5
-#define FP_SELC_MASK_YZ                6
-#define FP_SELC_MASK_XYZ       7
-
-#define FP_SELC(destidx,regmask,outmask,src0,src1,src2) \
-       (((destidx) << R300_ALU_DSTC_SHIFT) |           \
-        (FP_SELC_MASK_##regmask << 23) |               \
-        (FP_SELC_MASK_##outmask << 26) |               \
-        ((src0) << R300_ALU_SRC0C_SHIFT) |             \
-        ((src1) << R300_ALU_SRC1C_SHIFT) |             \
-        ((src2) << R300_ALU_SRC2C_SHIFT))
-
-#define FP_SELA_MASK_NO                0
-#define FP_SELA_MASK_W         1
-
-#define FP_SELA(destidx,regmask,outmask,src0,src1,src2) \
-       (((destidx) << R300_ALU_DSTA_SHIFT) |           \
-        (FP_SELA_MASK_##regmask << 23) |               \
-        (FP_SELA_MASK_##outmask << 24) |               \
-        ((src0) << R300_ALU_SRC0A_SHIFT) |             \
-        ((src1) << R300_ALU_SRC1A_SHIFT) |             \
-        ((src2) << R300_ALU_SRC2A_SHIFT))
-
-/* Produce unshifted argument selectors */
-#define FP_ARGC(source)        R300_ALU_ARGC_##source
-#define FP_ARGA(source) R300_ALU_ARGA_##source
-#define FP_ABS(arg) ((arg) | (1 << 6))
-#define FP_NEG(arg) ((arg) ^ (1 << 5))
-
-/* Produce instruction dword */
-#define FP_INSTRC(opcode,arg0,arg1,arg2) \
-       (R300_ALU_OUTC_##opcode |               \
-       ((arg0) << R300_ALU_ARG0C_SHIFT) |      \
-       ((arg1) << R300_ALU_ARG1C_SHIFT) |      \
-       ((arg2) << R300_ALU_ARG2C_SHIFT))
-
-#define FP_INSTRA(opcode,arg0,arg1,arg2) \
-       (R300_ALU_OUTA_##opcode |               \
-       ((arg0) << R300_ALU_ARG0A_SHIFT) |      \
-       ((arg1) << R300_ALU_ARG1A_SHIFT) |      \
-       ((arg2) << R300_ALU_ARG2A_SHIFT))
-
-#endif
-
-static void r300EmitClearState(GLcontext * ctx);
-
-static void r300ClearBuffer(r300ContextPtr r300, int flags,
-                           struct radeon_renderbuffer *rrb,
-                           struct radeon_renderbuffer *rrbd)
-{
-       BATCH_LOCALS(&r300->radeon);
-       GLcontext *ctx = r300->radeon.glCtx;
-       __DRIdrawablePrivate *dPriv = radeon_get_drawable(&r300->radeon);
-       GLuint cbpitch = 0;
-       r300ContextPtr rmesa = r300;
-
-       if (RADEON_DEBUG & RADEON_IOCTL)
-               fprintf(stderr, "%s: buffer %p (%i,%i %ix%i)\n",
-                       __FUNCTION__, rrb, dPriv->x, dPriv->y,
-                       dPriv->w, dPriv->h);
-
-       if (rrb) {
-               cbpitch = (rrb->pitch / rrb->cpp);
-               if (rrb->cpp == 4)
-                       cbpitch |= R300_COLOR_FORMAT_ARGB8888;
-               else
-                       cbpitch |= R300_COLOR_FORMAT_RGB565;
-
-               if (rrb->bo->flags & RADEON_BO_FLAGS_MACRO_TILE){
-                       cbpitch |= R300_COLOR_TILE_ENABLE;
-        }
-       }
-
-       /* TODO in bufmgr */
-       cp_wait(&r300->radeon, R300_WAIT_3D | R300_WAIT_3D_CLEAN);
-       end_3d(&rmesa->radeon);
-
-       if (flags & CLEARBUFFER_COLOR) {
-               assert(rrb != 0);
-               BEGIN_BATCH_NO_AUTOSTATE(6);
-               OUT_BATCH_REGSEQ(R300_RB3D_COLOROFFSET0, 1);
-               OUT_BATCH_RELOC(0, rrb->bo, 0, 0, RADEON_GEM_DOMAIN_VRAM, 0);
-               OUT_BATCH_REGVAL(R300_RB3D_COLORPITCH0, cbpitch);
-               END_BATCH();
-       }
-#if 1
-       if (flags & (CLEARBUFFER_DEPTH | CLEARBUFFER_STENCIL)) {
-               uint32_t zbpitch = (rrbd->pitch / rrbd->cpp);
-               if (rrbd->bo->flags & RADEON_BO_FLAGS_MACRO_TILE){
-                       zbpitch |= R300_DEPTHMACROTILE_ENABLE;
-        }
-               if (rrbd->bo->flags & RADEON_BO_FLAGS_MICRO_TILE){
-            zbpitch |= R300_DEPTHMICROTILE_TILED;
-        }
-               BEGIN_BATCH_NO_AUTOSTATE(6);
-               OUT_BATCH_REGSEQ(R300_ZB_DEPTHOFFSET, 1);
-               OUT_BATCH_RELOC(0, rrbd->bo, 0, 0, RADEON_GEM_DOMAIN_VRAM, 0);
-               OUT_BATCH_REGSEQ(R300_ZB_DEPTHPITCH, 1);
-               if (!r300->radeon.radeonScreen->kernel_mm)
-                       OUT_BATCH(zbpitch);
-               else
-                       OUT_BATCH_RELOC(zbpitch, rrbd->bo, zbpitch, 0, RADEON_GEM_DOMAIN_VRAM, 0);
-               END_BATCH();
-       }
-#endif
-       BEGIN_BATCH_NO_AUTOSTATE(6);
-       OUT_BATCH_REGSEQ(RB3D_COLOR_CHANNEL_MASK, 1);
-       if (flags & CLEARBUFFER_COLOR) {
-               OUT_BATCH((ctx->Color.ColorMask[BCOMP] ? RB3D_COLOR_CHANNEL_MASK_BLUE_MASK0 : 0) |
-                         (ctx->Color.ColorMask[GCOMP] ? RB3D_COLOR_CHANNEL_MASK_GREEN_MASK0 : 0) |
-                         (ctx->Color.ColorMask[RCOMP] ? RB3D_COLOR_CHANNEL_MASK_RED_MASK0 : 0) |
-                         (ctx->Color.ColorMask[ACOMP] ? RB3D_COLOR_CHANNEL_MASK_ALPHA_MASK0 : 0));
-       } else {
-               OUT_BATCH(0);
-       }
-
-
-       {
-               uint32_t t1, t2;
-
-               t1 = 0x0;
-               t2 = 0x0;
-
-               if (flags & CLEARBUFFER_DEPTH) {
-                       t1 |= R300_Z_ENABLE | R300_Z_WRITE_ENABLE;
-                       t2 |=
-                           (R300_ZS_ALWAYS << R300_Z_FUNC_SHIFT);
-               }
-
-               if (flags & CLEARBUFFER_STENCIL) {
-                       t1 |= R300_STENCIL_ENABLE;
-                       t2 |=
-                           (R300_ZS_ALWAYS <<
-                            R300_S_FRONT_FUNC_SHIFT) |
-                           (R300_ZS_REPLACE <<
-                            R300_S_FRONT_SFAIL_OP_SHIFT) |
-                           (R300_ZS_REPLACE <<
-                            R300_S_FRONT_ZPASS_OP_SHIFT) |
-                           (R300_ZS_REPLACE <<
-                            R300_S_FRONT_ZFAIL_OP_SHIFT);
-               }
-
-               OUT_BATCH_REGSEQ(R300_ZB_CNTL, 3);
-               OUT_BATCH(t1);
-               OUT_BATCH(t2);
-               OUT_BATCH(((ctx->Stencil.WriteMask[0] & R300_STENCILREF_MASK) <<
-                   R300_STENCILWRITEMASK_SHIFT) |
-                         (ctx->Stencil.Clear & R300_STENCILREF_MASK));
-               END_BATCH();
-       }
-
-       if (!rmesa->radeon.radeonScreen->kernel_mm) {
-               BEGIN_BATCH_NO_AUTOSTATE(9);
-               OUT_BATCH(cmdpacket3(r300->radeon.radeonScreen, R300_CMD_PACKET3_CLEAR));
-               OUT_BATCH_FLOAT32(dPriv->w / 2.0);
-               OUT_BATCH_FLOAT32(dPriv->h / 2.0);
-               OUT_BATCH_FLOAT32(ctx->Depth.Clear);
-               OUT_BATCH_FLOAT32(1.0);
-               OUT_BATCH_FLOAT32(ctx->Color.ClearColor[0]);
-               OUT_BATCH_FLOAT32(ctx->Color.ClearColor[1]);
-               OUT_BATCH_FLOAT32(ctx->Color.ClearColor[2]);
-               OUT_BATCH_FLOAT32(ctx->Color.ClearColor[3]);
-               END_BATCH();
-       } else {
-               OUT_BATCH(CP_PACKET3(R200_3D_DRAW_IMMD_2, 8));
-               OUT_BATCH(R300_PRIM_TYPE_POINT | R300_PRIM_WALK_RING |
-                         (1 << R300_PRIM_NUM_VERTICES_SHIFT));
-               OUT_BATCH_FLOAT32(dPriv->w / 2.0);
-               OUT_BATCH_FLOAT32(dPriv->h / 2.0);
-               OUT_BATCH_FLOAT32(ctx->Depth.Clear);
-               OUT_BATCH_FLOAT32(1.0);
-               OUT_BATCH_FLOAT32(ctx->Color.ClearColor[0]);
-               OUT_BATCH_FLOAT32(ctx->Color.ClearColor[1]);
-               OUT_BATCH_FLOAT32(ctx->Color.ClearColor[2]);
-               OUT_BATCH_FLOAT32(ctx->Color.ClearColor[3]);
-       }
-
-       r300EmitCacheFlush(rmesa);
-       cp_wait(&r300->radeon, R300_WAIT_3D | R300_WAIT_3D_CLEAN);
-
-       R300_STATECHANGE(r300, cb);
-       R300_STATECHANGE(r300, cmk);
-       R300_STATECHANGE(r300, zs);
-}
-
-static void r300EmitClearState(GLcontext * ctx)
-{
-       r300ContextPtr r300 = R300_CONTEXT(ctx);
-       BATCH_LOCALS(&r300->radeon);
-       __DRIdrawablePrivate *dPriv = radeon_get_drawable(&r300->radeon);
-       int i;
-       int has_tcl;
-       int is_r500 = 0;
-       GLuint vap_cntl;
-
-       has_tcl = r300->options.hw_tcl_enabled;
-
-       if (r300->radeon.radeonScreen->chip_family >= CHIP_FAMILY_RV515)
-               is_r500 = 1;
-
-       /* State atom dirty tracking is a little subtle here.
-        *
-        * On the one hand, we need to make sure base state is emitted
-        * here if we start with an empty batch buffer, otherwise clear
-        * works incorrectly with multiple processes. Therefore, the first
-        * BEGIN_BATCH cannot be a BEGIN_BATCH_NO_AUTOSTATE.
-        *
-        * On the other hand, implicit state emission clears the state atom
-        * dirty bits, so we have to call R300_STATECHANGE later than the
-        * first BEGIN_BATCH.
-        *
-        * The final trickiness is that, because we change state, we need
-        * to ensure that any stored swtcl primitives are flushed properly
-        * before we start changing state. See the R300_NEWPRIM in r300Clear
-        * for this.
-        */
-       BEGIN_BATCH(31);
-       OUT_BATCH_REGSEQ(R300_VAP_PROG_STREAM_CNTL_0, 1);
-       if (!has_tcl)
-               OUT_BATCH(((((0 << R300_DST_VEC_LOC_SHIFT) | R300_DATA_TYPE_FLOAT_4) << R300_DATA_TYPE_0_SHIFT) |
-                ((R300_LAST_VEC | (2 << R300_DST_VEC_LOC_SHIFT) | R300_DATA_TYPE_FLOAT_4) << R300_DATA_TYPE_1_SHIFT)));
-       else
-               OUT_BATCH(((((0 << R300_DST_VEC_LOC_SHIFT) | R300_DATA_TYPE_FLOAT_4) << R300_DATA_TYPE_0_SHIFT) |
-                ((R300_LAST_VEC | (1 << R300_DST_VEC_LOC_SHIFT) | R300_DATA_TYPE_FLOAT_4) << R300_DATA_TYPE_1_SHIFT)));
-
-       OUT_BATCH_REGVAL(R300_FG_FOG_BLEND, 0);
-       OUT_BATCH_REGVAL(R300_VAP_PROG_STREAM_CNTL_EXT_0,
-          ((((R300_SWIZZLE_SELECT_X << R300_SWIZZLE_SELECT_X_SHIFT) |
-              (R300_SWIZZLE_SELECT_Y << R300_SWIZZLE_SELECT_Y_SHIFT) |
-              (R300_SWIZZLE_SELECT_Z << R300_SWIZZLE_SELECT_Z_SHIFT) |
-              (R300_SWIZZLE_SELECT_W << R300_SWIZZLE_SELECT_W_SHIFT) |
-              ((R300_WRITE_ENA_X | R300_WRITE_ENA_Y | R300_WRITE_ENA_Z | R300_WRITE_ENA_W) << R300_WRITE_ENA_SHIFT))
-             << R300_SWIZZLE0_SHIFT) |
-            (((R300_SWIZZLE_SELECT_X << R300_SWIZZLE_SELECT_X_SHIFT) |
-              (R300_SWIZZLE_SELECT_Y << R300_SWIZZLE_SELECT_Y_SHIFT) |
-              (R300_SWIZZLE_SELECT_Z << R300_SWIZZLE_SELECT_Z_SHIFT) |
-              (R300_SWIZZLE_SELECT_W << R300_SWIZZLE_SELECT_W_SHIFT) |
-              ((R300_WRITE_ENA_X | R300_WRITE_ENA_Y | R300_WRITE_ENA_Z | R300_WRITE_ENA_W) << R300_WRITE_ENA_SHIFT))
-             << R300_SWIZZLE1_SHIFT)));
-
-       /* R300_VAP_INPUT_CNTL_0, R300_VAP_INPUT_CNTL_1 */
-       OUT_BATCH_REGSEQ(R300_VAP_VTX_STATE_CNTL, 2);
-       OUT_BATCH((R300_SEL_USER_COLOR_0 << R300_COLOR_0_ASSEMBLY_SHIFT));
-       OUT_BATCH(R300_INPUT_CNTL_POS | R300_INPUT_CNTL_COLOR | R300_INPUT_CNTL_TC0);
-
-       /* comes from fglrx startup of clear */
-       OUT_BATCH_REGSEQ(R300_SE_VTE_CNTL, 2);
-       OUT_BATCH(R300_VTX_W0_FMT | R300_VPORT_X_SCALE_ENA |
-                 R300_VPORT_X_OFFSET_ENA | R300_VPORT_Y_SCALE_ENA |
-                 R300_VPORT_Y_OFFSET_ENA | R300_VPORT_Z_SCALE_ENA |
-                 R300_VPORT_Z_OFFSET_ENA);
-       OUT_BATCH(0x8);
-
-       OUT_BATCH_REGVAL(R300_VAP_PSC_SGN_NORM_CNTL, 0xaaaaaaaa);
-
-       OUT_BATCH_REGSEQ(R300_VAP_OUTPUT_VTX_FMT_0, 2);
-       OUT_BATCH(R300_VAP_OUTPUT_VTX_FMT_0__POS_PRESENT |
-                 R300_VAP_OUTPUT_VTX_FMT_0__COLOR_0_PRESENT);
-       OUT_BATCH(0); /* no textures */
-
-       OUT_BATCH_REGVAL(R300_TX_ENABLE, 0);
-
-       OUT_BATCH_REGSEQ(R300_SE_VPORT_XSCALE, 6);
-       OUT_BATCH_FLOAT32(1.0);
-       OUT_BATCH_FLOAT32(dPriv->x);
-       OUT_BATCH_FLOAT32(1.0);
-       OUT_BATCH_FLOAT32(dPriv->y);
-       OUT_BATCH_FLOAT32(1.0);
-       OUT_BATCH_FLOAT32(0.0);
-
-       OUT_BATCH_REGVAL(R300_FG_ALPHA_FUNC, 0);
-
-       OUT_BATCH_REGSEQ(R300_RB3D_CBLEND, 2);
-       OUT_BATCH(0x0);
-       OUT_BATCH(0x0);
-       END_BATCH();
-
-       R300_STATECHANGE(r300, vir[0]);
-       R300_STATECHANGE(r300, fogs);
-       R300_STATECHANGE(r300, vir[1]);
-       R300_STATECHANGE(r300, vic);
-       R300_STATECHANGE(r300, vte);
-       R300_STATECHANGE(r300, vof);
-       R300_STATECHANGE(r300, txe);
-       R300_STATECHANGE(r300, vpt);
-       R300_STATECHANGE(r300, at);
-       R300_STATECHANGE(r300, bld);
-       R300_STATECHANGE(r300, ps);
-
-       if (has_tcl) {
-               R300_STATECHANGE(r300, vap_clip_cntl);
-
-               BEGIN_BATCH_NO_AUTOSTATE(2);
-               OUT_BATCH_REGVAL(R300_VAP_CLIP_CNTL, R300_PS_UCP_MODE_CLIP_AS_TRIFAN | R300_CLIP_DISABLE);
-               END_BATCH();
-        }
-
-       BEGIN_BATCH_NO_AUTOSTATE(2);
-       OUT_BATCH_REGVAL(R300_GA_POINT_SIZE,
-               ((dPriv->w * 6) << R300_POINTSIZE_X_SHIFT) |
-               ((dPriv->h * 6) << R300_POINTSIZE_Y_SHIFT));
-       END_BATCH();
-
-       if (!is_r500) {
-               R300_STATECHANGE(r300, ri);
-               R300_STATECHANGE(r300, rc);
-               R300_STATECHANGE(r300, rr);
-
-               BEGIN_BATCH(14);
-               OUT_BATCH_REGSEQ(R300_RS_IP_0, 8);
-               for (i = 0; i < 8; ++i)
-                       OUT_BATCH(R300_RS_SEL_T(1) | R300_RS_SEL_R(2) | R300_RS_SEL_Q(3));
-
-               OUT_BATCH_REGSEQ(R300_RS_COUNT, 2);
-               OUT_BATCH((1 << R300_IC_COUNT_SHIFT) | R300_HIRES_EN);
-               OUT_BATCH(0x0);
-
-               OUT_BATCH_REGVAL(R300_RS_INST_0, R300_RS_INST_COL_CN_WRITE);
-               END_BATCH();
-       } else {
-               R300_STATECHANGE(r300, ri);
-               R300_STATECHANGE(r300, rc);
-               R300_STATECHANGE(r300, rr);
-
-               BEGIN_BATCH(14);
-               OUT_BATCH_REGSEQ(R500_RS_IP_0, 8);
-               for (i = 0; i < 8; ++i) {
-                       OUT_BATCH((R500_RS_IP_PTR_K0 << R500_RS_IP_TEX_PTR_S_SHIFT) |
-                                 (R500_RS_IP_PTR_K0 << R500_RS_IP_TEX_PTR_T_SHIFT) |
-                                 (R500_RS_IP_PTR_K0 << R500_RS_IP_TEX_PTR_R_SHIFT) |
-                                 (R500_RS_IP_PTR_K1 << R500_RS_IP_TEX_PTR_Q_SHIFT));
-               }
-
-               OUT_BATCH_REGSEQ(R300_RS_COUNT, 2);
-               OUT_BATCH((1 << R300_IC_COUNT_SHIFT) | R300_HIRES_EN);
-               OUT_BATCH(0x0);
-
-               OUT_BATCH_REGVAL(R500_RS_INST_0, R500_RS_INST_COL_CN_WRITE);
-               END_BATCH();
-       }
-
-       if (!is_r500) {
-               R300_STATECHANGE(r300, fp);
-               R300_STATECHANGE(r300, fpi[0]);
-               R300_STATECHANGE(r300, fpi[1]);
-               R300_STATECHANGE(r300, fpi[2]);
-               R300_STATECHANGE(r300, fpi[3]);
-
-               BEGIN_BATCH(17);
-               OUT_BATCH_REGSEQ(R300_US_CONFIG, 3);
-               OUT_BATCH(0x0);
-               OUT_BATCH(0x0);
-               OUT_BATCH(0x0);
-               OUT_BATCH_REGSEQ(R300_US_CODE_ADDR_0, 4);
-               OUT_BATCH(0x0);
-               OUT_BATCH(0x0);
-               OUT_BATCH(0x0);
-               OUT_BATCH(R300_RGBA_OUT);
-
-               OUT_BATCH_REGVAL(R300_US_ALU_RGB_INST_0,
-                       FP_INSTRC(MAD, FP_ARGC(SRC0C_XYZ), FP_ARGC(ONE), FP_ARGC(ZERO)));
-               OUT_BATCH_REGVAL(R300_US_ALU_RGB_ADDR_0,
-                       FP_SELC(0, NO, XYZ, FP_TMP(0), 0, 0));
-               OUT_BATCH_REGVAL(R300_US_ALU_ALPHA_INST_0,
-                       FP_INSTRA(MAD, FP_ARGA(SRC0A), FP_ARGA(ONE), FP_ARGA(ZERO)));
-               OUT_BATCH_REGVAL(R300_US_ALU_ALPHA_ADDR_0,
-                       FP_SELA(0, NO, W, FP_TMP(0), 0, 0));
-               END_BATCH();
-       } else {
-               struct radeon_state_atom r500fp;
-               uint32_t _cmd[10];
-
-               R300_STATECHANGE(r300, fp);
-               R300_STATECHANGE(r300, r500fp);
-
-               BEGIN_BATCH(7);
-               OUT_BATCH_REGSEQ(R500_US_CONFIG, 2);
-               OUT_BATCH(R500_ZERO_TIMES_ANYTHING_EQUALS_ZERO);
-               OUT_BATCH(0x0);
-               OUT_BATCH_REGSEQ(R500_US_CODE_ADDR, 3);
-               OUT_BATCH(R500_US_CODE_START_ADDR(0) | R500_US_CODE_END_ADDR(1));
-               OUT_BATCH(R500_US_CODE_RANGE_ADDR(0) | R500_US_CODE_RANGE_SIZE(1));
-               OUT_BATCH(R500_US_CODE_OFFSET_ADDR(0));
-               END_BATCH();
-
-               r500fp.check = check_r500fp;
-               r500fp.cmd = _cmd;
-               r500fp.cmd[0] = cmdr500fp(r300->radeon.radeonScreen, 0, 1, 0, 0);
-               r500fp.cmd[1] = R500_INST_TYPE_OUT |
-                       R500_INST_TEX_SEM_WAIT |
-                       R500_INST_LAST |
-                       R500_INST_RGB_OMASK_R |
-                       R500_INST_RGB_OMASK_G |
-                       R500_INST_RGB_OMASK_B |
-                       R500_INST_ALPHA_OMASK |
-                       R500_INST_RGB_CLAMP |
-                       R500_INST_ALPHA_CLAMP;
-               r500fp.cmd[2] = R500_RGB_ADDR0(0) |
-                       R500_RGB_ADDR1(0) |
-                       R500_RGB_ADDR1_CONST |
-                       R500_RGB_ADDR2(0) |
-                       R500_RGB_ADDR2_CONST;
-               r500fp.cmd[3] = R500_ALPHA_ADDR0(0) |
-                       R500_ALPHA_ADDR1(0) |
-                       R500_ALPHA_ADDR1_CONST |
-                       R500_ALPHA_ADDR2(0) |
-                       R500_ALPHA_ADDR2_CONST;
-               r500fp.cmd[4] = R500_ALU_RGB_SEL_A_SRC0 |
-                       R500_ALU_RGB_R_SWIZ_A_R |
-                       R500_ALU_RGB_G_SWIZ_A_G |
-                       R500_ALU_RGB_B_SWIZ_A_B |
-                       R500_ALU_RGB_SEL_B_SRC0 |
-                       R500_ALU_RGB_R_SWIZ_B_R |
-                       R500_ALU_RGB_B_SWIZ_B_G |
-                       R500_ALU_RGB_G_SWIZ_B_B;
-               r500fp.cmd[5] = R500_ALPHA_OP_CMP |
-                       R500_ALPHA_SWIZ_A_A |
-                       R500_ALPHA_SWIZ_B_A;
-               r500fp.cmd[6] = R500_ALU_RGBA_OP_CMP |
-                       R500_ALU_RGBA_R_SWIZ_0 |
-                       R500_ALU_RGBA_G_SWIZ_0 |
-                       R500_ALU_RGBA_B_SWIZ_0 |
-                       R500_ALU_RGBA_A_SWIZ_0;
-
-               r500fp.cmd[7] = 0;
-               if (r300->radeon.radeonScreen->kernel_mm) {
-                       emit_r500fp(ctx, &r500fp);
-               } else {
-                       int dwords = r500fp.check(ctx,&r500fp);
-                       BEGIN_BATCH_NO_AUTOSTATE(dwords);
-                       OUT_BATCH_TABLE(r500fp.cmd, dwords);
-                       END_BATCH();
-               }
-
-       }
-
-       BEGIN_BATCH(2);
-       OUT_BATCH_REGVAL(R300_VAP_PVS_STATE_FLUSH_REG, 0);
-       END_BATCH();
-
-       if (has_tcl) {
-               vap_cntl = ((10 << R300_PVS_NUM_SLOTS_SHIFT) |
-                       (5 << R300_PVS_NUM_CNTLRS_SHIFT) |
-                       (12 << R300_VF_MAX_VTX_NUM_SHIFT));
-               if (r300->radeon.radeonScreen->chip_family >= CHIP_FAMILY_RV515)
-                       vap_cntl |= R500_TCL_STATE_OPTIMIZATION;
-       } else {
-               vap_cntl = ((10 << R300_PVS_NUM_SLOTS_SHIFT) |
-                       (5 << R300_PVS_NUM_CNTLRS_SHIFT) |
-                       (5 << R300_VF_MAX_VTX_NUM_SHIFT));
-       }
-
-       if (r300->radeon.radeonScreen->chip_family == CHIP_FAMILY_RV515)
-               vap_cntl |= (2 << R300_PVS_NUM_FPUS_SHIFT);
-       else if ((r300->radeon.radeonScreen->chip_family == CHIP_FAMILY_RV530) ||
-                (r300->radeon.radeonScreen->chip_family == CHIP_FAMILY_RV560) ||
-                (r300->radeon.radeonScreen->chip_family == CHIP_FAMILY_RV570))
-               vap_cntl |= (5 << R300_PVS_NUM_FPUS_SHIFT);
-       else if ((r300->radeon.radeonScreen->chip_family == CHIP_FAMILY_RV410) ||
-                (r300->radeon.radeonScreen->chip_family == CHIP_FAMILY_R420))
-               vap_cntl |= (6 << R300_PVS_NUM_FPUS_SHIFT);
-       else if ((r300->radeon.radeonScreen->chip_family == CHIP_FAMILY_R520) ||
-                (r300->radeon.radeonScreen->chip_family == CHIP_FAMILY_R580))
-               vap_cntl |= (8 << R300_PVS_NUM_FPUS_SHIFT);
-       else
-               vap_cntl |= (4 << R300_PVS_NUM_FPUS_SHIFT);
-
-       R300_STATECHANGE(r300, vap_cntl);
-
-       BEGIN_BATCH(2);
-       OUT_BATCH_REGVAL(R300_VAP_CNTL, vap_cntl);
-       END_BATCH();
-
-       if (has_tcl) {
-        struct radeon_state_atom vpu;
-        uint32_t _cmd[10];
-               R300_STATECHANGE(r300, pvs);
-               R300_STATECHANGE(r300, vap_flush);
-               R300_STATECHANGE(r300, vpi);
-
-               BEGIN_BATCH(4);
-               OUT_BATCH_REGSEQ(R300_VAP_PVS_CODE_CNTL_0, 3);
-               OUT_BATCH((0 << R300_PVS_FIRST_INST_SHIFT) |
-                         (0 << R300_PVS_XYZW_VALID_INST_SHIFT) |
-                         (1 << R300_PVS_LAST_INST_SHIFT));
-               OUT_BATCH((0 << R300_PVS_CONST_BASE_OFFSET_SHIFT) |
-                         (0 << R300_PVS_MAX_CONST_ADDR_SHIFT));
-               OUT_BATCH(1 << R300_PVS_LAST_VTX_SRC_INST_SHIFT);
-               END_BATCH();
-
-               vpu.check = check_vpu;
-               vpu.cmd = _cmd;
-               vpu.cmd[0] = cmdvpu(r300->radeon.radeonScreen, 0, 2);
-
-               vpu.cmd[1] = PVS_OP_DST_OPERAND(VE_ADD, GL_FALSE, GL_FALSE,
-                                         0, 0xf, PVS_DST_REG_OUT);
-               vpu.cmd[2] = PVS_SRC_OPERAND(0, PVS_SRC_SELECT_X, PVS_SRC_SELECT_Y,
-                                      PVS_SRC_SELECT_Z, PVS_SRC_SELECT_W,
-                                      PVS_SRC_REG_INPUT, NEGATE_NONE);
-               vpu.cmd[3] = PVS_SRC_OPERAND(0, PVS_SRC_SELECT_FORCE_0,
-                                      PVS_SRC_SELECT_FORCE_0,
-                                      PVS_SRC_SELECT_FORCE_0,
-                                      PVS_SRC_SELECT_FORCE_0,
-                                      PVS_SRC_REG_INPUT, NEGATE_NONE);
-               vpu.cmd[4] = 0x0;
-
-               vpu.cmd[5] = PVS_OP_DST_OPERAND(VE_ADD, GL_FALSE, GL_FALSE, 1, 0xf,
-                                         PVS_DST_REG_OUT);
-               vpu.cmd[6] = PVS_SRC_OPERAND(1, PVS_SRC_SELECT_X,
-                                      PVS_SRC_SELECT_Y, PVS_SRC_SELECT_Z,
-                                      PVS_SRC_SELECT_W, PVS_SRC_REG_INPUT,
-                                      NEGATE_NONE);
-               vpu.cmd[7] = PVS_SRC_OPERAND(1, PVS_SRC_SELECT_FORCE_0,
-                                      PVS_SRC_SELECT_FORCE_0,
-                                      PVS_SRC_SELECT_FORCE_0,
-                                      PVS_SRC_SELECT_FORCE_0,
-                                      PVS_SRC_REG_INPUT, NEGATE_NONE);
-               vpu.cmd[8] = 0x0;
-
-               if (r300->radeon.radeonScreen->kernel_mm) {
-                       int dwords = r300->hw.vap_flush.check(ctx,&r300->hw.vap_flush);
-                       BEGIN_BATCH_NO_AUTOSTATE(dwords);
-                       OUT_BATCH_TABLE(r300->hw.vap_flush.cmd, dwords);
-                       END_BATCH();
-                       emit_vpu(ctx, &vpu);
-               } else {
-                       int dwords = vpu.check(ctx,&vpu);
-                       BEGIN_BATCH_NO_AUTOSTATE(dwords);
-                       OUT_BATCH_TABLE(vpu.cmd, dwords);
-                       END_BATCH();
-               }
-
-       }
-}
-
-static int r300KernelClear(GLcontext *ctx, GLuint flags)
-{
-       r300ContextPtr r300 = R300_CONTEXT(ctx);
-       __DRIdrawablePrivate *dPriv = radeon_get_drawable(&r300->radeon);
-       struct radeon_framebuffer *rfb = dPriv->driverPrivate;
-       struct radeon_renderbuffer *rrb;
-       struct radeon_renderbuffer *rrbd;
-       int bits = 0, ret;
-
-       /* Make sure it fits there. */
-       radeon_cs_space_reset_bos(r300->radeon.cmdbuf.cs);
-
-       if (flags & BUFFER_BIT_COLOR0) {
-               rrb = radeon_get_renderbuffer(&rfb->base, BUFFER_COLOR0);
-               radeon_cs_space_add_persistent_bo(r300->radeon.cmdbuf.cs,
-                                                 rrb->bo, 0, RADEON_GEM_DOMAIN_VRAM);
-       }
-
-       if (flags & BUFFER_BIT_FRONT_LEFT) {
-               rrb = radeon_get_renderbuffer(&rfb->base, BUFFER_FRONT_LEFT);
-               radeon_cs_space_add_persistent_bo(r300->radeon.cmdbuf.cs,
-                                                 rrb->bo, 0, RADEON_GEM_DOMAIN_VRAM);
-       }
-
-       if (flags & BUFFER_BIT_BACK_LEFT) {
-               rrb = radeon_get_renderbuffer(&rfb->base, BUFFER_BACK_LEFT);
-               radeon_cs_space_add_persistent_bo(r300->radeon.cmdbuf.cs,
-                                                 rrb->bo, 0, RADEON_GEM_DOMAIN_VRAM);
-       }
-
-       rrbd = radeon_get_renderbuffer(&rfb->base, BUFFER_DEPTH);
-       if (rrbd) {
-               radeon_cs_space_add_persistent_bo(r300->radeon.cmdbuf.cs,
-                                                 rrbd->bo, 0, RADEON_GEM_DOMAIN_VRAM);
-       }
-
-       ret = radeon_cs_space_check(r300->radeon.cmdbuf.cs);
-       if (ret)
-         return -1;
-
-       rcommonEnsureCmdBufSpace(&r300->radeon, 421 * 3, __FUNCTION__);
-       if (flags || bits)
-               r300EmitClearState(ctx);
-
-       rrbd = radeon_get_renderbuffer(&rfb->base, BUFFER_DEPTH);
-       if (rrbd && (flags & BUFFER_BIT_DEPTH))
-               bits |= CLEARBUFFER_DEPTH;
-
-       if (rrbd && (flags & BUFFER_BIT_STENCIL))
-               bits |= CLEARBUFFER_STENCIL;
-
-       if (flags & BUFFER_BIT_COLOR0) {
-               rrb = radeon_get_renderbuffer(&rfb->base, BUFFER_COLOR0);
-               r300ClearBuffer(r300, CLEARBUFFER_COLOR, rrb, NULL);
-               bits = 0;
-       }
-
-       if (flags & BUFFER_BIT_FRONT_LEFT) {
-               rrb = radeon_get_renderbuffer(&rfb->base, BUFFER_FRONT_LEFT);
-               r300ClearBuffer(r300, bits | CLEARBUFFER_COLOR, rrb, rrbd);
-               bits = 0;
-       }
-
-       if (flags & BUFFER_BIT_BACK_LEFT) {
-               rrb = radeon_get_renderbuffer(&rfb->base, BUFFER_BACK_LEFT);
-               r300ClearBuffer(r300, bits | CLEARBUFFER_COLOR, rrb, rrbd);
-               bits = 0;
-       }
-
-       if (bits)
-               r300ClearBuffer(r300, bits, NULL, rrbd);
-
-       COMMIT_BATCH();
-       return 0;
-}
-
-/**
- * Buffer clear
- */
-static void r300Clear(GLcontext * ctx, GLbitfield mask)
-{
-       r300ContextPtr r300 = R300_CONTEXT(ctx);
-       __DRIdrawablePrivate *dPriv = radeon_get_drawable(&r300->radeon);
-       const GLuint colorMask = *((GLuint *) & ctx->Color.ColorMask);
-       GLbitfield swrast_mask = 0, tri_mask = 0;
-       int i, ret;
-       struct gl_framebuffer *fb = ctx->DrawBuffer;
-
-       if (RADEON_DEBUG & RADEON_IOCTL)
-               fprintf(stderr, "r300Clear\n");
-
-       if (!r300->radeon.radeonScreen->driScreen->dri2.enabled) {
-               LOCK_HARDWARE(&r300->radeon);
-               UNLOCK_HARDWARE(&r300->radeon);
-               if (dPriv->numClipRects == 0)
-                       return;
-       }
-
-       /* Flush swtcl vertices if necessary, because we will change hardware
-        * state during clear. See also the state-related comment in
-        * r300EmitClearState.
-        */
-       R300_NEWPRIM(r300);
-
-       if (colorMask == ~0)
-         tri_mask |= (mask & BUFFER_BITS_COLOR);
-       else
-         tri_mask |= (mask & (BUFFER_BIT_FRONT_LEFT | BUFFER_BIT_BACK_LEFT));
-
-
-       /* HW stencil */
-       if (mask & BUFFER_BIT_STENCIL) {
-               tri_mask |= BUFFER_BIT_STENCIL;
-       }
-
-       /* HW depth */
-       if (mask & BUFFER_BIT_DEPTH) {
-               tri_mask |= BUFFER_BIT_DEPTH;
-       }
-
-       /* If we're doing a tri pass for depth/stencil, include a likely color
-        * buffer with it.
-        */
-
-       for (i = 0; i < BUFFER_COUNT; i++) {
-         GLuint bufBit = 1 << i;
-         if ((tri_mask) & bufBit) {
-           if (!fb->Attachment[i].Renderbuffer->ClassID) {
-             tri_mask &= ~bufBit;
-             swrast_mask |= bufBit;
-           }
-         }
-       }
-
-       /* SW fallback clearing */
-       swrast_mask = mask & ~tri_mask;
-
-       ret = 0;
-       if (tri_mask) {
-               if (r300->radeon.radeonScreen->kernel_mm)
-                       radeonUserClear(ctx, tri_mask);
-               else {
-                       /* if kernel clear fails due to size restraints fallback */
-                       ret = r300KernelClear(ctx, tri_mask);
-                       if (ret < 0)
-                               swrast_mask |= tri_mask;
-               }
-       }
-
-       if (swrast_mask) {
-               if (RADEON_DEBUG & RADEON_FALLBACKS)
-                       fprintf(stderr, "%s: swrast clear, mask: %x\n",
-                               __FUNCTION__, swrast_mask);
-               _swrast_Clear(ctx, swrast_mask);
-       }
-}
-
-void r300InitIoctlFuncs(struct dd_function_table *functions)
-{
-       functions->Clear = r300Clear;
-       functions->Finish = radeonFinish;
-       functions->Flush = radeonFlush;
-}
diff --git a/src/mesa/drivers/dri/r300/r300_ioctl.h b/src/mesa/drivers/dri/r300/r300_ioctl.h
deleted file mode 100644 (file)
index 3abfa71..0000000
+++ /dev/null
@@ -1,44 +0,0 @@
-/*
-Copyright (C) The Weather Channel, Inc.  2002.  All Rights Reserved.
-
-The Weather Channel (TM) funded Tungsten Graphics to develop the
-initial release of the Radeon 8500 driver under the XFree86 license.
-This notice must be preserved.
-
-Permission is hereby granted, free of charge, to any person obtaining
-a copy of this software and associated documentation files (the
-"Software"), to deal in the Software without restriction, including
-without limitation the rights to use, copy, modify, merge, publish,
-distribute, sublicense, and/or sell copies of the Software, and to
-permit persons to whom the Software is furnished to do so, subject to
-the following conditions:
-
-The above copyright notice and this permission notice (including the
-next paragraph) shall be included in all copies or substantial
-portions of the Software.
-
-THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
-EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
-MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
-IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
-LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
-OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
-WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
-
-**************************************************************************/
-
-/*
- * Authors:
- *   Keith Whitwell <keith@tungstengraphics.com>
- *   Nicolai Haehnle <prefect_@gmx.net>
- */
-
-#ifndef __R300_IOCTL_H__
-#define __R300_IOCTL_H__
-
-#include "r300_context.h"
-#include "radeon_drm.h"
-
-extern void r300InitIoctlFuncs(struct dd_function_table *functions);
-
-#endif                         /* __R300_IOCTL_H__ */
index 4ae593cbe79a6df457585e88d3a8b96dbb9f5faf..02c94250a8ffd3a1fe53060eb7300b83fadfdb1c 100644 (file)
@@ -68,7 +68,6 @@ USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "tnl/tnl.h"
 #include "tnl/t_vp_build.h"
 #include "r300_context.h"
-#include "r300_ioctl.h"
 #include "r300_state.h"
 #include "r300_reg.h"
 #include "r300_tex.h"
index ac20c08e201a4fcd8fe702c0433ff98363336a89..23f81fe790182d6b91dade41d67712cb7c3034da 100644 (file)
@@ -55,7 +55,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "tnl/t_vp_build.h"
 
 #include "r300_context.h"
-#include "r300_ioctl.h"
 #include "r300_state.h"
 #include "r300_reg.h"
 #include "r300_emit.h"
@@ -1312,7 +1311,7 @@ static void r300SetupTextures(GLcontext * ctx)
                fprintf(stderr,
                        "Aiiee ! mtu=%d is greater than R300_MAX_TEXTURE_UNITS=%d\n",
                        mtu, R300_MAX_TEXTURE_UNITS);
-               _mesa_exit(-1);
+               exit(-1);
        }
 
        /* We cannot let disabled tmu offsets pass DRM */
index ee2c71e1a7f81382afc0801f0ce5d224d3ee5555..93983cee2053c781f8e40571f4416618981f0bb1 100644 (file)
@@ -124,7 +124,7 @@ void r300ChooseSwtclVertexFormat(GLcontext *ctx, GLuint *_InputsRead,  GLuint *_
        }
 
        if (ctx->Light.Enabled && ctx->Light.Model.TwoSide) {
-               VB->AttribPtr[VERT_ATTRIB_GENERIC0] = VB->ColorPtr[1];
+               VB->AttribPtr[VERT_ATTRIB_GENERIC0] = VB->BackfaceColorPtr;
                OutputsWritten |= 1 << VERT_RESULT_BFC0;
 #if MESA_LITTLE_ENDIAN
                EMIT_ATTR( _TNL_ATTRIB_GENERIC0, EMIT_4UB_4F_RGBA );
@@ -134,7 +134,7 @@ void r300ChooseSwtclVertexFormat(GLcontext *ctx, GLuint *_InputsRead,  GLuint *_
                ADD_ATTR(VERT_ATTRIB_GENERIC0, R300_DATA_TYPE_BYTE, SWTCL_OVM_COLOR2, SWIZZLE_XYZW, MASK_XYZW, 1);
 #endif
                if (fp_reads & FRAG_BIT_COL1) {
-                       VB->AttribPtr[VERT_ATTRIB_GENERIC1] = VB->SecondaryColorPtr[1];
+                       VB->AttribPtr[VERT_ATTRIB_GENERIC1] = VB->BackfaceSecondaryColorPtr;
                        GLuint swiz = MAKE_SWIZZLE4(SWIZZLE_X, SWIZZLE_Y, SWIZZLE_Z, SWIZZLE_ONE);
                        OutputsWritten |= 1 << VERT_RESULT_BFC1;
 #if MESA_LITTLE_ENDIAN
@@ -159,7 +159,7 @@ void r300ChooseSwtclVertexFormat(GLcontext *ctx, GLuint *_InputsRead,  GLuint *_
                int tex_id = rmesa->selected_fp->wpos_attr - FRAG_ATTRIB_TEX0;
 
                VB->AttribPtr[VERT_ATTRIB_TEX0 + tex_id] = VB->AttribPtr[VERT_ATTRIB_POS];
-               VB->TexCoordPtr[tex_id] = VB->AttribPtr[VERT_ATTRIB_POS];
+               VB->AttribPtr[_TNL_ATTRIB_TEX0 + tex_id] = VB->AttribPtr[VERT_ATTRIB_POS];
                RENDERINPUTS_SET(tnl->render_inputs_bitset, _TNL_ATTRIB_TEX0 + tex_id);
        }
 
@@ -167,7 +167,7 @@ void r300ChooseSwtclVertexFormat(GLcontext *ctx, GLuint *_InputsRead,  GLuint *_
                int tex_id = rmesa->selected_fp->fog_attr - FRAG_ATTRIB_TEX0;
 
                VB->AttribPtr[VERT_ATTRIB_TEX0 + tex_id] = VB->AttribPtr[VERT_ATTRIB_FOG];
-               VB->TexCoordPtr[tex_id] = VB->AttribPtr[VERT_ATTRIB_FOG];
+               VB->AttribPtr[_TNL_ATTRIB_TEX0 + tex_id] = VB->AttribPtr[VERT_ATTRIB_FOG];
                RENDERINPUTS_SET(tnl->render_inputs_bitset, _TNL_ATTRIB_TEX0 + tex_id);
        }
 
@@ -180,7 +180,7 @@ void r300ChooseSwtclVertexFormat(GLcontext *ctx, GLuint *_InputsRead,  GLuint *_
                GLuint swiz, format, hw_format;
                for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
                        if (fp_reads & FRAG_BIT_TEX(i)) {
-                               switch (VB->TexCoordPtr[i]->size) {
+                               switch (VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]->size) {
                                        case 1:
                                                format = EMIT_1F;
                                                hw_format = R300_DATA_TYPE_FLOAT_1;
@@ -215,7 +215,7 @@ void r300ChooseSwtclVertexFormat(GLcontext *ctx, GLuint *_InputsRead,  GLuint *_
 
        if (first_free_tex >= ctx->Const.MaxTextureUnits) {
                fprintf(stderr, "\tout of free texcoords to write fog coordinate\n");
-               _mesa_exit(-1);
+               exit(-1);
        }
 
        R300_NEWPRIM(rmesa);
@@ -665,11 +665,11 @@ void r300_swtcl_flush(GLcontext *ctx, uint32_t current_offset)
        r300EmitCacheFlush(rmesa);
 
        radeonEmitState(&rmesa->radeon);
-    r300_emit_scissor(ctx);
+       r300_emit_scissor(ctx);
        r300EmitVertexAOS(rmesa,
-                       rmesa->radeon.swtcl.vertex_size,
-                       first_elem(&rmesa->radeon.dma.reserved)->bo,
-                       current_offset);
+                         rmesa->radeon.swtcl.vertex_size,
+                         rmesa->radeon.swtcl.bo,
+                         current_offset);
 
        r300EmitVbufPrim(rmesa,
                   rmesa->radeon.swtcl.hw_primitive,
index 726b3ff98e176cc217c68d9b91dd640740a4b357..ac3d5b1bec34840ec691910b48ba6c4915c1c457 100644 (file)
@@ -48,7 +48,6 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "r300_context.h"
 #include "r300_state.h"
-#include "r300_ioctl.h"
 #include "radeon_mipmap_tree.h"
 #include "r300_tex.h"
 
index 8a653ea2d11b0b186e384056141f63f9e751562a..6ede0fe25c9047241a2bea540403b6f4e050f55d 100644 (file)
@@ -51,4 +51,6 @@ extern GLboolean r300ValidateBuffers(GLcontext * ctx);
 
 extern void r300InitTextureFuncs(struct dd_function_table *functions);
 
+int32_t r300TranslateTexFormat(gl_format mesaFormat);
+
 #endif                         /* __r300_TEX_H__ */
diff --git a/src/mesa/drivers/dri/r300/r300_texcopy.c b/src/mesa/drivers/dri/r300/r300_texcopy.c
new file mode 100644 (file)
index 0000000..ebc9c05
--- /dev/null
@@ -0,0 +1,168 @@
+/*
+ * Copyright (C) 2009 Maciej Cencora <m.cencora@gmail.com>
+ *
+ * All Rights Reserved.
+ *
+ * Permission is hereby granted, free of charge, to any person obtaining
+ * a copy of this software and associated documentation files (the
+ * "Software"), to deal in the Software without restriction, including
+ * without limitation the rights to use, copy, modify, merge, publish,
+ * distribute, sublicense, and/or sell copies of the Software, and to
+ * permit persons to whom the Software is furnished to do so, subject to
+ * the following conditions:
+ *
+ * The above copyright notice and this permission notice (including the
+ * next paragraph) shall be included in all copies or substantial
+ * portions of the Software.
+ *
+ * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
+ * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
+ * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.
+ * IN NO EVENT SHALL THE COPYRIGHT OWNER(S) AND/OR ITS SUPPLIERS BE
+ * LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION
+ * OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION
+ * WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
+ *
+ */
+
+#include "radeon_common.h"
+#include "r300_context.h"
+
+#include "main/image.h"
+#include "main/teximage.h"
+#include "main/texstate.h"
+#include "drivers/common/meta.h"
+
+#include "radeon_mipmap_tree.h"
+#include "r300_blit.h"
+#include <main/debug.h>
+
+// TODO:
+// need to pass correct pitch for small dst textures!
+static GLboolean
+do_copy_texsubimage(GLcontext *ctx,
+                    GLenum target, GLint level,
+                    struct radeon_tex_obj *tobj,
+                    radeon_texture_image *timg,
+                    GLint dstx, GLint dsty,
+                    GLint x, GLint y,
+                    GLsizei width, GLsizei height)
+{
+    struct r300_context *r300 = R300_CONTEXT(ctx);
+    struct radeon_renderbuffer *rrb;
+
+    if (_mesa_get_format_bits(timg->base.TexFormat, GL_DEPTH_BITS) > 0) {
+        rrb = radeon_get_depthbuffer(&r300->radeon);
+    } else {
+        rrb = radeon_get_colorbuffer(&r300->radeon);
+    }
+
+    if (!timg->mt) {
+        radeon_validate_texture_miptree(ctx, &tobj->base);
+    }
+
+    assert(rrb && rrb->bo);
+    assert(timg->mt->bo);
+    assert(timg->base.Width >= dstx + width);
+    assert(timg->base.Height >= dsty + height);
+
+    intptr_t src_offset = rrb->draw_offset;
+    intptr_t dst_offset = radeon_miptree_image_offset(timg->mt, _mesa_tex_target_to_face(target), level);
+
+    if (src_offset % 32 || dst_offset % 32) {
+        return GL_FALSE;
+    }
+
+    if (0) {
+        fprintf(stderr, "%s: copying to face %d, level %d\n",
+                __FUNCTION__, _mesa_tex_target_to_face(target), level);
+        fprintf(stderr, "to: x %d, y %d, offset %d\n", dstx, dsty, (uint32_t) dst_offset);
+        fprintf(stderr, "from (%dx%d) width %d, height %d, offset %d, pitch %d\n",
+                x, y, rrb->base.Width, rrb->base.Height, (uint32_t) src_offset, rrb->pitch/rrb->cpp);
+        fprintf(stderr, "src size %d, dst size %d\n", rrb->bo->size, timg->mt->bo->size);
+
+    }
+
+    /* blit from src buffer to texture */
+    return r300_blit(r300, rrb->bo, src_offset, rrb->base.Format, rrb->pitch/rrb->cpp,
+                     rrb->base.Width, rrb->base.Height, x, y,
+                     timg->mt->bo, dst_offset, timg->base.TexFormat,
+                     timg->base.Width, timg->base.Width, timg->base.Height,
+                     dstx, dsty, width, height, 1);
+}
+
+static void
+r300CopyTexImage2D(GLcontext *ctx, GLenum target, GLint level,
+                   GLenum internalFormat,
+                   GLint x, GLint y, GLsizei width, GLsizei height,
+                   GLint border)
+{
+    struct gl_texture_unit *texUnit = _mesa_get_current_tex_unit(ctx);
+    struct gl_texture_object *texObj =
+        _mesa_select_tex_object(ctx, texUnit, target);
+    struct gl_texture_image *texImage =
+        _mesa_select_tex_image(ctx, texObj, target, level);
+    int srcx, srcy, dstx, dsty;
+
+    if (border)
+        goto fail;
+
+    /* Setup or redefine the texture object, mipmap tree and texture
+     * image.  Don't populate yet.
+     */
+    ctx->Driver.TexImage2D(ctx, target, level, internalFormat,
+                           width, height, border,
+                           GL_RGBA, GL_UNSIGNED_BYTE, NULL,
+                           &ctx->DefaultPacking, texObj, texImage);
+
+    srcx = x;
+    srcy = y;
+    dstx = 0;
+    dsty = 0;
+    if (!_mesa_clip_copytexsubimage(ctx,
+                                    &dstx, &dsty,
+                                    &srcx, &srcy,
+                                    &width, &height)) {
+        return;
+    }
+
+    if (!do_copy_texsubimage(ctx, target, level,
+                             radeon_tex_obj(texObj), (radeon_texture_image *)texImage,
+                             0, 0, x, y, width, height)) {
+        goto fail;
+    }
+
+    return;
+
+fail:
+    _mesa_meta_CopyTexImage2D(ctx, target, level, internalFormat, x, y,
+                              width, height, border);
+}
+
+static void
+r300CopyTexSubImage2D(GLcontext *ctx, GLenum target, GLint level,
+                      GLint xoffset, GLint yoffset,
+                      GLint x, GLint y,
+                      GLsizei width, GLsizei height)
+{
+    struct gl_texture_unit *texUnit = _mesa_get_current_tex_unit(ctx);
+    struct gl_texture_object *texObj = _mesa_select_tex_object(ctx, texUnit, target);
+    struct gl_texture_image *texImage = _mesa_select_tex_image(ctx, texObj, target, level);
+
+    if (!do_copy_texsubimage(ctx, target, level,
+                             radeon_tex_obj(texObj), (radeon_texture_image *)texImage,
+                             xoffset, yoffset, x, y, width, height)) {
+
+       //DEBUG_FALLBACKS
+
+        _mesa_meta_CopyTexSubImage2D(ctx, target, level,
+                                     xoffset, yoffset, x, y, width, height);
+    }
+}
+
+
+void r300_init_texcopy_functions(struct dd_function_table *table)
+{
+    table->CopyTexImage2D = r300CopyTexImage2D;
+    table->CopyTexSubImage2D = r300CopyTexSubImage2D;
+}
\ No newline at end of file
index bbe8b1e95f9c998be2f5e0b7cb0231cbd7982483..78ff54574f623c5b142c0ccde812ec421304b46d 100644 (file)
@@ -46,19 +46,10 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 
 #include "r300_context.h"
 #include "r300_state.h"
-#include "r300_ioctl.h"
 #include "radeon_mipmap_tree.h"
 #include "r300_tex.h"
 #include "r300_reg.h"
 
-#define VALID_FORMAT(f) ( ((f) <= MESA_FORMAT_RGBA_DXT5                        \
-                          || ((f) >= MESA_FORMAT_RGBA_FLOAT32 &&       \
-                              (f) <= MESA_FORMAT_INTENSITY_FLOAT16))   \
-                         && tx_table[f].flag )
-
-#define _ASSIGN(entry, format)                         \
-       [ MESA_FORMAT_ ## entry ] = { format, 0, 1}
-
 /*
  * Note that the _REV formats are the same as the non-REV formats.  This is
  * because the REV and non-REV formats are identical as a byte string, but
@@ -68,67 +59,119 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
  * identically.  -- paulus
  */
 
-static const struct tx_table {
-       GLuint format, filter, flag;
-} tx_table[] = {
-       /* *INDENT-OFF* */
+int32_t r300TranslateTexFormat(gl_format mesaFormat)
+{
+       switch (mesaFormat)
+       {
 #ifdef MESA_LITTLE_ENDIAN
-       _ASSIGN(RGBA8888, R300_EASY_TX_FORMAT(Y, Z, W, X, W8Z8Y8X8)),
-       _ASSIGN(RGBA8888_REV, R300_EASY_TX_FORMAT(Z, Y, X, W, W8Z8Y8X8)),
-       _ASSIGN(ARGB8888, R300_EASY_TX_FORMAT(X, Y, Z, W, W8Z8Y8X8)),
-       _ASSIGN(ARGB8888_REV, R300_EASY_TX_FORMAT(W, Z, Y, X, W8Z8Y8X8)),
+               case MESA_FORMAT_RGBA8888:
+                       return R300_EASY_TX_FORMAT(Y, Z, W, X, W8Z8Y8X8);
+               case MESA_FORMAT_RGBA8888_REV:
+                       return R300_EASY_TX_FORMAT(Z, Y, X, W, W8Z8Y8X8);
+               case MESA_FORMAT_ARGB8888:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, W, W8Z8Y8X8);
+               case MESA_FORMAT_ARGB8888_REV:
+                       return R300_EASY_TX_FORMAT(W, Z, Y, X, W8Z8Y8X8);
 #else
-       _ASSIGN(RGBA8888, R300_EASY_TX_FORMAT(Z, Y, X, W, W8Z8Y8X8)),
-       _ASSIGN(RGBA8888_REV, R300_EASY_TX_FORMAT(Y, Z, W, X, W8Z8Y8X8)),
-       _ASSIGN(ARGB8888, R300_EASY_TX_FORMAT(W, Z, Y, X, W8Z8Y8X8)),
-       _ASSIGN(ARGB8888_REV, R300_EASY_TX_FORMAT(X, Y, Z, W, W8Z8Y8X8)),
+               case MESA_FORMAT_RGBA8888:
+                       return R300_EASY_TX_FORMAT(Z, Y, X, W, W8Z8Y8X8);
+               case MESA_FORMAT_RGBA8888_REV:
+                       return R300_EASY_TX_FORMAT(Y, Z, W, X, W8Z8Y8X8);
+               case MESA_FORMAT_ARGB8888:
+                       return R300_EASY_TX_FORMAT(W, Z, Y, X, W8Z8Y8X8);
+               case MESA_FORMAT_ARGB8888_REV:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, W, W8Z8Y8X8);
 #endif
-       _ASSIGN(XRGB8888, R300_EASY_TX_FORMAT(X, Y, Z, ONE, W8Z8Y8X8)),
-       _ASSIGN(RGB888, R300_EASY_TX_FORMAT(X, Y, Z, ONE, W8Z8Y8X8)),
-       _ASSIGN(RGB565, R300_EASY_TX_FORMAT(X, Y, Z, ONE, Z5Y6X5)),
-       _ASSIGN(RGB565_REV, R300_EASY_TX_FORMAT(X, Y, Z, ONE, Z5Y6X5)),
-       _ASSIGN(ARGB4444, R300_EASY_TX_FORMAT(X, Y, Z, W, W4Z4Y4X4)),
-       _ASSIGN(ARGB4444_REV, R300_EASY_TX_FORMAT(X, Y, Z, W, W4Z4Y4X4)),
-       _ASSIGN(ARGB1555, R300_EASY_TX_FORMAT(X, Y, Z, W, W1Z5Y5X5)),
-       _ASSIGN(ARGB1555_REV, R300_EASY_TX_FORMAT(X, Y, Z, W, W1Z5Y5X5)),
-       _ASSIGN(AL88, R300_EASY_TX_FORMAT(X, X, X, Y, Y8X8)),
-       _ASSIGN(AL88_REV, R300_EASY_TX_FORMAT(X, X, X, Y, Y8X8)),
-       _ASSIGN(RGB332, R300_EASY_TX_FORMAT(X, Y, Z, ONE, Z3Y3X2)),
-       _ASSIGN(A8, R300_EASY_TX_FORMAT(ZERO, ZERO, ZERO, X, X8)),
-       _ASSIGN(L8, R300_EASY_TX_FORMAT(X, X, X, ONE, X8)),
-       _ASSIGN(I8, R300_EASY_TX_FORMAT(X, X, X, X, X8)),
-       _ASSIGN(CI8, R300_EASY_TX_FORMAT(X, X, X, X, X8)),
-       _ASSIGN(YCBCR, R300_EASY_TX_FORMAT(X, Y, Z, ONE, G8R8_G8B8) | R300_TX_FORMAT_YUV_MODE),
-       _ASSIGN(YCBCR_REV, R300_EASY_TX_FORMAT(X, Y, Z, ONE, G8R8_G8B8) | R300_TX_FORMAT_YUV_MODE),
-       _ASSIGN(RGB_DXT1, R300_EASY_TX_FORMAT(X, Y, Z, ONE, DXT1)),
-       _ASSIGN(RGBA_DXT1, R300_EASY_TX_FORMAT(X, Y, Z, W, DXT1)),
-       _ASSIGN(RGBA_DXT3, R300_EASY_TX_FORMAT(X, Y, Z, W, DXT3)),
-       _ASSIGN(RGBA_DXT5, R300_EASY_TX_FORMAT(Y, Z, W, X, DXT5)),
-       _ASSIGN(RGBA_FLOAT32, R300_EASY_TX_FORMAT(Z, Y, X, W, FL_R32G32B32A32)),
-       _ASSIGN(RGBA_FLOAT16, R300_EASY_TX_FORMAT(Z, Y, X, W, FL_R16G16B16A16)),
-       _ASSIGN(RGB_FLOAT32, 0xffffffff),
-       _ASSIGN(RGB_FLOAT16, 0xffffffff),
-       _ASSIGN(ALPHA_FLOAT32, R300_EASY_TX_FORMAT(ZERO, ZERO, ZERO, X, FL_I32)),
-       _ASSIGN(ALPHA_FLOAT16, R300_EASY_TX_FORMAT(ZERO, ZERO, ZERO, X, FL_I16)),
-       _ASSIGN(LUMINANCE_FLOAT32, R300_EASY_TX_FORMAT(X, X, X, ONE, FL_I32)),
-       _ASSIGN(LUMINANCE_FLOAT16, R300_EASY_TX_FORMAT(X, X, X, ONE, FL_I16)),
-       _ASSIGN(LUMINANCE_ALPHA_FLOAT32, R300_EASY_TX_FORMAT(X, X, X, Y, FL_I32A32)),
-       _ASSIGN(LUMINANCE_ALPHA_FLOAT16, R300_EASY_TX_FORMAT(X, X, X, Y, FL_I16A16)),
-       _ASSIGN(INTENSITY_FLOAT32, R300_EASY_TX_FORMAT(X, X, X, X, FL_I32)),
-       _ASSIGN(INTENSITY_FLOAT16, R300_EASY_TX_FORMAT(X, X, X, X, FL_I16)),
-       _ASSIGN(Z16, R300_EASY_TX_FORMAT(X, X, X, X, X16)),
-       _ASSIGN(Z24_S8, R300_EASY_TX_FORMAT(X, X, X, X, X24_Y8)),
-       _ASSIGN(S8_Z24, R300_EASY_TX_FORMAT(Y, Y, Y, Y, X24_Y8)),
-       _ASSIGN(Z32, R300_EASY_TX_FORMAT(X, X, X, X, X32)),
-       /* EXT_texture_sRGB */
-       _ASSIGN(SRGBA8, R300_EASY_TX_FORMAT(Y, Z, W, X, W8Z8Y8X8) | R300_TX_FORMAT_GAMMA),
-       _ASSIGN(SLA8, R300_EASY_TX_FORMAT(X, X, X, Y, Y8X8) | R300_TX_FORMAT_GAMMA),
-       _ASSIGN(SL8, R300_EASY_TX_FORMAT(X, X, X, ONE, X8) | R300_TX_FORMAT_GAMMA),
-       /* *INDENT-ON* */
+               case MESA_FORMAT_XRGB8888:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, ONE, W8Z8Y8X8);
+               case MESA_FORMAT_RGB888:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, ONE, W8Z8Y8X8);
+               case MESA_FORMAT_RGB565:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, ONE, Z5Y6X5);
+               case MESA_FORMAT_RGB565_REV:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, ONE, Z5Y6X5);
+               case MESA_FORMAT_ARGB4444:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, W, W4Z4Y4X4);
+               case MESA_FORMAT_ARGB4444_REV:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, W, W4Z4Y4X4);
+               case MESA_FORMAT_ARGB1555:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, W, W1Z5Y5X5);
+               case MESA_FORMAT_ARGB1555_REV:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, W, W1Z5Y5X5);
+               case MESA_FORMAT_AL88:
+                       return R300_EASY_TX_FORMAT(X, X, X, Y, Y8X8);
+               case MESA_FORMAT_AL88_REV:
+                       return R300_EASY_TX_FORMAT(X, X, X, Y, Y8X8);
+               case MESA_FORMAT_RGB332:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, ONE, Z3Y3X2);
+               case MESA_FORMAT_A8:
+                       return R300_EASY_TX_FORMAT(ZERO, ZERO, ZERO, X, X8);
+               case MESA_FORMAT_L8:
+                       return R300_EASY_TX_FORMAT(X, X, X, ONE, X8);
+               case MESA_FORMAT_I8:
+                       return R300_EASY_TX_FORMAT(X, X, X, X, X8);
+               case MESA_FORMAT_CI8:
+                       return R300_EASY_TX_FORMAT(X, X, X, X, X8);
+               case MESA_FORMAT_YCBCR:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, ONE, G8R8_G8B8) | R300_TX_FORMAT_YUV_MODE;
+               case MESA_FORMAT_YCBCR_REV:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, ONE, G8R8_G8B8) | R300_TX_FORMAT_YUV_MODE;
+               case MESA_FORMAT_RGB_DXT1:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, ONE, DXT1);
+               case MESA_FORMAT_RGBA_DXT1:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, W, DXT1);
+               case MESA_FORMAT_RGBA_DXT3:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, W, DXT3);
+               case MESA_FORMAT_RGBA_DXT5:
+                       return R300_EASY_TX_FORMAT(Y, Z, W, X, DXT5);
+               case MESA_FORMAT_RGBA_FLOAT32:
+                       return R300_EASY_TX_FORMAT(Z, Y, X, W, FL_R32G32B32A32);
+               case MESA_FORMAT_RGBA_FLOAT16:
+                       return R300_EASY_TX_FORMAT(Z, Y, X, W, FL_R16G16B16A16);
+               case MESA_FORMAT_ALPHA_FLOAT32:
+                       return R300_EASY_TX_FORMAT(ZERO, ZERO, ZERO, X, FL_I32);
+               case MESA_FORMAT_ALPHA_FLOAT16:
+                       return R300_EASY_TX_FORMAT(ZERO, ZERO, ZERO, X, FL_I16);
+               case MESA_FORMAT_LUMINANCE_FLOAT32:
+                       return R300_EASY_TX_FORMAT(X, X, X, ONE, FL_I32);
+               case MESA_FORMAT_LUMINANCE_FLOAT16:
+                       return R300_EASY_TX_FORMAT(X, X, X, ONE, FL_I16);
+               case MESA_FORMAT_LUMINANCE_ALPHA_FLOAT32:
+                       return R300_EASY_TX_FORMAT(X, X, X, Y, FL_I32A32);
+               case MESA_FORMAT_LUMINANCE_ALPHA_FLOAT16:
+                       return R300_EASY_TX_FORMAT(X, X, X, Y, FL_I16A16);
+               case MESA_FORMAT_INTENSITY_FLOAT32:
+                       return R300_EASY_TX_FORMAT(X, X, X, X, FL_I32);
+               case MESA_FORMAT_INTENSITY_FLOAT16:
+                       return R300_EASY_TX_FORMAT(X, X, X, X, FL_I16);
+               case MESA_FORMAT_Z16:
+                       return R300_EASY_TX_FORMAT(X, X, X, X, X16);
+               case MESA_FORMAT_Z24_S8:
+                       return R300_EASY_TX_FORMAT(X, X, X, X, X24_Y8);
+               case MESA_FORMAT_S8_Z24:
+                       return R300_EASY_TX_FORMAT(Y, Y, Y, Y, X24_Y8);
+               case MESA_FORMAT_Z32:
+                       return R300_EASY_TX_FORMAT(X, X, X, X, X32);
+               /* EXT_texture_sRGB */
+               case MESA_FORMAT_SRGBA8:
+                       return R300_EASY_TX_FORMAT(Y, Z, W, X, W8Z8Y8X8) | R300_TX_FORMAT_GAMMA;
+               case MESA_FORMAT_SLA8:
+                       return R300_EASY_TX_FORMAT(X, X, X, Y, Y8X8) | R300_TX_FORMAT_GAMMA;
+               case MESA_FORMAT_SL8:
+                       return R300_EASY_TX_FORMAT(X, X, X, ONE, X8) | R300_TX_FORMAT_GAMMA;
+               case MESA_FORMAT_SRGB_DXT1:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, ONE, DXT1) | R300_TX_FORMAT_GAMMA;
+               case MESA_FORMAT_SRGBA_DXT1:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, W, DXT1) | R300_TX_FORMAT_GAMMA;
+               case MESA_FORMAT_SRGBA_DXT3:
+                       return R300_EASY_TX_FORMAT(X, Y, Z, W, DXT3) | R300_TX_FORMAT_GAMMA;
+               case MESA_FORMAT_SRGBA_DXT5:
+                       return R300_EASY_TX_FORMAT(Y, Z, W, X, DXT5) | R300_TX_FORMAT_GAMMA;
+               default:
+                       return -1;
+       }
 };
 
-#undef _ASSIGN
-
 void r300SetDepthTexMode(struct gl_texture_object *tObj)
 {
        static const GLuint formats[3][3] = {
@@ -205,19 +248,18 @@ static void setup_hardware_state(r300ContextPtr rmesa, radeonTexObj *t)
        const struct gl_texture_image *firstImage;
        firstImage = t->base.Image[0][t->minLod];
 
-       if (!t->image_override
-           && VALID_FORMAT(firstImage->TexFormat)) {
+       if (!t->image_override) {
                if (firstImage->_BaseFormat == GL_DEPTH_COMPONENT) {
                        r300SetDepthTexMode(&t->base);
                } else {
-                       t->pp_txformat = tx_table[firstImage->TexFormat].format;
+                       int32_t txformat = r300TranslateTexFormat(firstImage->TexFormat);
+                       if (txformat < 0) {
+                               _mesa_problem(rmesa->radeon.glCtx, "%s: Invalid format %s",
+                                                         __FUNCTION__, _mesa_get_format_name(firstImage->TexFormat));
+                               exit(1);
+                       }
+                       t->pp_txformat = (uint32_t) txformat;
                }
-
-               t->pp_txfilter |= tx_table[firstImage->TexFormat].filter;
-       } else if (!t->image_override) {
-               _mesa_problem(NULL, "unexpected texture format in %s",
-                             __FUNCTION__);
-               return;
        }
 
        if (t->image_override && t->bo)
@@ -357,18 +399,15 @@ void r300SetTexOffset(__DRIcontext * pDRICtx, GLint texname,
        switch (depth) {
        case 32:
                t->pp_txformat = R300_EASY_TX_FORMAT(X, Y, Z, W, W8Z8Y8X8);
-               t->pp_txfilter |= tx_table[2].filter;
                pitch_val /= 4;
                break;
        case 24:
        default:
                t->pp_txformat = R300_EASY_TX_FORMAT(X, Y, Z, ONE, W8Z8Y8X8);
-               t->pp_txfilter |= tx_table[4].filter;
                pitch_val /= 4;
                break;
        case 16:
                t->pp_txformat = R300_EASY_TX_FORMAT(X, Y, Z, ONE, Z5Y6X5);
-               t->pp_txfilter |= tx_table[5].filter;
                pitch_val /= 2;
                break;
        }
@@ -409,18 +448,7 @@ void r300SetTexBuffer2(__DRIcontext *pDRICtx, GLint target, GLint glx_texture_fo
            return;
        }
 
-       radeon_update_renderbuffers(pDRICtx, dPriv);
-       /* back & depth buffer are useless free them right away */
-       rb = (void*)rfb->base.Attachment[BUFFER_DEPTH].Renderbuffer;
-       if (rb && rb->bo) {
-               radeon_bo_unref(rb->bo);
-        rb->bo = NULL;
-       }
-       rb = (void*)rfb->base.Attachment[BUFFER_BACK_LEFT].Renderbuffer;
-       if (rb && rb->bo) {
-               radeon_bo_unref(rb->bo);
-               rb->bo = NULL;
-       }
+       radeon_update_renderbuffers(pDRICtx, dPriv, GL_TRUE);
        rb = rfb->color_rb[0];
        if (rb->bo == NULL) {
                /* Failed to BO for the buffer */
@@ -458,18 +486,15 @@ void r300SetTexBuffer2(__DRIcontext *pDRICtx, GLint target, GLint glx_texture_fo
                        t->pp_txformat = R300_EASY_TX_FORMAT(X, Y, Z, ONE, W8Z8Y8X8);
                else
                        t->pp_txformat = R300_EASY_TX_FORMAT(X, Y, Z, W, W8Z8Y8X8);
-               t->pp_txfilter |= tx_table[2].filter;
                pitch_val /= 4;
                break;
        case 3:
        default:
                t->pp_txformat = R300_EASY_TX_FORMAT(X, Y, Z, ONE, W8Z8Y8X8);
-               t->pp_txfilter |= tx_table[4].filter;
                pitch_val /= 4;
                break;
        case 2:
                t->pp_txformat = R300_EASY_TX_FORMAT(X, Y, Z, ONE, Z5Y6X5);
-               t->pp_txfilter |= tx_table[5].filter;
                pitch_val /= 2;
                break;
        }
index c2f96af2c1215de84e1a635c4e7793df2c5f254b..aa98a049aa4aad5d64ab00c5997263a3200c6f25 100644 (file)
@@ -365,7 +365,7 @@ static void r300EmitVertexProgram(r300ContextPtr r300, int dest, struct r300_ver
                        break;
                default:
                        fprintf(stderr, "%s:%s don't know how to handle dest %04x\n", __FILE__, __FUNCTION__, dest);
-                       _mesa_exit(-1);
+                       exit(-1);
        }
 }
 
index dbd233729c01113d579151970227f9bdf3ba5f37..45bbc3c071e75b5ee770ea6e6bdce038e484bf02 100644 (file)
@@ -74,6 +74,8 @@ WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
 #include "utils.h"
 #include "xmlpool.h"           /* for symbolic values of enum-type options */
 
+#define R600_ENABLE_GLSL_TEST 1
+
 #define need_GL_VERSION_2_0
 #define need_GL_ARB_occlusion_query
 #define need_GL_ARB_point_parameters
@@ -109,6 +111,7 @@ static const struct dri_extension card_extensions[] = {
   {"GL_ARB_texture_env_crossbar",      NULL},
   {"GL_ARB_texture_env_dot3",          NULL},
   {"GL_ARB_texture_mirrored_repeat",   NULL},
+  {"GL_ARB_texture_non_power_of_two",   NULL},
   {"GL_ARB_vertex_program",            GL_ARB_vertex_program_functions},
   {"GL_EXT_blend_equation_separate",   GL_EXT_blend_equation_separate_functions},
   {"GL_EXT_blend_func_separate",       GL_EXT_blend_func_separate_functions},
@@ -155,7 +158,11 @@ static const struct dri_extension mm_extensions[] = {
  * functions added by GL_ATI_separate_stencil.
  */
 static const struct dri_extension gl_20_extension[] = {
+#ifdef R600_ENABLE_GLSL_TEST
+    {"GL_ARB_shading_language_100",                    GL_VERSION_2_0_functions },
+#else
   {"GL_VERSION_2_0",                   GL_VERSION_2_0_functions },
+#endif /* R600_ENABLE_GLSL_TEST */
 };
 
 static const struct tnl_pipeline_stage *r600_pipeline[] = {
@@ -308,6 +315,14 @@ static void r600InitGLExtensions(GLcontext *ctx)
        if (r600->radeon.radeonScreen->kernel_mm)
          driInitExtensions(ctx, mm_extensions, GL_FALSE);
 
+#ifdef R600_ENABLE_GLSL_TEST
+    driInitExtensions(ctx, gl_20_extension, GL_TRUE);
+    _mesa_enable_2_0_extensions(ctx);
+    
+    /* glsl compiler has problem if this is not GL_TRUE */
+    ctx->Shader.EmitCondCodes = GL_TRUE;
+#endif /* R600_ENABLE_GLSL_TEST */
+
        if (driQueryOptionb
            (&r600->radeon.optionCache, "disable_stencil_two_side"))
                _mesa_disable_extension(ctx, "GL_EXT_stencil_two_side");
index 4ec315b78c7439a8f053bf96db3a3aeae4d1d19c..2a4a6e6ee146bd599348b6bf9f0618027ad8a59e 100644 (file)
@@ -917,18 +917,7 @@ void r600SetTexBuffer2(__DRIcontext *pDRICtx, GLint target, GLint glx_texture_fo
            return;
        }
 
-       radeon_update_renderbuffers(pDRICtx, dPriv);
-       /* back & depth buffer are useless free them right away */
-       rb = (void*)rfb->base.Attachment[BUFFER_DEPTH].Renderbuffer;
-       if (rb && rb->bo) {
-               radeon_bo_unref(rb->bo);
-        rb->bo = NULL;
-       }
-       rb = (void*)rfb->base.Attachment[BUFFER_BACK_LEFT].Renderbuffer;
-       if (rb && rb->bo) {
-               radeon_bo_unref(rb->bo);
-               rb->bo = NULL;
-       }
+       radeon_update_renderbuffers(pDRICtx, dPriv, GL_TRUE);
        rb = rfb->color_rb[0];
        if (rb->bo == NULL) {
                /* Failed to BO for the buffer */
index 67e0ee774631c681df7701bec1950c8ae6a840c6..e464c6191c35a60c35ec97e74619224954146e7f 100644 (file)
 
 #include "main/mtypes.h"
 #include "main/imports.h"
+#include "shader/prog_parameter.h"
 
 #include "radeon_debug.h"
 #include "r600_context.h"
 
 #include "r700_assembler.h"
 
+#define USE_CF_FOR_CONTINUE_BREAK 1
+#define USE_CF_FOR_POP_AFTER      1
+
+struct prog_instruction noise1_insts[12] = { 
+    {OPCODE_BGNSUB , {{13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {13, 0, 15, 0, 8, 1672, 0}, 0, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}, 
+    {OPCODE_MOV , {{0, 0, 0, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {0, 0, 2, 0, 8, 1672, 0}, 0, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}, 
+    {OPCODE_MOV , {{8, 0, 0, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {0, 0, 4, 0, 8, 1672, 0}, 0, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}, 
+    {OPCODE_MOV , {{8, 0, 585, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {0, 0, 8, 0, 8, 1672, 0}, 0, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}, 
+    {OPCODE_SGT , {{0, 0, 585, 0, 0, 0}, {8, 0, 1170, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {0, 1, 1, 0, 8, 1672, 0}, 1, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}, 
+    {OPCODE_IF , {{13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {13, 0, 15, 0, 7, 0, 0}, 0, 0, 0, 1, 0, 0, 0, 15, 0, 0, 0}, 
+    {OPCODE_MOV , {{0, 0, 1755, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {0, 0, 1, 0, 8, 1672, 0}, 0, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}, 
+    {OPCODE_RET , {{13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {13, 0, 15, 0, 8, 1672, 0}, 0, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}, 
+    {OPCODE_ENDIF , {{13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {13, 0, 15, 0, 8, 1672, 0}, 0, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}, 
+    {OPCODE_MOV , {{0, 0, 1170, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {0, 0, 1, 0, 8, 1672, 0}, 0, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}, 
+    {OPCODE_RET , {{13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {13, 0, 15, 0, 8, 1672, 0}, 0, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}, 
+    {OPCODE_ENDSUB , {{13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}, {13, 0, 1672, 0, 0, 0}}, {13, 0, 15, 0, 8, 1672, 0}, 0, 0, 0, 1, 0, 0, 0, -1, 0, 0, 0}
+};
+float noise1_const[2][4] = {
+    {0.300000f, 0.900000f, 0.500000f, 0.300000f}
+};
+
+COMPILED_SUB noise1_presub = {
+    &(noise1_insts[0]),
+    12, 
+    2, 
+    1, 
+    0, 
+    &(noise1_const[0]), 
+    SWIZZLE_X, 
+    SWIZZLE_X, 
+    SWIZZLE_X, 
+    SWIZZLE_X,
+    {0,0,0},
+    0 
+};
+
 BITS addrmode_PVSDST(PVSDST * pPVSDST)
 {
        return pPVSDST->addrmode0 | ((BITS)pPVSDST->addrmode1 << 1);
@@ -327,22 +364,27 @@ GLuint GetSurfaceFormat(GLenum eType, GLuint nChannels, GLuint * pClient_size)
     return(format);
 }
 
-unsigned int r700GetNumOperands(r700_AssemblerBase* pAsm)
+unsigned int r700GetNumOperands(GLuint opcode, GLuint nIsOp3) 
 {
-    if(pAsm->D.dst.op3)
+    if(nIsOp3 > 0)
     {
         return 3;
     }
 
-    switch (pAsm->D.dst.opcode)
+    switch (opcode)
     {
     case SQ_OP2_INST_ADD:
+    case SQ_OP2_INST_KILLE:
     case SQ_OP2_INST_KILLGT:
+    case SQ_OP2_INST_KILLGE:
+    case SQ_OP2_INST_KILLNE:
     case SQ_OP2_INST_MUL: 
     case SQ_OP2_INST_MAX:
     case SQ_OP2_INST_MIN:
     //case SQ_OP2_INST_MAX_DX10:
     //case SQ_OP2_INST_MIN_DX10:
+    case SQ_OP2_INST_SETE: 
+    case SQ_OP2_INST_SETNE:
     case SQ_OP2_INST_SETGT:
     case SQ_OP2_INST_SETGE:
     case SQ_OP2_INST_PRED_SETE:
@@ -358,6 +400,7 @@ unsigned int r700GetNumOperands(r700_AssemblerBase* pAsm)
     case SQ_OP2_INST_MOVA_FLOOR:
     case SQ_OP2_INST_FRACT:
     case SQ_OP2_INST_FLOOR:
+    case SQ_OP2_INST_TRUNC:
     case SQ_OP2_INST_EXP_IEEE:
     case SQ_OP2_INST_LOG_CLAMPED:
     case SQ_OP2_INST_LOG_IEEE:
@@ -369,7 +412,7 @@ unsigned int r700GetNumOperands(r700_AssemblerBase* pAsm)
         return 1;
         
     default: radeon_error(
-                   "Need instruction operand number for %x.\n", pAsm->D.dst.opcode);
+                   "Need instruction operand number for %x.\n", opcode); 
     };
 
     return 3;
@@ -383,103 +426,128 @@ int Init_r700_AssemblerBase(SHADER_PIPE_TYPE spt, r700_AssemblerBase* pAsm, R700
     pAsm->pR700Shader = pShader;
     pAsm->currentShaderType = spt;
 
-       pAsm->cf_last_export_ptr   = NULL;
+    pAsm->cf_last_export_ptr   = NULL;
+
+    pAsm->cf_current_export_clause_ptr = NULL;
+    pAsm->cf_current_alu_clause_ptr    = NULL;
+    pAsm->cf_current_tex_clause_ptr    = NULL;
+    pAsm->cf_current_vtx_clause_ptr    = NULL;
+    pAsm->cf_current_cf_clause_ptr     = NULL;
 
-       pAsm->cf_current_export_clause_ptr = NULL;
-       pAsm->cf_current_alu_clause_ptr    = NULL;
-       pAsm->cf_current_tex_clause_ptr    = NULL;
-       pAsm->cf_current_vtx_clause_ptr    = NULL;
-       pAsm->cf_current_cf_clause_ptr     = NULL;
+    // No clause has been created yet
+    pAsm->cf_current_clause_type = CF_EMPTY_CLAUSE;
 
-       // No clause has been created yet
-       pAsm->cf_current_clause_type = CF_EMPTY_CLAUSE;
+    pAsm->number_of_colorandz_exports = 0;
+    pAsm->number_of_exports           = 0;
+    pAsm->number_of_export_opcodes    = 0;
 
-       pAsm->number_of_colorandz_exports = 0;
-       pAsm->number_of_exports           = 0;
-       pAsm->number_of_export_opcodes    = 0;
+    pAsm->alu_x_opcode = 0;
 
+    pAsm->D2.bits = 0;
 
-       pAsm->D.bits = 0;
-       pAsm->S[0].bits = 0;
-       pAsm->S[1].bits = 0;
-       pAsm->S[2].bits = 0;
+    pAsm->D.bits = 0;
+    pAsm->S[0].bits = 0;
+    pAsm->S[1].bits = 0;
+    pAsm->S[2].bits = 0;
 
-       pAsm->uLastPosUpdate = 0; 
+    pAsm->uLastPosUpdate = 0; 
        
-       *(BITS *) &pAsm->fp_stOutFmt0 = 0;
+    *(BITS *) &pAsm->fp_stOutFmt0 = 0;
+
+    pAsm->uIIns = 0;
+    pAsm->uOIns = 0;
+    pAsm->number_used_registers = 0;
+    pAsm->uUsedConsts = 256; 
+
+
+    // Fragment programs
+    pAsm->uBoolConsts = 0;
+    pAsm->uIntConsts = 0;
+    pAsm->uInsts = 0;
+    pAsm->uConsts = 0;
 
-       pAsm->uIIns = 0;
-       pAsm->uOIns = 0;
-       pAsm->number_used_registers = 0;
-       pAsm->uUsedConsts = 256; 
+    pAsm->FCSP = 0;
+    pAsm->fc_stack[0].type = FC_NONE;
 
+    pAsm->aArgSubst[0] =
+    pAsm->aArgSubst[1] =
+    pAsm->aArgSubst[2] =
+    pAsm->aArgSubst[3] = (-1);
+
+    pAsm->uOutputs = 0;
 
-       // Fragment programs
-       pAsm->uBoolConsts = 0;
-       pAsm->uIntConsts = 0;
-       pAsm->uInsts = 0;
-       pAsm->uConsts = 0;
+    for (i=0; i<NUMBER_OF_OUTPUT_COLORS; i++) 
+    {
+        pAsm->color_export_register_number[i] = (-1);
+    }
 
-       pAsm->FCSP = 0;
-       pAsm->fc_stack[0].type = FC_NONE;
 
-       pAsm->branch_depth     = 0;
-       pAsm->max_branch_depth = 0;
+    pAsm->depth_export_register_number = (-1);
+    pAsm->stencil_export_register_number = (-1);
+    pAsm->coverage_to_mask_export_register_number = (-1);
+    pAsm->mask_export_register_number = (-1);
 
-       pAsm->aArgSubst[0] =
-       pAsm->aArgSubst[1] =
-       pAsm->aArgSubst[2] =
-       pAsm->aArgSubst[3] = (-1);
+    pAsm->starting_export_register_number = 0;
+    pAsm->starting_vfetch_register_number = 0;
+    pAsm->starting_temp_register_number   = 0;
+    pAsm->uFirstHelpReg = 0;
 
-       pAsm->uOutputs = 0;
+    pAsm->input_position_is_used = GL_FALSE;
+    pAsm->input_normal_is_used   = GL_FALSE;
 
-       for (i=0; i<NUMBER_OF_OUTPUT_COLORS; i++) 
-       {
-               pAsm->color_export_register_number[i] = (-1);
-       }
+    for (i=0; i<NUMBER_OF_INPUT_COLORS; i++) 
+    {
+        pAsm->input_color_is_used[ i ] = GL_FALSE;
+    }
 
+    for (i=0; i<NUMBER_OF_TEXTURE_UNITS; i++) 
+    {
+        pAsm->input_texture_unit_is_used[ i ] = GL_FALSE;
+    }
 
-       pAsm->depth_export_register_number = (-1);
-       pAsm->stencil_export_register_number = (-1);
-       pAsm->coverage_to_mask_export_register_number = (-1);
-       pAsm->mask_export_register_number = (-1);
+    for (i=0; i<VERT_ATTRIB_MAX; i++) 
+    {
+        pAsm->vfetch_instruction_ptr_array[ i ] = NULL;
+    }
 
-       pAsm->starting_export_register_number = 0;
-       pAsm->starting_vfetch_register_number = 0;
-       pAsm->starting_temp_register_number   = 0;
-       pAsm->uFirstHelpReg = 0;
+    pAsm->number_of_inputs = 0;
 
+    pAsm->is_tex = GL_FALSE;
+    pAsm->need_tex_barrier = GL_FALSE;
 
-       pAsm->input_position_is_used = GL_FALSE;
-       pAsm->input_normal_is_used   = GL_FALSE;
+    pAsm->subs              = NULL;
+    pAsm->unSubArraySize    = 0;
+    pAsm->unSubArrayPointer = 0;
+    pAsm->callers              = NULL;
+    pAsm->unCallerArraySize    = 0;
+    pAsm->unCallerArrayPointer = 0;
 
+    pAsm->CALLSP = 0;
+    pAsm->CALLSTACK[0].FCSP_BeforeEntry = 0;
+    pAsm->CALLSTACK[0].plstCFInstructions_local
+          = &(pAsm->pR700Shader->lstCFInstructions);
 
-       for (i=0; i<NUMBER_OF_INPUT_COLORS; i++) 
-       {
-               pAsm->input_color_is_used[ i ] = GL_FALSE;
-       }
+    pAsm->CALLSTACK[0].max = 0;
+    pAsm->CALLSTACK[0].current = 0;
 
-       for (i=0; i<NUMBER_OF_TEXTURE_UNITS; i++) 
-       {
-               pAsm->input_texture_unit_is_used[ i ] = GL_FALSE;
-       }
+    SetActiveCFlist(pAsm->pR700Shader, pAsm->CALLSTACK[0].plstCFInstructions_local);
 
-       for (i=0; i<VERT_ATTRIB_MAX; i++) 
-       {
-               pAsm->vfetch_instruction_ptr_array[ i ] = NULL;
-       }
+    pAsm->unCFflags = 0;
 
-       pAsm->number_of_inputs = 0;
+    pAsm->presubs           = NULL;
+    pAsm->unPresubArraySize = 0;
+    pAsm->unNumPresub       = 0;
+    pAsm->unCurNumILInsts   = 0;
 
-       pAsm->is_tex = GL_FALSE;
-       pAsm->need_tex_barrier = GL_FALSE;
+    pAsm->unVetTexBits      = 0;
 
-       return 0;
+    return 0;
 }
 
 GLboolean IsTex(gl_inst_opcode Opcode)
 {
-    if( (OPCODE_TEX==Opcode) || (OPCODE_TXP==Opcode) || (OPCODE_TXB==Opcode) )
+    if( (OPCODE_TEX==Opcode) || (OPCODE_TXP==Opcode) || (OPCODE_TXB==Opcode) ||
+        (OPCODE_DDX==Opcode) || (OPCODE_DDY==Opcode) )
     {
         return GL_TRUE;
     }
@@ -592,6 +660,31 @@ int check_current_clause(r700_AssemblerBase* pAsm,
     return GL_TRUE;
 }
 
+GLboolean add_cf_instruction(r700_AssemblerBase* pAsm)
+{
+    if(GL_FALSE == check_current_clause(pAsm, CF_OTHER_CLAUSE))
+    {
+        return GL_FALSE;
+    }
+
+    pAsm->cf_current_cf_clause_ptr = 
+      (R700ControlFlowGenericClause*) CALLOC_STRUCT(R700ControlFlowGenericClause);
+
+    if (pAsm->cf_current_cf_clause_ptr != NULL) 
+       {
+               Init_R700ControlFlowGenericClause(pAsm->cf_current_cf_clause_ptr);
+               AddCFInstruction( pAsm->pR700Shader, 
+                          (R700ControlFlowInstruction *)pAsm->cf_current_cf_clause_ptr );
+       }
+       else 
+       {
+        radeon_error("Could not allocate a new VFetch CF instruction.\n");
+               return GL_FALSE;
+       }
+
+    return GL_TRUE;
+}
+
 GLboolean add_vfetch_instruction(r700_AssemblerBase*     pAsm,
                                                                 R700VertexInstruction*  vertex_instruction_ptr)
 {
@@ -987,7 +1080,8 @@ GLboolean checkop2(r700_AssemblerBase* pAsm)
 
     checkop_init(pAsm);
 
-    if( (pILInst->SrcReg[0].File == PROGRAM_CONSTANT)    ||
+    if( (pILInst->SrcReg[0].File == PROGRAM_UNIFORM)     || 
+        (pILInst->SrcReg[0].File == PROGRAM_CONSTANT)    ||
         (pILInst->SrcReg[0].File == PROGRAM_LOCAL_PARAM) ||
         (pILInst->SrcReg[0].File == PROGRAM_ENV_PARAM)   ||
         (pILInst->SrcReg[0].File == PROGRAM_STATE_VAR) )
@@ -998,7 +1092,8 @@ GLboolean checkop2(r700_AssemblerBase* pAsm)
     {
         bSrcConst[0] = GL_FALSE;
     }
-    if( (pILInst->SrcReg[1].File == PROGRAM_CONSTANT)    ||
+    if( (pILInst->SrcReg[1].File == PROGRAM_UNIFORM)     || 
+        (pILInst->SrcReg[1].File == PROGRAM_CONSTANT)    ||
         (pILInst->SrcReg[1].File == PROGRAM_LOCAL_PARAM) ||
         (pILInst->SrcReg[1].File == PROGRAM_ENV_PARAM)   ||
         (pILInst->SrcReg[1].File == PROGRAM_STATE_VAR) )
@@ -1031,7 +1126,8 @@ GLboolean checkop3(r700_AssemblerBase* pAsm)
 
     checkop_init(pAsm);
 
-    if( (pILInst->SrcReg[0].File == PROGRAM_CONSTANT)    ||
+    if( (pILInst->SrcReg[0].File == PROGRAM_UNIFORM)     || 
+        (pILInst->SrcReg[0].File == PROGRAM_CONSTANT)    ||
         (pILInst->SrcReg[0].File == PROGRAM_LOCAL_PARAM) ||
         (pILInst->SrcReg[0].File == PROGRAM_ENV_PARAM)   ||
         (pILInst->SrcReg[0].File == PROGRAM_STATE_VAR) )
@@ -1042,7 +1138,8 @@ GLboolean checkop3(r700_AssemblerBase* pAsm)
     {
         bSrcConst[0] = GL_FALSE;
     }
-    if( (pILInst->SrcReg[1].File == PROGRAM_CONSTANT)    ||
+    if( (pILInst->SrcReg[1].File == PROGRAM_UNIFORM)     || 
+        (pILInst->SrcReg[1].File == PROGRAM_CONSTANT)    ||
         (pILInst->SrcReg[1].File == PROGRAM_LOCAL_PARAM) ||
         (pILInst->SrcReg[1].File == PROGRAM_ENV_PARAM)   ||
         (pILInst->SrcReg[1].File == PROGRAM_STATE_VAR) )
@@ -1053,7 +1150,8 @@ GLboolean checkop3(r700_AssemblerBase* pAsm)
     {
         bSrcConst[1] = GL_FALSE;
     }
-    if( (pILInst->SrcReg[2].File == PROGRAM_CONSTANT)    ||
+    if( (pILInst->SrcReg[2].File == PROGRAM_UNIFORM)     || 
+        (pILInst->SrcReg[2].File == PROGRAM_CONSTANT)    ||
         (pILInst->SrcReg[2].File == PROGRAM_LOCAL_PARAM) ||
         (pILInst->SrcReg[2].File == PROGRAM_ENV_PARAM)   ||
         (pILInst->SrcReg[2].File == PROGRAM_STATE_VAR) )
@@ -1153,6 +1251,7 @@ GLboolean assemble_src(r700_AssemblerBase *pAsm,
         case PROGRAM_LOCAL_PARAM:
         case PROGRAM_ENV_PARAM:
         case PROGRAM_STATE_VAR:
+        case PROGRAM_UNIFORM:
             if (1 == pILInst->SrcReg[src].RelAddr)
             {
                 setaddrmode_PVSSRC(&(pAsm->S[fld].src), ADDR_RELATIVE_A0);
@@ -1163,10 +1262,18 @@ GLboolean assemble_src(r700_AssemblerBase *pAsm,
             }
 
             pAsm->S[fld].src.rtype = SRC_REG_CONSTANT;
-            pAsm->S[fld].src.reg   = pILInst->SrcReg[src].Index;
+            if(pILInst->SrcReg[src].Index < 0)
+            {
+                WARN_ONCE("Negative register offsets not supported yet!\n");
+                pAsm->S[fld].src.reg  = 0;
+            } 
+            else
+            {
+                pAsm->S[fld].src.reg = pILInst->SrcReg[src].Index;
+            }
             break;      
         case PROGRAM_INPUT:
-            setaddrmode_PVSSRC(&(pAsm->S[fld].src), ADDR_ABSOLUTE);
+            setaddrmode_PVSSRC(&(pAsm->S[fld].src), ADDR_ABSOLUTE); 
             pAsm->S[fld].src.rtype = SRC_REG_INPUT;
             switch (pAsm->currentShaderType)
             {
@@ -1179,7 +1286,7 @@ GLboolean assemble_src(r700_AssemblerBase *pAsm,
             }
             break;      
         default:
-            radeon_error("Invalid source argument type\n");
+            radeon_error("Invalid source argument type : %d \n", pILInst->SrcReg[src].File);
             return GL_FALSE;
         }
     } 
@@ -1235,6 +1342,15 @@ GLboolean assemble_dst(r700_AssemblerBase *pAsm)
     pAsm->D.dst.writez = (pILInst->DstReg.WriteMask >> 2) & 0x1;
     pAsm->D.dst.writew = (pILInst->DstReg.WriteMask >> 3) & 0x1;
   
+    if(pILInst->SaturateMode == SATURATE_ZERO_ONE)
+    {
+        pAsm->D2.dst2.SaturateMode = 1;
+    }
+    else
+    {
+        pAsm->D2.dst2.SaturateMode = 0;
+    }
+
     return GL_TRUE;
 }
 
@@ -1294,6 +1410,7 @@ GLboolean tex_src(r700_AssemblerBase *pAsm)
     else
     {
     switch (pILInst->SrcReg[0].File) {
+        case PROGRAM_UNIFORM: 
         case PROGRAM_CONSTANT:
         case PROGRAM_LOCAL_PARAM:
         case PROGRAM_ENV_PARAM:
@@ -1306,36 +1423,65 @@ GLboolean tex_src(r700_AssemblerBase *pAsm)
             pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
             break;
         case PROGRAM_INPUT:
-            switch (pILInst->SrcReg[0].Index)
+            if(SPT_VP == pAsm->currentShaderType)
+            {
+                switch (pILInst->SrcReg[0].Index)
+                {
+                    case VERT_ATTRIB_TEX0:
+                    case VERT_ATTRIB_TEX1:
+                    case VERT_ATTRIB_TEX2:
+                    case VERT_ATTRIB_TEX3:
+                    case VERT_ATTRIB_TEX4:
+                    case VERT_ATTRIB_TEX5:
+                    case VERT_ATTRIB_TEX6:
+                    case VERT_ATTRIB_TEX7:
+                        bValidTexCoord = GL_TRUE;
+                        pAsm->S[0].src.reg   =
+                            pAsm->ucVP_AttributeMap[pILInst->SrcReg[0].Index];
+                        pAsm->S[0].src.rtype = SRC_REG_INPUT;
+                        break;
+                }
+            }
+            else
             {
-                case FRAG_ATTRIB_WPOS:
-                case FRAG_ATTRIB_COL0:
-                case FRAG_ATTRIB_COL1:
-                case FRAG_ATTRIB_FOGC:
-                case FRAG_ATTRIB_TEX0:
-                case FRAG_ATTRIB_TEX1:
-                case FRAG_ATTRIB_TEX2:
-               case FRAG_ATTRIB_TEX3:
-                case FRAG_ATTRIB_TEX4:
-                case FRAG_ATTRIB_TEX5:
-                case FRAG_ATTRIB_TEX6:
-                case FRAG_ATTRIB_TEX7:
-                    bValidTexCoord = GL_TRUE;
+                switch (pILInst->SrcReg[0].Index)
+                {
+                    case FRAG_ATTRIB_WPOS:
+                    case FRAG_ATTRIB_COL0:
+                    case FRAG_ATTRIB_COL1:
+                    case FRAG_ATTRIB_FOGC:
+                    case FRAG_ATTRIB_TEX0:
+                    case FRAG_ATTRIB_TEX1:
+                    case FRAG_ATTRIB_TEX2:
+                    case FRAG_ATTRIB_TEX3:
+                    case FRAG_ATTRIB_TEX4:
+                    case FRAG_ATTRIB_TEX5:
+                    case FRAG_ATTRIB_TEX6:
+                    case FRAG_ATTRIB_TEX7:
+                        bValidTexCoord = GL_TRUE;
+                        pAsm->S[0].src.reg   =
+                            pAsm->uiFP_AttributeMap[pILInst->SrcReg[0].Index];
+                        pAsm->S[0].src.rtype = SRC_REG_INPUT;
+                        break;
+                    case FRAG_ATTRIB_FACE:
+                        fprintf(stderr, "FRAG_ATTRIB_FACE unsupported\n");
+                        break;
+                    case FRAG_ATTRIB_PNTC:
+                        fprintf(stderr, "FRAG_ATTRIB_PNTC unsupported\n");
+                        break;
+                }
+
+                if( (pILInst->SrcReg[0].Index >= FRAG_ATTRIB_VAR0) ||
+                    (pILInst->SrcReg[0].Index < FRAG_ATTRIB_MAX) )
+                {
+                                   bValidTexCoord = GL_TRUE;
                     pAsm->S[0].src.reg   =
                         pAsm->uiFP_AttributeMap[pILInst->SrcReg[0].Index];
                     pAsm->S[0].src.rtype = SRC_REG_INPUT;
-                    break;
-                case FRAG_ATTRIB_FACE:
-                    fprintf(stderr, "FRAG_ATTRIB_FACE unsupported\n");
-                    break;
-                case FRAG_ATTRIB_PNTC:
-                    fprintf(stderr, "FRAG_ATTRIB_PNTC unsupported\n");
-                    break;
-                case FRAG_ATTRIB_VAR0:
-                    fprintf(stderr, "FRAG_ATTRIB_VAR0 unsupported\n");
-                    break;
+                }
             }
-        break;
+
+            break;
         }
     }
 
@@ -1380,8 +1526,17 @@ GLboolean assemble_tex_instruction(r700_AssemblerBase *pAsm, GLboolean normalize
     tex_instruction_ptr->m_Word0.f.tex_inst         = pAsm->D.dst.opcode;
     tex_instruction_ptr->m_Word0.f.bc_frac_mode     = 0x0;
     tex_instruction_ptr->m_Word0.f.fetch_whole_quad = 0x0;
+    tex_instruction_ptr->m_Word0.f.alt_const        = 0;
 
-    tex_instruction_ptr->m_Word0.f.resource_id      = texture_unit_source->reg;
+    if(SPT_VP == pAsm->currentShaderType)
+    {
+        tex_instruction_ptr->m_Word0.f.resource_id      = texture_unit_source->reg + VERT_ATTRIB_MAX;
+        pAsm->unVetTexBits |= 1 << texture_unit_source->reg;
+    }
+    else
+    {
+        tex_instruction_ptr->m_Word0.f.resource_id      = texture_unit_source->reg;
+    }
 
     tex_instruction_ptr->m_Word1.f.lod_bias     = 0x0;
     if (normalized) {
@@ -1400,7 +1555,6 @@ GLboolean assemble_tex_instruction(r700_AssemblerBase *pAsm, GLboolean normalize
     tex_instruction_ptr->m_Word2.f.offset_x   = 0x0;
     tex_instruction_ptr->m_Word2.f.offset_y   = 0x0;
     tex_instruction_ptr->m_Word2.f.offset_z   = 0x0;
-
     tex_instruction_ptr->m_Word2.f.sampler_id = texture_unit_source->reg;
 
     // dst
@@ -1517,6 +1671,10 @@ GLboolean assemble_alu_src(R700ALUInstruction*  alu_instruction_ptr,
         {
             src_sel = pSource->reg + CFILE_REGISTER_OFFSET;            
         }
+        else if (pSource->rtype == SRC_REC_LITERAL)
+        {
+            src_sel = SQ_ALU_SRC_LITERAL;            
+        }
         else
         {
             radeon_error("Source (%d) register type (%d) not one of TEMP, INPUT, or CONSTANT.\n",
@@ -1606,7 +1764,8 @@ GLboolean add_alu_instruction(r700_AssemblerBase* pAsm,
         return GL_FALSE;
     }
 
-    if ( pAsm->cf_current_alu_clause_ptr == NULL ||
+    if ( pAsm->alu_x_opcode != 0 ||
+         pAsm->cf_current_alu_clause_ptr == NULL ||
          ( (pAsm->cf_current_alu_clause_ptr != NULL) && 
            (pAsm->cf_current_alu_clause_ptr->m_Word1.f.count >= (GetCFMaxInstructions(pAsm->cf_current_alu_clause_ptr->m_ShaderInstType)-contiguous_slots_needed-1) )
          ) ) 
@@ -1636,9 +1795,17 @@ GLboolean add_alu_instruction(r700_AssemblerBase* pAsm,
         pAsm->cf_current_alu_clause_ptr->m_Word1.f.kcache_addr0 = 0x0;
         pAsm->cf_current_alu_clause_ptr->m_Word1.f.kcache_addr1 = 0x0;
 
-        //cf_current_alu_clause_ptr->m_Word1.f.count           = number_of_scalar_operations - 1;
         pAsm->cf_current_alu_clause_ptr->m_Word1.f.count           = 0x0;
-        pAsm->cf_current_alu_clause_ptr->m_Word1.f.cf_inst         = SQ_CF_INST_ALU;
+
+        if(pAsm->alu_x_opcode != 0)
+        {
+            pAsm->cf_current_alu_clause_ptr->m_Word1.f.cf_inst = pAsm->alu_x_opcode;
+            pAsm->alu_x_opcode = 0;
+        }
+        else
+        {
+            pAsm->cf_current_alu_clause_ptr->m_Word1.f.cf_inst = SQ_CF_INST_ALU;
+        }
 
         pAsm->cf_current_alu_clause_ptr->m_Word1.f.whole_quad_mode = 0x0;
 
@@ -1646,7 +1813,7 @@ GLboolean add_alu_instruction(r700_AssemblerBase* pAsm,
     }
     else 
     {
-        pAsm->cf_current_alu_clause_ptr->m_Word1.f.count++;
+        pAsm->cf_current_alu_clause_ptr->m_Word1.f.count += (GetInstructionSize(alu_instruction_ptr->m_ShaderInstType) / 2);
     }
 
     // If this clause constains any instruction that is forward dependent on a TEX instruction, 
@@ -1923,7 +2090,7 @@ GLboolean check_scalar(r700_AssemblerBase* pAsm,
 
     GLuint swizzle_key;
 
-    GLuint number_of_operands = r700GetNumOperands(pAsm);
+    GLuint number_of_operands = r700GetNumOperands(pAsm->D.dst.opcode, pAsm->D.dst.op3);
 
     for (src=0; src<number_of_operands; src++) 
     {
@@ -2012,7 +2179,7 @@ GLboolean check_vector(r700_AssemblerBase* pAsm,
 
     GLuint swizzle_key;
 
-    GLuint number_of_operands = r700GetNumOperands(pAsm);
+    GLuint number_of_operands = r700GetNumOperands(pAsm->D.dst.opcode, pAsm->D.dst.op3);
 
     for (src=0; src<number_of_operands; src++) 
     {
@@ -2045,7 +2212,7 @@ GLboolean check_vector(r700_AssemblerBase* pAsm,
         if( is_gpr(sel) ) 
         {
             if( GL_FALSE == cycle_for_vector_bank_swizzle(bank_swizzle, src, &cycle) )
-            {
+            {             
                 return GL_FALSE;
             }
 
@@ -2057,7 +2224,7 @@ GLboolean check_vector(r700_AssemblerBase* pAsm,
             else 
             {
                 if( GL_FALSE == reserve_gpr(pAsm, sel, chan, cycle) )
-                {
+                {                    
                     return GL_FALSE;
                 }
             }
@@ -2069,7 +2236,7 @@ GLboolean check_vector(r700_AssemblerBase* pAsm,
             if( is_cfile(sel) ) 
             {        
                 if( GL_FALSE == reserve_cfile(pAsm, sel, chan) )
-                {
+                {                    
                     return GL_FALSE;
                 }
             }
@@ -2081,6 +2248,10 @@ GLboolean check_vector(r700_AssemblerBase* pAsm,
 
 GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
 {
+    R700ALUInstruction            * alu_instruction_ptr;
+    R700ALUInstructionHalfLiteral * alu_instruction_ptr_hl;
+    R700ALUInstructionFullLiteral * alu_instruction_ptr_fl;
+
     GLuint    number_of_scalar_operations;
     GLboolean is_single_scalar_operation;
     GLuint    scalar_channel_index;
@@ -2089,7 +2260,7 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
     int    current_source_index;
     GLuint contiguous_slots_needed;
 
-    GLuint    uNumSrc = r700GetNumOperands(pAsm);
+    GLuint    uNumSrc = r700GetNumOperands(pAsm->D.dst.opcode, pAsm->D.dst.op3);
     //GLuint    channel_swizzle, j;
     //GLuint    chan_counter[4] = {0, 0, 0, 0};
     //PVSSRC *  pSource[3];
@@ -2146,23 +2317,44 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
 
     contiguous_slots_needed = 0;
 
-    if(GL_TRUE == is_reduction_opcode(&(pAsm->D)) 
+    if(!is_single_scalar_operation
     {
         contiguous_slots_needed = 4;
     }
 
+    contiguous_slots_needed += pAsm->D2.dst2.literal_slots;
+
     initialize(pAsm);    
 
     for (scalar_channel_index=0;
             scalar_channel_index < number_of_scalar_operations; 
                 scalar_channel_index++) 
     {
-        R700ALUInstruction* alu_instruction_ptr = (R700ALUInstruction*) CALLOC_STRUCT(R700ALUInstruction);
-        if (alu_instruction_ptr == NULL) 
-               {
-                       return GL_FALSE;
-               }
-        Init_R700ALUInstruction(alu_instruction_ptr);
+        if(scalar_channel_index == (number_of_scalar_operations-1))
+        {
+            switch(pAsm->D2.dst2.literal_slots)
+            {
+            case 0:
+                alu_instruction_ptr = (R700ALUInstruction*) CALLOC_STRUCT(R700ALUInstruction);
+                Init_R700ALUInstruction(alu_instruction_ptr);
+                break;
+            case 1:
+                alu_instruction_ptr_hl = (R700ALUInstructionHalfLiteral*) CALLOC_STRUCT(R700ALUInstructionHalfLiteral);
+                Init_R700ALUInstructionHalfLiteral(alu_instruction_ptr_hl, pAsm->C[0].f, pAsm->C[1].f);
+                alu_instruction_ptr = (R700ALUInstruction*)alu_instruction_ptr_hl;
+                break;
+            case 2:
+                alu_instruction_ptr_fl = (R700ALUInstructionFullLiteral*) CALLOC_STRUCT(R700ALUInstructionFullLiteral);
+                Init_R700ALUInstructionFullLiteral(alu_instruction_ptr_fl,pAsm->C[0].f, pAsm->C[1].f, pAsm->C[2].f, pAsm->C[3].f);
+                alu_instruction_ptr = (R700ALUInstruction*)alu_instruction_ptr_fl;
+            break;
+            };
+        }
+        else
+        {
+            alu_instruction_ptr = (R700ALUInstruction*) CALLOC_STRUCT(R700ALUInstruction);
+            Init_R700ALUInstruction(alu_instruction_ptr);
+        }
         
         //src 0
         current_source_index = 0;
@@ -2172,7 +2364,7 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
                                          current_source_index,
                                          pcurrent_source, 
                                          scalar_channel_index) )     
-        {
+        {            
             return GL_FALSE;
         }
    
@@ -2186,13 +2378,13 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
                                              current_source_index,
                                              pcurrent_source, 
                                              scalar_channel_index) ) 
-            {
+            {                
                 return GL_FALSE;
             }
         }
 
         //other bits
-        alu_instruction_ptr->m_Word0.f.index_mode = SQ_INDEX_AR_X;
+        alu_instruction_ptr->m_Word0.f.index_mode = pAsm->D2.dst2.index_mode;
 
         if(   (is_single_scalar_operation == GL_TRUE) 
            || (GL_TRUE == bSplitInst) )
@@ -2204,9 +2396,17 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
             alu_instruction_ptr->m_Word0.f.last = (scalar_channel_index == 3) ?  1 : 0;
         }
 
-        alu_instruction_ptr->m_Word0.f.pred_sel                = 0x0;
-        alu_instruction_ptr->m_Word1_OP2.f.update_pred         = 0x0;  
-        alu_instruction_ptr->m_Word1_OP2.f.update_execute_mask = 0x0;
+        alu_instruction_ptr->m_Word0.f.pred_sel = (pAsm->D.dst.pred_inv > 0) ? 1 : 0;
+        if(1 == pAsm->D.dst.predicated)
+        {
+            alu_instruction_ptr->m_Word1_OP2.f.update_pred         = 0x1;
+            alu_instruction_ptr->m_Word1_OP2.f.update_execute_mask = 0x1;
+        }
+        else
+        {
+            alu_instruction_ptr->m_Word1_OP2.f.update_pred         = 0x0;
+            alu_instruction_ptr->m_Word1_OP2.f.update_execute_mask = 0x0;
+        }
 
         // dst
         if( (pAsm->D.dst.rtype == DST_REG_TEMPORARY) || 
@@ -2215,7 +2415,7 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
             alu_instruction_ptr->m_Word1.f.dst_gpr  = pAsm->D.dst.reg;
         }
         else 
-        {
+        {            
             radeon_error("Only temp destination registers supported for ALU dest regs.\n");
             return GL_FALSE;
         }
@@ -2245,7 +2445,7 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
 
         alu_instruction_ptr->m_Word1.f.dst_chan = scalar_channel_index;
 
-        alu_instruction_ptr->m_Word1.f.clamp    = pAsm->pILInst[pAsm->uiCurInst].SaturateMode;
+        alu_instruction_ptr->m_Word1.f.clamp    = pAsm->D2.dst2.SaturateMode;
 
         if (pAsm->D.dst.op3) 
         {            
@@ -2272,8 +2472,8 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
             {
                 alu_instruction_ptr->m_Word1_OP2.f6.alu_inst           = pAsm->D.dst.opcode;
 
-                alu_instruction_ptr->m_Word1_OP2.f6.src0_abs           = 0x0;
-                alu_instruction_ptr->m_Word1_OP2.f6.src1_abs           = 0x0;
+                alu_instruction_ptr->m_Word1_OP2.f6.src0_abs           = pAsm->S[0].src.abs;
+                alu_instruction_ptr->m_Word1_OP2.f6.src1_abs           = pAsm->S[1].src.abs;
 
                 //alu_instruction_ptr->m_Word1_OP2.f6.update_execute_mask = 0x0;
                 //alu_instruction_ptr->m_Word1_OP2.f6.update_pred         = 0x0;
@@ -2301,8 +2501,8 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
             {
                 alu_instruction_ptr->m_Word1_OP2.f.alu_inst           = pAsm->D.dst.opcode;
 
-                alu_instruction_ptr->m_Word1_OP2.f.src0_abs           = 0x0;
-                alu_instruction_ptr->m_Word1_OP2.f.src1_abs           = 0x0;
+                alu_instruction_ptr->m_Word1_OP2.f.src0_abs           = pAsm->S[0].src.abs;
+                alu_instruction_ptr->m_Word1_OP2.f.src1_abs           = pAsm->S[1].src.abs;
 
                 //alu_instruction_ptr->m_Word1_OP2.f.update_execute_mask = 0x0;
                 //alu_instruction_ptr->m_Word1_OP2.f.update_pred         = 0x0;
@@ -2329,7 +2529,7 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
         }
 
         if(GL_FALSE == add_alu_instruction(pAsm, alu_instruction_ptr, contiguous_slots_needed) )
-        {
+        {            
             return GL_FALSE;
         }
 
@@ -2340,19 +2540,19 @@ GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm)
         if (is_single_scalar_operation) 
         {
             if(GL_FALSE == check_scalar(pAsm, alu_instruction_ptr) )
-            {
+            {                
                 return GL_FALSE;
             }
         }
         else 
         {
             if(GL_FALSE == check_vector(pAsm, alu_instruction_ptr) )
-            {
-                return 1;
+            {                
+                return GL_FALSE; 
             }
         }
 
-        contiguous_slots_needed = 0;
+        contiguous_slots_needed -= 1;
     }
 
     return GL_TRUE;
@@ -2403,11 +2603,14 @@ GLboolean next_ins(r700_AssemblerBase *pAsm)
     
     //reset for next inst.
     pAsm->D.bits    = 0;
+    pAsm->D2.bits   = 0;
     pAsm->S[0].bits = 0;
     pAsm->S[1].bits = 0;
     pAsm->S[2].bits = 0;
     pAsm->is_tex = GL_FALSE;
     pAsm->need_tex_barrier = GL_FALSE;
+    pAsm->D2.bits = 0;
+    pAsm->C[0].bits = pAsm->C[1].bits = pAsm->C[2].bits = pAsm->C[3].bits = 0;
     return GL_TRUE;
 }
 
@@ -2640,9 +2843,44 @@ GLboolean assemble_CMP(r700_AssemblerBase *pAsm)
     return GL_TRUE;
 }
 
-GLboolean assemble_COS(r700_AssemblerBase *pAsm)
+GLboolean assemble_TRIG(r700_AssemblerBase *pAsm, BITS opcode)
 {
-    return assemble_math_function(pAsm, SQ_OP2_INST_COS);
+    int tmp;
+    checkop1(pAsm);
+
+    tmp = gethelpr(pAsm);
+
+    pAsm->D.dst.opcode = SQ_OP2_INST_MUL;
+    setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
+    pAsm->D.dst.rtype  = DST_REG_TEMPORARY;
+    pAsm->D.dst.reg    = tmp;
+    pAsm->D.dst.writex = 1;
+
+    assemble_src(pAsm, 0, -1);
+
+    pAsm->S[1].src.rtype = SRC_REC_LITERAL;
+    setswizzle_PVSSRC(&(pAsm->S[1].src), SQ_SEL_X);
+    pAsm->D2.dst2.literal_slots = 1;
+    pAsm->C[0].f = 1/(3.1415926535 * 2);
+    pAsm->C[1].f = 0.0F;
+    next_ins(pAsm);
+
+    pAsm->D.dst.opcode = opcode;
+    pAsm->D.dst.math = 1;
+
+    assemble_dst(pAsm);
+
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
+    pAsm->S[0].src.reg   = tmp;
+    setswizzle_PVSSRC(&(pAsm->S[0].src), SQ_SEL_X);
+    noneg_PVSSRC(&(pAsm->S[0].src));
+
+    next_ins(pAsm);
+
+    //TODO - replicate if more channels set in WriteMask
+    return GL_TRUE;
+
 }
  
 GLboolean assemble_DOT(r700_AssemblerBase *pAsm)
@@ -2910,13 +3148,15 @@ GLboolean assemble_FRC(r700_AssemblerBase *pAsm)
     return GL_TRUE;
 }
  
-GLboolean assemble_KIL(r700_AssemblerBase *pAsm)
-{
-    /* TODO: doc says KILL has to be last(end) ALU clause */
-    
-    checkop1(pAsm);
+GLboolean assemble_KIL(r700_AssemblerBase *pAsm, GLuint opcode)
+{  
+    struct prog_instruction *pILInst = &(pAsm->pILInst[pAsm->uiCurInst]);
+
+    if(pILInst->Opcode == OPCODE_KIL)
+        checkop1(pAsm);
 
-    pAsm->D.dst.opcode = SQ_OP2_INST_KILLGT;  
+    pAsm->D.dst.opcode = opcode;  
+    //pAsm->D.dst.math = 1;
 
     setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
     pAsm->D.dst.rtype = DST_REG_TEMPORARY;
@@ -2929,21 +3169,34 @@ GLboolean assemble_KIL(r700_AssemblerBase *pAsm)
     setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
     pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
     pAsm->S[0].src.reg = 0;
-
     setswizzle_PVSSRC(&(pAsm->S[0].src), SQ_SEL_0);
     noneg_PVSSRC(&(pAsm->S[0].src));
 
-    if ( GL_FALSE == assemble_src(pAsm, 0, 1) )
+    if(pILInst->Opcode == OPCODE_KIL_NV)
     {
-        return GL_FALSE;
+        setaddrmode_PVSSRC(&(pAsm->S[1].src), ADDR_ABSOLUTE);
+        pAsm->S[1].src.rtype = SRC_REG_TEMPORARY;
+        pAsm->S[1].src.reg = 0;
+        setswizzle_PVSSRC(&(pAsm->S[1].src), SQ_SEL_1);
+        neg_PVSSRC(&(pAsm->S[1].src));
     }
-  
+    else
+    {
+        if( GL_FALSE == assemble_src(pAsm, 0, 1) )
+        {
+            return GL_FALSE;
+        }
+
+    }
+
     if ( GL_FALSE == next_ins(pAsm) )
     {
         return GL_FALSE;
     }
 
+    /* Doc says KILL has to be last(end) ALU clause */
     pAsm->pR700Shader->killIsUsed = GL_TRUE;
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU;
     
     return GL_TRUE;
 }
@@ -3007,6 +3260,7 @@ GLboolean assemble_LRP(r700_AssemblerBase *pAsm)
     {
         return GL_FALSE;
     }
+
     if( GL_FALSE == assemble_src(pAsm, 2, -1) ) 
     {
         return GL_FALSE;
@@ -3742,89 +3996,80 @@ GLboolean assemble_RSQ(r700_AssemblerBase *pAsm)
     return assemble_math_function(pAsm, SQ_OP2_INST_RECIPSQRT_IEEE);
 }
  
-GLboolean assemble_SIN(r700_AssemblerBase *pAsm) 
-{
-    return assemble_math_function(pAsm, SQ_OP2_INST_SIN);
-}
 GLboolean assemble_SCS(r700_AssemblerBase *pAsm) 
 {
     BITS tmp;
 
-       checkop1(pAsm);
-
-       tmp = gethelpr(pAsm);
-
-       // COS tmp.x,    a.x
-       pAsm->D.dst.opcode = SQ_OP2_INST_COS;
-       pAsm->D.dst.math = 1;
+    checkop1(pAsm);
 
-       setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
-       pAsm->D.dst.rtype = DST_REG_TEMPORARY;
-       pAsm->D.dst.reg = tmp;
-       pAsm->D.dst.writex = 1;
+    tmp = gethelpr(pAsm);
+    /* tmp.x = src /2*PI */
+    pAsm->D.dst.opcode = SQ_OP2_INST_MUL;
+    setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
+    pAsm->D.dst.rtype  = DST_REG_TEMPORARY;
+    pAsm->D.dst.reg    = tmp;
+    pAsm->D.dst.writex = 1;
 
-       if( GL_FALSE == assemble_src(pAsm, 0, -1) )
-       {
-               return GL_FALSE;
-       }
+    assemble_src(pAsm, 0, -1);
 
-       if ( GL_FALSE == next_ins(pAsm) )
-       {
-               return GL_FALSE;
-       }
+    pAsm->S[1].src.rtype = SRC_REC_LITERAL;
+    setswizzle_PVSSRC(&(pAsm->S[1].src), SQ_SEL_X);
+    pAsm->D2.dst2.literal_slots = 1;
+    pAsm->C[0].f = 1/(3.1415926535 * 2);
+    pAsm->C[1].f = 0.0F;
 
-       // SIN tmp.y,    a.x
-       pAsm->D.dst.opcode = SQ_OP2_INST_SIN;
-       pAsm->D.dst.math = 1;
+    next_ins(pAsm);
 
-       setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
-       pAsm->D.dst.rtype = DST_REG_TEMPORARY;
-       pAsm->D.dst.reg = tmp;
-       pAsm->D.dst.writey = 1;
+    // COS dst.x,    a.x
+    pAsm->D.dst.opcode = SQ_OP2_INST_COS;
+    pAsm->D.dst.math = 1;
 
-       if( GL_FALSE == assemble_src(pAsm, 0, -1) )
-       {
-               return GL_FALSE;
-       }
+    assemble_dst(pAsm);
+    /* mask y */
+    pAsm->D.dst.writey = 0;
 
-       if( GL_FALSE == next_ins(pAsm) )
-       {
-               return GL_FALSE;
-       }
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
+    pAsm->S[0].src.reg   = tmp;
+    setswizzle_PVSSRC(&(pAsm->S[0].src), SQ_SEL_X);
+    noneg_PVSSRC(&(pAsm->S[0].src));
 
-       // MOV dst.mask,     tmp
-       pAsm->D.dst.opcode = SQ_OP2_INST_MOV;
+    if ( GL_FALSE == next_ins(pAsm) )
+    {
+        return GL_FALSE;
+    }
 
-       if( GL_FALSE == assemble_dst(pAsm) )
-       {
-               return GL_FALSE;
-       }
+    // SIN dst.y,    a.x
+    pAsm->D.dst.opcode = SQ_OP2_INST_SIN;
+    pAsm->D.dst.math = 1;
 
-       setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
-       pAsm->S[0].src.rtype = DST_REG_TEMPORARY;
-       pAsm->S[0].src.reg = tmp;
+    assemble_dst(pAsm);
+    /* mask x */
+    pAsm->D.dst.writex = 0;
 
-       noswizzle_PVSSRC(&(pAsm->S[0].src));
-       pAsm->S[0].src.swizzlez = SQ_SEL_0;
-       pAsm->S[0].src.swizzlew = SQ_SEL_0;
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
+    pAsm->S[0].src.reg   = tmp;
+    setswizzle_PVSSRC(&(pAsm->S[0].src), SQ_SEL_X);
+    noneg_PVSSRC(&(pAsm->S[0].src));
 
-       if ( GL_FALSE == next_ins(pAsm) )
-       {
-               return GL_FALSE;
-       }
+    if( GL_FALSE == next_ins(pAsm) )
+    {
+        return GL_FALSE;
+    }
 
     return GL_TRUE;
 }
-GLboolean assemble_SGE(r700_AssemblerBase *pAsm
+
+GLboolean assemble_LOGIC(r700_AssemblerBase *pAsm, BITS opcode
 {
     if( GL_FALSE == checkop2(pAsm) )
     {
            return GL_FALSE;
     }
 
-    pAsm->D.dst.opcode = SQ_OP2_INST_SETGE;  
+    pAsm->D.dst.opcode = opcode;
+    //pAsm->D.dst.math   = 1;
 
     if( GL_FALSE == assemble_dst(pAsm) )
     {
@@ -3848,23 +4093,92 @@ GLboolean assemble_SGE(r700_AssemblerBase *pAsm)
 
     return GL_TRUE;
 }
-GLboolean assemble_SLT(r700_AssemblerBase *pAsm
+
+GLboolean assemble_LOGIC_PRED(r700_AssemblerBase *pAsm, BITS opcode
 {
-    if( GL_FALSE == checkop2(pAsm) )
-    {
-           return GL_FALSE;
-    }
+    struct prog_instruction *pILInst = &(pAsm->pILInst[pAsm->uiCurInst]);
 
-    pAsm->D.dst.opcode = SQ_OP2_INST_SETGT;  
+    pAsm->D.dst.opcode = opcode;
+    pAsm->D.dst.math   = 1;
+    pAsm->D.dst.predicated = 1;
 
-    if( GL_FALSE == assemble_dst(pAsm) )
-    {
-        return GL_FALSE;
-    }
-                
-    if( GL_FALSE == assemble_src(pAsm, 0, 1) )  
-    {
+    setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
+    pAsm->D.dst.rtype = DST_REG_TEMPORARY;
+    pAsm->D.dst.reg = pAsm->uHelpReg;
+    pAsm->D.dst.writex = 1;
+    pAsm->D.dst.writey = pAsm->D.dst.writez = pAsm->D.dst.writew = 0;
+
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
+    pAsm->S[0].src.reg = pAsm->last_cond_register + pAsm->starting_temp_register_number;
+    pAsm->S[0].src.swizzlex = pILInst->DstReg.CondSwizzle & 0x7;
+    noneg_PVSSRC(&(pAsm->S[0].src));
+
+    pAsm->S[1].src.rtype = SRC_REG_TEMPORARY;
+    pAsm->S[1].src.reg   = pAsm->uHelpReg;
+    setaddrmode_PVSSRC(&(pAsm->S[1].src), ADDR_ABSOLUTE);
+    noneg_PVSSRC(&(pAsm->S[1].src));
+    pAsm->S[1].src.swizzlex = SQ_SEL_0;
+    pAsm->S[1].src.swizzley = SQ_SEL_0;
+    pAsm->S[1].src.swizzlez = SQ_SEL_0;
+    pAsm->S[1].src.swizzlew = SQ_SEL_0;
+
+    if( GL_FALSE == next_ins(pAsm) ) 
+    {
+           return GL_FALSE;
+    }
+
+    return GL_TRUE;
+}
+GLboolean assemble_SGE(r700_AssemblerBase *pAsm) 
+{
+    if( GL_FALSE == checkop2(pAsm) )
+    {
+           return GL_FALSE;
+    }
+
+    pAsm->D.dst.opcode = SQ_OP2_INST_SETGE;  
+
+    if( GL_FALSE == assemble_dst(pAsm) )
+    {
+           return GL_FALSE;
+    }
+
+    if( GL_FALSE == assemble_src(pAsm, 0, -1) )
+    {
+           return GL_FALSE;
+    }
+
+    if( GL_FALSE == assemble_src(pAsm, 1, -1) )
+    {
+           return GL_FALSE;
+    }
+
+    if( GL_FALSE == next_ins(pAsm) ) 
+    {
+           return GL_FALSE;
+    }
+
+    return GL_TRUE;
+}
+GLboolean assemble_SLT(r700_AssemblerBase *pAsm) 
+{
+    if( GL_FALSE == checkop2(pAsm) )
+    {
+           return GL_FALSE;
+    }
+
+    pAsm->D.dst.opcode = SQ_OP2_INST_SETGT;  
+
+    if( GL_FALSE == assemble_dst(pAsm) )
+    {
+        return GL_FALSE;
+    }
+                
+    if( GL_FALSE == assemble_src(pAsm, 0, 1) )  
+    {
         return GL_FALSE;
     }
 
@@ -3895,6 +4209,7 @@ GLboolean assemble_TEX(r700_AssemblerBase *pAsm)
     
     switch (pAsm->pILInst[pAsm->uiCurInst].SrcReg[0].File)
     {
+    case PROGRAM_UNIFORM: 
     case PROGRAM_CONSTANT:
     case PROGRAM_LOCAL_PARAM:
     case PROGRAM_ENV_PARAM:
@@ -3992,24 +4307,6 @@ GLboolean assemble_TEX(r700_AssemblerBase *pAsm)
             return GL_FALSE;
         }
  
-        /* tmp1.z = ABS(tmp1.z) dont have abs support in assembler currently
-         * have to do explicit instruction
-         */
-        pAsm->D.dst.opcode = SQ_OP2_INST_MAX;
-        setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
-        pAsm->D.dst.rtype = DST_REG_TEMPORARY;
-        pAsm->D.dst.reg   = tmp1;
-        pAsm->D.dst.writez = 1;
-
-        setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
-        pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
-        pAsm->S[0].src.reg = tmp1;
-       noswizzle_PVSSRC(&(pAsm->S[0].src));
-        pAsm->S[1].bits = pAsm->S[0].bits;
-        flipneg_PVSSRC(&(pAsm->S[1].src));
-        
-        next_ins(pAsm);
-
         /* tmp1.z = RCP_e(|tmp1.z|) */
         pAsm->D.dst.opcode = SQ_OP2_INST_RECIP_IEEE;
         pAsm->D.dst.math = 1;
@@ -4022,13 +4319,13 @@ GLboolean assemble_TEX(r700_AssemblerBase *pAsm)
         pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
         pAsm->S[0].src.reg = tmp1;
         pAsm->S[0].src.swizzlex = SQ_SEL_Z;
+        pAsm->S[0].src.abs = 1;
 
         next_ins(pAsm);
 
         /* MULADD R0.x,  R0.x,  PS1,  (0x3FC00000, 1.5f).x
          * MULADD R0.y,  R0.y,  PS1,  (0x3FC00000, 1.5f).x
          * muladd has no writemask, have to use another temp 
-         * also no support for imm constants, so add 1 here
          */
         pAsm->D.dst.opcode = SQ_OP3_INST_MULADD;
         pAsm->D.dst.op3    = 1;
@@ -4045,30 +4342,12 @@ GLboolean assemble_TEX(r700_AssemblerBase *pAsm)
         pAsm->S[1].src.reg   = tmp1;
         setswizzle_PVSSRC(&(pAsm->S[1].src), SQ_SEL_Z);
         setaddrmode_PVSSRC(&(pAsm->S[2].src), ADDR_ABSOLUTE);
-        pAsm->S[2].src.rtype = SRC_REG_TEMPORARY;
+        /* immediate c 1.5 */
+        pAsm->D2.dst2.literal_slots = 1;
+        pAsm->C[0].f = 1.5F;
+        pAsm->S[2].src.rtype = SRC_REC_LITERAL;
         pAsm->S[2].src.reg   = tmp1;
-        setswizzle_PVSSRC(&(pAsm->S[2].src), SQ_SEL_1);
-
-        next_ins(pAsm);
-
-        /* ADD the remaining .5 */
-        pAsm->D.dst.opcode = SQ_OP2_INST_ADD;
-        setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
-        pAsm->D.dst.rtype = DST_REG_TEMPORARY;
-        pAsm->D.dst.reg   = tmp2;
-        pAsm->D.dst.writex = 1;
-        pAsm->D.dst.writey = 1;
-        pAsm->D.dst.writez = 0;
-        pAsm->D.dst.writew = 0;
-
-        setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
-        pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
-        pAsm->S[0].src.reg   = tmp2;
-        noswizzle_PVSSRC(&(pAsm->S[0].src));
-        setaddrmode_PVSSRC(&(pAsm->S[1].src), ADDR_ABSOLUTE);
-        pAsm->S[1].src.rtype = SRC_REG_TEMPORARY;
-        pAsm->S[1].src.reg   = 252; // SQ_ALU_SRC_0_5 
-        noswizzle_PVSSRC(&(pAsm->S[1].src));
+        setswizzle_PVSSRC(&(pAsm->S[2].src), SQ_SEL_X);
 
         next_ins(pAsm);
 
@@ -4093,22 +4372,32 @@ GLboolean assemble_TEX(r700_AssemblerBase *pAsm)
 
     }
 
-    if(pAsm->pILInst[pAsm->uiCurInst].Opcode == OPCODE_TXB)
-    {
-        pAsm->D.dst.opcode = SQ_TEX_INST_SAMPLE_L;
-    }
-    else
+    switch(pAsm->pILInst[pAsm->uiCurInst].Opcode)
     {
-        pAsm->D.dst.opcode = SQ_TEX_INST_SAMPLE;
+        case OPCODE_DDX:
+            /* will these need WQM(1) on CF inst ? */
+            pAsm->D.dst.opcode = SQ_TEX_INST_GET_GRADIENTS_H;
+            break;
+        case OPCODE_DDY:
+            pAsm->D.dst.opcode = SQ_TEX_INST_GET_GRADIENTS_V;
+            break;
+        case OPCODE_TXB:
+            pAsm->D.dst.opcode = SQ_TEX_INST_SAMPLE_L;
+            break;
+        default:
+            pAsm->D.dst.opcode = SQ_TEX_INST_SAMPLE;
     }
 
+    pAsm->is_tex = GL_TRUE;
+    if ( GL_TRUE == need_barrier )
+
     pAsm->is_tex = GL_TRUE;
     if ( GL_TRUE == need_barrier )
     {
         pAsm->need_tex_barrier = GL_TRUE;
     }
     // Set src1 to tex unit id
-    pAsm->S[1].src.reg   = pAsm->pILInst[pAsm->uiCurInst].TexSrcUnit;
+    pAsm->S[1].src.reg   = pAsm->SamplerUnits[pAsm->pILInst[pAsm->uiCurInst].TexSrcUnit];
     pAsm->S[1].src.rtype = SRC_REG_TEMPORARY;
 
     //No sw info from mesa compiler, so hard code here.
@@ -4265,137 +4554,1041 @@ GLboolean assemble_EXPORT(r700_AssemblerBase *pAsm)
     return GL_TRUE;
 }
 
-GLboolean assemble_IF(r700_AssemblerBase *pAsm)
+static inline void decreaseCurrent(r700_AssemblerBase *pAsm, GLuint uReason)
 {
-    return GL_TRUE;
+    switch (uReason)
+    {
+    case FC_PUSH_VPM:
+        pAsm->CALLSTACK[pAsm->CALLSP].current--;
+        break;
+    case FC_PUSH_WQM:
+        pAsm->CALLSTACK[pAsm->CALLSP].current -= 4;
+        break;
+    case FC_LOOP:
+        pAsm->CALLSTACK[pAsm->CALLSP].current -= 4;
+        break;
+    case FC_REP:
+        /* TODO : for 16 vp asic, should -= 2; */
+        pAsm->CALLSTACK[pAsm->CALLSP].current -= 1;
+        break;
+    };
 }
 
-GLboolean assemble_ENDIF(r700_AssemblerBase *pAsm)
+static inline void checkStackDepth(r700_AssemblerBase *pAsm, GLuint uReason, GLboolean bCheckMaxOnly)
 {
-    return GL_TRUE;
+    if(GL_TRUE == bCheckMaxOnly)
+    {
+        switch (uReason)
+        {
+        case FC_PUSH_VPM:
+            if((pAsm->CALLSTACK[pAsm->CALLSP].current + 1)
+                    > pAsm->CALLSTACK[pAsm->CALLSP].max)
+            {
+                pAsm->CALLSTACK[pAsm->CALLSP].max =
+                    pAsm->CALLSTACK[pAsm->CALLSP].current + 1;
+            }
+            break;
+        case FC_PUSH_WQM:
+            if((pAsm->CALLSTACK[pAsm->CALLSP].current + 4)
+                    > pAsm->CALLSTACK[pAsm->CALLSP].max)
+            {
+                pAsm->CALLSTACK[pAsm->CALLSP].max =
+                    pAsm->CALLSTACK[pAsm->CALLSP].current + 4;
+            }
+            break;
+        }
+        return;
+    }
+
+    switch (uReason)
+    {
+    case FC_PUSH_VPM:
+        pAsm->CALLSTACK[pAsm->CALLSP].current++;
+        break;
+    case FC_PUSH_WQM:
+        pAsm->CALLSTACK[pAsm->CALLSP].current += 4;
+        break;
+    case FC_LOOP:
+        pAsm->CALLSTACK[pAsm->CALLSP].current += 4;
+        break;
+    case FC_REP:
+        /* TODO : for 16 vp asic, should += 2; */
+        pAsm->CALLSTACK[pAsm->CALLSP].current += 1;
+        break;
+    };
+
+    if(pAsm->CALLSTACK[pAsm->CALLSP].current
+         > pAsm->CALLSTACK[pAsm->CALLSP].max)
+    {
+        pAsm->CALLSTACK[pAsm->CALLSP].max =
+            pAsm->CALLSTACK[pAsm->CALLSP].current;
+    }
 }
 
-GLboolean AssembleInstr(GLuint uiNumberInsts,
-                        struct prog_instruction *pILInst, 
-                                               r700_AssemblerBase *pR700AsmCode)
+GLboolean jumpToOffest(r700_AssemblerBase *pAsm, GLuint pops, GLint offset)
 {
-    GLuint i;
-
-    pR700AsmCode->pILInst = pILInst;
-       for(i=0; i<uiNumberInsts; i++)
+    if(GL_FALSE == add_cf_instruction(pAsm) )
     {
-        pR700AsmCode->uiCurInst = i;
+        return GL_FALSE;
+    }
 
-        switch (pILInst[i].Opcode)
-        {
-        case OPCODE_ABS: 
-            if ( GL_FALSE == assemble_ABS(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
-        case OPCODE_ADD: 
-        case OPCODE_SUB: 
-            if ( GL_FALSE == assemble_ADD(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = pops;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
 
-        case OPCODE_ARL: 
-            if ( GL_FALSE == assemble_ARL(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;
-        case OPCODE_ARR: 
-            radeon_error("Not yet implemented instruction OPCODE_ARR \n");
-            //if ( GL_FALSE == assemble_BAD("ARR") ) 
-                return GL_FALSE;
-            break;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_JUMP;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
 
-        case OPCODE_CMP: 
-            if ( GL_FALSE == assemble_CMP(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
-        case OPCODE_COS: 
-            if ( GL_FALSE == assemble_COS(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
 
-        case OPCODE_DP3: 
-        case OPCODE_DP4: 
-        case OPCODE_DPH: 
-            if ( GL_FALSE == assemble_DOT(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
+    pAsm->cf_current_cf_clause_ptr->m_Word0.f.addr = pAsm->cf_current_cf_clause_ptr->m_uIndex + offset;
 
-        case OPCODE_DST: 
-            if ( GL_FALSE == assemble_DST(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
+    return GL_TRUE;
+}
 
-        case OPCODE_EX2: 
-            if ( GL_FALSE == assemble_EX2(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
-        case OPCODE_EXP: 
-            if ( GL_FALSE == assemble_EXP(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;
+GLboolean pops(r700_AssemblerBase *pAsm, GLuint pops)
+{
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
 
-        case OPCODE_FLR:     
-            if ( GL_FALSE == assemble_FLR(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
-        //case OP_FLR_INT: 
-        //    if ( GL_FALSE == assemble_FLR_INT() ) 
-        //        return GL_FALSE;
-        //    break;  
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = pops;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
 
-        case OPCODE_FRC: 
-            if ( GL_FALSE == assemble_FRC(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_POP;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
 
-        case OPCODE_KIL: 
-            if ( GL_FALSE == assemble_KIL(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;
-        case OPCODE_LG2: 
-            if ( GL_FALSE == assemble_LG2(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
-        case OPCODE_LIT:
-            if ( GL_FALSE == assemble_LIT(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;
-        case OPCODE_LRP: 
-            if ( GL_FALSE == assemble_LRP(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
-        case OPCODE_LOG: 
-            if ( GL_FALSE == assemble_LOG(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+    pAsm->cf_current_cf_clause_ptr->m_Word0.f.addr             = pAsm->cf_current_cf_clause_ptr->m_uIndex + 1;
 
-        case OPCODE_MAD: 
-            if ( GL_FALSE == assemble_MAD(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
-        case OPCODE_MAX: 
-            if ( GL_FALSE == assemble_MAX(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
-        case OPCODE_MIN: 
-            if ( GL_FALSE == assemble_MIN(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
+    return GL_TRUE;
+}
 
-        case OPCODE_MOV: 
-            if ( GL_FALSE == assemble_MOV(pR700AsmCode) ) 
-                return GL_FALSE;
-            break;  
-        case OPCODE_MUL: 
-            if ( GL_FALSE == assemble_MUL(pR700AsmCode) ) 
-                return GL_FALSE;
-            break; 
+GLboolean assemble_IF(r700_AssemblerBase *pAsm, GLboolean bHasElse)
+{
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU_PUSH_BEFORE;
 
-        case OPCODE_POW: 
+    assemble_LOGIC_PRED(pAsm, SQ_OP2_INST_PRED_SETNE);
+
+
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    if(GL_TRUE != bHasElse)
+    {
+        pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count = 1; 
+    }
+    else
+    {
+        pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count = 0;
+    }
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_JUMP;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+
+    pAsm->FCSP++;
+       pAsm->fc_stack[pAsm->FCSP].type  = FC_IF;
+    pAsm->fc_stack[pAsm->FCSP].mid   = NULL;
+    pAsm->fc_stack[pAsm->FCSP].midLen= 0;
+    pAsm->fc_stack[pAsm->FCSP].first = pAsm->cf_current_cf_clause_ptr;
+
+#ifndef USE_CF_FOR_POP_AFTER
+    if(GL_TRUE != bHasElse)
+    {
+        pAsm->alu_x_opcode = SQ_CF_INST_ALU_POP_AFTER;
+    }
+#endif /* USE_CF_FOR_POP_AFTER */
+
+    checkStackDepth(pAsm, FC_PUSH_VPM, GL_FALSE); 
+
+    return GL_TRUE;
+}
+
+GLboolean assemble_ELSE(r700_AssemblerBase *pAsm)
+{
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 1; ///
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_ELSE;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+
+    pAsm->fc_stack[pAsm->FCSP].mid = (R700ControlFlowGenericClause **)_mesa_realloc( (void *)pAsm->fc_stack[pAsm->FCSP].mid,
+                                                                                     0,
+                                                                                     sizeof(R700ControlFlowGenericClause *) );
+    pAsm->fc_stack[pAsm->FCSP].mid[0] = pAsm->cf_current_cf_clause_ptr;
+    //pAsm->fc_stack[pAsm->FCSP].unNumMid = 1;
+
+#ifndef USE_CF_FOR_POP_AFTER
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU_POP_AFTER;
+#endif /* USE_CF_FOR_POP_AFTER */
+
+    pAsm->fc_stack[pAsm->FCSP].first->m_Word0.f.addr = pAsm->pR700Shader->plstCFInstructions_active->uNumOfNode - 1; 
+
+    return GL_TRUE;
+}
+
+GLboolean assemble_ENDIF(r700_AssemblerBase *pAsm)
+{
+#ifdef USE_CF_FOR_POP_AFTER
+    pops(pAsm, 1); 
+#endif /* USE_CF_FOR_POP_AFTER */
+
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU;
+
+    if(NULL == pAsm->fc_stack[pAsm->FCSP].mid)
+    {
+        /* no else in between */
+        pAsm->fc_stack[pAsm->FCSP].first->m_Word0.f.addr = pAsm->pR700Shader->plstCFInstructions_active->uNumOfNode;
+    }
+    else
+    {
+        pAsm->fc_stack[pAsm->FCSP].mid[0]->m_Word0.f.addr = pAsm->pR700Shader->plstCFInstructions_active->uNumOfNode;
+    }
+
+    if(NULL != pAsm->fc_stack[pAsm->FCSP].mid)
+    {
+        FREE(pAsm->fc_stack[pAsm->FCSP].mid);
+    }
+
+    if(pAsm->fc_stack[pAsm->FCSP].type != FC_IF)
+    {
+        radeon_error("if/endif in shader code are not paired. \n");
+        return GL_FALSE;
+    }
+    
+    pAsm->FCSP--;
+
+    decreaseCurrent(pAsm, FC_PUSH_VPM);
+
+    return GL_TRUE;
+}
+
+GLboolean assemble_BGNLOOP(r700_AssemblerBase *pAsm)
+{
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_LOOP_START_NO_AL;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+
+    pAsm->FCSP++;
+       pAsm->fc_stack[pAsm->FCSP].type  = FC_LOOP;
+    pAsm->fc_stack[pAsm->FCSP].mid   = NULL;
+    pAsm->fc_stack[pAsm->FCSP].unNumMid = 0;
+    pAsm->fc_stack[pAsm->FCSP].midLen   = 0;
+    pAsm->fc_stack[pAsm->FCSP].first    = pAsm->cf_current_cf_clause_ptr;
+
+    checkStackDepth(pAsm, FC_LOOP, GL_FALSE);
+
+    return GL_TRUE;
+}
+
+GLboolean assemble_BRK(r700_AssemblerBase *pAsm)
+{
+#ifdef USE_CF_FOR_CONTINUE_BREAK
+
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU_PUSH_BEFORE;
+
+    assemble_LOGIC_PRED(pAsm, SQ_OP2_INST_PRED_SETNE);
+    
+    unsigned int unFCSP;
+    for(unFCSP=pAsm->FCSP; unFCSP>0; unFCSP--)
+    {
+        if(FC_LOOP == pAsm->fc_stack[unFCSP].type)
+        {
+            break;
+        }
+    }
+    if(0 == FC_LOOP)
+    {
+        radeon_error("Break is not inside loop/endloop pair.\n");
+        return GL_FALSE;
+    }
+
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 1;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_LOOP_BREAK;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+
+    pAsm->fc_stack[unFCSP].mid = (R700ControlFlowGenericClause **)_mesa_realloc( 
+                                              (void *)pAsm->fc_stack[unFCSP].mid,
+                                              sizeof(R700ControlFlowGenericClause *) * pAsm->fc_stack[unFCSP].unNumMid,
+                                              sizeof(R700ControlFlowGenericClause *) * (pAsm->fc_stack[unFCSP].unNumMid + 1) );
+    pAsm->fc_stack[unFCSP].mid[pAsm->fc_stack[unFCSP].unNumMid] = pAsm->cf_current_cf_clause_ptr;
+    pAsm->fc_stack[unFCSP].unNumMid++;
+
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 1;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_POP;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+    pAsm->cf_current_cf_clause_ptr->m_Word0.f.addr             = pAsm->cf_current_cf_clause_ptr->m_uIndex + 1;
+
+    checkStackDepth(pAsm, FC_PUSH_VPM, GL_TRUE);
+
+#endif //USE_CF_FOR_CONTINUE_BREAK
+    return GL_TRUE;
+}
+
+GLboolean assemble_CONT(r700_AssemblerBase *pAsm)
+{
+#ifdef USE_CF_FOR_CONTINUE_BREAK
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU_PUSH_BEFORE;
+
+    assemble_LOGIC_PRED(pAsm, SQ_OP2_INST_PRED_SETNE);
+
+    unsigned int unFCSP;
+    for(unFCSP=pAsm->FCSP; unFCSP>0; unFCSP--)
+    {
+        if(FC_LOOP == pAsm->fc_stack[unFCSP].type)
+        {
+            break;
+        }
+    }
+    if(0 == FC_LOOP)
+    {
+        radeon_error("Continue is not inside loop/endloop pair.\n");
+        return GL_FALSE;
+    }
+
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 1;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_LOOP_CONTINUE;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+
+    pAsm->fc_stack[unFCSP].mid = (R700ControlFlowGenericClause **)_mesa_realloc( 
+                                              (void *)pAsm->fc_stack[unFCSP].mid,
+                                              sizeof(R700ControlFlowGenericClause *) * pAsm->fc_stack[unFCSP].unNumMid,
+                                              sizeof(R700ControlFlowGenericClause *) * (pAsm->fc_stack[unFCSP].unNumMid + 1) );
+    pAsm->fc_stack[unFCSP].mid[pAsm->fc_stack[unFCSP].unNumMid] = pAsm->cf_current_cf_clause_ptr;
+    pAsm->fc_stack[unFCSP].unNumMid++;
+
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 1;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_POP;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+    pAsm->cf_current_cf_clause_ptr->m_Word0.f.addr             = pAsm->cf_current_cf_clause_ptr->m_uIndex + 1;
+
+    checkStackDepth(pAsm, FC_PUSH_VPM, GL_TRUE);
+
+#endif /* USE_CF_FOR_CONTINUE_BREAK */
+
+    return GL_TRUE;
+}
+
+GLboolean assemble_ENDLOOP(r700_AssemblerBase *pAsm)
+{
+    GLuint i;
+
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_LOOP_END;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word0.f.addr   = pAsm->fc_stack[pAsm->FCSP].first->m_uIndex + 1;
+    pAsm->fc_stack[pAsm->FCSP].first->m_Word0.f.addr = pAsm->cf_current_cf_clause_ptr->m_uIndex + 1;
+
+#ifdef USE_CF_FOR_CONTINUE_BREAK
+    for(i=0; i<pAsm->fc_stack[pAsm->FCSP].unNumMid; i++)
+    {
+        pAsm->fc_stack[pAsm->FCSP].mid[i]->m_Word0.f.addr = pAsm->cf_current_cf_clause_ptr->m_uIndex;
+    }
+    if(NULL != pAsm->fc_stack[pAsm->FCSP].mid)
+    {
+        FREE(pAsm->fc_stack[pAsm->FCSP].mid);
+    }
+#endif
+
+    if(pAsm->fc_stack[pAsm->FCSP].type != FC_LOOP)
+    {
+        radeon_error("loop/endloop in shader code are not paired. \n");
+        return GL_FALSE;
+    }
+
+    GLuint unFCSP;
+    GLuint unIF = 0;
+    if((pAsm->unCFflags & HAS_CURRENT_LOOPRET) > 0)
+    {        
+        for(unFCSP=(pAsm->FCSP-1); unFCSP>pAsm->CALLSTACK[pAsm->CALLSP].FCSP_BeforeEntry; unFCSP--)
+        {
+            if(FC_LOOP == pAsm->fc_stack[unFCSP].type)
+            {
+                breakLoopOnFlag(pAsm, unFCSP);
+                break;
+            }
+            else if(FC_IF == pAsm->fc_stack[unFCSP].type)
+            {
+                unIF++;
+            }
+        }
+        if(unFCSP <= pAsm->CALLSTACK[pAsm->CALLSP].FCSP_BeforeEntry)
+        {            
+#ifdef USE_CF_FOR_POP_AFTER
+            returnOnFlag(pAsm, unIF); 
+#else
+            returnOnFlag(pAsm, 0);
+#endif /* USE_CF_FOR_POP_AFTER */
+            pAsm->unCFflags &= ~HAS_CURRENT_LOOPRET;
+        }
+    }
+
+    pAsm->FCSP--;
+
+    decreaseCurrent(pAsm, FC_LOOP);
+    
+    return GL_TRUE;
+}
+
+void add_return_inst(r700_AssemblerBase *pAsm)
+{
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+    //pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 1;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_RETURN;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+}
+
+GLboolean assemble_BGNSUB(r700_AssemblerBase *pAsm, GLint nILindex, GLuint uiIL_Shift)
+{
+    /* Put in sub */
+    if( (pAsm->unSubArrayPointer + 1) > pAsm->unSubArraySize )
+    {
+        pAsm->subs = (SUB_OFFSET*)_mesa_realloc( (void *)pAsm->subs,
+                                  sizeof(SUB_OFFSET) * pAsm->unSubArraySize,
+                                  sizeof(SUB_OFFSET) * (pAsm->unSubArraySize + 10) );
+        if(NULL == pAsm->subs)
+        {
+            return GL_FALSE;
+        }
+        pAsm->unSubArraySize += 10;
+    }
+
+    pAsm->subs[pAsm->unSubArrayPointer].subIL_Offset = nILindex + uiIL_Shift;
+    pAsm->subs[pAsm->unSubArrayPointer].lstCFInstructions_local.pHead=NULL;  
+    pAsm->subs[pAsm->unSubArrayPointer].lstCFInstructions_local.pTail=NULL;  
+    pAsm->subs[pAsm->unSubArrayPointer].lstCFInstructions_local.uNumOfNode=0;
+
+    pAsm->CALLSP++;
+    pAsm->CALLSTACK[pAsm->CALLSP].subDescIndex = pAsm->unSubArrayPointer;
+    pAsm->CALLSTACK[pAsm->CALLSP].FCSP_BeforeEntry = pAsm->FCSP;
+    pAsm->CALLSTACK[pAsm->CALLSP].plstCFInstructions_local
+                   = &(pAsm->subs[pAsm->unSubArrayPointer].lstCFInstructions_local);
+    pAsm->CALLSTACK[pAsm->CALLSP].max = 0;
+    pAsm->CALLSTACK[pAsm->CALLSP].current = 0;
+    SetActiveCFlist(pAsm->pR700Shader, 
+                    pAsm->CALLSTACK[pAsm->CALLSP].plstCFInstructions_local);
+
+    pAsm->unSubArrayPointer++;
+
+    /* start sub */
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU;
+
+    pAsm->FCSP++;
+    pAsm->fc_stack[pAsm->FCSP].type  = FC_REP;
+
+    checkStackDepth(pAsm, FC_REP, GL_FALSE);
+
+    return GL_TRUE;
+}
+
+GLboolean assemble_ENDSUB(r700_AssemblerBase *pAsm)
+{
+    if(pAsm->fc_stack[pAsm->FCSP].type != FC_REP)
+    {
+        radeon_error("BGNSUB/ENDSUB in shader code are not paired. \n");
+        return GL_FALSE;
+    }
+
+    /* copy max to sub structure */
+    pAsm->subs[pAsm->CALLSTACK[pAsm->CALLSP].subDescIndex].unStackDepthMax
+        = pAsm->CALLSTACK[pAsm->CALLSP].max;
+
+    decreaseCurrent(pAsm, FC_REP);
+
+    pAsm->CALLSP--;
+    SetActiveCFlist(pAsm->pR700Shader, 
+                    pAsm->CALLSTACK[pAsm->CALLSP].plstCFInstructions_local);
+    
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU;
+
+    pAsm->FCSP--;
+
+    return GL_TRUE;
+}
+
+GLboolean assemble_RET(r700_AssemblerBase *pAsm)
+{
+    GLuint unIF = 0;
+
+    if(pAsm->CALLSP > 0)
+    {   /* in sub */
+        GLuint unFCSP;        
+        for(unFCSP=pAsm->FCSP; unFCSP>pAsm->CALLSTACK[pAsm->CALLSP].FCSP_BeforeEntry; unFCSP--)
+        {
+            if(FC_LOOP == pAsm->fc_stack[unFCSP].type)
+            {
+                setRetInLoopFlag(pAsm, SQ_SEL_1);
+                breakLoopOnFlag(pAsm, unFCSP);
+                pAsm->unCFflags |= LOOPRET_FLAGS;
+
+                return GL_TRUE;
+            }
+            else if(FC_IF == pAsm->fc_stack[unFCSP].type)
+            {
+                unIF++;
+            }
+        }
+    }
+
+#ifdef USE_CF_FOR_POP_AFTER    
+    if(unIF > 0)
+    {
+        pops(pAsm, unIF);
+    }
+#endif /* USE_CF_FOR_POP_AFTER */
+
+    add_return_inst(pAsm);
+
+    return GL_TRUE;
+}
+
+GLboolean assemble_CAL(r700_AssemblerBase *pAsm, 
+                       GLint nILindex,
+                       GLuint uiIL_Shift,
+                       GLuint uiNumberInsts,
+                       struct prog_instruction *pILInst,
+                       PRESUB_DESC * pPresubDesc)
+{
+    GLint uiIL_Offset;
+
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU;
+
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.call_count       = 1;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_CALL;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+
+    /* Put in caller */
+    if( (pAsm->unCallerArrayPointer + 1) > pAsm->unCallerArraySize )
+    {
+        pAsm->callers = (CALLER_POINTER*)_mesa_realloc( (void *)pAsm->callers, 
+                       sizeof(CALLER_POINTER) * pAsm->unCallerArraySize, 
+                       sizeof(CALLER_POINTER) * (pAsm->unCallerArraySize + 10) );
+        if(NULL == pAsm->callers)
+        {
+            return GL_FALSE;
+        }
+        pAsm->unCallerArraySize += 10;
+    }
+    
+    uiIL_Offset = nILindex + uiIL_Shift;
+    pAsm->callers[pAsm->unCallerArrayPointer].subIL_Offset = uiIL_Offset; 
+    pAsm->callers[pAsm->unCallerArrayPointer].cf_ptr       = pAsm->cf_current_cf_clause_ptr;
+    
+    pAsm->callers[pAsm->unCallerArrayPointer].finale_cf_ptr  = NULL; 
+    pAsm->callers[pAsm->unCallerArrayPointer].prelude_cf_ptr = NULL; 
+
+    pAsm->unCallerArrayPointer++;
+
+    int j;
+    GLuint max;
+    GLuint unSubID;
+    GLboolean bRet;
+    for(j=0; j<pAsm->unSubArrayPointer; j++)
+    {
+        if(uiIL_Offset == pAsm->subs[j].subIL_Offset)
+        {   /* compiled before */
+
+            max = pAsm->subs[j].unStackDepthMax 
+                + pAsm->CALLSTACK[pAsm->CALLSP].current;
+            if(max > pAsm->CALLSTACK[pAsm->CALLSP].max)
+            {
+                pAsm->CALLSTACK[pAsm->CALLSP].max = max;
+            }
+            
+            pAsm->callers[pAsm->unCallerArrayPointer - 1].subDescIndex = j; 
+            return GL_TRUE;
+        }
+    }
+
+    pAsm->callers[pAsm->unCallerArrayPointer - 1].subDescIndex = pAsm->unSubArrayPointer;
+    unSubID = pAsm->unSubArrayPointer;
+
+    bRet = AssembleInstr(nILindex, uiIL_Shift, uiNumberInsts, pILInst, pAsm);
+
+    if(GL_TRUE == bRet)
+    {
+        max = pAsm->subs[unSubID].unStackDepthMax 
+            + pAsm->CALLSTACK[pAsm->CALLSP].current;
+        if(max > pAsm->CALLSTACK[pAsm->CALLSP].max)
+        {
+            pAsm->CALLSTACK[pAsm->CALLSP].max = max;
+        }
+
+        pAsm->subs[unSubID].pPresubDesc = pPresubDesc;
+    }
+
+    return bRet;
+}
+
+GLboolean setRetInLoopFlag(r700_AssemblerBase *pAsm, GLuint flagValue)
+{
+    GLfloat fLiteral[2] = {0.1, 0.0};
+
+    pAsm->D.dst.opcode   = SQ_OP2_INST_MOV;
+    pAsm->D.dst.op3      = 0;
+    pAsm->D.dst.rtype    = DST_REG_TEMPORARY;
+    pAsm->D.dst.reg      = pAsm->flag_reg_index;
+    pAsm->D.dst.writex   = 1;
+    pAsm->D.dst.writey   = 0;
+    pAsm->D.dst.writez   = 0;
+    pAsm->D.dst.writew   = 0;
+    pAsm->D2.dst2.literal_slots      = 1;
+    pAsm->D2.dst2.SaturateMode = SATURATE_OFF;
+    pAsm->D.dst.predicated     = 0;
+    /* in reloc where dislink flag init inst, only one slot alu inst is handled. */
+    pAsm->D.dst.math           = 1; /* TODO : not math really, but one channel op, more generic alu assembler needed */
+    pAsm->D2.dst2.index_mode = SQ_INDEX_LOOP; /* Check this ! */
+#if 0
+    pAsm->S[0].src.rtype = SRC_REC_LITERAL;
+    //pAsm->S[0].src.reg   = 0;
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    noneg_PVSSRC(&(pAsm->S[0].src));
+    pAsm->S[0].src.swizzlex = SQ_SEL_X;
+    pAsm->S[0].src.swizzley = SQ_SEL_Y;
+    pAsm->S[0].src.swizzlez = SQ_SEL_Z;
+    pAsm->S[0].src.swizzlew = SQ_SEL_W;
+
+    if( GL_FALSE == next_ins_literal(pAsm, &(fLiteral[0])) )
+    {
+        return GL_FALSE;
+    }
+#else
+    pAsm->S[0].src.rtype = DST_REG_TEMPORARY;
+    pAsm->S[0].src.reg   = 0;
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    noneg_PVSSRC(&(pAsm->S[0].src));
+    pAsm->S[0].src.swizzlex = flagValue;
+    pAsm->S[0].src.swizzley = flagValue;
+    pAsm->S[0].src.swizzlez = flagValue;
+    pAsm->S[0].src.swizzlew = flagValue;
+
+    if( GL_FALSE == next_ins(pAsm) )
+    {
+        return GL_FALSE;
+    }
+#endif
+
+    return GL_TRUE;
+}
+
+GLboolean testFlag(r700_AssemblerBase *pAsm)
+{
+    GLfloat fLiteral[2] = {0.1, 0.0};
+
+    //Test flag
+    GLuint tmp = gethelpr(pAsm);
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU_PUSH_BEFORE;
+
+    pAsm->D.dst.opcode   = SQ_OP2_INST_PRED_SETE;
+    pAsm->D.dst.math     = 1;
+    pAsm->D.dst.rtype    = DST_REG_TEMPORARY;
+    pAsm->D.dst.reg      = tmp;
+    pAsm->D.dst.writex   = 1;
+    pAsm->D.dst.writey   = 0;
+    pAsm->D.dst.writez   = 0;
+    pAsm->D.dst.writew   = 0;
+    pAsm->D2.dst2.literal_slots      = 1;
+    pAsm->D2.dst2.SaturateMode = SATURATE_OFF;
+    pAsm->D.dst.predicated     = 1;
+    pAsm->D2.dst2.index_mode = SQ_INDEX_LOOP; /* Check this ! */
+
+    pAsm->S[0].src.rtype = DST_REG_TEMPORARY;
+    pAsm->S[0].src.reg   = pAsm->flag_reg_index;
+    setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+    noneg_PVSSRC(&(pAsm->S[0].src));
+    pAsm->S[0].src.swizzlex = SQ_SEL_X;
+    pAsm->S[0].src.swizzley = SQ_SEL_Y;
+    pAsm->S[0].src.swizzlez = SQ_SEL_Z;
+    pAsm->S[0].src.swizzlew = SQ_SEL_W;
+#if 0
+    pAsm->S[1].src.rtype = SRC_REC_LITERAL;
+    //pAsm->S[1].src.reg   = 0;
+    setaddrmode_PVSSRC(&(pAsm->S[1].src), ADDR_ABSOLUTE);
+    noneg_PVSSRC(&(pAsm->S[1].src));
+    pAsm->S[1].src.swizzlex = SQ_SEL_X;
+    pAsm->S[1].src.swizzley = SQ_SEL_Y;
+    pAsm->S[1].src.swizzlez = SQ_SEL_Z;
+    pAsm->S[1].src.swizzlew = SQ_SEL_W;
+
+    if( GL_FALSE == next_ins_literal(pAsm, &(fLiteral[0])) )
+    {
+        return GL_FALSE;
+    }
+#else
+    pAsm->S[1].src.rtype = DST_REG_TEMPORARY;
+    pAsm->S[1].src.reg   = 0;
+    setaddrmode_PVSSRC(&(pAsm->S[1].src), ADDR_ABSOLUTE);
+    noneg_PVSSRC(&(pAsm->S[1].src));
+    pAsm->S[1].src.swizzlex = SQ_SEL_1;
+    pAsm->S[1].src.swizzley = SQ_SEL_1;
+    pAsm->S[1].src.swizzlez = SQ_SEL_1;
+    pAsm->S[1].src.swizzlew = SQ_SEL_1;
+
+    if( GL_FALSE == next_ins(pAsm) )
+    {
+        return GL_FALSE;
+    }
+#endif
+
+    checkStackDepth(pAsm, FC_PUSH_VPM, GL_TRUE);
+
+    return GL_TRUE;
+}
+
+GLboolean returnOnFlag(r700_AssemblerBase *pAsm, GLuint unIF)
+{
+    testFlag(pAsm);
+    jumpToOffest(pAsm, 1, 4);
+    setRetInLoopFlag(pAsm, SQ_SEL_0);
+    pops(pAsm, unIF + 1);
+    add_return_inst(pAsm);
+
+    return GL_TRUE;
+}
+
+GLboolean breakLoopOnFlag(r700_AssemblerBase *pAsm, GLuint unFCSP)
+{
+    testFlag(pAsm);
+    //break
+    if(GL_FALSE == add_cf_instruction(pAsm) )
+    {
+        return GL_FALSE;
+    }
+    
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.pop_count        = 1;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_const         = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cond             = SQ_CF_COND_ACTIVE;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.end_of_program   = 0x0;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.valid_pixel_mode = 0x0; 
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.cf_inst          = SQ_CF_INST_LOOP_BREAK;
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.whole_quad_mode  = 0x0;
+
+    pAsm->cf_current_cf_clause_ptr->m_Word1.f.barrier          = 0x1;
+
+    pAsm->fc_stack[unFCSP].mid = (R700ControlFlowGenericClause **)_mesa_realloc( 
+                                              (void *)pAsm->fc_stack[unFCSP].mid,
+                                              sizeof(R700ControlFlowGenericClause *) * pAsm->fc_stack[unFCSP].unNumMid,
+                                              sizeof(R700ControlFlowGenericClause *) * (pAsm->fc_stack[unFCSP].unNumMid + 1) );
+    pAsm->fc_stack[unFCSP].mid[pAsm->fc_stack[unFCSP].unNumMid] = pAsm->cf_current_cf_clause_ptr;
+    pAsm->fc_stack[unFCSP].unNumMid++;
+
+    pops(pAsm, 1);
+               
+    return GL_TRUE;
+}
+
+GLboolean AssembleInstr(GLuint uiFirstInst,
+                        GLuint uiIL_Shift,
+                        GLuint uiNumberInsts,
+                        struct prog_instruction *pILInst, 
+                                               r700_AssemblerBase *pR700AsmCode)
+{
+    GLuint i;
+
+    pR700AsmCode->pILInst = pILInst;
+       for(i=uiFirstInst; i<uiNumberInsts; i++)
+    {
+        pR700AsmCode->uiCurInst = i;
+
+#ifndef USE_CF_FOR_CONTINUE_BREAK
+        if(OPCODE_BRK == pILInst[i+1].Opcode)
+        {
+            switch(pILInst[i].Opcode)            
+            {
+            case OPCODE_SLE:
+                pILInst[i].Opcode = OPCODE_SGT;
+                break;
+            case OPCODE_SLT:
+                pILInst[i].Opcode = OPCODE_SGE;
+                break;
+            case OPCODE_SGE:
+                pILInst[i].Opcode = OPCODE_SLT;
+                break;
+            case OPCODE_SGT:
+                pILInst[i].Opcode = OPCODE_SLE;
+                break;
+            case OPCODE_SEQ:
+                pILInst[i].Opcode = OPCODE_SNE;
+                break;
+            case OPCODE_SNE:
+                pILInst[i].Opcode = OPCODE_SEQ;
+                break;
+            default:
+                break;
+            }
+        }
+#endif
+        if(pILInst[i].CondUpdate == 1)
+        {
+            /* remember dest register used for cond evaluation */
+            /* XXX also handle PROGRAM_OUTPUT registers here? */
+            pR700AsmCode->last_cond_register = pILInst[i].DstReg.Index; 
+        }
+
+        switch (pILInst[i].Opcode)
+        {
+        case OPCODE_ABS: 
+            if ( GL_FALSE == assemble_ABS(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+        case OPCODE_ADD: 
+        case OPCODE_SUB: 
+            if ( GL_FALSE == assemble_ADD(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+
+        case OPCODE_ARL: 
+            if ( GL_FALSE == assemble_ARL(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;
+        case OPCODE_ARR: 
+            radeon_error("Not yet implemented instruction OPCODE_ARR \n");
+            //if ( GL_FALSE == assemble_BAD("ARR") ) 
+                return GL_FALSE;
+            break;
+
+        case OPCODE_CMP: 
+            if ( GL_FALSE == assemble_CMP(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+        case OPCODE_COS: 
+            if ( GL_FALSE == assemble_TRIG(pR700AsmCode, SQ_OP2_INST_COS) ) 
+                return GL_FALSE;
+            break;  
+
+        case OPCODE_DP3: 
+        case OPCODE_DP4: 
+        case OPCODE_DPH: 
+            if ( GL_FALSE == assemble_DOT(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+
+        case OPCODE_DST: 
+            if ( GL_FALSE == assemble_DST(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+
+        case OPCODE_EX2: 
+            if ( GL_FALSE == assemble_EX2(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+        case OPCODE_EXP: 
+            if ( GL_FALSE == assemble_EXP(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;
+
+        case OPCODE_FLR:     
+            if ( GL_FALSE == assemble_FLR(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+        //case OP_FLR_INT: ;
+
+        //    if ( GL_FALSE == assemble_FLR_INT() ) 
+        //        return GL_FALSE;
+        //    break;  
+
+        case OPCODE_FRC: 
+            if ( GL_FALSE == assemble_FRC(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+
+        case OPCODE_KIL: 
+        case OPCODE_KIL_NV: 
+            if ( GL_FALSE == assemble_KIL(pR700AsmCode, SQ_OP2_INST_KILLGT) ) 
+                return GL_FALSE;
+            break;
+        case OPCODE_LG2: 
+            if ( GL_FALSE == assemble_LG2(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+        case OPCODE_LIT:
+            if ( GL_FALSE == assemble_LIT(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;
+        case OPCODE_LRP: 
+            if ( GL_FALSE == assemble_LRP(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+        case OPCODE_LOG: 
+            if ( GL_FALSE == assemble_LOG(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;
+
+        case OPCODE_MAD: 
+            if ( GL_FALSE == assemble_MAD(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+        case OPCODE_MAX: 
+            if ( GL_FALSE == assemble_MAX(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+        case OPCODE_MIN: 
+            if ( GL_FALSE == assemble_MIN(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+
+        case OPCODE_MOV: 
+            if ( GL_FALSE == assemble_MOV(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;  
+        case OPCODE_MUL: 
+            if ( GL_FALSE == assemble_MUL(pR700AsmCode) ) 
+                return GL_FALSE;
+            break;
+            
+        case OPCODE_NOISE1:
+            {                                               
+                callPreSub(pR700AsmCode, 
+                           GLSL_NOISE1,                         
+                           &noise1_presub,                                                  
+                           pILInst->DstReg.Index + pR700AsmCode->starting_temp_register_number, 
+                           1); 
+                radeon_error("noise1: not yet supported shader instruction\n");
+            };
+            break; 
+        case OPCODE_NOISE2: 
+            radeon_error("noise2: not yet supported shader instruction\n");
+            break; 
+        case OPCODE_NOISE3: 
+            radeon_error("noise3: not yet supported shader instruction\n");
+            break; 
+        case OPCODE_NOISE4: 
+            radeon_error("noise4: not yet supported shader instruction\n");
+            break; 
+
+        case OPCODE_POW: 
             if ( GL_FALSE == assemble_POW(pR700AsmCode) ) 
                 return GL_FALSE;
             break;  
@@ -4408,22 +5601,78 @@ GLboolean AssembleInstr(GLuint uiNumberInsts,
                 return GL_FALSE;
             break;  
         case OPCODE_SIN: 
-            if ( GL_FALSE == assemble_SIN(pR700AsmCode) ) 
+            if ( GL_FALSE == assemble_TRIG(pR700AsmCode, SQ_OP2_INST_SIN) ) 
                 return GL_FALSE;
             break;  
         case OPCODE_SCS: 
             if ( GL_FALSE == assemble_SCS(pR700AsmCode) ) 
                 return GL_FALSE;
-            break;  
+            break; 
+            
+        case OPCODE_SEQ:
+            if ( GL_FALSE == assemble_LOGIC(pR700AsmCode, SQ_OP2_INST_SETE) ) 
+            {
+                return GL_FALSE;
+            }
+            break;
+
+        case OPCODE_SGT: 
+            if ( GL_FALSE == assemble_LOGIC(pR700AsmCode, SQ_OP2_INST_SETGT) ) 
+            {
+                return GL_FALSE;
+            }
+            break;
 
         case OPCODE_SGE: 
             if ( GL_FALSE == assemble_SGE(pR700AsmCode) ) 
+            { 
                 return GL_FALSE;
-            break; 
+            }
+            break;
+        
+        /* NO LT, LE, TODO : use GE => LE, GT => LT : reverse 2 src order would be simpliest. Or use SQ_CF_COND_FALSE for SQ_CF_COND_ACTIVE.*/
         case OPCODE_SLT: 
-            if ( GL_FALSE == assemble_SLT(pR700AsmCode) ) 
+            {
+                struct prog_src_register SrcRegSave[2];
+                SrcRegSave[0] = pILInst[i].SrcReg[0];
+                SrcRegSave[1] = pILInst[i].SrcReg[1];
+                pILInst[i].SrcReg[0] = SrcRegSave[1];
+                pILInst[i].SrcReg[1] = SrcRegSave[0];
+                if ( GL_FALSE == assemble_LOGIC(pR700AsmCode, SQ_OP2_INST_SETGT) ) 
+                {
+                    pILInst[i].SrcReg[0] = SrcRegSave[0];
+                    pILInst[i].SrcReg[1] = SrcRegSave[1];
+                    return GL_FALSE;
+                }
+                pILInst[i].SrcReg[0] = SrcRegSave[0];
+                pILInst[i].SrcReg[1] = SrcRegSave[1];
+            }
+            break;
+
+        case OPCODE_SLE: 
+            {
+                struct prog_src_register SrcRegSave[2];
+                SrcRegSave[0] = pILInst[i].SrcReg[0];
+                SrcRegSave[1] = pILInst[i].SrcReg[1];
+                pILInst[i].SrcReg[0] = SrcRegSave[1];
+                pILInst[i].SrcReg[1] = SrcRegSave[0];
+                if ( GL_FALSE == assemble_LOGIC(pR700AsmCode, SQ_OP2_INST_SETGE) ) 
+                {
+                    pILInst[i].SrcReg[0] = SrcRegSave[0];
+                    pILInst[i].SrcReg[1] = SrcRegSave[1];
+                    return GL_FALSE;
+                }
+                pILInst[i].SrcReg[0] = SrcRegSave[0];
+                pILInst[i].SrcReg[1] = SrcRegSave[1];
+            }
+            break;
+
+        case OPCODE_SNE: 
+            if ( GL_FALSE == assemble_LOGIC(pR700AsmCode, SQ_OP2_INST_SETNE) ) 
+            {
                 return GL_FALSE;
-            break; 
+            }
+            break;
 
         //case OP_STP: 
         //    if ( GL_FALSE == assemble_STP(pR700AsmCode) ) 
@@ -4449,7 +5698,8 @@ GLboolean AssembleInstr(GLuint uiNumberInsts,
                 }
             }
             break;
-
+        case OPCODE_DDX:
+        case OPCODE_DDY:
         case OPCODE_TEX: 
         case OPCODE_TXB:  
         case OPCODE_TXP: 
@@ -4457,30 +5707,104 @@ GLboolean AssembleInstr(GLuint uiNumberInsts,
                 return GL_FALSE;
             break;
 
+        case OPCODE_TRUNC:
+            if ( GL_FALSE == assemble_math_function(pR700AsmCode, SQ_OP2_INST_TRUNC) )
+                return GL_FALSE;
+            break;
+
         case OPCODE_XPD: 
             if ( GL_FALSE == assemble_XPD(pR700AsmCode) ) 
                 return GL_FALSE;
             break;  
 
-        case OPCODE_IF   : 
-            if ( GL_FALSE == assemble_IF(pR700AsmCode) ) 
-                return GL_FALSE;
+        case OPCODE_IF:
+            {                
+                GLboolean bHasElse = GL_FALSE;
+
+                if(pILInst[pILInst[i].BranchTarget].Opcode == OPCODE_ELSE)
+                {
+                    bHasElse = GL_TRUE;
+                }
+
+                if ( GL_FALSE == assemble_IF(pR700AsmCode, bHasElse) ) 
+                {
+                    return GL_FALSE;
+                }
+            }
             break;
+
         case OPCODE_ELSE : 
-            radeon_error("Not yet implemented instruction OPCODE_ELSE \n");
-            //if ( GL_FALSE == assemble_BAD("ELSE") ) 
+            if ( GL_FALSE == assemble_ELSE(pR700AsmCode) ) 
                 return GL_FALSE;
             break;
+
         case OPCODE_ENDIF: 
             if ( GL_FALSE == assemble_ENDIF(pR700AsmCode) ) 
                 return GL_FALSE;
             break;
 
+        case OPCODE_BGNLOOP:
+            if( GL_FALSE == assemble_BGNLOOP(pR700AsmCode) )
+            {
+                return GL_FALSE;
+            }
+            break;
+
+        case OPCODE_BRK:
+            if( GL_FALSE == assemble_BRK(pR700AsmCode) )
+            {
+                return GL_FALSE;
+            }
+            break;
+
+        case OPCODE_CONT:
+            if( GL_FALSE == assemble_CONT(pR700AsmCode) )
+            {
+                return GL_FALSE;
+            }
+            break;
+
+        case OPCODE_ENDLOOP:
+            if( GL_FALSE == assemble_ENDLOOP(pR700AsmCode) )
+            {
+                return GL_FALSE;
+            }
+            break;
+
+        case OPCODE_BGNSUB:
+            if( GL_FALSE == assemble_BGNSUB(pR700AsmCode, i, uiIL_Shift) )
+            {
+                return GL_FALSE;
+            }
+            break;
+        
+        case OPCODE_RET:
+            if( GL_FALSE == assemble_RET(pR700AsmCode) )
+            {
+                return GL_FALSE;
+            }
+            break;
+        
+        case OPCODE_CAL:
+            if( GL_FALSE == assemble_CAL(pR700AsmCode, 
+                                         pILInst[i].BranchTarget,
+                                         uiIL_Shift,
+                                         uiNumberInsts,
+                                         pILInst,
+                                         NULL) )
+            {
+                return GL_FALSE;
+            }
+            break;
+
         //case OPCODE_EXPORT: 
         //    if ( GL_FALSE == assemble_EXPORT() ) 
         //        return GL_FALSE;
         //    break;
 
+        case OPCODE_ENDSUB:
+            return assemble_ENDSUB(pR700AsmCode);
+
         case OPCODE_END: 
                        //pR700AsmCode->uiCurInst = i;
                        //This is to remaind that if in later exoort there is depth/stencil
@@ -4497,6 +5821,417 @@ GLboolean AssembleInstr(GLuint uiNumberInsts,
     return GL_TRUE;
 }
 
+GLboolean InitShaderProgram(r700_AssemblerBase * pAsm)
+{
+    setRetInLoopFlag(pAsm, SQ_SEL_0);
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU;
+    return GL_TRUE;
+}
+
+GLboolean RelocProgram(r700_AssemblerBase * pAsm, struct gl_program * pILProg)
+{
+    GLuint i;
+    GLuint unCFoffset;
+    TypedShaderList * plstCFmain;
+    TypedShaderList * plstCFsub;
+
+    R700ShaderInstruction *        pInst;
+    R700ControlFlowGenericClause * pCFInst;
+
+    R700ControlFlowALUClause * pCF_ALU;
+    R700ALUInstruction       * pALU;
+    GLuint                     unConstOffset = 0;
+    GLuint                     unRegOffset;
+    GLuint                     unMinRegIndex;
+
+    plstCFmain = pAsm->CALLSTACK[0].plstCFInstructions_local;
+
+    /* remove flags init if they are not used */
+    if((pAsm->unCFflags & HAS_LOOPRET) == 0)
+    {
+        R700ControlFlowALUClause * pCF_ALU;
+        pInst = plstCFmain->pHead;
+        while(pInst)
+        {
+            if(SIT_CF_ALU == pInst->m_ShaderInstType)
+            {
+                pCF_ALU = (R700ControlFlowALUClause *)pInst;
+                if(0 == pCF_ALU->m_Word1.f.count)
+                {
+                    pCF_ALU->m_Word1.f.cf_inst = SQ_CF_INST_NOP;
+                }
+                else
+                {
+                    R700ALUInstruction * pALU = pCF_ALU->m_pLinkedALUInstruction;
+                    
+                    pALU->m_pLinkedALUClause = NULL;
+                    pALU = (R700ALUInstruction *)(pALU->pNextInst);
+                    pALU->m_pLinkedALUClause = pCF_ALU;
+                    pCF_ALU->m_pLinkedALUInstruction = pALU;
+
+                    pCF_ALU->m_Word1.f.count--;
+                }
+                break;
+            }
+            pInst = pInst->pNextInst;
+        };
+    }
+
+    if(pAsm->CALLSTACK[0].max > 0)
+    {
+        pAsm->pR700Shader->uStackSize = ((pAsm->CALLSTACK[0].max + 3)>>2) + 2;
+    }
+
+    if(0 == pAsm->unSubArrayPointer)
+    {
+        return GL_TRUE;
+    }
+
+    unCFoffset = plstCFmain->uNumOfNode;
+
+    if(NULL != pILProg->Parameters)
+    {        
+        unConstOffset = pILProg->Parameters->NumParameters;
+    }
+
+    /* Reloc subs */
+    for(i=0; i<pAsm->unSubArrayPointer; i++)
+    {
+        pAsm->subs[i].unCFoffset = unCFoffset;
+        plstCFsub = &(pAsm->subs[i].lstCFInstructions_local);
+
+        pInst = plstCFsub->pHead;
+
+        /* reloc instructions */
+        while(pInst)
+        {
+            if(SIT_CF_GENERIC == pInst->m_ShaderInstType)
+            {
+                pCFInst = (R700ControlFlowGenericClause *)pInst;
+
+                switch (pCFInst->m_Word1.f.cf_inst)
+                {
+                case SQ_CF_INST_POP:
+                case SQ_CF_INST_JUMP:
+                case SQ_CF_INST_ELSE:
+                case SQ_CF_INST_LOOP_END:
+                case SQ_CF_INST_LOOP_START:
+                case SQ_CF_INST_LOOP_START_NO_AL:
+                case SQ_CF_INST_LOOP_CONTINUE:
+                case SQ_CF_INST_LOOP_BREAK:
+                    pCFInst->m_Word0.f.addr += unCFoffset;
+                    break;
+                default:
+                    break;
+                }
+            }  
+            
+            pInst->m_uIndex += unCFoffset;
+
+            pInst = pInst->pNextInst;
+        };
+
+        if(NULL != pAsm->subs[i].pPresubDesc)
+        {
+            GLuint                     uNumSrc;            
+            
+            unMinRegIndex  = pAsm->subs[i].pPresubDesc->pCompiledSub->MinRegIndex;
+            unRegOffset    = pAsm->subs[i].pPresubDesc->maxStartReg;            
+            unConstOffset += pAsm->subs[i].pPresubDesc->unConstantsStart;
+
+            pInst = plstCFsub->pHead;
+            while(pInst)
+            {
+                if(SIT_CF_ALU == pInst->m_ShaderInstType)
+                {
+                    pCF_ALU = (R700ControlFlowALUClause *)pInst;
+
+                    pALU = pCF_ALU->m_pLinkedALUInstruction;
+                    for(int j=0; j<=pCF_ALU->m_Word1.f.count; j++)
+                    {
+                        pALU->m_Word1.f.dst_gpr = pALU->m_Word1.f.dst_gpr + unRegOffset - unMinRegIndex;
+
+                        if(pALU->m_Word0.f.src0_sel < SQ_ALU_SRC_GPR_SIZE)
+                        {   
+                            pALU->m_Word0.f.src0_sel = pALU->m_Word0.f.src0_sel + unRegOffset - unMinRegIndex;
+                        }
+                        else if(pALU->m_Word0.f.src0_sel >= SQ_ALU_SRC_CFILE_BASE)
+                        {   
+                            pALU->m_Word0.f.src0_sel += unConstOffset;
+                        }
+
+                        if( ((pALU->m_Word1.val >> SQ_ALU_WORD1_OP3_ALU_INST_SHIFT) & 0x0000001F) 
+                            >= SQ_OP3_INST_MUL_LIT )
+                        {   /* op3 : 3 srcs */
+                            if(pALU->m_Word1_OP3.f.src2_sel < SQ_ALU_SRC_GPR_SIZE)
+                            {   
+                                pALU->m_Word1_OP3.f.src2_sel = pALU->m_Word1_OP3.f.src2_sel + unRegOffset - unMinRegIndex;
+                            }
+                            else if(pALU->m_Word1_OP3.f.src2_sel >= SQ_ALU_SRC_CFILE_BASE)
+                            {   
+                                pALU->m_Word1_OP3.f.src2_sel += unConstOffset;
+                            }    
+                            if(pALU->m_Word0.f.src1_sel < SQ_ALU_SRC_GPR_SIZE)
+                            {   
+                                pALU->m_Word0.f.src1_sel = pALU->m_Word0.f.src1_sel + unRegOffset - unMinRegIndex;
+                            }
+                            else if(pALU->m_Word0.f.src1_sel >= SQ_ALU_SRC_CFILE_BASE)
+                            {   
+                                pALU->m_Word0.f.src1_sel += unConstOffset;
+                            }                                 
+                        }
+                        else
+                        {
+                            if(pAsm->bR6xx)
+                            {
+                                uNumSrc = r700GetNumOperands(pALU->m_Word1_OP2.f6.alu_inst, 0);
+                            }
+                            else
+                            {
+                                uNumSrc = r700GetNumOperands(pALU->m_Word1_OP2.f.alu_inst, 0);
+                            }
+                            if(2 == uNumSrc)
+                            {   /* 2 srcs */
+                                if(pALU->m_Word0.f.src1_sel < SQ_ALU_SRC_GPR_SIZE)
+                                {   
+                                    pALU->m_Word0.f.src1_sel = pALU->m_Word0.f.src1_sel + unRegOffset - unMinRegIndex;
+                                }
+                                else if(pALU->m_Word0.f.src1_sel >= SQ_ALU_SRC_CFILE_BASE)
+                                {   
+                                    pALU->m_Word0.f.src1_sel += unConstOffset;
+                                }                                  
+                            }                            
+                        }
+                        pALU = (R700ALUInstruction*)(pALU->pNextInst);
+                    }                    
+                }             
+                pInst = pInst->pNextInst;
+            };
+        }
+
+        /* Put sub into main */
+        plstCFmain->pTail->pNextInst = plstCFsub->pHead;
+        plstCFmain->pTail            = plstCFsub->pTail;
+        plstCFmain->uNumOfNode      += plstCFsub->uNumOfNode;
+
+        unCFoffset += plstCFsub->uNumOfNode;
+    }
+
+    /* reloc callers */
+    for(i=0; i<pAsm->unCallerArrayPointer; i++)
+    {
+        pAsm->callers[i].cf_ptr->m_Word0.f.addr
+            = pAsm->subs[pAsm->callers[i].subDescIndex].unCFoffset; 
+
+        if(NULL != pAsm->subs[pAsm->callers[i].subDescIndex].pPresubDesc)
+        {                 
+            unMinRegIndex = pAsm->subs[pAsm->callers[i].subDescIndex].pPresubDesc->pCompiledSub->MinRegIndex;
+            unRegOffset = pAsm->subs[pAsm->callers[i].subDescIndex].pPresubDesc->maxStartReg;
+
+            if(NULL != pAsm->callers[i].prelude_cf_ptr)
+            {                
+                pCF_ALU = (R700ControlFlowALUClause * )(pAsm->callers[i].prelude_cf_ptr);
+                pALU = pCF_ALU->m_pLinkedALUInstruction;
+                for(int j=0; j<=pCF_ALU->m_Word1.f.count; j++)
+                {
+                    pALU->m_Word1.f.dst_gpr = pALU->m_Word1.f.dst_gpr + unRegOffset - unMinRegIndex;
+                    pALU = (R700ALUInstruction*)(pALU->pNextInst);
+                }
+            }
+            if(NULL != pAsm->callers[i].finale_cf_ptr)
+            {
+                pCF_ALU = (R700ControlFlowALUClause * )(pAsm->callers[i].finale_cf_ptr);
+                pALU = pCF_ALU->m_pLinkedALUInstruction;
+                for(int j=0; j<=pCF_ALU->m_Word1.f.count; j++)
+                {
+                    pALU->m_Word0.f.src0_sel = pALU->m_Word0.f.src0_sel + unRegOffset - unMinRegIndex;
+                    pALU = (R700ALUInstruction*)(pALU->pNextInst);
+                }
+            }
+        }
+    }
+
+    return GL_TRUE;
+}
+
+GLboolean callPreSub(r700_AssemblerBase* pAsm, 
+                         LOADABLE_SCRIPT_SIGNITURE scriptSigniture,                          
+                         COMPILED_SUB * pCompiledSub,                                               
+                         GLshort uOutReg,
+                         GLshort uNumValidSrc)
+{
+    /* save assemble context */
+    GLuint starting_temp_register_number_save;
+    GLuint number_used_registers_save;
+    GLuint uFirstHelpReg_save;
+    GLuint uHelpReg_save;
+    GLuint uiCurInst_save;
+    struct prog_instruction *pILInst_save;
+    PRESUB_DESC * pPresubDesc;
+    GLboolean     bRet;
+    int i;
+
+    R700ControlFlowGenericClause* prelude_cf_ptr = NULL;
+
+    /* copy srcs to presub inputs */  
+    pAsm->alu_x_opcode = SQ_CF_INST_ALU;
+    for(i=0; i<uNumValidSrc; i++)
+    {
+        pAsm->D.dst.opcode = SQ_OP2_INST_MOV;
+        setaddrmode_PVSDST(&(pAsm->D.dst), ADDR_ABSOLUTE);
+        pAsm->D.dst.rtype = DST_REG_TEMPORARY;
+        pAsm->D.dst.reg   = pCompiledSub->srcRegIndex[i];
+        pAsm->D.dst.writex = 1;
+        pAsm->D.dst.writey = 1;
+        pAsm->D.dst.writez = 1;
+        pAsm->D.dst.writew = 1;
+
+        if( GL_FALSE == assemble_src(pAsm, i, 0) )
+        {
+            return GL_FALSE;
+        }
+
+        next_ins(pAsm);
+    }
+    if(uNumValidSrc > 0)
+    {
+        prelude_cf_ptr     = pAsm->cf_current_alu_clause_ptr;
+        pAsm->alu_x_opcode = SQ_CF_INST_ALU;
+    }
+
+    /* browse thro existing presubs. */
+    for(i=0; i<pAsm->unNumPresub; i++)
+    {
+        if(pAsm->presubs[i].sptSigniture == scriptSigniture)
+        {
+            break;
+        }
+    }
+
+    if(i == pAsm->unNumPresub)
+    {   /* not loaded yet */
+        /* save assemble context */
+        number_used_registers_save         = pAsm->number_used_registers;
+        uFirstHelpReg_save                 = pAsm->uFirstHelpReg;
+        uHelpReg_save                      = pAsm->uHelpReg;
+        starting_temp_register_number_save = pAsm->starting_temp_register_number;
+        pILInst_save                       = pAsm->pILInst;
+        uiCurInst_save                     = pAsm->uiCurInst;
+
+        /* alloc in presub */
+        if( (pAsm->unNumPresub + 1) > pAsm->unPresubArraySize )
+        {
+            pAsm->presubs = (PRESUB_DESC*)_mesa_realloc( (void *)pAsm->presubs,
+                                      sizeof(PRESUB_DESC) * pAsm->unPresubArraySize,
+                                      sizeof(PRESUB_DESC) * (pAsm->unPresubArraySize + 4) );
+            if(NULL == pAsm->presubs)
+            {
+                radeon_error("No memeory to allocate built in shader function description structures. \n");
+                return GL_FALSE;
+            }
+            pAsm->unPresubArraySize += 4;
+        }
+        
+        pPresubDesc = &(pAsm->presubs[i]);
+        pPresubDesc->sptSigniture = scriptSigniture;
+
+        /* constants offsets need to be final resolved at reloc. */
+        if(0 == pAsm->unNumPresub)
+        {
+            pPresubDesc->unConstantsStart = 0; 
+        }
+        else
+        {
+            pPresubDesc->unConstantsStart =  pAsm->presubs[i-1].unConstantsStart
+                                           + pAsm->presubs[i-1].pCompiledSub->NumParameters;
+        }
+
+        pPresubDesc->pCompiledSub = pCompiledSub;
+
+        pPresubDesc->subIL_Shift = pAsm->unCurNumILInsts;
+        pPresubDesc->maxStartReg  = uFirstHelpReg_save;
+        pAsm->unCurNumILInsts    += pCompiledSub->NumInstructions;
+
+        pAsm->unNumPresub++;
+
+        /* setup new assemble context */
+        pAsm->starting_temp_register_number = 0;
+        pAsm->number_used_registers = pCompiledSub->NumTemporaries;
+        pAsm->uFirstHelpReg         = pAsm->number_used_registers;
+        pAsm->uHelpReg              = pAsm->uFirstHelpReg;
+
+        bRet = assemble_CAL(pAsm, 
+                            0, 
+                            pPresubDesc->subIL_Shift, 
+                            pCompiledSub->NumInstructions,
+                            pCompiledSub->Instructions,
+                            pPresubDesc);
+
+        
+        pPresubDesc->number_used_registers = pAsm->number_used_registers;        
+
+        /* restore assemble context */
+        pAsm->number_used_registers         = number_used_registers_save; 
+        pAsm->uFirstHelpReg                 = uFirstHelpReg_save;
+        pAsm->uHelpReg                      = uHelpReg_save;
+        pAsm->starting_temp_register_number = starting_temp_register_number_save;
+        pAsm->pILInst                       = pILInst_save; 
+        pAsm->uiCurInst                     = uiCurInst_save;
+    }
+    else
+    {   /* was loaded */
+        pPresubDesc = &(pAsm->presubs[i]);  
+        
+        bRet = assemble_CAL(pAsm, 
+                            0, 
+                            pPresubDesc->subIL_Shift, 
+                            pCompiledSub->NumInstructions,
+                            pCompiledSub->Instructions,
+                            pPresubDesc);
+    }
+
+    if(GL_FALSE == bRet)
+    {
+        radeon_error("Shader presub assemble failed. \n");
+    }
+    else
+    {
+        /* copy presub output to real dst */ 
+        pAsm->alu_x_opcode = SQ_CF_INST_ALU;
+        pAsm->D.dst.opcode = SQ_OP2_INST_MOV;
+
+        if( GL_FALSE == assemble_dst(pAsm) )
+        {
+            return GL_FALSE;
+        }
+
+        setaddrmode_PVSSRC(&(pAsm->S[0].src), ADDR_ABSOLUTE);
+        pAsm->S[0].src.rtype = SRC_REG_TEMPORARY;
+        pAsm->S[0].src.reg   = pCompiledSub->dstRegIndex;
+        pAsm->S[0].src.swizzlex = pCompiledSub->outputSwizzleX;
+        pAsm->S[0].src.swizzley = pCompiledSub->outputSwizzleY;
+        pAsm->S[0].src.swizzlez = pCompiledSub->outputSwizzleZ;
+        pAsm->S[0].src.swizzlew = pCompiledSub->outputSwizzleW;
+
+        next_ins(pAsm);        
+
+        pAsm->callers[pAsm->unCallerArrayPointer - 1].finale_cf_ptr  = pAsm->cf_current_alu_clause_ptr;
+        pAsm->callers[pAsm->unCallerArrayPointer - 1].prelude_cf_ptr = prelude_cf_ptr;
+        pAsm->alu_x_opcode = SQ_CF_INST_ALU;
+    }
+
+    if( (pPresubDesc->number_used_registers + pAsm->uFirstHelpReg) > pAsm->number_used_registers )
+    {
+        pAsm->number_used_registers = pPresubDesc->number_used_registers + pAsm->uFirstHelpReg;
+    }
+    if(pAsm->uFirstHelpReg > pPresubDesc->maxStartReg)
+    {
+        pPresubDesc->maxStartReg = pAsm->uFirstHelpReg;
+    }
+
+    return bRet;
+}
+
 GLboolean Process_Export(r700_AssemblerBase* pAsm,
                          GLuint type,
                          GLuint export_starting_index,
@@ -4791,6 +6526,25 @@ GLboolean Process_Vertex_Exports(r700_AssemblerBase *pR700AsmCode,
             export_starting_index++;
                }
        }
+    
+    for(i=VERT_RESULT_VAR0; i<VERT_RESULT_MAX; i++)
+       {
+        unBit = 1 << i;
+        if(OutputsWritten & unBit)
+               {
+            if( GL_FALSE == Process_Export(pR700AsmCode,
+                                          SQ_EXPORT_PARAM, 
+                                          export_starting_index, 
+                                          1, 
+                                          pR700AsmCode->ucVP_OutputMap[i],
+                                          GL_FALSE) )
+            {                
+                return GL_FALSE;
+            }
+
+            export_starting_index++;
+               }
+    }
 
     // At least one param should be exported
     if (export_count) 
@@ -4825,6 +6579,21 @@ GLboolean Clean_Up_Assembler(r700_AssemblerBase *pR700AsmCode)
 {
     FREE(pR700AsmCode->pucOutMask);
     FREE(pR700AsmCode->pInstDeps);
+
+    if(NULL != pR700AsmCode->subs)
+    {
+        FREE(pR700AsmCode->subs);
+    }
+    if(NULL != pR700AsmCode->callers)
+    {
+        FREE(pR700AsmCode->callers);
+    }
+
+    if(NULL != pR700AsmCode->presubs)
+    {
+        FREE(pR700AsmCode->presubs);
+    }
+
     return GL_TRUE;
 }
 
index c66db502a1704f98aa00127c85a29f1154a0737b..dbd9860f7d0b578dc3486ca4b4909a4aadb1412b 100644 (file)
 #include "r700_shaderinst.h"
 #include "r700_shader.h"
 
+typedef enum LOADABLE_SCRIPT_SIGNITURE
+{
+    GLSL_NOISE1 = 0x10000001,
+    GLSL_NOISE2 = 0x10000002,
+    GLSL_NOISE3 = 0x10000003,
+    GLSL_NOISE4 = 0x10000004
+}LOADABLE_SCRIPT_SIGNITURE;
+
+typedef struct COMPILED_SUB
+{
+    struct  prog_instruction *Instructions;
+    GLuint  NumInstructions;
+    GLuint  NumTemporaries;
+    GLuint  NumParameters;
+    GLuint  MinRegIndex;
+    GLfloat (*ParameterValues)[4];
+    GLbyte  outputSwizzleX; 
+    GLbyte  outputSwizzleY;
+    GLbyte  outputSwizzleZ;
+    GLbyte  outputSwizzleW;
+    GLshort srcRegIndex[3];
+    GLushort dstRegIndex;
+}COMPILED_SUB;
+
+typedef struct PRESUB_DESCtag 
+{
+    LOADABLE_SCRIPT_SIGNITURE sptSigniture;
+    GLint  subIL_Shift;
+    struct prog_src_register InReg[3];
+    struct prog_dst_register OutReg;
+
+    GLushort maxStartReg;
+    GLushort number_used_registers;
+
+    GLuint   unConstantsStart;
+
+    COMPILED_SUB * pCompiledSub;
+} PRESUB_DESC;
+
 typedef enum SHADER_PIPE_TYPE 
 {
     SPT_VP = 0,
@@ -72,7 +111,8 @@ typedef enum SrcRegisterType
     SRC_REG_INPUT          = 1,
     SRC_REG_CONSTANT       = 2,
     SRC_REG_ALT_TEMPORARY  = 3,
-    NUMBER_OF_SRC_REG_TYPE = 4
+    SRC_REC_LITERAL        = 4, 
+    NUMBER_OF_SRC_REG_TYPE = 5
 } SrcRegisterType;
 
 typedef enum DstRegisterType 
@@ -111,16 +151,24 @@ typedef struct PVSDSTtag
        BITS addrmode1:1; //32
 } PVSDST;
 
+typedef struct PVSINSTtag
+{
+    BITS literal_slots      :2; 
+    BITS SaturateMode :2; 
+    BITS index_mode   :3;
+} PVSINST;
+
 typedef struct PVSSRCtag 
 {
-       BITS rtype:4;            
+       BITS rtype:3;            
        BITS addrmode0:1;        
-       BITS reg:10;      //15     (8)
+       BITS reg:10;      //14     (8)
        BITS swizzlex:3;
        BITS swizzley:3;
        BITS swizzlez:3;
-       BITS swizzlew:3;  //27        
+       BITS swizzlew:3;  //26        
 
+       BITS abs:1;
        BITS negx:1;
        BITS negy:1;
        BITS negz:1;
@@ -148,6 +196,7 @@ typedef union PVSDWORDtag
 {
        BITS    bits;
        PVSDST  dst;
+    PVSINST dst2;
        PVSSRC  src;
        PVSMATH math;
        float   f;
@@ -251,6 +300,8 @@ enum
     FC_IF = 1,
     FC_LOOP = 2,
     FC_REP = 3,
+    FC_PUSH_VPM = 4,
+    FC_PUSH_WQM = 5,
 
     COND_NONE = 0,
     COND_BOOL = 1,
@@ -263,22 +314,56 @@ enum
 
 typedef struct FC_LEVEL 
 {
-       unsigned int           first; ///< first fc instruction on level (if, rep, loop)
-       unsigned int*          mid; ///< middle instructions - else or all breaks on this level
-       unsigned int           midLen;
-       unsigned int           type;
-       unsigned int           cond;
-       unsigned int           inv;
-       unsigned int           bpush; ///< 1 if first instruction does branch stack push
-                        int           id; ///< id of bool or int variable
+    R700ControlFlowGenericClause *  first;
+    R700ControlFlowGenericClause ** mid;
+    unsigned int unNumMid;
+    unsigned int midLen;
+    unsigned int type;
+    unsigned int cond;
+    unsigned int inv;
+    int id; ///< id of bool or int variable
 } FC_LEVEL;
 
 typedef struct VTX_FETCH_METHOD 
 {
-       GLboolean bEnableMini;
-       GLuint mega_fetch_remainder;
+    GLboolean bEnableMini;
+    GLuint mega_fetch_remainder;
 } VTX_FETCH_METHOD;
 
+typedef struct SUB_OFFSET
+{
+    GLint  subIL_Offset;
+    GLuint unCFoffset;
+    GLuint unStackDepthMax;
+    PRESUB_DESC *   pPresubDesc;
+    TypedShaderList lstCFInstructions_local;
+} SUB_OFFSET;
+
+typedef struct CALLER_POINTER
+{
+    GLint  subIL_Offset;
+    GLint  subDescIndex;
+    R700ControlFlowGenericClause* cf_ptr;
+
+    R700ControlFlowGenericClause* prelude_cf_ptr;
+    R700ControlFlowGenericClause* finale_cf_ptr;
+} CALLER_POINTER;
+
+#define SQ_MAX_CALL_DEPTH 0x00000020
+
+typedef struct CALL_LEVEL
+{
+    unsigned int      FCSP_BeforeEntry;
+    GLint             subDescIndex;
+    GLushort          current;
+    GLushort          max;
+    TypedShaderList * plstCFInstructions_local;
+} CALL_LEVEL;
+
+#define HAS_CURRENT_LOOPRET 0x1L
+#define HAS_LOOPRET         0x2L
+#define LOOPRET_FLAGS       HAS_LOOPRET | HAS_CURRENT_LOOPRET
+
 typedef struct r700_AssemblerBase 
 {
        R700ControlFlowSXClause*      cf_last_export_ptr;
@@ -294,14 +379,19 @@ typedef struct r700_AssemblerBase
        // No clause has been created yet
        CF_CLAUSE_TYPE cf_current_clause_type;
 
+    BITS alu_x_opcode;
+
        GLuint number_of_exports;
        GLuint number_of_colorandz_exports;
        GLuint number_of_export_opcodes;
 
        PVSDWORD D;
+    PVSDWORD D2;
        PVSDWORD S[3];
+        PVSDWORD C[4];
 
        unsigned int uLastPosUpdate;
+       unsigned int last_cond_register;
 
        OUT_FRAGMENT_FMT_0     fp_stOutFmt0;
 
@@ -310,6 +400,8 @@ typedef struct r700_AssemblerBase
        unsigned int number_used_registers;
        unsigned int uUsedConsts; 
 
+    unsigned int flag_reg_index;
+
        // Fragment programs
        unsigned int uiFP_AttributeMap[FRAG_ATTRIB_MAX];
        unsigned int uiFP_OutputMap[FRAG_RESULT_MAX];
@@ -330,9 +422,6 @@ typedef struct r700_AssemblerBase
        unsigned int FCSP;
        FC_LEVEL fc_stack[32];
 
-       unsigned int branch_depth;
-       unsigned int max_branch_depth;
-
        //-----------------------------------------------------------------------------------
        // ArgSubst used in Assemble_Source() function
        //-----------------------------------------------------------------------------------
@@ -373,11 +462,31 @@ typedef struct r700_AssemblerBase
     SHADER_PIPE_TYPE currentShaderType;
     struct prog_instruction * pILInst;
     GLuint             uiCurInst;
+    GLubyte SamplerUnits[MAX_SAMPLERS];
     GLboolean   bR6xx;
     /* helper to decide which type of instruction to assemble */
     GLboolean is_tex;
     /* we inserted helper intructions and need barrier on next TEX ins */ 
     GLboolean need_tex_barrier; 
+
+    SUB_OFFSET     * subs;
+    GLuint           unSubArraySize;
+    GLuint           unSubArrayPointer;
+    CALLER_POINTER * callers;
+    GLuint           unCallerArraySize;
+    GLuint           unCallerArrayPointer;
+    unsigned int     CALLSP;
+    CALL_LEVEL       CALLSTACK[SQ_MAX_CALL_DEPTH];
+
+    GLuint unCFflags;
+
+    PRESUB_DESC * presubs;
+    GLuint        unPresubArraySize;
+    GLuint        unNumPresub;
+    GLuint        unCurNumILInsts;
+
+    GLuint    unVetTexBits;
+
 } r700_AssemblerBase;
 
 //Internal use
@@ -399,7 +508,7 @@ BITS is_depth_component_exported(OUT_FRAGMENT_FMT_0* pFPOutFmt) ;
 GLboolean is_reduction_opcode(PVSDWORD * dest);
 GLuint GetSurfaceFormat(GLenum eType, GLuint nChannels, GLuint * pClient_size);
 
-unsigned int r700GetNumOperands(r700_AssemblerBase* pAsm);
+unsigned int r700GetNumOperands(GLuint opcode, GLuint nIsOp3);
 
 GLboolean IsTex(gl_inst_opcode Opcode);
 GLboolean IsAlu(gl_inst_opcode Opcode);
@@ -446,6 +555,10 @@ GLboolean assemble_alu_src(R700ALUInstruction*  alu_instruction_ptr,
 GLboolean add_alu_instruction(r700_AssemblerBase* pAsm,
                               R700ALUInstruction* alu_instruction_ptr,
                               GLuint              contiguous_slots_needed);
+
+GLboolean add_cf_instruction(r700_AssemblerBase* pAsm);
+void add_return_inst(r700_AssemblerBase *pAsm);
+
 void get_src_properties(R700ALUInstruction*  alu_instruction_ptr,
                         int                  source_index,
                         BITS*                psrc_sel,
@@ -467,13 +580,20 @@ GLboolean check_vector(r700_AssemblerBase* pAsm,
                        R700ALUInstruction* alu_instruction_ptr);
 GLboolean assemble_alu_instruction(r700_AssemblerBase *pAsm);
 GLboolean next_ins(r700_AssemblerBase *pAsm);
+
+GLboolean pops(r700_AssemblerBase *pAsm, GLuint pops);
+GLboolean jumpToOffest(r700_AssemblerBase *pAsm, GLuint pops, GLint offset);
+GLboolean setRetInLoopFlag(r700_AssemblerBase *pAsm, GLuint flagValue);
+GLboolean testFlag(r700_AssemblerBase *pAsm);
+GLboolean breakLoopOnFlag(r700_AssemblerBase *pAsm, GLuint unFCSP);
+GLboolean returnOnFlag(r700_AssemblerBase *pAsm, GLuint unIF);
+
 GLboolean assemble_math_function(r700_AssemblerBase* pAsm, BITS opcode);
 GLboolean assemble_ABS(r700_AssemblerBase *pAsm);
 GLboolean assemble_ADD(r700_AssemblerBase *pAsm);
 GLboolean assemble_ARL(r700_AssemblerBase *pAsm);
 GLboolean assemble_BAD(char *opcode_str);
 GLboolean assemble_CMP(r700_AssemblerBase *pAsm);
-GLboolean assemble_COS(r700_AssemblerBase *pAsm);
 GLboolean assemble_DOT(r700_AssemblerBase *pAsm);
 GLboolean assemble_DST(r700_AssemblerBase *pAsm);
 GLboolean assemble_EX2(r700_AssemblerBase *pAsm);
@@ -481,7 +601,7 @@ GLboolean assemble_EXP(r700_AssemblerBase *pAsm);
 GLboolean assemble_FLR(r700_AssemblerBase *pAsm);
 GLboolean assemble_FLR_INT(r700_AssemblerBase *pAsm);
 GLboolean assemble_FRC(r700_AssemblerBase *pAsm);
-GLboolean assemble_KIL(r700_AssemblerBase *pAsm);
+GLboolean assemble_KIL(r700_AssemblerBase *pAsm, GLuint opcode);
 GLboolean assemble_LG2(r700_AssemblerBase *pAsm);
 GLboolean assemble_LRP(r700_AssemblerBase *pAsm);
 GLboolean assemble_LOG(r700_AssemblerBase *pAsm);
@@ -494,17 +614,37 @@ GLboolean assemble_MUL(r700_AssemblerBase *pAsm);
 GLboolean assemble_POW(r700_AssemblerBase *pAsm);
 GLboolean assemble_RCP(r700_AssemblerBase *pAsm);
 GLboolean assemble_RSQ(r700_AssemblerBase *pAsm);
-GLboolean assemble_SIN(r700_AssemblerBase *pAsm);
 GLboolean assemble_SCS(r700_AssemblerBase *pAsm);
 GLboolean assemble_SGE(r700_AssemblerBase *pAsm);
+
+GLboolean assemble_LOGIC(r700_AssemblerBase *pAsm, BITS opcode);
+GLboolean assemble_LOGIC_PRED(r700_AssemblerBase *pAsm, BITS opcode); 
+GLboolean assemble_TRIG(r700_AssemblerBase *pAsm, BITS opcode);
+
 GLboolean assemble_SLT(r700_AssemblerBase *pAsm);
 GLboolean assemble_STP(r700_AssemblerBase *pAsm);
 GLboolean assemble_TEX(r700_AssemblerBase *pAsm);
 GLboolean assemble_XPD(r700_AssemblerBase *pAsm);
 GLboolean assemble_EXPORT(r700_AssemblerBase *pAsm);
-GLboolean assemble_IF(r700_AssemblerBase *pAsm);
+GLboolean assemble_IF(r700_AssemblerBase *pAsm, GLboolean bHasElse);
+GLboolean assemble_ELSE(r700_AssemblerBase *pAsm);
 GLboolean assemble_ENDIF(r700_AssemblerBase *pAsm);
 
+GLboolean assemble_BGNLOOP(r700_AssemblerBase *pAsm);
+GLboolean assemble_BRK(r700_AssemblerBase *pAsm);
+GLboolean assemble_COND(r700_AssemblerBase *pAsm);
+GLboolean assemble_ENDLOOP(r700_AssemblerBase *pAsm);
+
+GLboolean assemble_BGNSUB(r700_AssemblerBase *pAsm, GLint nILindex, GLuint uiIL_Shift);
+GLboolean assemble_ENDSUB(r700_AssemblerBase *pAsm);
+GLboolean assemble_RET(r700_AssemblerBase *pAsm);
+GLboolean assemble_CAL(r700_AssemblerBase *pAsm, 
+                       GLint nILindex,
+                       GLuint uiIL_Offest,
+                       GLuint uiNumberInsts,
+                       struct prog_instruction *pILInst,
+                       PRESUB_DESC * pPresubDesc);
+
 GLboolean Process_Export(r700_AssemblerBase* pAsm,
                          GLuint type, 
                          GLuint export_starting_index,
@@ -514,14 +654,25 @@ GLboolean Process_Export(r700_AssemblerBase* pAsm,
 GLboolean Move_Depth_Exports_To_Correct_Channels(r700_AssemblerBase *pAsm, 
                                                  BITS depth_channel_select);
 
+GLboolean callPreSub(r700_AssemblerBase* pAsm, 
+                     LOADABLE_SCRIPT_SIGNITURE scriptSigniture,
+                     /* struct prog_instruction ** pILInstParent, */
+                     COMPILED_SUB * pCompiledSub,                                            
+                     GLshort uOutReg,
+                     GLshort uNumValidSrc);
 
 //Interface
-GLboolean AssembleInstr(GLuint uiNumberInsts,
+GLboolean AssembleInstr(GLuint uiFirstInst,
+                        GLuint uiIL_Shift,
+                        GLuint uiNumberInsts,
                         struct prog_instruction *pILInst, 
                                                r700_AssemblerBase *pR700AsmCode);
 GLboolean Process_Fragment_Exports(r700_AssemblerBase *pR700AsmCode, GLbitfield OutputsWritten);  
 GLboolean Process_Vertex_Exports(r700_AssemblerBase *pR700AsmCode, GLbitfield OutputsWritten);
 
+GLboolean RelocProgram(r700_AssemblerBase * pAsm, struct gl_program * pILProg);
+GLboolean InitShaderProgram(r700_AssemblerBase * pAsm);
+
 int       Init_r700_AssemblerBase(SHADER_PIPE_TYPE spt, r700_AssemblerBase* pAsm, R700_Shader* pShader);
 GLboolean Clean_Up_Assembler(r700_AssemblerBase *pR700AsmCode);
 
index 02c56b98d9b224a044a118583248c74fa9ab8149..c124e02184f473691a6e31788d25890fb050d025 100644 (file)
@@ -45,6 +45,9 @@ static void r700SendTexState(GLcontext *ctx, struct radeon_state_atom *atom)
 {
        context_t         *context = R700_CONTEXT(ctx);
        R700_CHIP_CONTEXT *r700 = (R700_CHIP_CONTEXT*)(&context->hw);
+
+    struct r700_vertex_program *vp = context->selected_vp;
+
        struct radeon_bo *bo = NULL;
        unsigned int i;
        BATCH_LOCALS(&context->radeon);
@@ -52,7 +55,7 @@ static void r700SendTexState(GLcontext *ctx, struct radeon_state_atom *atom)
        radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
        for (i = 0; i < R700_TEXTURE_NUMBERUNITS; i++) {
-               if (ctx->Texture.Unit[i]._ReallyEnabled) {
+               if (ctx->Texture.Unit[i]._ReallyEnabled) {            
                        radeonTexObj *t = r700->textures[i];
                        uint32_t offset;
                        if (t) {
@@ -71,7 +74,16 @@ static void r700SendTexState(GLcontext *ctx, struct radeon_state_atom *atom)
 
                                        BEGIN_BATCH_NO_AUTOSTATE(9 + 4);
                                        R600_OUT_BATCH(CP_PACKET3(R600_IT_SET_RESOURCE, 7));
-                                       R600_OUT_BATCH(i * 7);
+
+                    if( (1<<i) & vp->r700AsmCode.unVetTexBits )                    
+                    {   /* vs texture */                                     
+                        R600_OUT_BATCH((i + VERT_ATTRIB_MAX + SQ_FETCH_RESOURCE_VS_OFFSET) * FETCH_RESOURCE_STRIDE);
+                    }
+                    else
+                    {
+                                           R600_OUT_BATCH(i * 7);
+                    }
+
                                        R600_OUT_BATCH(r700->textures[i]->SQ_TEX_RESOURCE0);
                                        R600_OUT_BATCH(r700->textures[i]->SQ_TEX_RESOURCE1);
                                        R600_OUT_BATCH(r700->textures[i]->SQ_TEX_RESOURCE2);
@@ -95,21 +107,35 @@ static void r700SendTexState(GLcontext *ctx, struct radeon_state_atom *atom)
        }
 }
 
+#define SAMPLER_STRIDE                 3
+
 static void r700SendTexSamplerState(GLcontext *ctx, struct radeon_state_atom *atom)
 {
        context_t         *context = R700_CONTEXT(ctx);
        R700_CHIP_CONTEXT *r700 = (R700_CHIP_CONTEXT*)(&context->hw);
        unsigned int i;
+
+    struct r700_vertex_program *vp = context->selected_vp;
+
        BATCH_LOCALS(&context->radeon);
        radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
        for (i = 0; i < R700_TEXTURE_NUMBERUNITS; i++) {
-               if (ctx->Texture.Unit[i]._ReallyEnabled) {
+               if (ctx->Texture.Unit[i]._ReallyEnabled) {            
                        radeonTexObj *t = r700->textures[i];
                        if (t) {
                                BEGIN_BATCH_NO_AUTOSTATE(5);
                                R600_OUT_BATCH(CP_PACKET3(R600_IT_SET_SAMPLER, 3));
-                               R600_OUT_BATCH(i * 3);
+
+                if( (1<<i) & vp->r700AsmCode.unVetTexBits )                    
+                {   /* vs texture */
+                    R600_OUT_BATCH((i+SQ_TEX_SAMPLER_VS_OFFSET) * SAMPLER_STRIDE); //work 1
+                }
+                else
+                {
+                                   R600_OUT_BATCH(i * 3);
+                }
+
                                R600_OUT_BATCH(r700->textures[i]->SQ_TEX_SAMPLER0);
                                R600_OUT_BATCH(r700->textures[i]->SQ_TEX_SAMPLER1);
                                R600_OUT_BATCH(r700->textures[i]->SQ_TEX_SAMPLER2);
@@ -446,68 +472,77 @@ static void r700SendRenderTargetState(GLcontext *ctx, struct radeon_state_atom *
 
 static void r700SendPSState(GLcontext *ctx, struct radeon_state_atom *atom)
 {
-       context_t *context = R700_CONTEXT(ctx);
-       R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
-       struct radeon_bo * pbo;
-       BATCH_LOCALS(&context->radeon);
-       radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
+    context_t *context = R700_CONTEXT(ctx);
+    R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
+    struct radeon_bo * pbo;
+    BATCH_LOCALS(&context->radeon);
+    radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
-       pbo = (struct radeon_bo *)r700GetActiveFpShaderBo(GL_CONTEXT(context));
+    pbo = (struct radeon_bo *)r700GetActiveFpShaderBo(GL_CONTEXT(context));
 
-       if (!pbo)
-               return;
+    if (!pbo)
+           return;
 
-       r700SyncSurf(context, pbo, RADEON_GEM_DOMAIN_GTT, 0, SH_ACTION_ENA_bit);
+    r700SyncSurf(context, pbo, RADEON_GEM_DOMAIN_GTT, 0, SH_ACTION_ENA_bit);
 
-        BEGIN_BATCH_NO_AUTOSTATE(3 + 2);
-       R600_OUT_BATCH_REGSEQ(SQ_PGM_START_PS, 1);
-       R600_OUT_BATCH(r700->ps.SQ_PGM_START_PS.u32All);
-       R600_OUT_BATCH_RELOC(r700->ps.SQ_PGM_START_PS.u32All,
-                            pbo,
-                            r700->ps.SQ_PGM_START_PS.u32All,
-                            RADEON_GEM_DOMAIN_GTT, 0, 0);
-       END_BATCH();
+    BEGIN_BATCH_NO_AUTOSTATE(3 + 2);
+    R600_OUT_BATCH_REGSEQ(SQ_PGM_START_PS, 1);
+    R600_OUT_BATCH(r700->ps.SQ_PGM_START_PS.u32All);
+    R600_OUT_BATCH_RELOC(r700->ps.SQ_PGM_START_PS.u32All,
+                        pbo,
+                        r700->ps.SQ_PGM_START_PS.u32All,
+                        RADEON_GEM_DOMAIN_GTT, 0, 0);
+    END_BATCH();
 
-        BEGIN_BATCH_NO_AUTOSTATE(9);
-       R600_OUT_BATCH_REGVAL(SQ_PGM_RESOURCES_PS, r700->ps.SQ_PGM_RESOURCES_PS.u32All);
-       R600_OUT_BATCH_REGVAL(SQ_PGM_EXPORTS_PS, r700->ps.SQ_PGM_EXPORTS_PS.u32All);
-       R600_OUT_BATCH_REGVAL(SQ_PGM_CF_OFFSET_PS, r700->ps.SQ_PGM_CF_OFFSET_PS.u32All);
-        END_BATCH();
+    BEGIN_BATCH_NO_AUTOSTATE(9);
+    R600_OUT_BATCH_REGVAL(SQ_PGM_RESOURCES_PS, r700->ps.SQ_PGM_RESOURCES_PS.u32All);
+    R600_OUT_BATCH_REGVAL(SQ_PGM_EXPORTS_PS, r700->ps.SQ_PGM_EXPORTS_PS.u32All);
+    R600_OUT_BATCH_REGVAL(SQ_PGM_CF_OFFSET_PS, r700->ps.SQ_PGM_CF_OFFSET_PS.u32All);
+    END_BATCH();
 
-       COMMIT_BATCH();
+    BEGIN_BATCH_NO_AUTOSTATE(3);
+    R600_OUT_BATCH_REGVAL(SQ_LOOP_CONST_0, 0x01000FFF);
+    END_BATCH();
+
+    COMMIT_BATCH();
 
 }
 
 static void r700SendVSState(GLcontext *ctx, struct radeon_state_atom *atom)
 {
-       context_t *context = R700_CONTEXT(ctx);
-       R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
-       struct radeon_bo * pbo;
-       BATCH_LOCALS(&context->radeon);
-       radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
+    context_t *context = R700_CONTEXT(ctx);
+    R700_CHIP_CONTEXT *r700 = R700_CONTEXT_STATES(context);
+    struct radeon_bo * pbo;
+    BATCH_LOCALS(&context->radeon);
+    radeon_print(RADEON_STATE, RADEON_VERBOSE, "%s\n", __func__);
 
-       pbo = (struct radeon_bo *)r700GetActiveVpShaderBo(GL_CONTEXT(context));
+    pbo = (struct radeon_bo *)r700GetActiveVpShaderBo(GL_CONTEXT(context));
 
-       if (!pbo)
-               return;
+    if (!pbo)
+           return;
 
-       r700SyncSurf(context, pbo, RADEON_GEM_DOMAIN_GTT, 0, SH_ACTION_ENA_bit);
+    r700SyncSurf(context, pbo, RADEON_GEM_DOMAIN_GTT, 0, SH_ACTION_ENA_bit);
 
-        BEGIN_BATCH_NO_AUTOSTATE(3 + 2);
-       R600_OUT_BATCH_REGSEQ(SQ_PGM_START_VS, 1);
-       R600_OUT_BATCH(r700->vs.SQ_PGM_START_VS.u32All);
-       R600_OUT_BATCH_RELOC(r700->vs.SQ_PGM_START_VS.u32All,
-                            pbo,
-                            r700->vs.SQ_PGM_START_VS.u32All,
-                            RADEON_GEM_DOMAIN_GTT, 0, 0);
-       END_BATCH();
+    BEGIN_BATCH_NO_AUTOSTATE(3 + 2);
+    R600_OUT_BATCH_REGSEQ(SQ_PGM_START_VS, 1);
+    R600_OUT_BATCH(r700->vs.SQ_PGM_START_VS.u32All);
+    R600_OUT_BATCH_RELOC(r700->vs.SQ_PGM_START_VS.u32All,
+                        pbo,
+                        r700->vs.SQ_PGM_START_VS.u32All,
+                        RADEON_GEM_DOMAIN_GTT, 0, 0);
+    END_BATCH();
 
-        BEGIN_BATCH_NO_AUTOSTATE(6);
-       R600_OUT_BATCH_REGVAL(SQ_PGM_RESOURCES_VS, r700->vs.SQ_PGM_RESOURCES_VS.u32All);
-       R600_OUT_BATCH_REGVAL(SQ_PGM_CF_OFFSET_VS, r700->vs.SQ_PGM_CF_OFFSET_VS.u32All);
-        END_BATCH();
+    BEGIN_BATCH_NO_AUTOSTATE(6);
+    R600_OUT_BATCH_REGVAL(SQ_PGM_RESOURCES_VS, r700->vs.SQ_PGM_RESOURCES_VS.u32All);
+    R600_OUT_BATCH_REGVAL(SQ_PGM_CF_OFFSET_VS, r700->vs.SQ_PGM_CF_OFFSET_VS.u32All);
+    END_BATCH();
 
-       COMMIT_BATCH();
+    BEGIN_BATCH_NO_AUTOSTATE(3);
+    R600_OUT_BATCH_REGVAL((SQ_LOOP_CONST_0 + 32*4), 0x0100000F);
+    //R600_OUT_BATCH_REGVAL((SQ_LOOP_CONST_0 + (SQ_LOOP_CONST_vs<2)), 0x0100000F);
+    END_BATCH();
+
+    COMMIT_BATCH();
 }
 
 static void r700SendFSState(GLcontext *ctx, struct radeon_state_atom *atom)
@@ -1309,8 +1344,8 @@ void r600InitAtoms(context_t *context)
        ALLOC_STATE(spi, always, (59 + R700_MAX_SHADER_EXPORTS), r700SendSPIState);
        ALLOC_STATE(vpt, always, 16, r700SendViewportState);
        ALLOC_STATE(fs, always, 18, r700SendFSState);
-       ALLOC_STATE(vs, always, 18, r700SendVSState);
-       ALLOC_STATE(ps, always, 21, r700SendPSState);
+       ALLOC_STATE(vs, always, 21, r700SendVSState);
+       ALLOC_STATE(ps, always, 24, r700SendPSState);
        ALLOC_STATE(vs_consts, vs_consts, (2 + (R700_MAX_DX9_CONSTS * 4)), r700SendVSConsts);
        ALLOC_STATE(ps_consts, ps_consts, (2 + (R700_MAX_DX9_CONSTS * 4)), r700SendPSConsts);
        ALLOC_STATE(vtx, vtx, (6 + (VERT_ATTRIB_MAX * 18)), r700SendVTXState);
index c6546ab00c2e674daa81dd18e7949d33fc173103..526d3843d16e7669fdbc4081ea494ef4439e8201 100644 (file)
@@ -57,6 +57,10 @@ void r700Clear(GLcontext * ctx, GLbitfield mask)
 
     radeon_print(RADEON_RENDER, RADEON_VERBOSE, "%s %x\n", __func__, mask);
 
+    if (mask & (BUFFER_BIT_FRONT_LEFT | BUFFER_BIT_FRONT_RIGHT)) {
+        context->radeon.front_buffer_dirty = GL_TRUE;
+    }
+
     if( GL_TRUE == r700ClearFast(context, mask) )
     {
         return;
index ccafd433bfaaa0b32c0f5939702198f395b1bc4d..ce2d9fdf79e14548947368c38ca98fd23e5b5990 100644 (file)
@@ -34,6 +34,7 @@
 #include "main/imports.h"
 #include "shader/prog_parameter.h"
 #include "shader/prog_statevars.h"
+#include "shader/program.h"
 
 #include "r600_context.h"
 #include "r600_cmdbuf.h"
 
 #include "r700_debug.h"
 
+void insert_wpos_code(GLcontext *ctx, struct gl_fragment_program *fprog)
+{
+    static const gl_state_index winstate[STATE_LENGTH]
+         = { STATE_INTERNAL, STATE_FB_SIZE, 0, 0, 0};
+    struct prog_instruction *newInst, *inst;
+    GLint  win_size;  /* state reference */
+    GLuint wpos_temp; /* temp register */
+    int i, j;
+
+    /* PARAM win_size = STATE_FB_SIZE */
+    win_size = _mesa_add_state_reference(fprog->Base.Parameters, winstate);
+
+    wpos_temp = fprog->Base.NumTemporaries++;
+
+    /* scan program where WPOS is used and replace with wpos_temp */
+    inst = fprog->Base.Instructions;
+    for (i = 0; i < fprog->Base.NumInstructions; i++) {
+        for (j=0; j < 3; j++) {
+            if(inst->SrcReg[j].File == PROGRAM_INPUT && 
+               inst->SrcReg[j].Index == FRAG_ATTRIB_WPOS) {
+                inst->SrcReg[j].File = PROGRAM_TEMPORARY;
+                inst->SrcReg[j].Index = wpos_temp;
+            }
+        }
+        inst++;
+    }
+
+    _mesa_insert_instructions(&(fprog->Base), 0, 1);
+
+    newInst = fprog->Base.Instructions;
+    /* invert wpos.y
+     * wpos_temp.xyzw = wpos.x-yzw + winsize.0y00 */
+    newInst[0].Opcode = OPCODE_ADD;
+    newInst[0].DstReg.File = PROGRAM_TEMPORARY;
+    newInst[0].DstReg.Index = wpos_temp;
+    newInst[0].DstReg.WriteMask = WRITEMASK_XYZW;
+
+    newInst[0].SrcReg[0].File = PROGRAM_INPUT;
+    newInst[0].SrcReg[0].Index = FRAG_ATTRIB_WPOS;
+    newInst[0].SrcReg[0].Swizzle = SWIZZLE_XYZW;
+    newInst[0].SrcReg[0].Negate = NEGATE_Y;
+
+    newInst[0].SrcReg[1].File = PROGRAM_STATE_VAR;
+    newInst[0].SrcReg[1].Index = win_size;
+    newInst[0].SrcReg[1].Swizzle = MAKE_SWIZZLE4(SWIZZLE_ZERO, SWIZZLE_Y, SWIZZLE_ZERO, SWIZZLE_ZERO);
+
+}
+
 //TODO : Validate FP input with VP output.
 void Map_Fragment_Program(r700_AssemblerBase         *pAsm,
-                                                 struct gl_fragment_program *mesa_fp)
+                                                 struct gl_fragment_program *mesa_fp,
+                          GLcontext *ctx) 
 {
        unsigned int unBit;
     unsigned int i;
     GLuint       ui;
 
+    /* match fp inputs with vp exports. */
+    struct r700_vertex_program_cont *vpc =
+                      (struct r700_vertex_program_cont *)ctx->VertexProgram._Current;
+    GLbitfield OutputsWritten = vpc->mesa_program.Base.OutputsWritten;
+    
        pAsm->number_used_registers = 0;
 
 //Input mapping : mesa_fp->Base.InputsRead set the flag, set in 
@@ -61,32 +116,99 @@ void Map_Fragment_Program(r700_AssemblerBase         *pAsm,
                pAsm->uiFP_AttributeMap[FRAG_ATTRIB_WPOS] = pAsm->number_used_registers++;
        }
 
-       unBit = 1 << FRAG_ATTRIB_COL0;
-       if(mesa_fp->Base.InputsRead & unBit)
+    unBit = 1 << VERT_RESULT_COL0;
+       if(OutputsWritten & unBit)
        {
                pAsm->uiFP_AttributeMap[FRAG_ATTRIB_COL0] = pAsm->number_used_registers++;
        }
 
-       unBit = 1 << FRAG_ATTRIB_COL1;
-       if(mesa_fp->Base.InputsRead & unBit)
+       unBit = 1 << VERT_RESULT_COL1;
+       if(OutputsWritten & unBit)
        {
                pAsm->uiFP_AttributeMap[FRAG_ATTRIB_COL1] = pAsm->number_used_registers++;
        }
 
-        unBit = 1 << FRAG_ATTRIB_FOGC;
-        if(mesa_fp->Base.InputsRead & unBit)
-        {
-                pAsm->uiFP_AttributeMap[FRAG_ATTRIB_FOGC] = pAsm->number_used_registers++;
-        }
+    unBit = 1 << VERT_RESULT_FOGC;
+    if(OutputsWritten & unBit)
+    {
+        pAsm->uiFP_AttributeMap[FRAG_ATTRIB_FOGC] = pAsm->number_used_registers++;
+    }
 
        for(i=0; i<8; i++)
        {
-               unBit = 1 << (FRAG_ATTRIB_TEX0 + i);
-               if(mesa_fp->Base.InputsRead & unBit)
+               unBit = 1 << (VERT_RESULT_TEX0 + i);
+               if(OutputsWritten & unBit)
                {
                        pAsm->uiFP_AttributeMap[FRAG_ATTRIB_TEX0 + i] = pAsm->number_used_registers++;
                }
        }
+/* order has been taken care of */ 
+#if 1
+    for(i=VERT_RESULT_VAR0; i<VERT_RESULT_MAX; i++)
+       {
+        unBit = 1 << i;
+        if(OutputsWritten & unBit)
+               {
+            pAsm->uiFP_AttributeMap[i-VERT_RESULT_VAR0+FRAG_ATTRIB_VAR0] = pAsm->number_used_registers++;
+        }
+    }
+#else
+    if( (mesa_fp->Base.InputsRead >> FRAG_ATTRIB_VAR0) > 0 )
+    {
+           struct r700_vertex_program_cont *vpc =
+                      (struct r700_vertex_program_cont *)ctx->VertexProgram._Current;
+        struct gl_program_parameter_list * VsVarying = vpc->mesa_program.Base.Varying;
+        struct gl_program_parameter_list * PsVarying = mesa_fp->Base.Varying;
+        struct gl_program_parameter      * pVsParam;
+        struct gl_program_parameter      * pPsParam;
+        GLuint j, k;
+        GLuint unMaxVarying = 0;
+
+        for(i=0; i<VsVarying->NumParameters; i++)
+        {
+            pAsm->uiFP_AttributeMap[i + FRAG_ATTRIB_VAR0] = 0;
+        }
+
+        for(i=FRAG_ATTRIB_VAR0; i<FRAG_ATTRIB_MAX; i++)
+           {
+            unBit = 1 << i;
+            if(mesa_fp->Base.InputsRead & unBit)
+                   {
+                j = i - FRAG_ATTRIB_VAR0;
+                pPsParam = PsVarying->Parameters + j;
+
+                for(k=0; k<VsVarying->NumParameters; k++)
+                {                                      
+                    pVsParam = VsVarying->Parameters + k;
+
+                               if( strcmp(pPsParam->Name, pVsParam->Name) == 0)
+                    {
+                        pAsm->uiFP_AttributeMap[i] = pAsm->number_used_registers + k;                  
+                        if(k > unMaxVarying)
+                        {
+                            unMaxVarying = k;
+                        }
+                        break;
+                    }
+                }
+                   }
+        }
+
+        pAsm->number_used_registers += unMaxVarying + 1;
+    }
+#endif
+    unBit = 1 << FRAG_ATTRIB_FACE;
+    if(mesa_fp->Base.InputsRead & unBit)
+    {
+        pAsm->uiFP_AttributeMap[FRAG_ATTRIB_FACE] = pAsm->number_used_registers++;
+    }
+
+    unBit = 1 << FRAG_ATTRIB_PNTC;
+    if(mesa_fp->Base.InputsRead & unBit)
+    {
+        pAsm->uiFP_AttributeMap[FRAG_ATTRIB_PNTC] = pAsm->number_used_registers++;
+    }
 
 /* Map temporary registers (GPRs) */
     pAsm->starting_temp_register_number = pAsm->number_used_registers;
@@ -127,6 +249,8 @@ void Map_Fragment_Program(r700_AssemblerBase         *pAsm,
         pAsm->pucOutMask[ui] = 0x0;
     }
 
+    pAsm->flag_reg_index = pAsm->number_used_registers++;
+
     pAsm->uFirstHelpReg = pAsm->number_used_registers;
 }
 
@@ -233,22 +357,41 @@ GLboolean Find_Instruction_Dependencies_fp(struct r700_fragment_program *fp,
 }
 
 GLboolean r700TranslateFragmentShader(struct r700_fragment_program *fp,
-                                                            struct gl_fragment_program   *mesa_fp)
+                                                            struct gl_fragment_program   *mesa_fp,
+                                 GLcontext *ctx) 
 {
        GLuint    number_of_colors_exported;
        GLboolean z_enabled = GL_FALSE;
        GLuint    unBit;
+       int i;
 
     //Init_Program
        Init_r700_AssemblerBase( SPT_FP, &(fp->r700AsmCode), &(fp->r700Shader) );
-       Map_Fragment_Program(&(fp->r700AsmCode), mesa_fp);
+
+    if(mesa_fp->Base.InputsRead & FRAG_BIT_WPOS)
+    {
+        insert_wpos_code(ctx, mesa_fp);
+    }
+
+    Map_Fragment_Program(&(fp->r700AsmCode), mesa_fp, ctx); 
 
     if( GL_FALSE == Find_Instruction_Dependencies_fp(fp, mesa_fp) )
        {
                return GL_FALSE;
     }
+
+    InitShaderProgram(&(fp->r700AsmCode));
        
-       if( GL_FALSE == AssembleInstr(mesa_fp->Base.NumInstructions,
+    for(i=0; i < MAX_SAMPLERS; i++)
+    {
+         fp->r700AsmCode.SamplerUnits[i] = fp->mesa_program.Base.SamplerUnits[i];
+    }
+
+    fp->r700AsmCode.unCurNumILInsts = mesa_fp->Base.NumInstructions;
+
+       if( GL_FALSE == AssembleInstr(0,
+                                  0,
+                                  mesa_fp->Base.NumInstructions,
                                   &(mesa_fp->Base.Instructions[0]), 
                                   &(fp->r700AsmCode)) )
        {
@@ -260,6 +403,11 @@ GLboolean r700TranslateFragmentShader(struct r700_fragment_program *fp,
         return GL_FALSE;
     }
 
+    if( GL_FALSE == RelocProgram(&(fp->r700AsmCode), &(mesa_fp->Base)) )
+    {
+        return GL_FALSE;
+    }
+
     fp->r700Shader.nRegs = (fp->r700AsmCode.number_used_registers == 0) ? 0 
                          : (fp->r700AsmCode.number_used_registers - 1);
 
@@ -300,7 +448,7 @@ void r700SelectFragmentShader(GLcontext *ctx)
     }
 
     if (GL_FALSE == fp->translated)
-           r700TranslateFragmentShader(fp, &(fp->mesa_program));
+           r700TranslateFragmentShader(fp, &(fp->mesa_program), ctx); 
 }
 
 void * r700GetActiveFpShaderBo(GLcontext * ctx)
@@ -325,6 +473,7 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
     unsigned int unNumOfReg;
     unsigned int unBit;
     GLuint exportCount;
+    GLboolean point_sprite = GL_FALSE;
 
     if(GL_FALSE == fp->loaded)
     {
@@ -378,6 +527,50 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
         CLEARbit(r700->SPI_INPUT_Z.u32All, PROVIDE_Z_TO_SPI_bit);
     }
 
+    if (mesa_fp->Base.InputsRead & (1 << FRAG_ATTRIB_FACE))
+    {
+        ui += 1;
+        SETfield(r700->SPI_PS_IN_CONTROL_0.u32All, ui, NUM_INTERP_shift, NUM_INTERP_mask);
+        SETbit(r700->SPI_PS_IN_CONTROL_1.u32All, FRONT_FACE_ENA_bit);
+        SETbit(r700->SPI_PS_IN_CONTROL_1.u32All, FRONT_FACE_ALL_BITS_bit);
+        SETfield(r700->SPI_PS_IN_CONTROL_1.u32All, pAsm->uiFP_AttributeMap[FRAG_ATTRIB_FACE], FRONT_FACE_ADDR_shift, FRONT_FACE_ADDR_mask);
+    }
+    else
+    {
+        CLEARbit(r700->SPI_PS_IN_CONTROL_1.u32All, FRONT_FACE_ENA_bit);
+    }
+
+    /* see if we need any point_sprite replacements */
+    for (i = VERT_RESULT_TEX0; i<= VERT_RESULT_TEX7; i++)
+    {
+        if(ctx->Point.CoordReplace[i - VERT_RESULT_TEX0] == GL_TRUE)
+            point_sprite = GL_TRUE;
+    }
+
+    if ((mesa_fp->Base.InputsRead & (1 << FRAG_ATTRIB_PNTC)) || point_sprite)
+    {
+        /* for FRAG_ATTRIB_PNTC we need to increase num_interp */
+        if(mesa_fp->Base.InputsRead & (1 << FRAG_ATTRIB_PNTC))
+        {
+            ui++;
+            SETfield(r700->SPI_PS_IN_CONTROL_0.u32All, ui, NUM_INTERP_shift, NUM_INTERP_mask);
+        }
+        SETbit(r700->SPI_INTERP_CONTROL_0.u32All, PNT_SPRITE_ENA_bit);
+        SETfield(r700->SPI_INTERP_CONTROL_0.u32All, SPI_PNT_SPRITE_SEL_S, PNT_SPRITE_OVRD_X_shift, PNT_SPRITE_OVRD_X_mask);
+        SETfield(r700->SPI_INTERP_CONTROL_0.u32All, SPI_PNT_SPRITE_SEL_T, PNT_SPRITE_OVRD_Y_shift, PNT_SPRITE_OVRD_Y_mask);
+        SETfield(r700->SPI_INTERP_CONTROL_0.u32All, SPI_PNT_SPRITE_SEL_0, PNT_SPRITE_OVRD_Z_shift, PNT_SPRITE_OVRD_Z_mask);
+        SETfield(r700->SPI_INTERP_CONTROL_0.u32All, SPI_PNT_SPRITE_SEL_1, PNT_SPRITE_OVRD_W_shift, PNT_SPRITE_OVRD_W_mask);
+        if(ctx->Point.SpriteOrigin == GL_LOWER_LEFT)
+            SETbit(r700->SPI_INTERP_CONTROL_0.u32All, PNT_SPRITE_TOP_1_bit);
+        else
+            CLEARbit(r700->SPI_INTERP_CONTROL_0.u32All, PNT_SPRITE_TOP_1_bit);
+    }
+    else
+    {
+        CLEARbit(r700->SPI_INTERP_CONTROL_0.u32All, PNT_SPRITE_ENA_bit);
+    }
+
+
     ui = (unNumOfReg < ui) ? ui : unNumOfReg;
 
     SETfield(r700->ps.SQ_PGM_RESOURCES_PS.u32All, ui, NUM_GPRS_shift, NUM_GPRS_mask);
@@ -394,6 +587,13 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
              EXPORT_MODE_shift, EXPORT_MODE_mask);
 
     // emit ps input map
+    struct r700_vertex_program_cont *vpc =
+                      (struct r700_vertex_program_cont *)ctx->VertexProgram._Current;
+    GLbitfield OutputsWritten = vpc->mesa_program.Base.OutputsWritten;
+    
+    for(ui = 0; ui < R700_MAX_SHADER_EXPORTS; ui++)
+        r700->SPI_PS_INPUT_CNTL[ui].u32All = 0;
+
     unBit = 1 << FRAG_ATTRIB_WPOS;
     if(mesa_fp->Base.InputsRead & unBit)
     {
@@ -407,8 +607,8 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
                     CLEARbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
     }
 
-    unBit = 1 << FRAG_ATTRIB_COL0;
-    if(mesa_fp->Base.InputsRead & unBit)
+    unBit = 1 << VERT_RESULT_COL0;
+    if(OutputsWritten & unBit)
     {
            ui = pAsm->uiFP_AttributeMap[FRAG_ATTRIB_COL0];
            SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, SEL_CENTROID_bit);
@@ -420,8 +620,8 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
                    CLEARbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
     }
 
-    unBit = 1 << FRAG_ATTRIB_COL1;
-    if(mesa_fp->Base.InputsRead & unBit)
+    unBit = 1 << VERT_RESULT_COL1;
+    if(OutputsWritten & unBit)
     {
            ui = pAsm->uiFP_AttributeMap[FRAG_ATTRIB_COL1];
            SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, SEL_CENTROID_bit);
@@ -433,8 +633,8 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
                    CLEARbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
     }
 
-    unBit = 1 << FRAG_ATTRIB_FOGC;
-    if(mesa_fp->Base.InputsRead & unBit)
+    unBit = 1 << VERT_RESULT_FOGC;
+    if(OutputsWritten & unBit)
     {
             ui = pAsm->uiFP_AttributeMap[FRAG_ATTRIB_FOGC];
             SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, SEL_CENTROID_bit);
@@ -448,17 +648,67 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
 
     for(i=0; i<8; i++)
     {
-           unBit = 1 << (FRAG_ATTRIB_TEX0 + i);
-           if(mesa_fp->Base.InputsRead & unBit)
+           unBit = 1 << (VERT_RESULT_TEX0 + i);
+           if(OutputsWritten & unBit)
            {
                    ui = pAsm->uiFP_AttributeMap[FRAG_ATTRIB_TEX0 + i];
                    SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, SEL_CENTROID_bit);
                    SETfield(r700->SPI_PS_INPUT_CNTL[ui].u32All, ui,
                             SEMANTIC_shift, SEMANTIC_mask);
                    CLEARbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
+                   /* ARB_point_sprite */
+                   if(ctx->Point.CoordReplace[i] == GL_TRUE)
+                   {
+                            SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, PT_SPRITE_TEX_bit);
+                   }
            }
     }
 
+    unBit = 1 << FRAG_ATTRIB_FACE;
+    if(mesa_fp->Base.InputsRead & unBit)
+    {
+            ui = pAsm->uiFP_AttributeMap[FRAG_ATTRIB_FACE];
+            SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, SEL_CENTROID_bit);
+            SETfield(r700->SPI_PS_INPUT_CNTL[ui].u32All, ui,
+                     SEMANTIC_shift, SEMANTIC_mask);
+            if (r700->SPI_INTERP_CONTROL_0.u32All & FLAT_SHADE_ENA_bit)
+                    SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
+            else
+                    CLEARbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
+    }
+    unBit = 1 << FRAG_ATTRIB_PNTC;
+    if(mesa_fp->Base.InputsRead & unBit)
+    {
+            ui = pAsm->uiFP_AttributeMap[FRAG_ATTRIB_PNTC];
+            SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, SEL_CENTROID_bit);
+            SETfield(r700->SPI_PS_INPUT_CNTL[ui].u32All, ui,
+                     SEMANTIC_shift, SEMANTIC_mask);
+            if (r700->SPI_INTERP_CONTROL_0.u32All & FLAT_SHADE_ENA_bit)
+                    SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
+            else
+                    CLEARbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
+            SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, PT_SPRITE_TEX_bit);
+    }
+
+
+
+
+    for(i=VERT_RESULT_VAR0; i<VERT_RESULT_MAX; i++)
+       {
+        unBit = 1 << i;
+        if(OutputsWritten & unBit)
+               {
+            ui = pAsm->uiFP_AttributeMap[i-VERT_RESULT_VAR0+FRAG_ATTRIB_VAR0];
+            SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, SEL_CENTROID_bit);
+            SETfield(r700->SPI_PS_INPUT_CNTL[ui].u32All, ui,
+                            SEMANTIC_shift, SEMANTIC_mask);
+            if (r700->SPI_INTERP_CONTROL_0.u32All & FLAT_SHADE_ENA_bit)
+                       SETbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
+            else
+                       CLEARbit(r700->SPI_PS_INPUT_CNTL[ui].u32All, FLAT_SHADE_bit);
+        }
+    }
+
     exportCount = (r700->ps.SQ_PGM_EXPORTS_PS.u32All & EXPORT_MODE_mask) / (1 << EXPORT_MODE_shift);
     if (r700->CB_SHADER_CONTROL.u32All != ((1 << exportCount) - 1))
     {
@@ -469,7 +719,8 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
     /* sent out shader constants. */
     paramList = fp->mesa_program.Base.Parameters;
 
-    if(NULL != paramList) {
+    if(NULL != paramList) 
+    {
            _mesa_load_state_parameters(ctx, paramList);
 
            if (paramList->NumParameters > R700_MAX_DX9_CONSTS)
@@ -482,14 +733,33 @@ GLboolean r700SetupFragmentProgram(GLcontext * ctx)
            unNumParamData = paramList->NumParameters;
 
            for(ui=0; ui<unNumParamData; ui++) {
-                   r700->ps.consts[ui][0].f32All = paramList->ParameterValues[ui][0];
-                   r700->ps.consts[ui][1].f32All = paramList->ParameterValues[ui][1];
-                   r700->ps.consts[ui][2].f32All = paramList->ParameterValues[ui][2];
-                   r700->ps.consts[ui][3].f32All = paramList->ParameterValues[ui][3];
+                       r700->ps.consts[ui][0].f32All = paramList->ParameterValues[ui][0];
+                       r700->ps.consts[ui][1].f32All = paramList->ParameterValues[ui][1];
+                       r700->ps.consts[ui][2].f32All = paramList->ParameterValues[ui][2];
+                       r700->ps.consts[ui][3].f32All = paramList->ParameterValues[ui][3];
            }
     } else
            r700->ps.num_consts = 0;
 
+    COMPILED_SUB * pCompiledSub;
+    GLuint uj;
+    GLuint unConstOffset = r700->ps.num_consts;
+    for(ui=0; ui<pAsm->unNumPresub; ui++)
+    {
+        pCompiledSub = pAsm->presubs[ui].pCompiledSub;
+
+        r700->ps.num_consts += pCompiledSub->NumParameters;
+
+        for(uj=0; uj<pCompiledSub->NumParameters; uj++)
+        {
+            r700->ps.consts[uj + unConstOffset][0].f32All = pCompiledSub->ParameterValues[uj][0];
+                   r700->ps.consts[uj + unConstOffset][1].f32All = pCompiledSub->ParameterValues[uj][1];
+                   r700->ps.consts[uj + unConstOffset][2].f32All = pCompiledSub->ParameterValues[uj][2];
+                   r700->ps.consts[uj + unConstOffset][3].f32All = pCompiledSub->ParameterValues[uj][3];
+        }
+        unConstOffset += pCompiledSub->NumParameters;
+    }
+
     return GL_TRUE;
 }
 
index cbb108d21280556ac722e05ac54456e672bc05d8..39c59c9201df9e56ae9f590b43e23ad85c3baf3a 100644 (file)
@@ -48,13 +48,17 @@ struct r700_fragment_program
 };
 
 /* Internal */
+void insert_wpos_code(GLcontext *ctx, struct gl_fragment_program *fprog);
+
 void Map_Fragment_Program(r700_AssemblerBase         *pAsm,
-                         struct gl_fragment_program *mesa_fp);
+                         struct gl_fragment_program *mesa_fp,
+                          GLcontext *ctx); 
 GLboolean Find_Instruction_Dependencies_fp(struct r700_fragment_program *fp,
                                           struct gl_fragment_program   *mesa_fp);
 
 GLboolean r700TranslateFragmentShader(struct r700_fragment_program *fp,
-                                     struct gl_fragment_program   *mesa_vp);
+                                     struct gl_fragment_program   *mesa_vp,
+                                      GLcontext *ctx); 
 
 /* Interface */
 extern void r700SelectFragmentShader(GLcontext *ctx);
index 47f89c91f87cd57da04b193ab4fbb0c1ec62cb23..eab27cbd8429aeb9582051b48722a7c557e034f6 100644 (file)
@@ -526,6 +526,9 @@ static void r700ConvertAttrib(GLcontext *ctx, int count,
 
     radeonAllocDmaRegion(&context->radeon, &attr->bo, &attr->bo_offset, 
                          sizeof(GLfloat) * input->Size * count, 32);
+
+    radeon_bo_map(attr->bo, 1);
+
     dst_ptr = (GLfloat *)ADD_POINTERS(attr->bo->ptr, attr->bo_offset);
 
     assert(src_ptr != NULL);
@@ -559,6 +562,8 @@ static void r700ConvertAttrib(GLcontext *ctx, int count,
             break;
     }
 
+    radeon_bo_unmap(attr->bo);
+
     if (mapped_named_bo) 
     {
         ctx->Driver.UnmapBuffer(ctx, GL_ARRAY_BUFFER, input->BufferObj);
@@ -577,6 +582,8 @@ static void r700AlignDataToDword(GLcontext *ctx,
 
     radeonAllocDmaRegion(&context->radeon, &attr->bo, &attr->bo_offset, size, 32);
 
+    radeon_bo_map(attr->bo, 1);
+
     if (!input->BufferObj->Pointer) 
     {
         ctx->Driver.MapBuffer(ctx, GL_ARRAY_BUFFER, GL_READ_ONLY_ARB, input->BufferObj);
@@ -596,6 +603,7 @@ static void r700AlignDataToDword(GLcontext *ctx,
         }
     }
 
+    radeon_bo_unmap(attr->bo);
     if (mapped_named_bo) 
     {
         ctx->Driver.UnmapBuffer(ctx, GL_ARRAY_BUFFER, input->BufferObj);
@@ -664,14 +672,18 @@ static void r700SetupStreams(GLcontext *ctx, const struct gl_client_array *input
 
                 radeonAllocDmaRegion(&context->radeon, &context->stream_desc[index].bo, 
                                      &context->stream_desc[index].bo_offset, size, 32);
+
+                radeon_bo_map(context->stream_desc[index].bo, 1);
                 assert(context->stream_desc[index].bo->ptr != NULL);
+
+
                 dst = (uint32_t *)ADD_POINTERS(context->stream_desc[index].bo->ptr, 
                                                context->stream_desc[index].bo_offset);
 
                 switch (context->stream_desc[index].dwords) 
                 {
                 case 1:                     
-                    radeonEmitVec4(dst, input[i]->Ptr, input[i]->StrideB, local_count);                         
+                    radeonEmitVec4(dst, input[i]->Ptr, input[i]->StrideB, local_count);
                     break;
                 case 2: 
                     radeonEmitVec8(dst, input[i]->Ptr, input[i]->StrideB, local_count); 
@@ -686,6 +698,7 @@ static void r700SetupStreams(GLcontext *ctx, const struct gl_client_array *input
                     assert(0); 
                     break;
                 }
+               radeon_bo_unmap(context->stream_desc[index].bo);
             }
         }
 
@@ -757,6 +770,7 @@ static void r700FixupIndexBuffer(GLcontext *ctx, const struct _mesa_index_buffer
        radeonAllocDmaRegion(&context->radeon, &context->ind_buf.bo,
                             &context->ind_buf.bo_offset, size, 4);
 
+       radeon_bo_map(context->ind_buf.bo, 1);
        assert(context->ind_buf.bo->ptr != NULL);
        out = (GLuint *)ADD_POINTERS(context->ind_buf.bo->ptr, context->ind_buf.bo_offset);
 
@@ -770,6 +784,7 @@ static void r700FixupIndexBuffer(GLcontext *ctx, const struct _mesa_index_buffer
             *out++ = in[i];
         }
 
+       radeon_bo_unmap(context->ind_buf.bo);
 #if MESA_BIG_ENDIAN
     }
     else
@@ -780,6 +795,7 @@ static void r700FixupIndexBuffer(GLcontext *ctx, const struct _mesa_index_buffer
        radeonAllocDmaRegion(&context->radeon, &context->ind_buf.bo,
                             &context->ind_buf.bo_offset, size, 4);
 
+       radeon_bo_map(context->ind_buf.bo, 1);
        assert(context->ind_buf.bo->ptr != NULL);
        out = (GLuint *)ADD_POINTERS(context->ind_buf.bo->ptr, context->ind_buf.bo_offset);
 
@@ -792,6 +808,7 @@ static void r700FixupIndexBuffer(GLcontext *ctx, const struct _mesa_index_buffer
         {
             *out++ = in[i];
         }
+       radeon_bo_unmap(context->ind_buf.bo);
 #endif
     }
 
@@ -837,11 +854,13 @@ static void r700SetupIndexBuffer(GLcontext *ctx, const struct _mesa_index_buffer
 
        radeonAllocDmaRegion(&context->radeon, &context->ind_buf.bo,
                             &context->ind_buf.bo_offset, size, 4);
+       radeon_bo_map(context->ind_buf.bo, 1);
        assert(context->ind_buf.bo->ptr != NULL);
        dst_ptr = ADD_POINTERS(context->ind_buf.bo->ptr, context->ind_buf.bo_offset);
 
         _mesa_memcpy(dst_ptr, src_ptr, size);
 
+       radeon_bo_unmap(context->ind_buf.bo);
         context->ind_buf.is_32bit = (mesa_ind_buf->type == GL_UNSIGNED_INT);
         context->ind_buf.count = mesa_ind_buf->count;
 
index 955ea4e4e1da925398c4f72de2ba4f9f04937ba5..2eed1acc2f55f2f26139291e67a4299221bd530f 100644 (file)
@@ -159,13 +159,18 @@ void Init_R700_Shader(R700_Shader * pShader)
        pShader->lstVTXInstructions.uNumOfNode=0;
 }
 
+void SetActiveCFlist(R700_Shader *pShader, TypedShaderList * plstCF)
+{
+    pShader->plstCFInstructions_active = plstCF;
+}
+
 void AddCFInstruction(R700_Shader *pShader, R700ControlFlowInstruction *pCFInst)
 {
     R700ControlFlowSXClause*  pSXClause; 
     R700ControlFlowSMXClause* pSMXClause;
 
-    pCFInst->m_uIndex = pShader->lstCFInstructions.uNumOfNode;
-    AddInstToList(&(pShader->lstCFInstructions)
+    pCFInst->m_uIndex = pShader->plstCFInstructions_active->uNumOfNode;
+    AddInstToList(pShader->plstCFInstructions_active
                   (R700ShaderInstruction*)pCFInst);
     pShader->uShaderBinaryDWORDSize += GetInstructionSize(pCFInst->m_ShaderInstType);
 
index c6a058617ec070bcb2fb4110e51140535f4ff24e..0599ffd901f8542b61502a65917ad484bec68ff2 100644 (file)
@@ -109,6 +109,7 @@ typedef struct R700_Shader
     GLuint  uStackSize;
     GLuint  uMaxCallDepth;
 
+    TypedShaderList * plstCFInstructions_active;
        TypedShaderList lstCFInstructions;
        TypedShaderList lstALUInstructions;
        TypedShaderList lstTEXInstructions;
@@ -132,13 +133,13 @@ void TakeInstOutFromList(TypedShaderList * plstCFInstructions, R700ShaderInstruc
 void ResolveLinks(R700_Shader *pShader);
 void Assemble(R700_Shader *pShader);
 
-
 //Interface
 void Init_R700_Shader(R700_Shader * pShader);
 void AddCFInstruction(R700_Shader *pShader, R700ControlFlowInstruction *pCFInst);
 void AddVTXInstruction(R700_Shader *pShader, R700VertexInstruction *pVTXInst);
 void AddTEXInstruction(R700_Shader *pShader, R700TextureInstruction *pTEXInst);
 void AddALUInstruction(R700_Shader *pShader, R700ALUInstruction *pALUInst);
+void SetActiveCFlist(R700_Shader *pShader, TypedShaderList * plstCF);
 
 void LoadProgram(R700_Shader *pShader);
 void UpdateShaderRegisters(R700_Shader *pShader);
index 2829cca0a3cb1c87af4f583d6dac42ece0591420..cdb9a570f7ccf036442b979fee927eeff5631a06 100644 (file)
 #define SQ_FETCH_RESOURCE_VS_OFFSET    0x000000a0
 #define SQ_FETCH_RESOURCE_VS_COUNT     0x000000b0
 
+//richard dec.10 glsl
+#define SQ_TEX_SAMPLER_PS_OFFSET       0x00000000
+#define SQ_TEX_SAMPLER_PS_COUNT        0x00000012
+#define SQ_TEX_SAMPLER_VS_OFFSET       0x00000012
+#define SQ_TEX_SAMPLER_VS_COUNT        0x00000012
+//-------------------
+
 #define SHADERINST_TYPEMASK_CF  0x10
 #define SHADERINST_TYPEMASK_ALU 0x20
 #define SHADERINST_TYPEMASK_TEX 0x40
index ffc6068bd88cbbf12076ced3ed4272b7a9b01c0c..90fac078ff0bcae338c8e18e33d047fb28fbf169 100644 (file)
@@ -111,6 +111,15 @@ unsigned int Map_Vertex_Output(r700_AssemblerBase       *pAsm,
                }
        }
 
+    for(i=VERT_RESULT_VAR0; i<VERT_RESULT_MAX; i++)
+       {
+               unBit = 1 << i;
+               if(mesa_vp->Base.OutputsWritten & unBit)
+               {
+                       pAsm->ucVP_OutputMap[i] = unTotal++;
+               }
+       }
+
        return (unTotal - unStart);
 }
 
@@ -235,6 +244,8 @@ void Map_Vertex_Program(GLcontext *ctx,
         pAsm->number_used_registers += mesa_vp->Base.NumTemporaries;
     }
 
+    pAsm->flag_reg_index = pAsm->number_used_registers++;
+
     pAsm->uFirstHelpReg = pAsm->number_used_registers;
 }
 
@@ -324,7 +335,18 @@ struct r700_vertex_program* r700TranslateVertexShader(GLcontext *ctx,
                return NULL;
        }
 
-       if(GL_FALSE == AssembleInstr(vp->mesa_program->Base.NumInstructions,
+    InitShaderProgram(&(vp->r700AsmCode));
+
+    for(i=0; i < MAX_SAMPLERS; i++)
+    {
+        vp->r700AsmCode.SamplerUnits[i] = vp->mesa_program->Base.SamplerUnits[i];
+    }
+
+    vp->r700AsmCode.unCurNumILInsts = vp->mesa_program->Base.NumInstructions;
+
+       if(GL_FALSE == AssembleInstr(0,
+                                 0,
+                                 vp->mesa_program->Base.NumInstructions,
                                  &(vp->mesa_program->Base.Instructions[0]),
                                  &(vp->r700AsmCode)) )
        {
@@ -336,6 +358,11 @@ struct r700_vertex_program* r700TranslateVertexShader(GLcontext *ctx,
         return NULL;
     }
 
+    if( GL_FALSE == RelocProgram(&(vp->r700AsmCode), &(vp->mesa_program->Base)) )
+    {
+        return GL_FALSE;
+    }
+
     vp->r700Shader.nRegs = (vp->r700AsmCode.number_used_registers == 0) ? 0 
                          : (vp->r700AsmCode.number_used_registers - 1);
 
@@ -612,6 +639,12 @@ GLboolean r700SetupVertexProgram(GLcontext * ctx)
     paramList = vp->mesa_program->Base.Parameters;
 
     if(NULL != paramList) {
+        /* vp->mesa_program was cloned, not updated by glsl shader api. */
+        /* _mesa_reference_program has already checked glsl shProg is ok and set ctx->VertexProgem._Current */
+        /* so, use ctx->VertexProgem._Current */       
+        struct gl_program_parameter_list *paramListOrginal = 
+                         paramListOrginal = ctx->VertexProgram._Current->Base.Parameters;
+         
            _mesa_load_state_parameters(ctx, paramList);
 
            if (paramList->NumParameters > R700_MAX_DX9_CONSTS)
@@ -624,13 +657,42 @@ GLboolean r700SetupVertexProgram(GLcontext * ctx)
            unNumParamData = paramList->NumParameters;
 
            for(ui=0; ui<unNumParamData; ui++) {
-                   r700->vs.consts[ui][0].f32All = paramList->ParameterValues[ui][0];
-                   r700->vs.consts[ui][1].f32All = paramList->ParameterValues[ui][1];
-                   r700->vs.consts[ui][2].f32All = paramList->ParameterValues[ui][2];
-                   r700->vs.consts[ui][3].f32All = paramList->ParameterValues[ui][3];
+            if(paramList->Parameters[ui].Type == PROGRAM_UNIFORM) 
+            {
+                r700->vs.consts[ui][0].f32All = paramListOrginal->ParameterValues[ui][0];
+                       r700->vs.consts[ui][1].f32All = paramListOrginal->ParameterValues[ui][1];
+                       r700->vs.consts[ui][2].f32All = paramListOrginal->ParameterValues[ui][2];
+                       r700->vs.consts[ui][3].f32All = paramListOrginal->ParameterValues[ui][3];
+            }
+            else
+            {
+                       r700->vs.consts[ui][0].f32All = paramList->ParameterValues[ui][0];
+                       r700->vs.consts[ui][1].f32All = paramList->ParameterValues[ui][1];
+                       r700->vs.consts[ui][2].f32All = paramList->ParameterValues[ui][2];
+                       r700->vs.consts[ui][3].f32All = paramList->ParameterValues[ui][3];
+            }
            }
     } else
            r700->vs.num_consts = 0;
 
+    COMPILED_SUB * pCompiledSub;
+    GLuint uj;
+    GLuint unConstOffset = r700->vs.num_consts;
+    for(ui=0; ui<vp->r700AsmCode.unNumPresub; ui++)
+    {
+        pCompiledSub = vp->r700AsmCode.presubs[ui].pCompiledSub;
+
+        r700->vs.num_consts += pCompiledSub->NumParameters;
+
+        for(uj=0; uj<pCompiledSub->NumParameters; uj++)
+        {
+            r700->vs.consts[uj + unConstOffset][0].f32All = pCompiledSub->ParameterValues[uj][0];
+                   r700->vs.consts[uj + unConstOffset][1].f32All = pCompiledSub->ParameterValues[uj][1];
+                   r700->vs.consts[uj + unConstOffset][2].f32All = pCompiledSub->ParameterValues[uj][2];
+                   r700->vs.consts[uj + unConstOffset][3].f32All = pCompiledSub->ParameterValues[uj][3];
+        }
+        unConstOffset += pCompiledSub->NumParameters;
+    }
+
     return GL_TRUE;
 }
index 9b64c216859741c16fedf5009a8068513c6ea6ff..c0b3165ddad97fed8dceb2cd82621e97c94cadc6 100644 (file)
@@ -641,6 +641,27 @@ void radeonCopySubBuffer(__DRIdrawablePrivate * dPriv,
        }
 }
 
+/**
+ * Check if we're about to draw into the front color buffer.
+ * If so, set the intel->front_buffer_dirty field to true.
+ */
+void
+radeon_check_front_buffer_rendering(GLcontext *ctx)
+{
+       radeonContextPtr radeon = RADEON_CONTEXT(ctx);
+       const struct gl_framebuffer *fb = ctx->DrawBuffer;
+
+       if (fb->Name == 0) {
+               /* drawing to window system buffer */
+               if (fb->_NumColorDrawBuffers > 0) {
+                       if (fb->_ColorDrawBufferIndexes[0] == BUFFER_FRONT_LEFT) {
+                               radeon->front_buffer_dirty = GL_TRUE;
+                       }
+               }
+       }
+}
+
+
 void radeon_draw_buffer(GLcontext *ctx, struct gl_framebuffer *fb)
 {
        radeonContextPtr radeon = RADEON_CONTEXT(ctx);
@@ -817,7 +838,7 @@ void radeonDrawBuffer( GLcontext *ctx, GLenum mode )
        */
                if (!was_front_buffer_rendering && radeon->is_front_buffer_rendering) {
                        radeon_update_renderbuffers(radeon->dri.context,
-                               radeon->dri.context->driDrawablePriv);
+                               radeon->dri.context->driDrawablePriv, GL_FALSE);
       }
        }
 
@@ -834,7 +855,7 @@ void radeonReadBuffer( GLcontext *ctx, GLenum mode )
 
                if (!was_front_buffer_reading && rmesa->is_front_buffer_reading) {
                        radeon_update_renderbuffers(rmesa->dri.context,
-                                                   rmesa->dri.context->driReadablePriv);
+                                                   rmesa->dri.context->driReadablePriv, GL_FALSE);
                }
        }
        /* nothing, until we implement h/w glRead/CopyPixels or CopyTexImage */
@@ -885,9 +906,9 @@ void radeon_viewport(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei he
                if (radeon->is_front_buffer_rendering) {
                        ctx->Driver.Flush(ctx);
                }
-               radeon_update_renderbuffers(driContext, driContext->driDrawablePriv);
+               radeon_update_renderbuffers(driContext, driContext->driDrawablePriv, GL_FALSE);
                if (driContext->driDrawablePriv != driContext->driReadablePriv)
-                       radeon_update_renderbuffers(driContext, driContext->driReadablePriv);
+                       radeon_update_renderbuffers(driContext, driContext->driReadablePriv, GL_FALSE);
        }
 
        old_viewport = ctx->Driver.Viewport;
@@ -1095,7 +1116,7 @@ void radeonFlush(GLcontext *ctx)
           then no point flushing anything at all.
        */
        if (!radeon->dma.flush && !radeon->cmdbuf.cs->cdw && is_empty_list(&radeon->dma.reserved))
-               return;
+               goto flush_front;
 
        if (radeon->dma.flush)
                radeon->dma.flush( ctx );
@@ -1103,6 +1124,7 @@ void radeonFlush(GLcontext *ctx)
        if (radeon->cmdbuf.cs->cdw)
                rcommonFlushCmdBuf(radeon, __FUNCTION__);
 
+flush_front:
        if ((ctx->DrawBuffer->Name == 0) && radeon->front_buffer_dirty) {
                __DRIscreen *const screen = radeon->radeonScreen->driScreen;
 
@@ -1208,7 +1230,7 @@ int rcommonFlushCmdBuf(radeonContextPtr rmesa, const char *caller)
                fprintf(stderr, "drmRadeonCmdBuffer: %d. Kernel failed to "
                                "parse or rejected command stream. See dmesg "
                                "for more info.\n", ret);
-               _mesa_exit(ret);
+               exit(ret);
        }
 
        return ret;
index 0608fe2418c5b7daa1804baa2678c37f4069956a..faad145cc4978fcfce30b6469a0e5849c80b9935 100644 (file)
@@ -43,6 +43,8 @@ radeon_renderbuffer_set_bo(struct radeon_renderbuffer *rb,
                           struct radeon_bo *bo);
 struct radeon_renderbuffer *
 radeon_create_renderbuffer(gl_format format, __DRIdrawablePrivate *driDrawPriv);
+
+void radeon_check_front_buffer_rendering(GLcontext *ctx);
 static inline struct radeon_renderbuffer *radeon_renderbuffer(struct gl_renderbuffer *rb)
 {
        struct radeon_renderbuffer *rrb = (struct radeon_renderbuffer *)rb;
index 71f70d724b9458b9af7675b4997c626783ef0af0..5c68bf5df6c3eb7c6108ab999948400490b7221d 100644 (file)
@@ -499,7 +499,8 @@ radeon_bits_per_pixel(const struct radeon_renderbuffer *rb)
 }
 
 void
-radeon_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable)
+radeon_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable,
+                           GLboolean front_only)
 {
        unsigned int attachments[10];
        __DRIbuffer *buffers = NULL;
@@ -525,7 +526,7 @@ radeon_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable)
                struct radeon_renderbuffer *stencil_rb;
 
                i = 0;
-               if ((radeon->is_front_buffer_rendering ||
+               if ((front_only || radeon->is_front_buffer_rendering ||
                     radeon->is_front_buffer_reading ||
                     !draw->color_rb[1])
                    && draw->color_rb[0]) {
@@ -533,23 +534,25 @@ radeon_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable)
                        attachments[i++] = radeon_bits_per_pixel(draw->color_rb[0]);
                }
 
-               if (draw->color_rb[1]) {
-                       attachments[i++] = __DRI_BUFFER_BACK_LEFT;
-                       attachments[i++] = radeon_bits_per_pixel(draw->color_rb[1]);
-               }
+               if (!front_only) {
+                       if (draw->color_rb[1]) {
+                               attachments[i++] = __DRI_BUFFER_BACK_LEFT;
+                               attachments[i++] = radeon_bits_per_pixel(draw->color_rb[1]);
+                       }
 
-               depth_rb = radeon_get_renderbuffer(&draw->base, BUFFER_DEPTH);
-               stencil_rb = radeon_get_renderbuffer(&draw->base, BUFFER_STENCIL);
-
-               if ((depth_rb != NULL) && (stencil_rb != NULL)) {
-                       attachments[i++] = __DRI_BUFFER_DEPTH_STENCIL;
-                       attachments[i++] = radeon_bits_per_pixel(depth_rb);
-               } else if (depth_rb != NULL) {
-                       attachments[i++] = __DRI_BUFFER_DEPTH;
-                       attachments[i++] = radeon_bits_per_pixel(depth_rb);
-               } else if (stencil_rb != NULL) {
-                       attachments[i++] = __DRI_BUFFER_STENCIL;
-                       attachments[i++] = radeon_bits_per_pixel(stencil_rb);
+                       depth_rb = radeon_get_renderbuffer(&draw->base, BUFFER_DEPTH);
+                       stencil_rb = radeon_get_renderbuffer(&draw->base, BUFFER_STENCIL);
+
+                       if ((depth_rb != NULL) && (stencil_rb != NULL)) {
+                               attachments[i++] = __DRI_BUFFER_DEPTH_STENCIL;
+                               attachments[i++] = radeon_bits_per_pixel(depth_rb);
+                       } else if (depth_rb != NULL) {
+                               attachments[i++] = __DRI_BUFFER_DEPTH;
+                               attachments[i++] = radeon_bits_per_pixel(depth_rb);
+                       } else if (stencil_rb != NULL) {
+                               attachments[i++] = __DRI_BUFFER_STENCIL;
+                               attachments[i++] = radeon_bits_per_pixel(stencil_rb);
+                       }
                }
 
                buffers = (*screen->dri2.loader->getBuffersWithFormat)(drawable,
@@ -562,12 +565,14 @@ radeon_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable)
                i = 0;
                if (draw->color_rb[0])
                        attachments[i++] = __DRI_BUFFER_FRONT_LEFT;
-               if (draw->color_rb[1])
-                       attachments[i++] = __DRI_BUFFER_BACK_LEFT;
-               if (radeon_get_renderbuffer(&draw->base, BUFFER_DEPTH))
-                       attachments[i++] = __DRI_BUFFER_DEPTH;
-               if (radeon_get_renderbuffer(&draw->base, BUFFER_STENCIL))
-                       attachments[i++] = __DRI_BUFFER_STENCIL;
+               if (!front_only) {
+                       if (draw->color_rb[1])
+                               attachments[i++] = __DRI_BUFFER_BACK_LEFT;
+                       if (radeon_get_renderbuffer(&draw->base, BUFFER_DEPTH))
+                               attachments[i++] = __DRI_BUFFER_DEPTH;
+                       if (radeon_get_renderbuffer(&draw->base, BUFFER_STENCIL))
+                               attachments[i++] = __DRI_BUFFER_STENCIL;
+               }
 
                buffers = (*screen->dri2.loader->getBuffers)(drawable,
                                                                 &drawable->w,
@@ -735,9 +740,9 @@ GLboolean radeonMakeCurrent(__DRIcontextPrivate * driContextPriv,
        readfb = driReadPriv->driverPrivate;
 
        if (driContextPriv->driScreenPriv->dri2.enabled) {
-               radeon_update_renderbuffers(driContextPriv, driDrawPriv);
+               radeon_update_renderbuffers(driContextPriv, driDrawPriv, GL_FALSE);
                if (driDrawPriv != driReadPriv)
-                       radeon_update_renderbuffers(driContextPriv, driReadPriv);
+                       radeon_update_renderbuffers(driContextPriv, driReadPriv, GL_FALSE);
                _mesa_reference_renderbuffer(&radeon->state.color.rb,
                        &(radeon_get_renderbuffer(&drfb->base, BUFFER_BACK_LEFT)->base));
                _mesa_reference_renderbuffer(&radeon->state.depth.rb,
index 629874894282e8f1be100b1e80489935c2d2a545..0739496e0323af37cf5674407bc779413fac8963 100644 (file)
@@ -328,6 +328,7 @@ struct radeon_swtcl_info {
        GLuint vertex_attr_count;
 
        GLuint emit_prediction;
+        struct radeon_bo *bo;
 };
 
 #define RADEON_MAX_AOS_ARRAYS          16
@@ -585,7 +586,8 @@ GLboolean radeonInitContext(radeonContextPtr radeon,
 
 void radeonCleanupContext(radeonContextPtr radeon);
 GLboolean radeonUnbindContext(__DRIcontextPrivate * driContextPriv);
-void radeon_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable);
+void radeon_update_renderbuffers(__DRIcontext *context, __DRIdrawable *drawable,
+                                GLboolean front_only);
 GLboolean radeonMakeCurrent(__DRIcontextPrivate * driContextPriv,
                            __DRIdrawablePrivate * driDrawPriv,
                            __DRIdrawablePrivate * driReadPriv);
index 232972d7ec748a3803a686686cd642ac917176d4..d31e4e47ddbbae1401bd5e14b7c373a76fc0eb42 100644 (file)
@@ -151,6 +151,7 @@ void rcommon_emit_vector(GLcontext * ctx, struct radeon_aos *aos,
        aos->components = size;
        aos->count = count;
 
+       radeon_bo_map(aos->bo, 1);
        out = (uint32_t*)((char*)aos->bo->ptr + aos->offset);
        switch (size) {
        case 1: radeonEmitVec4(out, data, stride, count); break;
@@ -161,6 +162,7 @@ void rcommon_emit_vector(GLcontext * ctx, struct radeon_aos *aos,
                assert(0);
                break;
        }
+       radeon_bo_unmap(aos->bo);
 }
 
 void radeon_init_dma(radeonContextPtr rmesa)
@@ -183,10 +185,6 @@ void radeonRefillCurrentDmaRegion(radeonContextPtr rmesa, int size)
                        __FUNCTION__, size, rmesa->dma.minimum_size);
 
 
-       /* unmap old reserved bo */
-       if (!is_empty_list(&rmesa->dma.reserved))
-               radeon_bo_unmap(first_elem(&rmesa->dma.reserved)->bo);
-
        if (is_empty_list(&rmesa->dma.free)
              || last_elem(&rmesa->dma.free)->bo->size < size) {
                dma_bo = CALLOC_STRUCT(radeon_dma_bo);
@@ -223,8 +221,6 @@ again_alloc:
         /* Cmd buff have been flushed in radeon_revalidate_bos */
                goto again_alloc;
        }
-
-       radeon_bo_map(first_elem(&rmesa->dma.reserved)->bo, 1);
 }
 
 /* Allocates a region from rmesa->dma.current.  If there isn't enough
@@ -281,7 +277,6 @@ void radeonFreeDmaRegions(radeonContextPtr rmesa)
 
        foreach_s(dma_bo, temp, &rmesa->dma.reserved) {
                remove_from_list(dma_bo);
-               radeon_bo_unmap(dma_bo->bo);
                radeon_bo_unref(dma_bo->bo);
                FREE(dma_bo);
        }
@@ -361,9 +356,6 @@ void radeonReleaseDmaRegions(radeonContextPtr rmesa)
                insert_at_tail(&rmesa->dma.free, dma_bo);
        }
 
-       /* unmap the last dma region */
-       if (!is_empty_list(&rmesa->dma.reserved))
-               radeon_bo_unmap(first_elem(&rmesa->dma.reserved)->bo);
        /* move reserved to wait list */
        foreach_s(dma_bo, temp, &rmesa->dma.reserved) {
                /* free objects that are too small to be used because of large request */
@@ -397,11 +389,12 @@ void rcommon_flush_last_swtcl_prim( GLcontext *ctx  )
        radeonContextPtr rmesa = RADEON_CONTEXT(ctx);
        struct radeon_dma *dma = &rmesa->dma;
                
-
        if (RADEON_DEBUG & RADEON_IOCTL)
                fprintf(stderr, "%s\n", __FUNCTION__);
        dma->flush = NULL;
 
+       radeon_bo_unmap(rmesa->swtcl.bo);
+
        if (!is_empty_list(&dma->reserved)) {
            GLuint current_offset = dma->current_used;
 
@@ -416,6 +409,8 @@ void rcommon_flush_last_swtcl_prim( GLcontext *ctx  )
            }
            rmesa->swtcl.numverts = 0;
        }
+       radeon_bo_unref(rmesa->swtcl.bo);
+       rmesa->swtcl.bo = NULL;
 }
 /* Alloc space in the current dma region.
  */
@@ -426,6 +421,7 @@ rcommonAllocDmaLowVerts( radeonContextPtr rmesa, int nverts, int vsize )
        void *head;
        if (RADEON_DEBUG & RADEON_IOCTL)
                fprintf(stderr, "%s\n", __FUNCTION__);
+
        if(is_empty_list(&rmesa->dma.reserved)
              ||rmesa->dma.current_vertexptr + bytes > first_elem(&rmesa->dma.reserved)->bo->size) {
                if (rmesa->dma.flush) {
@@ -449,7 +445,13 @@ rcommonAllocDmaLowVerts( radeonContextPtr rmesa, int nverts, int vsize )
                 rmesa->swtcl.numverts * rmesa->swtcl.vertex_size * 4 ==
                 rmesa->dma.current_vertexptr );
 
-       head = (first_elem(&rmesa->dma.reserved)->bo->ptr + rmesa->dma.current_vertexptr);
+       if (!rmesa->swtcl.bo) {
+               rmesa->swtcl.bo = first_elem(&rmesa->dma.reserved)->bo;
+               radeon_bo_ref(rmesa->swtcl.bo);
+               radeon_bo_map(rmesa->swtcl.bo, 1);
+       }
+
+       head = (rmesa->swtcl.bo->ptr + rmesa->dma.current_vertexptr);
        rmesa->dma.current_vertexptr += bytes;
        rmesa->swtcl.numverts += nverts;
        return head;
index fc21069a92c1f7e34d57c57bc3a81d6c728bba08..a536436d55f669a52855739ce35684080ebc92a3 100644 (file)
@@ -166,8 +166,9 @@ radeon_alloc_renderbuffer_storage(GLcontext * ctx, struct gl_renderbuffer *rb,
      uint32_t size;
      uint32_t pitch = ((cpp * width + 63) & ~63) / cpp;
 
-     fprintf(stderr,"Allocating %d x %d radeon RBO (pitch %d)\n", width,
-         height, pitch);
+     if (RADEON_DEBUG & RADEON_MEMORY)
+            fprintf(stderr,"Allocating %d x %d radeon RBO (pitch %d)\n", width,
+                    height, pitch);
 
      size = pitch * height * cpp;
      rrb->pitch = pitch * cpp;
index a0106d00fa227bc95a1ea6267277e90ddd5b320c..13fd6f99719057c148006950d628013bd2b5422c 100644 (file)
@@ -575,6 +575,10 @@ static void radeonClear( GLcontext *ctx, GLbitfield mask )
    GLuint color_mask = 0;
    GLuint orig_mask = mask;
 
+   if (mask & (BUFFER_BIT_FRONT_LEFT | BUFFER_BIT_FRONT_RIGHT)) {
+      rmesa->radeon.front_buffer_dirty = GL_TRUE;
+   }
+
    if ( RADEON_DEBUG & RADEON_IOCTL ) {
       fprintf( stderr, "radeonClear\n");
    }
index 08e1c5d00df44d1c043b9bc83fdef8d0ac101f35..d810e6080eb786d38f4bb440c705ce1e25f6d117 100644 (file)
@@ -76,12 +76,14 @@ static void emit_vecfog(GLcontext *ctx, struct radeon_aos *aos,
 
    /* Emit the data
     */
+   radeon_bo_map(aos->bo, 1);
    out = (uint32_t*)((char*)aos->bo->ptr + aos->offset);
    for (i = 0; i < count; i++) {
       out[0] = radeonComputeFogBlendFactor( ctx, *(GLfloat *)data );
       out++;
       data += stride;
    }
+   radeon_bo_unmap(aos->bo);
 }
 
 static void emit_s0_vec(uint32_t *out, GLvoid *data, int stride, int count)
@@ -151,6 +153,7 @@ static void emit_tex_vector(GLcontext *ctx, struct radeon_aos *aos,
 
    /* Emit the data
     */
+   radeon_bo_map(aos->bo, 1);
    out = (uint32_t*)((char*)aos->bo->ptr + aos->offset);
    switch (size) {
    case 1:
@@ -170,6 +173,7 @@ static void emit_tex_vector(GLcontext *ctx, struct radeon_aos *aos,
       exit(1);
       break;
    }
+   radeon_bo_unmap(aos->bo);
 }
 
 
@@ -196,12 +200,12 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
       if (!rmesa->tcl.obj.buf) 
        rcommon_emit_vector( ctx, 
                             &(rmesa->tcl.aos[nr]),
-                            (char *)VB->ObjPtr->data,
-                            VB->ObjPtr->size,
-                            VB->ObjPtr->stride,
+                            (char *)VB->AttribPtr[_TNL_ATTRIB_POS]->data,
+                            VB->AttribPtr[_TNL_ATTRIB_POS]->size,
+                            VB->AttribPtr[_TNL_ATTRIB_POS]->stride,
                             count);
 
-      switch( VB->ObjPtr->size ) {
+      switch( VB->AttribPtr[_TNL_ATTRIB_POS]->size ) {
       case 4: vfmt |= RADEON_CP_VC_FRMT_W0;
       case 3: vfmt |= RADEON_CP_VC_FRMT_Z;
       case 2: vfmt |= RADEON_CP_VC_FRMT_XY;
@@ -216,9 +220,9 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
       if (!rmesa->tcl.norm.buf)
         rcommon_emit_vector( ctx, 
                              &(rmesa->tcl.aos[nr]),
-                             (char *)VB->NormalPtr->data,
+                             (char *)VB->AttribPtr[_TNL_ATTRIB_NORMAL]->data,
                              3,
-                             VB->NormalPtr->stride,
+                             VB->AttribPtr[_TNL_ATTRIB_NORMAL]->stride,
                              count);
 
       vfmt |= RADEON_CP_VC_FRMT_N0;
@@ -227,9 +231,9 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
 
    if (inputs & VERT_BIT_COLOR0) {
       int emitsize;
-      if (VB->ColorPtr[0]->size == 4 &&
-         (VB->ColorPtr[0]->stride != 0 ||
-          VB->ColorPtr[0]->data[0][3] != 1.0)) {
+      if (VB->AttribPtr[_TNL_ATTRIB_COLOR0]->size == 4 &&
+         (VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride != 0 ||
+          VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data[0][3] != 1.0)) {
         vfmt |= RADEON_CP_VC_FRMT_FPCOLOR | RADEON_CP_VC_FRMT_FPALPHA;
         emitsize = 4;
       }
@@ -242,9 +246,9 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
       if (!rmesa->tcl.rgba.buf)
        rcommon_emit_vector( ctx,
                             &(rmesa->tcl.aos[nr]),
-                            (char *)VB->ColorPtr[0]->data,
+                            (char *)VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data,
                             emitsize,
-                            VB->ColorPtr[0]->stride,
+                            VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride,
                             count);
 
       nr++;
@@ -256,9 +260,9 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
 
        rcommon_emit_vector( ctx,
                             &(rmesa->tcl.aos[nr]),
-                            (char *)VB->SecondaryColorPtr[0]->data,
+                            (char *)VB->AttribPtr[_TNL_ATTRIB_COLOR1]->data,
                             3,
-                            VB->SecondaryColorPtr[0]->stride,
+                            VB->AttribPtr[_TNL_ATTRIB_COLOR1]->stride,
                             count);
       }
 
@@ -273,8 +277,8 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
       if (!rmesa->tcl.fog.buf)
         emit_vecfog( ctx,
                      &(rmesa->tcl.aos[nr]),
-                     (char *)VB->FogCoordPtr->data,
-                     VB->FogCoordPtr->stride,
+                     (char *)VB->AttribPtr[_TNL_ATTRIB_FOG]->data,
+                     VB->AttribPtr[_TNL_ATTRIB_FOG]->stride,
                      count);
 
       vfmt |= RADEON_CP_VC_FRMT_FPFOG;
@@ -290,24 +294,24 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
         if (!rmesa->tcl.tex[unit].buf)
            emit_tex_vector( ctx,
                             &(rmesa->tcl.aos[nr]),
-                            (char *)VB->TexCoordPtr[unit]->data,
-                            VB->TexCoordPtr[unit]->size,
-                            VB->TexCoordPtr[unit]->stride,
+                            (char *)VB->AttribPtr[_TNL_ATTRIB_TEX0 + unit]->data,
+                            VB->AttribPtr[_TNL_ATTRIB_TEX0 + unit]->size,
+                            VB->AttribPtr[_TNL_ATTRIB_TEX0 + unit]->stride,
                             count );
         nr++;
 
         vfmt |= RADEON_ST_BIT(unit);
          /* assume we need the 3rd coord if texgen is active for r/q OR at least
            3 coords are submitted. This may not be 100% correct */
-         if (VB->TexCoordPtr[unit]->size >= 3) {
+         if (VB->AttribPtr[_TNL_ATTRIB_TEX0 + unit]->size >= 3) {
            vtx |= RADEON_Q_BIT(unit);
            vfmt |= RADEON_Q_BIT(unit);
         }
         if ( (ctx->Texture.Unit[unit].TexGenEnabled & (R_BIT | Q_BIT)) )
            vtx |= RADEON_Q_BIT(unit);
-        else if ((VB->TexCoordPtr[unit]->size >= 3) &&
+        else if ((VB->AttribPtr[_TNL_ATTRIB_TEX0 + unit]->size >= 3) &&
                  ((ctx->Texture.Unit[unit]._ReallyEnabled & (TEXTURE_CUBE_BIT)) == 0)) {
-           GLuint swaptexmatcol = (VB->TexCoordPtr[unit]->size - 3);
+           GLuint swaptexmatcol = (VB->AttribPtr[_TNL_ATTRIB_TEX0 + unit]->size - 3);
            if (((rmesa->NeedTexMatrix >> unit) & 1) &&
                 (swaptexmatcol != ((rmesa->TexMatColSwap >> unit) & 1)))
               radeonUploadTexMatrix( rmesa, unit, swaptexmatcol ) ;
index 515783135d6f074872c3eff293fa36ff53a87a09..d764ccb982695191a9a3a19414c59bb97ef6a1fc 100644 (file)
@@ -56,18 +56,18 @@ static void TAG(emit)( GLcontext *ctx,
 
    radeon_print(RADEON_SWRENDER, RADEON_VERBOSE, "%s\n", __FUNCTION__);
 
-   coord = (GLuint (*)[4])VB->ObjPtr->data;
-   coord_stride = VB->ObjPtr->stride;
+   coord = (GLuint (*)[4])VB->AttribPtr[_TNL_ATTRIB_POS]->data;
+   coord_stride = VB->AttribPtr[_TNL_ATTRIB_POS]->stride;
 
    if (DO_TEX2) {
-      if (VB->TexCoordPtr[2]) {
+      if (VB->AttribPtr[_TNL_ATTRIB_TEX2]) {
         const GLuint t2 = GET_TEXSOURCE(2);
-        tc2 = (GLuint (*)[4])VB->TexCoordPtr[t2]->data;
-        tc2_stride = VB->TexCoordPtr[t2]->stride;
-        if (DO_PTEX && VB->TexCoordPtr[t2]->size < 3) {
+        tc2 = (GLuint (*)[4])VB->AttribPtr[_TNL_ATTRIB_TEX0 + t2]->data;
+        tc2_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t2]->stride;
+        if (DO_PTEX && VB->AttribPtr[_TNL_ATTRIB_TEX0 + t2]->size < 3) {
            fill_tex |= (1<<2);
         }
-        else if (DO_PTEX && VB->TexCoordPtr[t2]->size < 4) {
+        else if (DO_PTEX && VB->AttribPtr[_TNL_ATTRIB_TEX0 + t2]->size < 4) {
            rqcoordsnoswap |= (1<<2);
         }
       } else {
@@ -77,14 +77,14 @@ static void TAG(emit)( GLcontext *ctx,
    }
 
    if (DO_TEX1) {
-      if (VB->TexCoordPtr[1]) {
+      if (VB->AttribPtr[_TNL_ATTRIB_TEX1]) {
         const GLuint t1 = GET_TEXSOURCE(1);
-        tc1 = (GLuint (*)[4])VB->TexCoordPtr[t1]->data;
-        tc1_stride = VB->TexCoordPtr[t1]->stride;
-        if (DO_PTEX && VB->TexCoordPtr[t1]->size < 3) {
+        tc1 = (GLuint (*)[4])VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->data;
+        tc1_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->stride;
+        if (DO_PTEX && VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->size < 3) {
            fill_tex |= (1<<1);
         }
-        else if (DO_PTEX && VB->TexCoordPtr[t1]->size < 4) {
+        else if (DO_PTEX && VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->size < 4) {
            rqcoordsnoswap |= (1<<1);
         }
       } else {
@@ -94,14 +94,14 @@ static void TAG(emit)( GLcontext *ctx,
    }
 
    if (DO_TEX0) {
-      if (VB->TexCoordPtr[0]) {
+      if (VB->AttribPtr[_TNL_ATTRIB_TEX0]) {
         const GLuint t0 = GET_TEXSOURCE(0);
-        tc0_stride = VB->TexCoordPtr[t0]->stride;
-        tc0 = (GLuint (*)[4])VB->TexCoordPtr[t0]->data;
-        if (DO_PTEX && VB->TexCoordPtr[t0]->size < 3) {
+        tc0_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->stride;
+        tc0 = (GLuint (*)[4])VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->data;
+        if (DO_PTEX && VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->size < 3) {
            fill_tex |= (1<<0);
         }
-        else if (DO_PTEX && VB->TexCoordPtr[t0]->size < 4) {
+        else if (DO_PTEX && VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->size < 4) {
            rqcoordsnoswap |= (1<<0);
         }
       } else {
@@ -112,9 +112,9 @@ static void TAG(emit)( GLcontext *ctx,
    }
 
    if (DO_NORM) {
-      if (VB->NormalPtr) {
-        norm_stride = VB->NormalPtr->stride;
-        norm = (GLuint (*)[4])VB->NormalPtr->data;
+      if (VB->AttribPtr[_TNL_ATTRIB_NORMAL]) {
+        norm_stride = VB->AttribPtr[_TNL_ATTRIB_NORMAL]->stride;
+        norm = (GLuint (*)[4])VB->AttribPtr[_TNL_ATTRIB_NORMAL]->data;
       } else {
         norm_stride = 0;
         norm = (GLuint (*)[4])&ctx->Current.Attrib[VERT_ATTRIB_NORMAL];
@@ -122,9 +122,9 @@ static void TAG(emit)( GLcontext *ctx,
    }
 
    if (DO_RGBA) {
-      if (VB->ColorPtr[0]) {
-        col = VB->ColorPtr[0]->data;
-        col_stride = VB->ColorPtr[0]->stride;
+      if (VB->AttribPtr[_TNL_ATTRIB_COLOR0]) {
+        col = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+        col_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
       } else {
         col = (GLfloat (*)[4])ctx->Current.Attrib[VERT_ATTRIB_COLOR0];
         col_stride = 0;
@@ -132,9 +132,9 @@ static void TAG(emit)( GLcontext *ctx,
    }
 
    if (DO_SPEC_OR_FOG) {
-      if (VB->SecondaryColorPtr[0]) {
-        spec = VB->SecondaryColorPtr[0]->data;
-        spec_stride = VB->SecondaryColorPtr[0]->stride;
+      if (VB->AttribPtr[_TNL_ATTRIB_COLOR1]) {
+        spec = VB->AttribPtr[_TNL_ATTRIB_COLOR1]->data;
+        spec_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR1]->stride;
       } else {
         spec = (GLfloat (*)[4])ctx->Current.Attrib[VERT_ATTRIB_COLOR1];
         spec_stride = 0;
@@ -142,9 +142,9 @@ static void TAG(emit)( GLcontext *ctx,
    }
 
    if (DO_SPEC_OR_FOG) {
-      if (VB->FogCoordPtr) {
-        fog = VB->FogCoordPtr->data;
-        fog_stride = VB->FogCoordPtr->stride;
+      if (VB->AttribPtr[_TNL_ATTRIB_FOG]) {
+        fog = VB->AttribPtr[_TNL_ATTRIB_FOG]->data;
+        fog_stride = VB->AttribPtr[_TNL_ATTRIB_FOG]->stride;
       } else {
         fog = (GLfloat (*)[4])ctx->Current.Attrib[VERT_ATTRIB_FOG];
         fog_stride = 0;
index 78ec1193026bb0e26171803b70c69c7daa1f82a7..98f96ff2a768962b3f4de84ebafca8cfd484f744 100644 (file)
@@ -326,7 +326,7 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
 
    if (1) {
       req |= RADEON_CP_VC_FRMT_Z;
-      if (VB->ObjPtr->size == 4) {
+      if (VB->AttribPtr[_TNL_ATTRIB_POS]->size == 4) {
         req |= RADEON_CP_VC_FRMT_W0;
       }
    }
@@ -348,15 +348,15 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
         req |= RADEON_ST_BIT(unit);
         /* assume we need the 3rd coord if texgen is active for r/q OR at least
            3 coords are submitted. This may not be 100% correct */
-        if (VB->TexCoordPtr[unit]->size >= 3) {
+        if (VB->AttribPtr[_TNL_ATTRIB_TEX0 + unit]->size >= 3) {
            req |= RADEON_Q_BIT(unit);
            vtx |= RADEON_Q_BIT(unit);
         }
         if ( (ctx->Texture.Unit[unit].TexGenEnabled & (R_BIT | Q_BIT)) )
            vtx |= RADEON_Q_BIT(unit);
-        else if ((VB->TexCoordPtr[unit]->size >= 3) &&
+        else if ((VB->AttribPtr[_TNL_ATTRIB_TEX0 + unit]->size >= 3) &&
                  ((ctx->Texture.Unit[unit]._ReallyEnabled & (TEXTURE_CUBE_BIT)) == 0)) {
-           GLuint swaptexmatcol = (VB->TexCoordPtr[unit]->size - 3);
+           GLuint swaptexmatcol = (VB->AttribPtr[_TNL_ATTRIB_TEX0 + unit]->size - 3);
            if (((rmesa->NeedTexMatrix >> unit) & 1) &&
                 (swaptexmatcol != ((rmesa->TexMatColSwap >> unit) & 1)))
               radeonUploadTexMatrix( rmesa, unit, swaptexmatcol ) ;
@@ -390,19 +390,19 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
     * this, add more vertex code (for obj-2, obj-3) or preferably move
     * to maos.  
     */
-   if (VB->ObjPtr->size < 3 || 
-       (VB->ObjPtr->size == 3 && 
+   if (VB->AttribPtr[_TNL_ATTRIB_POS]->size < 3 ||
+       (VB->AttribPtr[_TNL_ATTRIB_POS]->size == 3 &&
        (setup_tab[i].vertex_format & RADEON_CP_VC_FRMT_W0))) {
 
       _math_trans_4f( rmesa->tcl.ObjClean.data,
-                     VB->ObjPtr->data,
-                     VB->ObjPtr->stride,
+                     VB->AttribPtr[_TNL_ATTRIB_POS]->data,
+                     VB->AttribPtr[_TNL_ATTRIB_POS]->stride,
                      GL_FLOAT,
-                     VB->ObjPtr->size,
+                     VB->AttribPtr[_TNL_ATTRIB_POS]->size,
                      0,
                      VB->Count );
 
-      switch (VB->ObjPtr->size) {
+      switch (VB->AttribPtr[_TNL_ATTRIB_POS]->size) {
       case 1:
            _mesa_vector4f_clean_elem(&rmesa->tcl.ObjClean, VB->Count, 1);
       case 2:
@@ -416,14 +416,14 @@ void radeonEmitArrays( GLcontext *ctx, GLuint inputs )
         break;
       }
 
-      VB->ObjPtr = &rmesa->tcl.ObjClean;
+      VB->AttribPtr[_TNL_ATTRIB_POS] = &rmesa->tcl.ObjClean;
    }
 
 
-
+   radeon_bo_map(rmesa->radeon.tcl.aos[0].bo, 1);
    setup_tab[i].emit( ctx, 0, VB->Count, 
                      rmesa->radeon.tcl.aos[0].bo->ptr + rmesa->radeon.tcl.aos[0].offset);
-
+   radeon_bo_unmap(rmesa->radeon.tcl.aos[0].bo);
    //   rmesa->radeon.tcl.aos[0].size = setup_tab[i].vertex_size;
    rmesa->radeon.tcl.aos[0].stride = setup_tab[i].vertex_size;
    rmesa->tcl.vertex_format = setup_tab[i].vertex_format;
index d7b5d71d21f6a8bec436a4910a23f517947f973f..033f26db2a1b69a7ca92c044380cdbe019aa5cd1 100644 (file)
@@ -68,6 +68,19 @@ static unsigned get_compressed_image_size(
        return rowStride * ((height + blockHeight - 1) / blockHeight);
 }
 
+static int find_next_power_of_two(GLuint value)
+{
+       int i, tmp;
+
+       i = 0;
+       tmp = value - 1;
+       while (tmp) {
+               tmp >>= 1;
+               i++;
+       }
+       return (1 << i);
+}
+
 /**
  * Compute sizes and fill in offset and blit information for the given
  * image (determined by \p face and \p level).
@@ -80,25 +93,28 @@ static void compute_tex_image_offset(radeonContextPtr rmesa, radeon_mipmap_tree
 {
        radeon_mipmap_level *lvl = &mt->levels[level];
        uint32_t row_align;
+       GLuint height;
+
+       height = find_next_power_of_two(lvl->height);
 
        /* Find image size in bytes */
        if (_mesa_is_format_compressed(mt->mesaFormat)) {
                lvl->rowstride = get_aligned_compressed_row_stride(mt->mesaFormat, lvl->width, rmesa->texture_compressed_row_align);
-               lvl->size = get_compressed_image_size(mt->mesaFormat, lvl->rowstride, lvl->height);
+               lvl->size = get_compressed_image_size(mt->mesaFormat, lvl->rowstride, height);
        } else if (mt->target == GL_TEXTURE_RECTANGLE_NV) {
                row_align = rmesa->texture_rect_row_align - 1;
                lvl->rowstride = (_mesa_format_row_stride(mt->mesaFormat, lvl->width) + row_align) & ~row_align;
-               lvl->size = lvl->rowstride * lvl->height;
+               lvl->size = lvl->rowstride * height;
        } else if (mt->tilebits & RADEON_TXO_MICRO_TILE) {
                /* tile pattern is 16 bytes x2. mipmaps stay 32 byte aligned,
                 * though the actual offset may be different (if texture is less than
                 * 32 bytes width) to the untiled case */
                lvl->rowstride = (_mesa_format_row_stride(mt->mesaFormat, lvl->width) * 2 + 31) & ~31;
-               lvl->size = lvl->rowstride * ((lvl->height + 1) / 2) * lvl->depth;
+               lvl->size = lvl->rowstride * ((height + 1) / 2) * lvl->depth;
        } else {
                row_align = rmesa->texture_row_align - 1;
                lvl->rowstride = (_mesa_format_row_stride(mt->mesaFormat, lvl->width) + row_align) & ~row_align;
-               lvl->size = lvl->rowstride * lvl->height * lvl->depth;
+               lvl->size = lvl->rowstride * height * lvl->depth;
        }
        assert(lvl->size > 0);
 
@@ -110,7 +126,7 @@ static void compute_tex_image_offset(radeonContextPtr rmesa, radeon_mipmap_tree
        if (RADEON_DEBUG & RADEON_TEXTURE)
          fprintf(stderr,
                  "level %d, face %d: rs:%d %dx%d at %d\n",
-                 level, face, lvl->rowstride, lvl->width, lvl->height, lvl->faces[face].offset);
+                 level, face, lvl->rowstride, lvl->width, height, lvl->faces[face].offset);
 }
 
 static GLuint minify(GLuint size, GLuint levels)
index 665f2b6f51607459d5b8bb8e5aa68e8c4c8ed5ac..8db3d2b143b51531acb383da9c9dd4011e8cf5a3 100644 (file)
@@ -827,18 +827,21 @@ static void map_unmap_rb(struct gl_renderbuffer *rb, int flag)
 }
 
 static void
-radeon_map_unmap_buffers(GLcontext *ctx, GLboolean map)
+radeon_map_unmap_framebuffer(GLcontext *ctx, struct gl_framebuffer *fb,
+                            GLboolean map)
 {
        GLuint i, j;
 
        /* color draw buffers */
        for (j = 0; j < ctx->DrawBuffer->_NumColorDrawBuffers; j++)
-               map_unmap_rb(ctx->DrawBuffer->_ColorDrawBuffers[j], map);
+               map_unmap_rb(fb->_ColorDrawBuffers[j], map);
+
+       map_unmap_rb(fb->_ColorReadBuffer, map);
 
        /* check for render to textures */
        for (i = 0; i < BUFFER_COUNT; i++) {
                struct gl_renderbuffer_attachment *att =
-                       ctx->DrawBuffer->Attachment + i;
+                       fb->Attachment + i;
                struct gl_texture_object *tex = att->Texture;
                if (tex) {
                        /* Render to texture. Note that a mipmapped texture need not
@@ -854,15 +857,15 @@ radeon_map_unmap_buffers(GLcontext *ctx, GLboolean map)
                                radeon_teximage_unmap(image);
                }
        }
-
-       map_unmap_rb(ctx->ReadBuffer->_ColorReadBuffer, map);
-
+       
        /* depth buffer (Note wrapper!) */
-       if (ctx->DrawBuffer->_DepthBuffer)
-               map_unmap_rb(ctx->DrawBuffer->_DepthBuffer->Wrapped, map);
+       if (fb->_DepthBuffer)
+               map_unmap_rb(fb->_DepthBuffer->Wrapped, map);
+
+       if (fb->_StencilBuffer)
+               map_unmap_rb(fb->_StencilBuffer->Wrapped, map);
 
-       if (ctx->DrawBuffer->_StencilBuffer)
-               map_unmap_rb(ctx->DrawBuffer->_StencilBuffer->Wrapped, map);
+       radeon_check_front_buffer_rendering(ctx);
 }
 
 static void radeonSpanRenderStart(GLcontext * ctx)
@@ -887,23 +890,30 @@ static void radeonSpanRenderStart(GLcontext * ctx)
                        ctx->Driver.MapTexture(ctx, ctx->Texture.Unit[i]._Current);
        }
 
-       radeon_map_unmap_buffers(ctx, 1);
+       radeon_map_unmap_framebuffer(ctx, ctx->DrawBuffer, GL_TRUE);
+       if (ctx->ReadBuffer != ctx->DrawBuffer)
+               radeon_map_unmap_framebuffer(ctx, ctx->ReadBuffer, GL_TRUE);
 }
 
 static void radeonSpanRenderFinish(GLcontext * ctx)
 {
        radeonContextPtr rmesa = RADEON_CONTEXT(ctx);
        int i;
+
        _swrast_flush(ctx);
-       if (!rmesa->radeonScreen->driScreen->dri2.enabled) {
-               UNLOCK_HARDWARE(rmesa);
-       }
+
        for (i = 0; i < ctx->Const.MaxTextureImageUnits; i++) {
                if (ctx->Texture.Unit[i]._ReallyEnabled)
                        ctx->Driver.UnmapTexture(ctx, ctx->Texture.Unit[i]._Current);
        }
 
-       radeon_map_unmap_buffers(ctx, 0);
+       radeon_map_unmap_framebuffer(ctx, ctx->DrawBuffer, GL_FALSE);
+       if (ctx->ReadBuffer != ctx->DrawBuffer)
+               radeon_map_unmap_framebuffer(ctx, ctx->ReadBuffer, GL_FALSE);
+
+       if (!rmesa->radeonScreen->driScreen->dri2.enabled) {
+               UNLOCK_HARDWARE(rmesa);
+       }
 }
 
 void radeonInitSpanFuncs(GLcontext * ctx)
index e61f59eaeaf89146c8fa13939a023e7079d01326..8bf1bfbc5752d4c38840dd4d926dbd80a8a16806 100644 (file)
@@ -179,7 +179,7 @@ static void radeonSetVertexFormat( GLcontext *ctx )
 
       for (i = 0; i < ctx->Const.MaxTextureUnits; i++) {
         if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX(i) )) {
-           GLuint sz = VB->TexCoordPtr[i]->size;
+           GLuint sz = VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]->size;
 
            switch (sz) {
            case 1:
@@ -309,7 +309,7 @@ void r100_swtcl_flush(GLcontext *ctx, uint32_t current_offset)
    radeonEmitState(&rmesa->radeon);
    radeonEmitVertexAOS( rmesa,
                        rmesa->radeon.swtcl.vertex_size,
-                       first_elem(&rmesa->radeon.dma.reserved)->bo,
+                       rmesa->radeon.swtcl.bo,
                        current_offset);
 
                      
index 3cbe3b472548b792a532c4c8895fadb92dc32044..84ddcfd4fd36c9462cc84ec887ad5510ce71ea0d 100644 (file)
@@ -672,24 +672,13 @@ void radeonSetTexBuffer2(__DRIcontext *pDRICtx, GLint target, GLint glx_texture_
            return;
        }
 
-       radeon_update_renderbuffers(pDRICtx, dPriv);
-       /* back & depth buffer are useless free them right away */
-       rb = (void*)rfb->base.Attachment[BUFFER_DEPTH].Renderbuffer;
-       if (rb && rb->bo) {
-               radeon_bo_unref(rb->bo);
-        rb->bo = NULL;
-       }
-       rb = (void*)rfb->base.Attachment[BUFFER_BACK_LEFT].Renderbuffer;
-       if (rb && rb->bo) {
-               radeon_bo_unref(rb->bo);
-               rb->bo = NULL;
-       }
+       radeon_update_renderbuffers(pDRICtx, dPriv, GL_TRUE);
        rb = rfb->color_rb[0];
        if (rb->bo == NULL) {
                /* Failed to BO for the buffer */
                return;
        }
-       
+
        _mesa_lock_texture(radeon->glCtx, texObj);
        if (t->bo) {
                radeon_bo_unref(t->bo);
index 32c74f9467e6d0994fcbf5483ad483b376dfbcbc..8221edf38727228bd1e7d95adf0405b2debf8b92 100644 (file)
@@ -252,13 +252,13 @@ static GLboolean run_texnorm_stage( GLcontext *ctx,
          const GLboolean normalizeS = (texObj->WrapS == GL_REPEAT);
          const GLboolean normalizeT = (reallyEnabled & TEXTURE_2D_BIT) &&
             (texObj->WrapT == GL_REPEAT);
-         const GLfloat *in = (GLfloat *)VB->TexCoordPtr[i]->data;
-         const GLint instride = VB->TexCoordPtr[i]->stride;
+         const GLfloat *in = (GLfloat *)VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]->data;
+         const GLint instride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]->stride;
          GLfloat (*out)[4] = store->texcoord[i].data;
          GLint j;
 
          if (!ctx->Texture.Unit[i]._ReallyEnabled ||
-             VB->TexCoordPtr[i]->size == 4)
+             VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]->size == 4)
             /* Never try to normalize homogenous tex coords! */
             continue;
 
@@ -297,7 +297,7 @@ static GLboolean run_texnorm_stage( GLcontext *ctx,
          }
 
          if (normalizeS || normalizeT)
-            VB->AttribPtr[VERT_ATTRIB_TEX0+i] = VB->TexCoordPtr[i] = &store->texcoord[i];
+            VB->AttribPtr[_TNL_ATTRIB_TEX0 + i] = &store->texcoord[i];
       }
    }
 
index 0714101dc05abf229e67ee97ed5897bb79d3a60c..9a92541ef7ed60a1180c248db5c87a09452d3022 100644 (file)
@@ -880,13 +880,13 @@ static GLboolean savageCheckPTexHack( GLcontext *ctx )
 
    RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
 
-   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 ) && VB->TexCoordPtr[0]->size == 4) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 ) && VB->AttribPtr[_TNL_ATTRIB_TEX0]->size == 4) {
       if (!RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_ATTRIB_TEX1, _TNL_LAST_TEX ))
         return GL_TRUE; /* apply ptex hack */
       else
         FALLBACK(ctx, SAVAGE_FALLBACK_PROJ_TEXTURE, GL_TRUE);
    }
-   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 ) && VB->TexCoordPtr[1]->size == 4)
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 ) && VB->AttribPtr[_TNL_ATTRIB_TEX1]->size == 4)
       FALLBACK(ctx, SAVAGE_FALLBACK_PROJ_TEXTURE, GL_TRUE);
 
    return GL_FALSE; /* don't apply ptex hack */
@@ -977,13 +977,13 @@ static INLINE GLuint savageChooseVertexFormat_s3d( GLcontext *ctx )
    if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 )) {
       if (imesa->ptexHack)
         EMIT_ATTR( _TNL_ATTRIB_TEX0, EMIT_3F_XYW, SAVAGE_EMIT_STQ0, SAVAGE_SKIP_ST0);
-      else if (VB->TexCoordPtr[0]->size == 4)
+      else if (VB->AttribPtr[_TNL_ATTRIB_TEX0]->size == 4)
         assert (0); /* should be caught by savageCheckPTexHack */
-      else if (VB->TexCoordPtr[0]->size >= 2)
+      else if (VB->AttribPtr[_TNL_ATTRIB_TEX0]->size >= 2)
         /* The chromium menu emits some 3D tex coords even though no
          * 3D texture is enabled. Ignore the 3rd coordinate. */
         EMIT_ATTR( _TNL_ATTRIB_TEX0, EMIT_2F, SAVAGE_EMIT_ST0, SAVAGE_SKIP_ST0 );
-      else if (VB->TexCoordPtr[0]->size == 1) {
+      else if (VB->AttribPtr[_TNL_ATTRIB_TEX0]->size == 1) {
         EMIT_ATTR( _TNL_ATTRIB_TEX0, EMIT_1F, SAVAGE_EMIT_S0, SAVAGE_SKIP_S0 );
         EMIT_PAD( 4 );
       } else
@@ -1026,9 +1026,9 @@ static INLINE GLuint savageChooseVertexFormat_s4( GLcontext *ctx )
    if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 )) {
       if (imesa->ptexHack)
         NEED_ATTR( SAVAGE_EMIT_STQ0, SAVAGE_SKIP_ST0);
-      else if (VB->TexCoordPtr[0]->size == 4)
+      else if (VB->AttribPtr[_TNL_ATTRIB_TEX0]->size == 4)
         assert (0); /* should be caught by savageCheckPTexHack */
-      else if (VB->TexCoordPtr[0]->size >= 2)
+      else if (VB->AttribPtr[_TNL_ATTRIB_TEX0]->size >= 2)
         /* The chromium menu emits some 3D tex coords even though no
          * 3D texture is enabled. Ignore the 3rd coordinate. */
         NEED_ATTR( SAVAGE_EMIT_ST0, SAVAGE_SKIP_ST0 );
@@ -1036,10 +1036,10 @@ static INLINE GLuint savageChooseVertexFormat_s4( GLcontext *ctx )
         NEED_ATTR( SAVAGE_EMIT_S0, SAVAGE_SKIP_S0 );
    }
    if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 )) {
-      if (VB->TexCoordPtr[1]->size == 4)
+      if (VB->AttribPtr[_TNL_ATTRIB_TEX1]->size == 4)
         /* projective textures are not supported by the hardware */
         assert (0); /* should be caught by savageCheckPTexHack */
-      else if (VB->TexCoordPtr[1]->size >= 2)
+      else if (VB->AttribPtr[_TNL_ATTRIB_TEX1]->size >= 2)
         NEED_ATTR( SAVAGE_EMIT_ST1, SAVAGE_SKIP_ST1 );
       else
         NEED_ATTR( SAVAGE_EMIT_S1, SAVAGE_SKIP_S1 );
index 4fa2e41e44f30e0a6a0317ba016ac81ca6dafcc2..4690274c3c05f8d7f43921ff6a012bdd4e7e5629 100644 (file)
@@ -904,14 +904,14 @@ static void sisRenderStart( GLcontext *ctx )
 
    /* projective textures are not supported by the hardware */
    if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 )) {
-      if (VB->TexCoordPtr[0]->size > 2)
+      if (VB->AttribPtr[_TNL_ATTRIB_TEX0]->size > 2)
         tex_fallback = GL_TRUE;
       EMIT_ATTR(_TNL_ATTRIB_TEX0, EMIT_2F);
       AGPParseSet |= SiS_PS_HAS_UV0;
    }
    /* Will only hit tex1 on SiS300 */
    if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 )) {
-      if (VB->TexCoordPtr[1]->size > 2)
+      if (VB->AttribPtr[_TNL_ATTRIB_TEX1]->size > 2)
         tex_fallback = GL_TRUE;
       EMIT_ATTR(_TNL_ATTRIB_TEX1, EMIT_2F);
       AGPParseSet |= SiS_PS_HAS_UV1;
index 492880223284224ea4a4b00ea860b0bb21feb446..c200ba3255a2a4fcf131f2e29bded178d19a6486 100644 (file)
@@ -69,11 +69,11 @@ static void interp_extras( GLcontext *ctx,
 
    /*fprintf(stderr, "%s\n", __FUNCTION__);*/
 
-   if (VB->ColorPtr[1]) {
+   if (VB->BackfaceColorPtr) {
       INTERP_4F( t,
-                   GET_COLOR(VB->ColorPtr[1], dst),
-                   GET_COLOR(VB->ColorPtr[1], out),
-                   GET_COLOR(VB->ColorPtr[1], in) );
+                GET_COLOR(VB->BackfaceColorPtr, dst),
+                GET_COLOR(VB->BackfaceColorPtr, out),
+                GET_COLOR(VB->BackfaceColorPtr, in) );
    }
 
    if (VB->EdgeFlag) {
@@ -88,9 +88,9 @@ static void copy_pv_extras( GLcontext *ctx, GLuint dst, GLuint src )
 {
    struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
-   if (VB->ColorPtr[1]) {
-        COPY_4FV( GET_COLOR(VB->ColorPtr[1], dst), 
-                    GET_COLOR(VB->ColorPtr[1], src) );
+   if (VB->BackfaceColorPtr) {
+      COPY_4FV( GET_COLOR(VB->BackfaceColorPtr, dst),
+               GET_COLOR(VB->BackfaceColorPtr, src) );
    }
 
    setup_tab[TDFX_CONTEXT(ctx)->SetupIndex].copy_pv(ctx, dst, src);
index 9b780761f42f13892a2fb9e610b544676a96ec7d..19baf7d0d25ca59c55cf5e7da3aa189be23e853b 100644 (file)
@@ -58,32 +58,32 @@ static void TAG(emit)( GLcontext *ctx,
 /*     fprintf(stderr, "%s\n", __FUNCTION__); */
 
    if (IND & TDFX_TEX0_BIT) {
-      tc0_stride = VB->TexCoordPtr[tmu0_source]->stride;
-      tc0 = VB->TexCoordPtr[tmu0_source]->data;
+      tc0_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu0_source]->stride;
+      tc0 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu0_source]->data;
       u0scale = fxMesa->sScale0;
       v0scale = fxMesa->tScale0;
       if (IND & TDFX_PTEX_BIT)
-        tc0_size = VB->TexCoordPtr[tmu0_source]->size;
+        tc0_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu0_source]->size;
    }
 
    if (IND & TDFX_TEX1_BIT) {
-      tc1 = VB->TexCoordPtr[tmu1_source]->data;
-      tc1_stride = VB->TexCoordPtr[tmu1_source]->stride;
+      tc1 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu1_source]->data;
+      tc1_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu1_source]->stride;
       u1scale = fxMesa->sScale1;
       v1scale = fxMesa->tScale1;
       if (IND & TDFX_PTEX_BIT)
-        tc1_size = VB->TexCoordPtr[tmu1_source]->size;
+        tc1_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu1_source]->size;
    }
    
    if (IND & TDFX_RGBA_BIT) {
-      col = VB->ColorPtr[0]->data;
-      col_stride = VB->ColorPtr[0]->stride;
-      col_size = VB->ColorPtr[0]->size;
+      col = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+      col_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
+      col_size = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->size;
    }
    
    if (IND & TDFX_FOGC_BIT) {
-      fog = VB->FogCoordPtr->data;
-      fog_stride = VB->FogCoordPtr->stride;
+      fog = VB->AttribPtr[_TNL_ATTRIB_FOG]->data;
+      fog_stride = VB->AttribPtr[_TNL_ATTRIB_FOG]->stride;
    }
 
    {
@@ -168,14 +168,14 @@ static GLboolean TAG(check_tex_sizes)( GLcontext *ctx )
       struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
       if (IND & TDFX_TEX1_BIT) {
-        if (VB->TexCoordPtr[0] == 0)
-           VB->TexCoordPtr[0] = VB->TexCoordPtr[1];
+        if (VB->AttribPtr[_TNL_ATTRIB_TEX0] == 0)
+           VB->AttribPtr[_TNL_ATTRIB_TEX0] = VB->AttribPtr[_TNL_ATTRIB_TEX1];
         
-        if (VB->TexCoordPtr[1]->size == 4)
+        if (VB->AttribPtr[_TNL_ATTRIB_TEX1]->size == 4)
            return GL_FALSE;
       }
 
-      if (VB->TexCoordPtr[0]->size == 4)
+      if (VB->AttribPtr[_TNL_ATTRIB_TEX0]->size == 4)
         return GL_FALSE;
    }
 
index e2f1f02c99efe413809da6c4babb166b8049113f..01359d51ea641c4dbcbc78c155138f783824693b 100644 (file)
@@ -833,13 +833,13 @@ static GLboolean viaCheckPTexHack( GLcontext *ctx )
 
    RENDERINPUTS_COPY( index_bitset, tnl->render_inputs_bitset );
 
-   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 ) && VB->TexCoordPtr[0]->size == 4) {
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX0 ) && VB->AttribPtr[_TNL_ATTRIB_TEX0]->size == 4) {
       if (!RENDERINPUTS_TEST_RANGE( index_bitset, _TNL_ATTRIB_TEX1, _TNL_LAST_TEX ))
         ptexHack = GL_TRUE; 
       else
         fallback = GL_TRUE;
    }
-   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 ) && VB->TexCoordPtr[1]->size == 4)
+   if (RENDERINPUTS_TEST( index_bitset, _TNL_ATTRIB_TEX1 ) && VB->AttribPtr[_TNL_ATTRIB_TEX1]->size == 4)
       fallback = GL_TRUE;
 
    FALLBACK(VIA_CONTEXT(ctx), VIA_FALLBACK_PROJ_TEXTURE, fallback);
index a863b028adc4e7770d2f44d8ceb37a37458df933..9dd4f1e9c37223fc69d7dff214abbdf860e26694 100644 (file)
@@ -275,8 +275,9 @@ fxDDTexEnv(GLcontext * ctx, GLenum target, GLenum pname,
    /* apply any lod biasing right now */
    if (pname == GL_TEXTURE_LOD_BIAS_EXT) {
       GLfloat bias = *param;
-      CLAMP_SELF(bias, -ctx->Const.MaxTextureLodBias,
-                        ctx->Const.MaxTextureLodBias - 0.25);
+      bias = CLAMP(bias,
+                  -ctx->Const.MaxTextureLodBias,
+                  ctx->Const.MaxTextureLodBias - 0.25);
 
       grTexLodBiasValue(GR_TMU0, bias);
 
index 1dc5f9891a1613c876e13bbf7a13fa96b9cd8ae4..cc9ad0e8b8349d5ff3537068bd09f4db6bf40146 100644 (file)
@@ -104,24 +104,24 @@ static void interp_extras( GLcontext *ctx,
 {
    struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
-   if (VB->ColorPtr[1]) {
-      /* If stride is zero, ColorPtr[1] is constant across the VB, so
+   if (VB->BackfaceColorPtr) {
+      /* If stride is zero, BackfaceColorPtr is constant across the VB, so
        * there is no point interpolating between two values as they will
        * be identical.  This case is handled in t_dd_tritmp.h
        */
-      if (VB->ColorPtr[1]->stride) {
-        assert(VB->ColorPtr[1]->stride == 4 * sizeof(GLfloat));
+      if (VB->BackfaceColorPtr->stride) {
+        assert(VB->BackfaceColorPtr->stride == 4 * sizeof(GLfloat));
         INTERP_4F( t,
-                   GET_COLOR(VB->ColorPtr[1], dst),
-                   GET_COLOR(VB->ColorPtr[1], out),
-                   GET_COLOR(VB->ColorPtr[1], in) );
+                   GET_COLOR(VB->BackfaceColorPtr, dst),
+                   GET_COLOR(VB->BackfaceColorPtr, out),
+                   GET_COLOR(VB->BackfaceColorPtr, in) );
       }
 
-      if (VB->SecondaryColorPtr[1]) {
+      if (VB->BackfaceSecondaryColorPtr) {
         INTERP_3F( t,
-                   GET_COLOR(VB->SecondaryColorPtr[1], dst),
-                   GET_COLOR(VB->SecondaryColorPtr[1], out),
-                   GET_COLOR(VB->SecondaryColorPtr[1], in) );
+                   GET_COLOR(VB->BackfaceSecondaryColorPtr, dst),
+                   GET_COLOR(VB->BackfaceSecondaryColorPtr, out),
+                   GET_COLOR(VB->BackfaceSecondaryColorPtr, in) );
       }
    }
 
@@ -137,13 +137,13 @@ static void copy_pv_extras( GLcontext *ctx, GLuint dst, GLuint src )
 {
    struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
-   if (VB->ColorPtr[1]) {
-        COPY_4FV( GET_COLOR(VB->ColorPtr[1], dst),
-                  GET_COLOR(VB->ColorPtr[1], src) );
+   if (VB->BackfaceColorPtr) {
+        COPY_4FV( GET_COLOR(VB->BackfaceColorPtr, dst),
+                  GET_COLOR(VB->BackfaceColorPtr, src) );
 
-        if (VB->SecondaryColorPtr[1]) {
-           COPY_3FV( GET_COLOR(VB->SecondaryColorPtr[1], dst),
-                     GET_COLOR(VB->SecondaryColorPtr[1], src) );
+        if (VB->BackfaceSecondaryColorPtr) {
+           COPY_3FV( GET_COLOR(VB->BackfaceSecondaryColorPtr, dst),
+                     GET_COLOR(VB->BackfaceSecondaryColorPtr, src) );
         }
    }
 
index f7970c78e290872e7527731ccc2bbaab132fb74d..f7893c1573fb2943292834cbdce36e9516d2a642 100644 (file)
@@ -62,37 +62,37 @@ static void TAG(emit)( GLcontext *ctx,
    }
 
    if (IND & SETUP_TMU0) {
-      tc0 = VB->TexCoordPtr[tmu0_source]->data;
-      tc0_stride = VB->TexCoordPtr[tmu0_source]->stride;
+      tc0 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu0_source]->data;
+      tc0_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu0_source]->stride;
       u0scale = fxMesa->s0scale;
       v0scale = fxMesa->t0scale;
       if (IND & SETUP_PTEX)
-        tc0_size = VB->TexCoordPtr[tmu0_source]->size;
+        tc0_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu0_source]->size;
    }
 
    if (IND & SETUP_TMU1) {
-      tc1 = VB->TexCoordPtr[tmu1_source]->data;
-      tc1_stride = VB->TexCoordPtr[tmu1_source]->stride;
+      tc1 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu1_source]->data;
+      tc1_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu1_source]->stride;
       u1scale = fxMesa->s1scale; /* wrong if tmu1_source == 0, possible? */
       v1scale = fxMesa->t1scale;
       if (IND & SETUP_PTEX)
-        tc1_size = VB->TexCoordPtr[tmu1_source]->size;
+        tc1_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + tmu1_source]->size;
    }
    
    if (IND & SETUP_RGBA) {
-      col = VB->ColorPtr[0]->data;
-      col_stride = VB->ColorPtr[0]->stride;
-      col_size = VB->ColorPtr[0]->size;
+      col = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+      col_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
+      col_size = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->size;
    }
 
    if (IND & SETUP_SPEC) {
-      spec = VB->SecondaryColorPtr[0]->data;
-      spec_stride = VB->SecondaryColorPtr[0]->stride;
+      spec = VB->AttribPtr[_TNL_ATTRIB_COLOR1]->data;
+      spec_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR1]->stride;
    }
 
    if (IND & SETUP_FOGC) {
-      fog = VB->FogCoordPtr->data;
-      fog_stride = VB->FogCoordPtr->stride;
+      fog = VB->AttribPtr[_TNL_ATTRIB_FOG]->data;
+      fog_stride = VB->AttribPtr[_TNL_ATTRIB_FOG]->stride;
    }
 
    if (start) {
@@ -220,14 +220,15 @@ static GLboolean TAG(check_tex_sizes)( GLcontext *ctx )
       struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
       if (IND & SETUP_TMU1) {
-        if (VB->TexCoordPtr[0] == 0)
-           VB->TexCoordPtr[0] = VB->TexCoordPtr[1];
+        if (VB->AttribPtr[_TNL_ATTRIB_TEX0] == 0)
+           VB->AttribPtr[_TNL_ATTRIB_TEX0] = VB->AttribPtr[_TNL_ATTRIB_TEX1];
         
-        if (VB->TexCoordPtr[1]->size == 4)
+        if (VB->AttribPtr[_TNL_ATTRIB_TEX1]->size == 4)
            return GL_FALSE;
       }
 
-      if (VB->TexCoordPtr[0] && VB->TexCoordPtr[0]->size == 4)
+      if (VB->AttribPtr[_TNL_ATTRIB_TEX0] &&
+         VB->AttribPtr[_TNL_ATTRIB_TEX0]->size == 4)
         return GL_FALSE;
    }
 
index 92d4149466525d1fd590e14851287aa06207b042..9010bbd130a5ed10bf67e5f3b1b3c02d4c4860a7 100644 (file)
@@ -21,7 +21,11 @@ INCLUDE_DIRS = \
 
 # Standalone osmesa needs to be linked with core Mesa APIs
 ifeq ($(DRIVER_DIRS), osmesa)
-CORE_MESA = $(TOP)/src/mesa/libmesa.a $(TOP)/src/mesa/libglapi.a
+CORE_MESA = \
+       $(TOP)/src/mesa/libmesa.a \
+       $(TOP)/src/mesa/libglapi.a \
+       $(TOP)/src/glsl/cl/libglslcl.a \
+       $(TOP)/src/glsl/pp/libglslpp.a
 else
 CORE_MESA =
 endif
index c99ba0bba551edbdae2af5618b9bb947a450deaa..0b373814feeacd63510273cde20b8ad233d83b23 100644 (file)
                GLfloat         ex,ey,fx,fy,cc;                                                 \
                /* Get vars for later */                                                        \
                VB              = &TNL_CONTEXT(ctx)->vb;                                        \
-               vbcolor = (GLchan (*)[4])VB->ColorPtr[1]->data;         \
-               if (VB->SecondaryColorPtr[1]) {                                         \
-                       vbspec = (GLchan (*)[4])VB->SecondaryColorPtr[1]->data; \
+               vbcolor = (GLchan (*)[4])VB->BackfaceColorPtr->data;    \
+               if (VB->BackfaceSecondaryColorPtr) {                    \
+                       vbspec = (GLchan (*)[4])VB->BackfaceSecondaryColorPtr->data;    \
                } else {                                                                                                        \
                        vbspec = NULL;                                                                                  \
                }                                                                                                                       \
        DWORD                                   dwColor;
 
 #define GLD_SETUP_3D_VERTEX(v)                                 \
-       p4f                             = VB->ObjPtr->data;                     \
+       p4f = VB->AttribPtr[_TNL_ATTRIB_POS]->data;             \
        pV->Position.x  = p4f[##v][0];                          \
        pV->Position.y  = p4f[##v][1];                          \
        pV->Position.z  = p4f[##v][2];
 
 #define GLD_SETUP_SMOOTH_COLOUR_3D(v)                                                                                                                  \
-       p4f                     = (GLfloat (*)[4])VB->ColorPtr[0]->data;                                                                                \
+       p4f = (GLfloat (*)[4])VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;                                                                          \
        pV->Diffuse     = D3DCOLOR_COLORVALUE(p4f[##v][0], p4f[##v][1], p4f[##v][2], p4f[##v][3]);
 
 
 #define GLD_SETUP_GET_FLAT_COLOUR_3D(v)                                                                                                        \
-       p4f             = (GLfloat (*)[4])VB->ColorPtr[0]->data;                                                                                \
+       p4f = (GLfloat (*)[4])VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;  \
        dwColor = D3DCOLOR_COLORVALUE(p4f[##v][0], p4f[##v][1], p4f[##v][2], p4f[##v][3]);
 
 #define GLD_SETUP_USE_FLAT_COLOUR_3D                   \
        pV->Diffuse = dwColor;
 
 #define GLD_SETUP_TEX0_3D(v)                                           \
-       if (VB->TexCoordPtr[0]) {                                               \
-               tc                              = VB->TexCoordPtr[0]->data;     \
+       if (VB->AttribPtr[_TNL_ATTRIB_TEX0]) {                          \
+               tc = VB->AttribPtr[_TNL_ATTRIB_TEX0]->data;             \
                pV->TexUnit0.x  = tc[##v][0];                           \
                pV->TexUnit0.y  = tc[##v][1];                           \
        }
 
 #define GLD_SETUP_TEX1_3D(v)                                           \
-       if (VB->TexCoordPtr[1]) {                                               \
-               tc                              = VB->TexCoordPtr[1]->data;     \
+       if (VB->AttribPtr[_TNL_ATTRIB_TEX1]) {                          \
+               tc = VB->AttribPtr[_TNL_ATTRIB_TEX1]->data;             \
                pV->TexUnit1.x  = tc[##v][0];                           \
                pV->TexUnit1.y  = tc[##v][1];                           \
        }
index a85620dde8d2452a94d6467f2f443c49c385257e..c39775cad32897879ee8aba99ad4c8e9ebdcc29d 100644 (file)
@@ -151,7 +151,7 @@ static GLboolean gld_d3d_render_stage_run(
 #if 0
    // For debugging: Useful to see if an app passes colour data in
    // an unusual format.
-   switch (VB->ColorPtr[0]->Type) {
+   switch (VB->AttribPtr[_TNL_ATTRIB_COLOR0]->Type) {
    case GL_FLOAT:
           ddlogMessage(GLDLOG_SYSTEM, "ColorPtr: GL_FLOAT\n");
           break;
index a5b5462f039c0208a70424df8eade38fbebf3cc9..990922580aaf14726633ca28dcfd8a9f62e1578f 100644 (file)
                GLfloat         ex,ey,fx,fy,cc;                                                 \
                /* Get vars for later */                                                        \
                VB              = &TNL_CONTEXT(ctx)->vb;                                        \
-               vbcolor = (GLchan (*)[4])VB->ColorPtr[1]->data;         \
-               if (VB->SecondaryColorPtr[1]) {                                         \
-                       vbspec = (GLchan (*)[4])VB->SecondaryColorPtr[1]->data; \
+               vbcolor = (GLchan (*)[4])VB->BackfaceColorPtr->data;    \
+               if (VB->BackfaceSecondaryColorPtr) {                    \
+                       vbspec = (GLchan (*)[4])VB->BackfaceSecondaryColorPtr->data;    \
                } else {                                                                                                        \
                        vbspec = NULL;                                                                                  \
                }                                                                                                                       \
        DWORD                                   dwColor;
 
 #define GLD_SETUP_3D_VERTEX(v)                                 \
-       p4f                             = VB->ObjPtr->data;                     \
+       p4f = VB->AttribPtr[_TNL_ATTRIB_POS]->data;             \
        pV->Position.x  = p4f[##v][0];                          \
        pV->Position.y  = p4f[##v][1];                          \
        pV->Position.z  = p4f[##v][2];
 
 #define GLD_SETUP_SMOOTH_COLOUR_3D(v)                                                                                                                  \
-       p4f                     = (GLfloat (*)[4])VB->ColorPtr[0]->data;                                                                                \
+       p4f = (GLfloat (*)[4])VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;  \
        pV->Diffuse     = D3DCOLOR_COLORVALUE(p4f[##v][0], p4f[##v][1], p4f[##v][2], p4f[##v][3]);
 
 
 #define GLD_SETUP_GET_FLAT_COLOUR_3D(v)                                                                                                        \
-       p4f             = (GLfloat (*)[4])VB->ColorPtr[0]->data;                                                                                \
+       p4f = (GLfloat (*)[4])VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;  \
        dwColor = D3DCOLOR_COLORVALUE(p4f[##v][0], p4f[##v][1], p4f[##v][2], p4f[##v][3]);
 
 #define GLD_SETUP_USE_FLAT_COLOUR_3D                   \
        pV->Diffuse = dwColor;
 
 #define GLD_SETUP_TEX0_3D(v)                                           \
-       if (VB->TexCoordPtr[0]) {                                               \
-               tc                              = VB->TexCoordPtr[0]->data;     \
+       if (VB->AttribPtr[_TNL_ATTRIB_TEX0]) {                          \
+               tc = VB->TnlAttribPtr[_TNL_ATTRIB_TEX0]->data;          \
                pV->TexUnit0.x  = tc[##v][0];                           \
                pV->TexUnit0.y  = tc[##v][1];                           \
        }
 
 #define GLD_SETUP_TEX1_3D(v)                                           \
-       if (VB->TexCoordPtr[1]) {                                               \
-               tc                              = VB->TexCoordPtr[1]->data;     \
+       if (VB->TnlAttribPtr[_TNL_ATTRIB_TEX1]) {                       \
+               tc = VB->TnlAttribPtr[_TNL_ATTRIB_TEX1]->data;          \
                pV->TexUnit1.x  = tc[##v][0];                           \
                pV->TexUnit1.y  = tc[##v][1];                           \
        }
index cafbf4f5c50a089bc8dfcaf5f2679484d4b347e5..265c81fb4a48a0d6de07ee7e9d3c4bc9a49a3c1b 100644 (file)
@@ -149,7 +149,7 @@ static GLboolean gld_d3d_render_stage_run(
 #if 0
    // For debugging: Useful to see if an app passes colour data in
    // an unusual format.
-   switch (VB->ColorPtr[0]->Type) {
+   switch (VB->AttribPtr[_TNL_ATTRIB_COLOR0]->Type) {
    case GL_FLOAT:
           ddlogMessage(GLDLOG_SYSTEM, "ColorPtr: GL_FLOAT\n");
           break;
index 403a9d5f86e3ab124b648279be84695b67f079a4..fd4dd4ed75104f91962aeeb909605a1ffdf777d6 100644 (file)
                GLfloat         ex,ey,fx,fy,cc;                                                 \
                /* Get vars for later */                                                        \
                VB              = &TNL_CONTEXT(ctx)->vb;                                        \
-               vbcolor = (GLchan (*)[4])VB->ColorPtr[1]->data;         \
-               if (VB->SecondaryColorPtr[1]) {                                         \
-                       vbspec = (GLchan (*)[4])VB->SecondaryColorPtr[1]->data; \
+               vbcolor = (GLchan (*)[4])VB->BackfaceColorPtr->data;    \
+               if (VB->BackfaceSecondaryColorPtr) {                    \
+                       vbspec = (GLchan (*)[4])VB->BackfaceSecondaryColorPtr->data;    \
                } else {                                                                                                        \
                        vbspec = NULL;                                                                                  \
                }                                                                                                                       \
        DWORD                                   dwColor;
 
 #define GLD_SETUP_3D_VERTEX(v)                                 \
-       p4f                             = VB->ObjPtr->data;                     \
+       p4f = VB->AttribPtr[_TNL_ATTRIB_POS]->data;             \
        pV->Position.x  = p4f[##v][0];                          \
        pV->Position.y  = p4f[##v][1];                          \
        pV->Position.z  = p4f[##v][2];
 
 #define GLD_SETUP_SMOOTH_COLOUR_3D(v)                                                                                                                  \
-       p4f                     = (GLfloat (*)[4])VB->ColorPtr[0]->data;                                                                                \
+       p4f = (GLfloat (*)[4])VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;                                                                          \
        pV->Diffuse     = D3DCOLOR_COLORVALUE(p4f[##v][0], p4f[##v][1], p4f[##v][2], p4f[##v][3]);
 
 
 #define GLD_SETUP_GET_FLAT_COLOUR_3D(v)                                                                                                        \
-       p4f             = (GLfloat (*)[4])VB->ColorPtr[0]->data;                                                                                \
+       p4f = (GLfloat (*)[4])VB->AttribPtr[_TNL_ATTRIB_COLOR00]->data; \
        dwColor = D3DCOLOR_COLORVALUE(p4f[##v][0], p4f[##v][1], p4f[##v][2], p4f[##v][3]);
 
 #define GLD_SETUP_USE_FLAT_COLOUR_3D                   \
        pV->Diffuse = dwColor;
 
 #define GLD_SETUP_TEX0_3D(v)                                           \
-       if (VB->TexCoordPtr[0]) {                                               \
-               tc                              = VB->TexCoordPtr[0]->data;     \
+       if (VB->AttribPtr[_TNL_ATTRIB_TEX0]) {                          \
+               tc = VB->AttribPtr[_TNL_ATTRIB_TEX0]->data;             \
                pV->TexUnit0.x  = tc[##v][0];                           \
                pV->TexUnit0.y  = tc[##v][1];                           \
        }
 
 #define GLD_SETUP_TEX1_3D(v)                                           \
-       if (VB->TexCoordPtr[1]) {                                               \
-               tc                              = VB->TexCoordPtr[1]->data;     \
+       if (VB->AttribPtr[_TNL_ATTRIB_TEX1]) {                          \
+               tc = VB->AttribPtr[_TNL_ATTRIB_TEX1]->data;             \
                pV->TexUnit1.x  = tc[##v][0];                           \
                pV->TexUnit1.y  = tc[##v][1];                           \
        }
index 4fa6bcaf1abd1f9598ff271d969ffc78a750f5d4..91a68b3f2d9ce93ef71c231fec0303c8799ed65f 100644 (file)
@@ -149,7 +149,7 @@ static GLboolean gld_d3d_render_stage_run(
 #if 0
    // For debugging: Useful to see if an app passes colour data in
    // an unusual format.
-   switch (VB->ColorPtr[0]->Type) {
+   switch (VB->AttribPtr[_TNL_ATTRIB_COLOR0]->Type) {
    case GL_FLOAT:
           ddlogMessage(GLDLOG_SYSTEM, "ColorPtr: GL_FLOAT\n");
           break;
index c5048970cca03fb0b62652d6109a1330cf859a55..2eac1cc2ed9bfcc3eeee501961011ca901cd6c80 100644 (file)
 /*@{*/
 #define MAX_VERTEX_GENERIC_ATTRIBS 16
 #define MAX_VERTEX_TEXTURE_IMAGE_UNITS MAX_TEXTURE_IMAGE_UNITS
-#define MAX_COMBINED_TEXTURE_IMAGE_UNITS MAX_TEXTURE_IMAGE_UNITS
+#define MAX_COMBINED_TEXTURE_IMAGE_UNITS (MAX_VERTEX_TEXTURE_IMAGE_UNITS + \
+                                         MAX_TEXTURE_IMAGE_UNITS)
 /*@}*/
 
 
index 87eae9663923920ff246ea67c8ce513f34a24dc5..5c20ce017f79bb076134174a101df9559a038516 100644 (file)
@@ -571,6 +571,7 @@ _mesa_init_constants(GLcontext *ctx)
 
 #if FEATURE_ARB_vertex_shader
    ctx->Const.MaxVertexTextureImageUnits = MAX_VERTEX_TEXTURE_IMAGE_UNITS;
+   ctx->Const.MaxCombinedTextureImageUnits = MAX_COMBINED_TEXTURE_IMAGE_UNITS;
    ctx->Const.MaxVarying = MAX_VARYING;
 #endif
 
index 27ed921761c2421ea9aa53244d2192a378611d61..6dadf5c079b8d12e7e59f5189343c4137916acae 100644 (file)
@@ -182,7 +182,7 @@ struct dd_function_table {
     * 
     * This is called by the \c _mesa_store_tex[sub]image[123]d() fallback
     * functions.  The driver should examine \p internalFormat and return a
-    * pointer to an appropriate gl_texture_format.
+    * gl_format value.
     */
    GLuint (*ChooseTextureFormat)( GLcontext *ctx, GLint internalFormat,
                                      GLenum srcFormat, GLenum srcType );
@@ -537,11 +537,6 @@ struct dd_function_table {
    GLboolean (*IsTextureResident)( GLcontext *ctx,
                                    struct gl_texture_object *t );
 
-   /**
-    * Called by glActiveTextureARB() to set current texture unit.
-    */
-   void (*ActiveTexture)( GLcontext *ctx, GLuint texUnitNumber );
-
    /**
     * Called when the texture's color lookup table is changed.
     * 
@@ -766,13 +761,13 @@ struct dd_function_table {
 
    /* May return NULL if MESA_MAP_NOWAIT_BIT is set in access:
     */
-   void * (*MapBufferRange)( GLcontext *ctx, GLenum target,
-                             GLintptr offset, GLsizeiptr length, GLbitfield access,
+   void * (*MapBufferRange)( GLcontext *ctx, GLenum target, GLintptr offset,
+                             GLsizeiptr length, GLbitfield access,
                              struct gl_buffer_object *obj);
 
-   void (*FlushMappedBufferRange) (GLcontext *ctx, GLenum target, 
-                                   GLintptr offset, GLsizeiptr length,
-                                   struct gl_buffer_object *obj);
+   void (*FlushMappedBufferRange)(GLcontext *ctx, GLenum target, 
+                                  GLintptr offset, GLsizeiptr length,
+                                  struct gl_buffer_object *obj);
 
    GLboolean (*UnmapBuffer)( GLcontext *ctx, GLenum target,
                             struct gl_buffer_object *obj );
@@ -787,7 +782,8 @@ struct dd_function_table {
    struct gl_framebuffer * (*NewFramebuffer)(GLcontext *ctx, GLuint name);
    struct gl_renderbuffer * (*NewRenderbuffer)(GLcontext *ctx, GLuint name);
    void (*BindFramebuffer)(GLcontext *ctx, GLenum target,
-                           struct gl_framebuffer *fb, struct gl_framebuffer *fbread);
+                           struct gl_framebuffer *drawFb,
+                           struct gl_framebuffer *readFb);
    void (*FramebufferRenderbuffer)(GLcontext *ctx, 
                                    struct gl_framebuffer *fb,
                                    GLenum attachment,
index fe2416d8946d26ddcff5c5d191913afe6d495759..2d1db29cbfe556398c737b6c9c6fa23af8ce44e9 100644 (file)
@@ -357,7 +357,7 @@ static struct ureg get_temp( struct tnl_program *p )
    int bit = _mesa_ffs( ~p->temp_in_use );
    if (!bit) {
       _mesa_problem(NULL, "%s: out of temporaries\n", __FILE__);
-      _mesa_exit(1);
+      exit(1);
    }
 
    if ((GLuint) bit > p->program->Base.NumTemporaries)
@@ -523,7 +523,6 @@ static void emit_dst( struct prog_dst_register *dst,
    dst->CondMask = COND_TR;  /* always pass cond test */
    dst->CondSwizzle = SWIZZLE_NOOP;
    dst->CondSrc = 0;
-   dst->pad = 0;
    /* Check that bitfield sizes aren't exceeded */
    ASSERT(dst->Index == reg.idx);
 }
index 3f6b03c88aee124b043cecebcb99a4f5d11682b7..3d32649bade8ad52bf9e845ca55208e5ddfd2d09 100644 (file)
@@ -1877,7 +1877,7 @@ _mesa_GetBooleanv( GLenum pname, GLboolean *params )
          break;
       case GL_MAX_COMBINED_TEXTURE_IMAGE_UNITS_ARB:
          CHECK_EXT1(ARB_vertex_shader, "GetBooleanv");
-         params[0] = INT_TO_BOOLEAN(MAX_COMBINED_TEXTURE_IMAGE_UNITS);
+         params[0] = INT_TO_BOOLEAN(ctx->Const.MaxCombinedTextureImageUnits);
          break;
       case GL_CURRENT_PROGRAM:
          CHECK_EXT1(ARB_shader_objects, "GetBooleanv");
@@ -3712,7 +3712,7 @@ _mesa_GetFloatv( GLenum pname, GLfloat *params )
          break;
       case GL_MAX_COMBINED_TEXTURE_IMAGE_UNITS_ARB:
          CHECK_EXT1(ARB_vertex_shader, "GetFloatv");
-         params[0] = (GLfloat)(MAX_COMBINED_TEXTURE_IMAGE_UNITS);
+         params[0] = (GLfloat)(ctx->Const.MaxCombinedTextureImageUnits);
          break;
       case GL_CURRENT_PROGRAM:
          CHECK_EXT1(ARB_shader_objects, "GetFloatv");
@@ -5547,7 +5547,7 @@ _mesa_GetIntegerv( GLenum pname, GLint *params )
          break;
       case GL_MAX_COMBINED_TEXTURE_IMAGE_UNITS_ARB:
          CHECK_EXT1(ARB_vertex_shader, "GetIntegerv");
-         params[0] = MAX_COMBINED_TEXTURE_IMAGE_UNITS;
+         params[0] = ctx->Const.MaxCombinedTextureImageUnits;
          break;
       case GL_CURRENT_PROGRAM:
          CHECK_EXT1(ARB_shader_objects, "GetIntegerv");
@@ -7383,7 +7383,7 @@ _mesa_GetInteger64v( GLenum pname, GLint64 *params )
          break;
       case GL_MAX_COMBINED_TEXTURE_IMAGE_UNITS_ARB:
          CHECK_EXT1(ARB_vertex_shader, "GetInteger64v");
-         params[0] = (GLint64)(MAX_COMBINED_TEXTURE_IMAGE_UNITS);
+         params[0] = (GLint64)(ctx->Const.MaxCombinedTextureImageUnits);
          break;
       case GL_CURRENT_PROGRAM:
          CHECK_EXT1(ARB_shader_objects, "GetInteger64v");
index 697c4cfd92ef8574e946cfac3a0c735c434f9f2e..01170a42a72d0782b14395ee937cb1b08ad202d5 100644 (file)
@@ -1006,7 +1006,7 @@ StateVars = [
        ( "GL_MAX_VERTEX_TEXTURE_IMAGE_UNITS_ARB", GLint,
          ["ctx->Const.MaxVertexTextureImageUnits"], "", ["ARB_vertex_shader"] ),
        ( "GL_MAX_COMBINED_TEXTURE_IMAGE_UNITS_ARB", GLint,
-         ["MAX_COMBINED_TEXTURE_IMAGE_UNITS"], "", ["ARB_vertex_shader"] ),
+         ["ctx->Const.MaxCombinedTextureImageUnits"], "", ["ARB_vertex_shader"] ),
 
        # GL_ARB_shader_objects
        # Actually, this token isn't part of GL_ARB_shader_objects, but is
index 46ffb929b6cd81c8c96f2c782847a3fdb05a1841..def045269c58647c0bd005fd9ab9875d55d24b36 100644 (file)
@@ -460,7 +460,7 @@ _mesa_inv_sqrtf(float n)
 #if 0 /* not used, see below -BP */
         float r3, x3, y3;
 #endif
-        union { float f; unsigned int i; } u;
+        fi_type u;
         unsigned int magic;
 
         /*
@@ -629,11 +629,15 @@ _mesa_ffsll(int64_t val)
 unsigned int
 _mesa_bitcount(unsigned int n)
 {
+#if defined(__GNUC__)
+   return __builtin_popcount(n);
+#else
    unsigned int bits;
    for (bits = 0; n > 0; n = n >> 1) {
       bits += (n & 1);
    }
    return bits;
+#endif
 }
 
 
@@ -645,10 +649,10 @@ _mesa_bitcount(unsigned int n)
 GLhalfARB
 _mesa_float_to_half(float val)
 {
-   const int flt = *((int *) (void *) &val);
-   const int flt_m = flt & 0x7fffff;
-   const int flt_e = (flt >> 23) & 0xff;
-   const int flt_s = (flt >> 31) & 0x1;
+   const fi_type fi = {val};
+   const int flt_m = fi.i & 0x7fffff;
+   const int flt_e = (fi.i >> 23) & 0xff;
+   const int flt_s = (fi.i >> 31) & 0x1;
    int s, e, m = 0;
    GLhalfARB result;
    
@@ -735,7 +739,8 @@ _mesa_half_to_float(GLhalfARB val)
    const int m = val & 0x3ff;
    const int e = (val >> 10) & 0x1f;
    const int s = (val >> 15) & 0x1;
-   int flt_m, flt_e, flt_s, flt;
+   int flt_m, flt_e, flt_s;
+   fi_type fi;
    float result;
 
    /* sign bit */
@@ -770,8 +775,8 @@ _mesa_half_to_float(GLhalfARB val)
       flt_m = m << 13;
    }
 
-   flt = (flt_s << 31) | (flt_e << 23) | flt_m;
-   result = *((float *) (void *) &flt);
+   fi.i = (flt_s << 31) | (flt_e << 23) | flt_m;
+   result = fi.f;
    return result;
 }
 
@@ -1234,13 +1239,3 @@ _mesa_debug( const GLcontext *ctx, const char *fmtString, ... )
 }
 
 /*@}*/
-
-
-/**
- * Wrapper for exit().
- */
-void
-_mesa_exit( int status )
-{
-   exit(status);
-}
index 7d4012a85609b1addab2e0e2793d661f1a8d2802..b01fe5b0ab98d55f551b5be93d56114c3bab58c2 100644 (file)
@@ -618,10 +618,6 @@ _mesa_error( __GLcontext *ctx, GLenum error, const char *fmtString, ... );
 extern void
 _mesa_debug( const __GLcontext *ctx, const char *fmtString, ... );
 
-extern void 
-_mesa_exit( int status );
-
-
 #ifdef __cplusplus
 }
 #endif
index 3d9a1aba98d13a8e37e545e9f982f9a61dd54928..55578adf836c874e7ec606a57e9b5f0228059570 100644 (file)
@@ -202,17 +202,12 @@ do {                                \
 #endif
 
 /**
- * Copy a 4-element float vector (avoid using FPU registers)
- * XXX Could use two 64-bit moves on 64-bit systems
+ * Copy a 4-element float vector
+ * memcpy seems to be most efficient
  */
 #define COPY_4FV( DST, SRC )                  \
 do {                                          \
-   const GLuint *_s = (const GLuint *) (SRC); \
-   GLuint *_d = (GLuint *) (DST);             \
-   _d[0] = _s[0];                             \
-   _d[1] = _s[1];                             \
-   _d[2] = _s[2];                             \
-   _d[3] = _s[3];                             \
+   _mesa_memcpy(DST, SRC, sizeof(GLfloat) * 4);       \
 } while (0)
 
 /** Copy \p SZ elements into a 4-element vector */
@@ -631,12 +626,6 @@ do {                                    \
 /** Clamp X to [MIN,MAX] */
 #define CLAMP( X, MIN, MAX )  ( (X)<(MIN) ? (MIN) : ((X)>(MAX) ? (MAX) : (X)) )
 
-/** Assign X to CLAMP(X, MIN, MAX) */
-#define CLAMP_SELF(x, mn, mx)  \
-   ( (x)<(mn) ? ((x) = (mn)) : ((x)>(mx) ? ((x)=(mx)) : (x)) )
-
-
-
 /** Minimum of two values: */
 #define MIN2( A, B )   ( (A)<(B) ? (A) : (B) )
 
index cde2f5fe061afad5f4cbf99b4280e94d049342d0..b52c84b491a11e772027c344c32b58d4909d51a7 100644 (file)
@@ -2319,6 +2319,7 @@ struct gl_constants
    GLuint MaxTextureCoordUnits;
    GLuint MaxTextureImageUnits;
    GLuint MaxVertexTextureImageUnits;
+   GLuint MaxCombinedTextureImageUnits;
    GLuint MaxTextureUnits;           /**< = MIN(CoordUnits, ImageUnits) */
    GLfloat MaxTextureMaxAnisotropy;  /**< GL_EXT_texture_filter_anisotropic */
    GLfloat MaxTextureLodBias;        /**< GL_EXT_texture_lod_bias */
index f439d4addb32937004dd6fbb6f3cf59838beb48d..499b7330d0a35ffcfe0a67b834a7f843f09c151d 100644 (file)
@@ -606,7 +606,7 @@ static struct ureg get_temp( struct texenv_fragment_program *p )
 
    if (!bit) {
       _mesa_problem(NULL, "%s: out of temporaries\n", __FILE__);
-      _mesa_exit(1);
+      exit(1);
    }
 
    if ((GLuint) bit > p->program->Base.NumTemporaries)
@@ -634,7 +634,7 @@ static struct ureg get_tex_temp( struct texenv_fragment_program *p )
 
    if (!bit) {
       _mesa_problem(NULL, "%s: out of temporaries\n", __FILE__);
-      _mesa_exit(1);
+      exit(1);
    }
 
    if ((GLuint) bit > p->program->Base.NumTemporaries)
@@ -1040,8 +1040,6 @@ static struct ureg emit_combine( struct texenv_fragment_program *p,
 
    assert(nr <= MAX_COMBINER_TERMS);
 
-   tmp = undef; /* silence warning (bug 5318) */
-
    for (i = 0; i < nr; i++)
       src[i] = emit_combine_source( p, mask, unit, opt[i].Source, opt[i].Operand );
 
index 23765d275305b1b5a5ebbdd56e107e17c725e560..bd7cc8d278033ecc621b2a68a13b7c54b5612db9 100644 (file)
@@ -555,7 +555,8 @@ _mesa_get_compressed_teximage(GLcontext *ctx, GLenum target, GLint level,
       _mesa_get_format_block_size(texImage->TexFormat, &bw, &bh);
       for (i = 0; i < (texImage->Height + bh - 1) / bh; i++) {
          memcpy((GLubyte *)img + i * row_stride,
-                (GLubyte *)texImage->Data + i * row_stride_stored, row_stride);
+                (GLubyte *)texImage->Data + i * row_stride_stored,
+                row_stride);
       }
    }
 
index 237b6eb00d4391d4e7e97c6c1390026c5c418888..09fe7b85ba013e7390c84b8e12696bc8247c73cb 100644 (file)
@@ -940,7 +940,8 @@ _mesa_DeleteTextures( GLsizei n, const GLuint *textures)
 /**
  * Convert a GL texture target enum such as GL_TEXTURE_2D or GL_TEXTURE_3D
  * into the corresponding Mesa texture target index.
- * Return -1 if target is invalid.
+ * Note that proxy targets are not valid here.
+ * \return TEXTURE_x_INDEX or -1 if target is invalid
  */
 static GLint
 target_enum_to_index(GLenum target)
index e2432be6ca497cee2cb0e4e5ec9345d2fb404d53..cf603d46d86e370f0f24b8c6dca682b01c814b7b 100644 (file)
@@ -467,7 +467,6 @@ update_wrapper(GLcontext *ctx, const struct gl_renderbuffer_attachment *att)
 {
    struct texture_renderbuffer *trb
       = (struct texture_renderbuffer *) att->Renderbuffer;
-   gl_format texFormat;
 
    (void) ctx;
    ASSERT(trb);
@@ -490,8 +489,6 @@ update_wrapper(GLcontext *ctx, const struct gl_renderbuffer_attachment *att)
       trb->Zoffset = att->Zoffset;
    }
 
-   texFormat = trb->TexImage->TexFormat;
-
    trb->Base.Width = trb->TexImage->Width;
    trb->Base.Height = trb->TexImage->Height;
    trb->Base.InternalFormat = trb->TexImage->InternalFormat;
index 9c758ce3b9015e4ac26451cf91367f202cf7bc54..dc55cb7ccc962ed99a06053c1d18bebf223d6b8e 100644 (file)
@@ -1,6 +1,6 @@
 /*
  * Mesa 3-D graphics library
- * Version:  7.7
+ * Version:  7.8
  *
  * Copyright (C) 1999-2008  Brian Paul   All Rights Reserved.
  * Copyright (C) 2009  VMware, Inc.  All Rights Reserved.
@@ -30,9 +30,9 @@
 
 /* Mesa version */
 #define MESA_MAJOR 7
-#define MESA_MINOR 7
+#define MESA_MINOR 8
 #define MESA_PATCH 0
-#define MESA_VERSION_STRING "7.7"
+#define MESA_VERSION_STRING "7.8-devel"
 
 /* To make version comparison easy */
 #define MESA_VERSION(a,b,c) (((a) << 16) + ((b) << 8) + (c))
index 89c632e7d59d1370379551a592c65bf0ecaab472..f9b26d804717469797b9ef68442ac69becfdea6c 100644 (file)
@@ -230,7 +230,7 @@ static int test_norm_function( normal_func func, int mtype, long *cycles )
          case VAR:
             break;
          default:
-            _mesa_exit(1);
+            exit(1);
          }
       }
    }
index dd732b6666bc07afa011c52929d68fca17f15b6a..a09be71020ea90bcec8629347669c9ad3755236b 100644 (file)
@@ -56,7 +56,6 @@ having three separate program parameter arrays.
 #include "main/context.h"
 #include "main/macros.h"
 #include "main/mtypes.h"
-#include "shader/grammar/grammar_mesa.h"
 #include "arbprogparse.h"
 #include "program.h"
 #include "programopt.h"
index ac087d415c5df2f829c5bf80bd8d8e946b978fa0..e04a05b22fff0ca5ccf7c6d2300293f11b52f3b0 100644 (file)
@@ -290,9 +290,11 @@ _mesa_DeleteFragmentShaderATI(GLuint id)
 
       /* The ID is immediately available for re-use now */
       _mesa_HashRemove(ctx->Shared->ATIShaders, id);
-      prog->RefCount--;
-      if (prog->RefCount <= 0) {
-         _mesa_free(prog);
+      if (prog) {
+        prog->RefCount--;
+        if (prog->RefCount <= 0) {
+           _mesa_free(prog);
+        }
       }
    }
 }
index 19bafd483021ff3e0eed7d4f40a86c9faaa98777..59730020d0c0b5e5e2ff9e419fabb2efa3c26d31 100644 (file)
@@ -16,7 +16,7 @@
 
 VPATH = RCS
 
-INCDIR = [---.include],[.grammar],[-.main],[-.glapi],[.slang]
+INCDIR = [---.include],[-.main],[-.glapi],[.slang]
 LIBDIR = [---.lib]
 CFLAGS = /include=($(INCDIR),[])/define=(PTHREADS=1,"__extension__=")/name=(as_is,short)/float=ieee/ieee=denorm
 
@@ -64,8 +64,6 @@ all :
        $(MMS)$(MMSQUALIFIERS) $(LIBDIR)$(GL_LIB)
        set def [.slang]
        $(MMS)$(MMSQUALIFIERS)
-       set def [-.grammar]
-       $(MMS)$(MMSQUALIFIERS)
        set def [-]
 
 # Make the library
diff --git a/src/mesa/shader/grammar/grammar.c b/src/mesa/shader/grammar/grammar.c
deleted file mode 100644 (file)
index a977596..0000000
+++ /dev/null
@@ -1,3178 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  6.6
- *
- * Copyright (C) 1999-2006  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/**
- * \file grammar.c
- * syntax parsing engine
- * \author Michal Krol
- */
-
-#ifndef GRAMMAR_PORT_BUILD
-#error Do not build this file directly, build your grammar_XXX.c instead, which includes this file
-#endif
-
-/*
-*/
-
-/*
-    INTRODUCTION
-    ------------
-
-    The task is to check the syntax of an input string. Input string is a stream of ASCII
-    characters terminated with a null-character ('\0'). Checking it using C language is
-    difficult and hard to implement without bugs. It is hard to maintain and make changes when
-    the syntax changes.
-
-    This is because of a high redundancy of the C code. Large blocks of code are duplicated with
-    only small changes. Even use of macros does not solve the problem because macros cannot
-    erase the complexity of the problem.
-
-    The resolution is to create a new language that will be highly oriented to our task. Once
-    we describe a particular syntax, we are done. We can then focus on the code that implements
-    the language. The size and complexity of it is relatively small than the code that directly
-    checks the syntax.
-
-    First, we must implement our new language. Here, the language is implemented in C, but it
-    could also be implemented in any other language. The code is listed below. We must take
-    a good care that it is bug free. This is simple because the code is simple and clean.
-
-    Next, we must describe the syntax of our new language in itself. Once created and checked
-    manually that it is correct, we can use it to check another scripts.
-
-    Note that our new language loading code does not have to check the syntax. It is because we
-    assume that the script describing itself is correct, and other scripts can be syntactically
-    checked by the former script. The loading code must only do semantic checking which leads us to
-    simple resolving references.
-
-    THE LANGUAGE
-    ------------
-
-    Here I will describe the syntax of the new language (further called "Synek"). It is mainly a
-    sequence of declarations terminated by a semicolon. The declaration consists of a symbol,
-    which is an identifier, and its definition. A definition is in turn a sequence of specifiers
-    connected with ".and" or ".or" operator. These operators cannot be mixed together in a one
-    definition. Specifier can be a symbol, string, character, character range or a special
-    keyword ".true" or ".false".
-
-    On the very beginning of the script there is a declaration of a root symbol and is in the form:
-        .syntax <root_symbol>;
-    The <root_symbol> must be on of the symbols in declaration sequence. The syntax is correct if
-    the root symbol evaluates to true. A symbol evaluates to true if the definition associated with
-    the symbol evaluates to true. Definition evaluation depends on the operator used to connect
-    specifiers in the definition. If ".and" operator is used, definition evaluates to true if and
-    only if all the specifiers evaluate to true. If ".or" operator is used, definition evalutes to
-    true if any of the specifiers evaluates to true. If definition contains only one specifier,
-    it is evaluated as if it was connected with ".true" keyword by ".and" operator.
-
-    If specifier is a ".true" keyword, it always evaluates to true.
-
-    If specifier is a ".false" keyword, it always evaluates to false. Specifier evaluates to false
-    when it does not evaluate to true.
-
-    Character range specifier is in the form:
-        '<first_character>' - '<second_character>'
-    If specifier is a character range, it evaluates to true if character in the stream is greater
-    or equal to <first_character> and less or equal to <second_character>. In that situation 
-    the stream pointer is advanced to point to next character in the stream. All C-style escape
-    sequences are supported although trigraph sequences are not. The comparisions are performed
-    on 8-bit unsigned integers.
-
-    Character specifier is in the form:
-        '<single_character>'
-    It evaluates to true if the following character range specifier evaluates to true:
-        '<single_character>' - '<single_character>'
-
-    String specifier is in the form:
-        "<string>"
-    Let N be the number of characters in <string>. Let <string>[i] designate i-th character in
-    <string>. Then the string specifier evaluates to true if and only if for i in the range [0, N)
-    the following character specifier evaluates to true:
-        '<string>[i]'
-    If <string>[i] is a quotation mark, '<string>[i]' is replaced with '\<string>[i]'.
-
-    Symbol specifier can be optionally preceded by a ".loop" keyword in the form:
-        .loop <symbol>                  (1)
-    where <symbol> is defined as follows:
-        <symbol> <definition>;          (2)
-    Construction (1) is replaced by the following code:
-        <symbol$1>
-    and declaration (2) is replaced by the following:
-        <symbol$1> <symbol$2> .or .true;
-        <symbol$2> <symbol> .and <symbol$1>;
-        <symbol> <definition>;
-
-    Synek supports also a register mechanizm. User can, in its SYN file, declare a number of
-    registers that can be accessed in the syn body. Each reg has its name and a default value.
-    The register is one byte wide. The C code can change the default value by calling
-    grammar_set_reg8() with grammar id, register name and a new value. As we know, each rule is
-    a sequence of specifiers joined with .and or .or operator. And now each specifier can be
-    prefixed with a condition expression in a form ".if (<reg_name> <operator> <hex_literal>)"
-    where <operator> can be == or !=. If the condition evaluates to false, the specifier
-    evaluates to .false. Otherwise it evalutes to the specifier.
-
-    ESCAPE SEQUENCES
-    ----------------
-
-    Synek supports all escape sequences in character specifiers. The mapping table is listed below.
-    All occurences of the characters in the first column are replaced with the corresponding
-    character in the second column.
-
-        Escape sequence         Represents
-    ------------------------------------------------------------------------------------------------
-        \a                      Bell (alert)
-        \b                      Backspace
-        \f                      Formfeed
-        \n                      New line
-        \r                      Carriage return
-        \t                      Horizontal tab
-        \v                      Vertical tab
-        \'                      Single quotation mark
-        \"                      Double quotation mark
-        \\                      Backslash
-        \?                      Literal question mark
-        \ooo                    ASCII character in octal notation
-        \xhhh                   ASCII character in hexadecimal notation
-    ------------------------------------------------------------------------------------------------
-
-    RAISING ERRORS
-    --------------
-
-    Any specifier can be followed by a special construction that is executed when the specifier
-    evaluates to false. The construction is in the form:
-        .error <ERROR_TEXT>
-    <ERROR_TEXT> is an identifier declared earlier by error text declaration. The declaration is
-    in the form:
-        .errtext <ERROR_TEXT> "<error_desc>"
-    When specifier evaluates to false and this construction is present, parsing is stopped
-    immediately and <error_desc> is returned as a result of parsing. The error position is also
-    returned and it is meant as an offset from the beggining of the stream to the character that
-    was valid so far. Example:
-
-        (**** syntax script ****)
-
-        .syntax program;
-        .errtext MISSING_SEMICOLON      "missing ';'"
-        program         declaration .and .loop space .and ';' .error MISSING_SEMICOLON .and
-                        .loop space .and '\0';
-        declaration     "declare" .and .loop space .and identifier;
-        space           ' ';
-
-        (**** sample code ****)
-
-        declare foo ,
-
-    In the example above checking the sample code will result in error message "missing ';'" and
-    error position 12. The sample code is not correct. Note the presence of '\0' specifier to
-    assure that there is no code after semicolon - only spaces.
-    <error_desc> can optionally contain identifier surrounded by dollar signs $. In such a case,
-    the identifier and dollar signs are replaced by a string retrieved by invoking symbol with
-    the identifier name. The starting position is the error position. The lenght of the resulting
-    string is the position after invoking the symbol.
-
-    PRODUCTION
-    ----------
-
-    Synek not only checks the syntax but it can also produce (emit) bytes associated with specifiers
-    that evaluate to true. That is, every specifier and optional error construction can be followed
-    by a number of emit constructions that are in the form:
-        .emit <parameter>
-    <paramater> can be a HEX number, identifier, a star * or a dollar $. HEX number is preceded by
-    0x or 0X. If <parameter> is an identifier, it must be earlier declared by emit code declaration
-    in the form:
-        .emtcode <identifier> <hex_number>
-
-    When given specifier evaluates to true, all emits associated with the specifier are output
-    in order they were declared. A star means that last-read character should be output instead
-    of constant value. Example:
-
-        (**** syntax script ****)
-
-        .syntax foobar;
-        .emtcode WORD_FOO       0x01
-        .emtcode WORD_BAR       0x02
-        foobar      FOO .emit WORD_FOO .or BAR .emit WORD_BAR .or .true .emit 0x00;
-        FOO         "foo" .and SPACE;
-        BAR         "bar" .and SPACE;
-        SPACE       ' ' .or '\0';
-
-        (**** sample text 1 ****)
-
-        foo
-
-        (**** sample text 2 ****)
-
-        foobar
-
-    For both samples the result will be one-element array. For first sample text it will be
-    value 1, for second - 0. Note that every text will be accepted because of presence of
-    .true as an alternative.
-
-    Another example:
-
-        (**** syntax script ****)
-
-        .syntax declaration;
-        .emtcode VARIABLE       0x01
-        declaration     "declare" .and .loop space .and
-                        identifier .emit VARIABLE .and          (1)
-                        .true .emit 0x00 .and                   (2)
-                        .loop space .and ';';
-        space           ' ' .or '\t';
-        identifier      .loop id_char .emit *;                  (3)
-        id_char         'a'-'z' .or 'A'-'Z' .or '_';
-
-        (**** sample code ****)
-
-        declare    fubar;
-
-    In specifier (1) symbol <identifier> is followed by .emit VARIABLE. If it evaluates to
-    true, VARIABLE constant and then production of the symbol is output. Specifier (2) is used
-    to terminate the string with null to signal when the string ends. Specifier (3) outputs
-    all characters that make declared identifier. The result of sample code will be the
-    following array:
-        { 1, 'f', 'u', 'b', 'a', 'r', 0 }
-
-    If .emit is followed by dollar $, it means that current position should be output. Current
-    position is a 32-bit unsigned integer distance from the very beginning of the parsed string to
-    first character consumed by the specifier associated with the .emit instruction. Current
-    position is stored in the output buffer in Little-Endian convention (the lowest byte comes
-    first).
-*/
-
-#include <stdio.h>
-
-static void mem_free (void **);
-
-/*
-    internal error messages
-*/
-static const byte *OUT_OF_MEMORY =          (byte *) "internal error 1001: out of physical memory";
-static const byte *UNRESOLVED_REFERENCE =   (byte *) "internal error 1002: unresolved reference '$'";
-static const byte *INVALID_GRAMMAR_ID =     (byte *) "internal error 1003: invalid grammar object";
-static const byte *INVALID_REGISTER_NAME =  (byte *) "internal error 1004: invalid register name: '$'";
-/*static const byte *DUPLICATE_IDENTIFIER =   (byte *) "internal error 1005: identifier '$' already defined";*/
-static const byte *UNREFERENCED_IDENTIFIER =(byte *) "internal error 1006: unreferenced identifier '$'";
-
-static const byte *error_message = NULL;    /* points to one of the error messages above */
-static byte *error_param = NULL;        /* this is inserted into error_message in place of $ */
-static int error_position = -1;
-
-static byte *unknown = (byte *) "???";
-
-static void clear_last_error (void)
-{
-    /* reset error message */
-    error_message = NULL;
-
-    /* free error parameter - if error_param is a "???" don't free it - it's static */
-    if (error_param != unknown)
-        mem_free ((void **) (void *) &error_param);
-    else
-        error_param = NULL;
-
-    /* reset error position */
-    error_position = -1;
-}
-
-static void set_last_error (const byte *msg, byte *param, int pos)
-{
-    /* error message can be set only once */
-    if (error_message != NULL)
-    {
-        mem_free ((void **) (void *) &param);
-        return;
-    }
-
-    error_message = msg;
-
-    /* if param is NULL, set error_param to unknown ("???") */
-    /* note: do not try to strdup the "???" - it may be that we are here because of */
-    /* out of memory error so strdup can fail */
-    if (param != NULL)
-        error_param = param;
-    else
-        error_param = unknown;
-
-    error_position = pos;
-}
-
-/*
-    memory management routines
-*/
-static void *mem_alloc (size_t size)
-{
-    void *ptr = grammar_alloc_malloc (size);
-    if (ptr == NULL)
-        set_last_error (OUT_OF_MEMORY, NULL, -1);
-    return ptr;
-}
-
-static void *mem_copy (void *dst, const void *src, size_t size)
-{
-    return grammar_memory_copy (dst, src, size);
-}
-
-static void mem_free (void **ptr)
-{
-    grammar_alloc_free (*ptr);
-    *ptr = NULL;
-}
-
-static void *mem_realloc (void *ptr, size_t old_size, size_t new_size)
-{
-    void *ptr2 = grammar_alloc_realloc (ptr, old_size, new_size);
-    if (ptr2 == NULL)
-        set_last_error (OUT_OF_MEMORY, NULL, -1);
-    return ptr2;
-}
-
-static byte *str_copy_n (byte *dst, const byte *src, size_t max_len)
-{
-    return grammar_string_copy_n (dst, src, max_len);
-}
-
-static byte *str_duplicate (const byte *str)
-{
-    byte *new_str = grammar_string_duplicate (str);
-    if (new_str == NULL)
-        set_last_error (OUT_OF_MEMORY, NULL, -1);
-    return new_str;
-}
-
-static int str_equal (const byte *str1, const byte *str2)
-{
-    return grammar_string_compare (str1, str2) == 0;
-}
-
-static int str_equal_n (const byte *str1, const byte *str2, unsigned int n)
-{
-    return grammar_string_compare_n (str1, str2, n) == 0;
-}
-
-static int
-str_length (const byte *str)
-{
-   return (int) (grammar_string_length (str));
-}
-
-/*
-    useful macros
-*/
-#define GRAMMAR_IMPLEMENT_LIST_APPEND(_Ty)\
-    static void _Ty##_append (_Ty **x, _Ty *nx) {\
-        while (*x) x = &(**x).next;\
-        *x = nx;\
-    }
-
-/*
-    string to byte map typedef
-*/
-typedef struct map_byte_
-{
-    byte *key;
-    byte data;
-    struct map_byte_ *next;
-} map_byte;
-
-static void map_byte_create (map_byte **ma)
-{
-    *ma = (map_byte *) mem_alloc (sizeof (map_byte));
-    if (*ma)
-    {
-        (**ma).key = NULL;
-        (**ma).data = '\0';
-        (**ma).next = NULL;
-    }
-}
-
-static void map_byte_destroy (map_byte **ma)
-{
-    if (*ma)
-    {
-        map_byte_destroy (&(**ma).next);
-        mem_free ((void **) &(**ma).key);
-        mem_free ((void **) ma);
-    }
-}
-
-GRAMMAR_IMPLEMENT_LIST_APPEND(map_byte)
-
-/*
-    searches the map for the specified key,
-    returns pointer to the element with the specified key if it exists
-    returns NULL otherwise
-*/
-static map_byte *map_byte_locate (map_byte **ma, const byte *key)
-{
-    while (*ma)
-    {
-        if (str_equal ((**ma).key, key))
-            return *ma;
-
-        ma = &(**ma).next;
-    }
-
-    set_last_error (UNRESOLVED_REFERENCE, str_duplicate (key), -1);
-    return NULL;
-}
-
-/*
-    searches the map for specified key,
-    if the key is matched, *data is filled with data associated with the key,
-    returns 0 if the key is matched,
-    returns 1 otherwise
-*/
-static int map_byte_find (map_byte **ma, const byte *key, byte *data)
-{
-    map_byte *found = map_byte_locate (ma, key);
-    if (found != NULL)
-    {
-        *data = found->data;
-
-        return 0;
-    }
-
-    return 1;
-}
-
-/*
-    regbyte context typedef
-
-    Each regbyte consists of its name and a default value. These are static and created at
-    grammar script compile-time, for example the following line:
-        .regbyte vertex_blend      0x00
-    adds a new regbyte named "vertex_blend" to the static list and initializes it to 0.
-    When the script is executed, this regbyte can be accessed by name for read and write. When a
-    particular regbyte is written, a new regbyte_ctx entry is added to the top of the regbyte_ctx
-    stack. The new entry contains information abot which regbyte it references and its new value.
-    When a given regbyte is accessed for read, the stack is searched top-down to find an
-    entry that references the regbyte. The first matching entry is used to return the current
-    value it holds. If no entry is found, the default value is returned.
-*/
-typedef struct regbyte_ctx_
-{
-    map_byte *m_regbyte;
-    byte m_current_value;
-    struct regbyte_ctx_ *m_prev;
-} regbyte_ctx;
-
-static void regbyte_ctx_create (regbyte_ctx **re)
-{
-    *re = (regbyte_ctx *) mem_alloc (sizeof (regbyte_ctx));
-    if (*re)
-    {
-        (**re).m_regbyte = NULL;
-        (**re).m_prev = NULL;
-    }
-}
-
-static void regbyte_ctx_destroy (regbyte_ctx **re)
-{
-    if (*re)
-    {
-        mem_free ((void **) re);
-    }
-}
-
-static byte regbyte_ctx_extract (regbyte_ctx **re, map_byte *reg)
-{
-    /* first lookup in the register stack */
-    while (*re != NULL)
-    {
-        if ((**re).m_regbyte == reg)
-            return (**re).m_current_value;
-
-        re = &(**re).m_prev;
-    }
-
-    /* if not found - return the default value */
-    return reg->data;
-}
-
-/*
-    emit type typedef
-*/
-typedef enum emit_type_
-{
-    et_byte,            /* explicit number */
-    et_stream,          /* eaten character */
-    et_position         /* current position */
-} emit_type;
-
-/*
-    emit destination typedef
-*/
-typedef enum emit_dest_
-{
-    ed_output,          /* write to the output buffer */
-    ed_regbyte          /* write a particular regbyte */
-} emit_dest;
-
-/*
-    emit typedef
-*/
-typedef struct emit_
-{
-    emit_dest m_emit_dest;
-    emit_type m_emit_type;      /* ed_output */
-    byte m_byte;                /* et_byte */
-    map_byte *m_regbyte;        /* ed_regbyte */
-    byte *m_regname;            /* ed_regbyte - temporary */
-    struct emit_ *m_next;
-} emit;
-
-static void emit_create (emit **em)
-{
-    *em = (emit *) mem_alloc (sizeof (emit));
-    if (*em)
-    {
-        (**em).m_emit_dest = ed_output;
-        (**em).m_emit_type = et_byte;
-        (**em).m_byte = '\0';
-        (**em).m_regbyte = NULL;
-        (**em).m_regname = NULL;
-        (**em).m_next = NULL;
-    }
-}
-
-static void emit_destroy (emit **em)
-{
-    if (*em)
-    {
-        emit_destroy (&(**em).m_next);
-        mem_free ((void **) &(**em).m_regname);
-        mem_free ((void **) em);
-    }
-}
-
-static unsigned int emit_size (emit *_E)
-{
-    unsigned int n = 0;
-
-    while (_E != NULL)
-    {
-        if (_E->m_emit_dest == ed_output)
-        {
-            if (_E->m_emit_type == et_position)
-                n += 4;     /* position is a 32-bit unsigned integer */
-            else
-                n++;
-        }
-        _E = _E->m_next;
-    }
-
-    return n;
-}
-
-static int emit_push (emit *_E, byte *_P, byte c, unsigned int _Pos, regbyte_ctx **_Ctx)
-{
-    while (_E != NULL)
-    {
-        if (_E->m_emit_dest == ed_output)
-        {
-            if (_E->m_emit_type == et_byte)
-                *_P++ = _E->m_byte;
-            else if (_E->m_emit_type == et_stream)
-                *_P++ = c;
-            else /* _Em->type == et_position */
-            {
-                *_P++ = (byte) (_Pos);
-                *_P++ = (byte) (_Pos >> 8);
-                *_P++ = (byte) (_Pos >> 16);
-                *_P++ = (byte) (_Pos >> 24);
-            }
-        }
-        else
-        {
-            regbyte_ctx *new_rbc;
-            regbyte_ctx_create (&new_rbc);
-            if (new_rbc == NULL)
-                return 1;
-
-            new_rbc->m_prev = *_Ctx;
-            new_rbc->m_regbyte = _E->m_regbyte;
-            *_Ctx = new_rbc;
-
-            if (_E->m_emit_type == et_byte)
-                new_rbc->m_current_value = _E->m_byte;
-            else if (_E->m_emit_type == et_stream)
-                new_rbc->m_current_value = c;
-        }
-
-        _E = _E->m_next;
-    }
-
-    return 0;
-}
-
-/*
-    error typedef
-*/
-typedef struct error_
-{
-    byte *m_text;
-    byte *m_token_name;
-    struct rule_ *m_token;
-} error;
-
-static void error_create (error **er)
-{
-    *er = (error *) mem_alloc (sizeof (error));
-    if (*er)
-    {
-        (**er).m_text = NULL;
-        (**er).m_token_name = NULL;
-        (**er).m_token = NULL;
-    }
-}
-
-static void error_destroy (error **er)
-{
-    if (*er)
-    {
-        mem_free ((void **) &(**er).m_text);
-        mem_free ((void **) &(**er).m_token_name);
-        mem_free ((void **) er);
-    }
-}
-
-struct dict_;
-
-static byte *
-error_get_token (error *, struct dict_ *, const byte *, int);
-
-/*
-    condition operand type typedef
-*/
-typedef enum cond_oper_type_
-{
-    cot_byte,               /* constant 8-bit unsigned integer */
-    cot_regbyte             /* pointer to byte register containing the current value */
-} cond_oper_type;
-
-/*
-    condition operand typedef
-*/
-typedef struct cond_oper_
-{
-    cond_oper_type m_type;
-    byte m_byte;            /* cot_byte */
-    map_byte *m_regbyte;    /* cot_regbyte */
-    byte *m_regname;        /* cot_regbyte - temporary */
-} cond_oper;
-
-/*
-    condition type typedef
-*/
-typedef enum cond_type_
-{
-    ct_equal,
-    ct_not_equal
-} cond_type;
-
-/*
-    condition typedef
-*/
-typedef struct cond_
-{
-    cond_type m_type;
-    cond_oper m_operands[2];
-} cond;
-
-static void cond_create (cond **co)
-{
-    *co = (cond *) mem_alloc (sizeof (cond));
-    if (*co)
-    {
-        (**co).m_operands[0].m_regname = NULL;
-        (**co).m_operands[1].m_regname = NULL;
-    }
-}
-
-static void cond_destroy (cond **co)
-{
-    if (*co)
-    {
-        mem_free ((void **) &(**co).m_operands[0].m_regname);
-        mem_free ((void **) &(**co).m_operands[1].m_regname);
-        mem_free ((void **) co);
-    }
-}
-
-/*
-    specifier type typedef
-*/
-typedef enum spec_type_
-{
-    st_false,
-    st_true,
-    st_byte,
-    st_byte_range,
-    st_string,
-    st_identifier,
-    st_identifier_loop,
-    st_debug
-} spec_type;
-
-/*
-    specifier typedef
-*/
-typedef struct spec_
-{
-    spec_type m_spec_type;
-    byte m_byte[2];                 /* st_byte, st_byte_range */
-    byte *m_string;                 /* st_string */
-    struct rule_ *m_rule;           /* st_identifier, st_identifier_loop */
-    emit *m_emits;
-    error *m_errtext;
-    cond *m_cond;
-    struct spec_ *next;
-} spec;
-
-static void spec_create (spec **sp)
-{
-    *sp = (spec *) mem_alloc (sizeof (spec));
-    if (*sp)
-    {
-        (**sp).m_spec_type = st_false;
-        (**sp).m_byte[0] = '\0';
-        (**sp).m_byte[1] = '\0';
-        (**sp).m_string = NULL;
-        (**sp).m_rule = NULL;
-        (**sp).m_emits = NULL;
-        (**sp).m_errtext = NULL;
-        (**sp).m_cond = NULL;
-        (**sp).next = NULL;
-    }
-}
-
-static void spec_destroy (spec **sp)
-{
-    if (*sp)
-    {
-        spec_destroy (&(**sp).next);
-        emit_destroy (&(**sp).m_emits);
-        error_destroy (&(**sp).m_errtext);
-        mem_free ((void **) &(**sp).m_string);
-        cond_destroy (&(**sp).m_cond);
-        mem_free ((void **) sp);
-    }
-}
-
-GRAMMAR_IMPLEMENT_LIST_APPEND(spec)
-
-/*
-    operator typedef
-*/
-typedef enum oper_
-{
-    op_none,
-    op_and,
-    op_or
-} oper;
-
-/*
-    rule typedef
-*/
-typedef struct rule_
-{
-    oper m_oper;
-    spec *m_specs;
-    struct rule_ *next;
-    int m_referenced;
-} rule;
-
-static void rule_create (rule **ru)
-{
-    *ru = (rule *) mem_alloc (sizeof (rule));
-    if (*ru)
-    {
-        (**ru).m_oper = op_none;
-        (**ru).m_specs = NULL;
-        (**ru).next = NULL;
-        (**ru).m_referenced = 0;
-    }
-}
-
-static void rule_destroy (rule **ru)
-{
-    if (*ru)
-    {
-        rule_destroy (&(**ru).next);
-        spec_destroy (&(**ru).m_specs);
-        mem_free ((void **) ru);
-    }
-}
-
-GRAMMAR_IMPLEMENT_LIST_APPEND(rule)
-
-/*
-    returns unique grammar id
-*/
-static grammar next_valid_grammar_id (void)
-{
-    static grammar id = 0;
-
-    return ++id;
-}
-
-/*
-    dictionary typedef
-*/
-typedef struct dict_
-{
-    rule *m_rulez;
-    rule *m_syntax;
-    rule *m_string;
-    map_byte *m_regbytes;
-    grammar m_id;
-    struct dict_ *next;
-} dict;
-
-static void dict_create (dict **di)
-{
-    *di = (dict *) mem_alloc (sizeof (dict));
-    if (*di)
-    {
-        (**di).m_rulez = NULL;
-        (**di).m_syntax = NULL;
-        (**di).m_string = NULL;
-        (**di).m_regbytes = NULL;
-        (**di).m_id = next_valid_grammar_id ();
-        (**di).next = NULL;
-    }
-}
-
-static void dict_destroy (dict **di)
-{
-    if (*di)
-    {
-        rule_destroy (&(**di).m_rulez);
-        map_byte_destroy (&(**di).m_regbytes);
-        mem_free ((void **) di);
-    }
-}
-
-GRAMMAR_IMPLEMENT_LIST_APPEND(dict)
-
-static void dict_find (dict **di, grammar key, dict **data)
-{
-    while (*di)
-    {
-        if ((**di).m_id == key)
-        {
-            *data = *di;
-            return;
-        }
-
-        di = &(**di).next;
-    }
-
-    *data = NULL;
-}
-
-static dict *g_dicts = NULL;
-
-/*
-    byte array typedef
-*/
-typedef struct barray_
-{
-    byte *data;
-    unsigned int len;
-} barray;
-
-static void barray_create (barray **ba)
-{
-    *ba = (barray *) mem_alloc (sizeof (barray));
-    if (*ba)
-    {
-        (**ba).data = NULL;
-        (**ba).len = 0;
-    }
-}
-
-static void barray_destroy (barray **ba)
-{
-    if (*ba)
-    {
-        mem_free ((void **) &(**ba).data);
-        mem_free ((void **) ba);
-    }
-}
-
-/*
-    reallocates byte array to requested size,
-    returns 0 on success,
-    returns 1 otherwise
-*/
-static int barray_resize (barray **ba, unsigned int nlen)
-{
-    byte *new_pointer;
-
-    if (nlen == 0)
-    {
-        mem_free ((void **) &(**ba).data);
-        (**ba).data = NULL;
-        (**ba).len = 0;
-
-        return 0;
-    }
-    else
-    {
-        new_pointer = (byte *) mem_realloc ((**ba).data, (**ba).len * sizeof (byte),
-            nlen * sizeof (byte));
-        if (new_pointer)
-        {
-            (**ba).data = new_pointer;
-            (**ba).len = nlen;
-
-            return 0;
-        }
-    }
-
-    return 1;
-}
-
-/*
-    adds byte array pointed by *nb to the end of array pointed by *ba,
-    returns 0 on success,
-    returns 1 otherwise
-*/
-static int barray_append (barray **ba, barray **nb)
-{
-    const unsigned int len = (**ba).len;
-
-    if (barray_resize (ba, (**ba).len + (**nb).len))
-        return 1;
-
-    mem_copy ((**ba).data + len, (**nb).data, (**nb).len);
-
-    return 0;
-}
-
-/*
-    adds emit chain pointed by em to the end of array pointed by *ba,
-    returns 0 on success,
-    returns 1 otherwise
-*/
-static int barray_push (barray **ba, emit *em, byte c, unsigned int pos, regbyte_ctx **rbc)
-{
-    unsigned int count = emit_size (em);
-
-    if (barray_resize (ba, (**ba).len + count))
-        return 1;
-
-    return emit_push (em, (**ba).data + ((**ba).len - count), c, pos, rbc);
-}
-
-/*
-    byte pool typedef
-*/
-typedef struct bytepool_
-{
-    byte *_F;
-    unsigned int _Siz;
-} bytepool;
-
-static void bytepool_destroy (bytepool **by)
-{
-    if (*by != NULL)
-    {
-        mem_free ((void **) &(**by)._F);
-        mem_free ((void **) by);
-    }
-}
-
-static void bytepool_create (bytepool **by, int len)
-{
-    *by = (bytepool *) (mem_alloc (sizeof (bytepool)));
-    if (*by != NULL)
-    {
-        (**by)._F = (byte *) (mem_alloc (sizeof (byte) * len));
-        (**by)._Siz = len;
-
-        if ((**by)._F == NULL)
-            bytepool_destroy (by);
-    }
-}
-
-static int bytepool_reserve (bytepool *by, unsigned int n)
-{
-    byte *_P;
-
-    if (n <= by->_Siz)
-        return 0;
-
-    /* byte pool can only grow and at least by doubling its size */
-    n = n >= by->_Siz * 2 ? n : by->_Siz * 2;
-
-    /* reallocate the memory and adjust pointers to the new memory location */
-    _P = (byte *) (mem_realloc (by->_F, sizeof (byte) * by->_Siz, sizeof (byte) * n));
-    if (_P != NULL)
-    {
-        by->_F = _P;
-        by->_Siz = n;
-        return 0;
-    }
-
-    return 1;
-}
-
-/*
-    string to string map typedef
-*/
-typedef struct map_str_
-{
-    byte *key;
-    byte *data;
-    struct map_str_ *next;
-} map_str;
-
-static void map_str_create (map_str **ma)
-{
-    *ma = (map_str *) mem_alloc (sizeof (map_str));
-    if (*ma)
-    {
-        (**ma).key = NULL;
-        (**ma).data = NULL;
-        (**ma).next = NULL;
-    }
-}
-
-static void map_str_destroy (map_str **ma)
-{
-    if (*ma)
-    {
-        map_str_destroy (&(**ma).next);
-        mem_free ((void **) &(**ma).key);
-        mem_free ((void **) &(**ma).data);
-        mem_free ((void **) ma);
-    }
-}
-
-GRAMMAR_IMPLEMENT_LIST_APPEND(map_str)
-
-/*
-    searches the map for specified key,
-    if the key is matched, *data is filled with data associated with the key,
-    returns 0 if the key is matched,
-    returns 1 otherwise
-*/
-static int map_str_find (map_str **ma, const byte *key, byte **data)
-{
-    while (*ma)
-    {
-        if (str_equal ((**ma).key, key))
-        {
-            *data = str_duplicate ((**ma).data);
-            if (*data == NULL)
-                return 1;
-
-            return 0;
-        }
-
-        ma = &(**ma).next;
-    }
-
-    set_last_error (UNRESOLVED_REFERENCE, str_duplicate (key), -1);
-    return 1;
-}
-
-/*
-    string to rule map typedef
-*/
-typedef struct map_rule_
-{
-    byte *key;
-    rule *data;
-    struct map_rule_ *next;
-} map_rule;
-
-static void map_rule_create (map_rule **ma)
-{
-    *ma = (map_rule *) mem_alloc (sizeof (map_rule));
-    if (*ma)
-    {
-        (**ma).key = NULL;
-        (**ma).data = NULL;
-        (**ma).next = NULL;
-    }
-}
-
-static void map_rule_destroy (map_rule **ma)
-{
-    if (*ma)
-    {
-        map_rule_destroy (&(**ma).next);
-        mem_free ((void **) &(**ma).key);
-        mem_free ((void **) ma);
-    }
-}
-
-GRAMMAR_IMPLEMENT_LIST_APPEND(map_rule)
-
-/*
-    searches the map for specified key,
-    if the key is matched, *data is filled with data associated with the key,
-    returns 0 if the is matched,
-    returns 1 otherwise
-*/
-static int map_rule_find (map_rule **ma, const byte *key, rule **data)
-{
-    while (*ma)
-    {
-        if (str_equal ((**ma).key, key))
-        {
-            *data = (**ma).data;
-
-            return 0;
-        }
-
-        ma = &(**ma).next;
-    }
-
-    set_last_error (UNRESOLVED_REFERENCE, str_duplicate (key), -1);
-    return 1;
-}
-
-/*
-    returns 1 if given character is a white space,
-    returns 0 otherwise
-*/
-static int is_space (byte c)
-{
-    return c == ' ' || c == '\t' || c == '\n' || c == '\r';
-}
-
-/*
-    advances text pointer by 1 if character pointed by *text is a space,
-    returns 1 if a space has been eaten,
-    returns 0 otherwise
-*/
-static int eat_space (const byte **text)
-{
-    if (is_space (**text))
-    {
-        (*text)++;
-
-        return 1;
-    }
-
-    return 0;
-}
-
-/*
-    returns 1 if text points to C-style comment start string,
-    returns 0 otherwise
-*/
-static int is_comment_start (const byte *text)
-{
-    return text[0] == '/' && text[1] == '*';
-}
-
-/*
-    advances text pointer to first character after C-style comment block - if any,
-    returns 1 if C-style comment block has been encountered and eaten,
-    returns 0 otherwise
-*/
-static int eat_comment (const byte **text)
-{
-    if (is_comment_start (*text))
-    {
-        /* *text points to comment block - skip two characters to enter comment body */
-        *text += 2;
-        /* skip any character except consecutive '*' and '/' */
-        while (!((*text)[0] == '*' && (*text)[1] == '/'))
-            (*text)++;
-        /* skip those two terminating characters */
-        *text += 2;
-
-        return 1;
-    }
-
-    return 0;
-}
-
-/*
-    advances text pointer to first character that is neither space nor C-style comment block
-*/
-static void eat_spaces (const byte **text)
-{
-    while (eat_space (text) || eat_comment (text))
-        ;
-}
-
-/*
-    resizes string pointed by *ptr to successfully add character c to the end of the string,
-    returns 0 on success,
-    returns 1 otherwise
-*/
-static int string_grow (byte **ptr, unsigned int *len, byte c)
-{
-    /* reallocate the string in 16-byte increments */
-    if ((*len & 0x0F) == 0x0F || *ptr == NULL)
-    {
-        byte *tmp = (byte *) mem_realloc (*ptr, ((*len + 1) & ~0x0F) * sizeof (byte),
-            ((*len + 1 + 0x10) & ~0x0F) * sizeof (byte));
-        if (tmp == NULL)
-            return 1;
-
-        *ptr = tmp;
-    }
-
-    if (c)
-    {
-        /* append given character */
-        (*ptr)[*len] = c;
-        (*len)++;
-    }
-    (*ptr)[*len] = '\0';
-
-    return 0;
-}
-
-/*
-    returns 1 if given character is a valid identifier character a-z, A-Z, 0-9 or _
-    returns 0 otherwise
-*/
-static int is_identifier (byte c)
-{
-    return (c >= 'a' && c <= 'z') || (c >= 'A' && c <= 'Z') || (c >= '0' && c <= '9') || c == '_';
-}
-
-/*
-    copies characters from *text to *id until non-identifier character is encountered,
-    assumes that *id points to NULL object - caller is responsible for later freeing the string,
-    text pointer is advanced to point past the copied identifier,
-    returns 0 if identifier was successfully copied,
-    returns 1 otherwise
-*/
-static int get_identifier (const byte **text, byte **id)
-{
-    const byte *t = *text;
-    byte *p = NULL;
-    unsigned int len = 0;
-
-    if (string_grow (&p, &len, '\0'))
-        return 1;
-
-    /* loop while next character in buffer is valid for identifiers */
-    while (is_identifier (*t))
-    {
-        if (string_grow (&p, &len, *t++))
-        {
-            mem_free ((void **) (void *) &p);
-            return 1;
-        }
-    }
-
-    *text = t;
-    *id = p;
-
-    return 0;
-}
-
-/*
-    converts sequence of DEC digits pointed by *text until non-DEC digit is encountered,
-    advances text pointer past the converted sequence,
-    returns the converted value
-*/
-static unsigned int dec_convert (const byte **text)
-{
-    unsigned int value = 0;
-
-    while (**text >= '0' && **text <= '9')
-    {
-        value = value * 10 + **text - '0';
-        (*text)++;
-    }
-
-    return value;
-}
-
-/*
-    returns 1 if given character is HEX digit 0-9, A-F or a-f,
-    returns 0 otherwise
-*/
-static int is_hex (byte c)
-{
-    return (c >= '0' && c <= '9') || (c >= 'A' && c <= 'F') || (c >= 'a' && c <= 'f');
-}
-
-/*
-    returns value of passed character as if it was HEX digit
-*/
-static unsigned int hex2dec (byte c)
-{
-    if (c >= '0' && c <= '9')
-        return c - '0';
-    if (c >= 'A' && c <= 'F')
-        return c - 'A' + 10;
-    return c - 'a' + 10;
-}
-
-/*
-    converts sequence of HEX digits pointed by *text until non-HEX digit is encountered,
-    advances text pointer past the converted sequence,
-    returns the converted value
-*/
-static unsigned int hex_convert (const byte **text)
-{
-    unsigned int value = 0;
-
-    while (is_hex (**text))
-    {
-        value = value * 0x10 + hex2dec (**text);
-        (*text)++;
-    }
-
-    return value;
-}
-
-/*
-    returns 1 if given character is OCT digit 0-7,
-    returns 0 otherwise
-*/
-static int is_oct (byte c)
-{
-    return c >= '0' && c <= '7';
-}
-
-/*
-    returns value of passed character as if it was OCT digit
-*/
-static int oct2dec (byte c)
-{
-    return c - '0';
-}
-
-static byte get_escape_sequence (const byte **text)
-{
-    int value = 0;
-
-    /* skip '\' character */
-    (*text)++;
-
-    switch (*(*text)++)
-    {
-    case '\'':
-        return '\'';
-    case '"':
-        return '\"';
-    case '?':
-        return '\?';
-    case '\\':
-        return '\\';
-    case 'a':
-        return '\a';
-    case 'b':
-        return '\b';
-    case 'f':
-        return '\f';
-    case 'n':
-        return '\n';
-    case 'r':
-        return '\r';
-    case 't':
-        return '\t';
-    case 'v':
-        return '\v';
-    case 'x':
-        return (byte) hex_convert (text);
-    }
-
-    (*text)--;
-    if (is_oct (**text))
-    {
-        value = oct2dec (*(*text)++);
-        if (is_oct (**text))
-        {
-            value = value * 010 + oct2dec (*(*text)++);
-            if (is_oct (**text))
-                value = value * 010 + oct2dec (*(*text)++);
-        }
-    }
-
-    return (byte) value;
-}
-
-/*
-    copies characters from *text to *str until " or ' character is encountered,
-    assumes that *str points to NULL object - caller is responsible for later freeing the string,
-    assumes that *text points to " or ' character that starts the string,
-    text pointer is advanced to point past the " or ' character,
-    returns 0 if string was successfully copied,
-    returns 1 otherwise
-*/
-static int get_string (const byte **text, byte **str)
-{
-    const byte *t = *text;
-    byte *p = NULL;
-    unsigned int len = 0;
-    byte term_char;
-
-    if (string_grow (&p, &len, '\0'))
-        return 1;
-
-    /* read " or ' character that starts the string */
-    term_char = *t++;
-    /* while next character is not the terminating character */
-    while (*t && *t != term_char)
-    {
-        byte c;
-
-        if (*t == '\\')
-            c = get_escape_sequence (&t);
-        else
-            c = *t++;
-
-        if (string_grow (&p, &len, c))
-        {
-            mem_free ((void **) (void *) &p);
-            return 1;
-        }
-    }
-    /* skip " or ' character that ends the string */
-    t++;
-
-    *text = t;
-    *str = p;
-    return 0;
-}
-
-/*
-    gets emit code, the syntax is:
-    ".emtcode" " " <symbol> " " (("0x" | "0X") <hex_value>) | <dec_value> | <character>
-    assumes that *text already points to <symbol>,
-    returns 0 if emit code is successfully read,
-    returns 1 otherwise
-*/
-static int get_emtcode (const byte **text, map_byte **ma)
-{
-    const byte *t = *text;
-    map_byte *m = NULL;
-
-    map_byte_create (&m);
-    if (m == NULL)
-        return 1;
-
-    if (get_identifier (&t, &m->key))
-    {
-        map_byte_destroy (&m);
-        return 1;
-    }
-    eat_spaces (&t);
-
-    if (*t == '\'')
-    {
-        byte *c;
-
-        if (get_string (&t, &c))
-        {
-            map_byte_destroy (&m);
-            return 1;
-        }
-
-        m->data = (byte) c[0];
-        mem_free ((void **) (void *) &c);
-    }
-    else if (t[0] == '0' && (t[1] == 'x' || t[1] == 'X'))
-    {
-        /* skip HEX "0x" or "0X" prefix */
-        t += 2;
-        m->data = (byte) hex_convert (&t);
-    }
-    else
-    {
-        m->data = (byte) dec_convert (&t);
-    }
-
-    eat_spaces (&t);
-
-    *text = t;
-    *ma = m;
-    return 0;
-}
-
-/*
-    gets regbyte declaration, the syntax is:
-    ".regbyte" " " <symbol> " " (("0x" | "0X") <hex_value>) | <dec_value> | <character>
-    assumes that *text already points to <symbol>,
-    returns 0 if regbyte is successfully read,
-    returns 1 otherwise
-*/
-static int get_regbyte (const byte **text, map_byte **ma)
-{
-    /* pass it to the emtcode parser as it has the same syntax starting at <symbol> */
-    return get_emtcode (text, ma);
-}
-
-/*
-    returns 0 on success,
-    returns 1 otherwise
-*/
-static int get_errtext (const byte **text, map_str **ma)
-{
-    const byte *t = *text;
-    map_str *m = NULL;
-
-    map_str_create (&m);
-    if (m == NULL)
-        return 1;
-
-    if (get_identifier (&t, &m->key))
-    {
-        map_str_destroy (&m);
-        return 1;
-    }
-    eat_spaces (&t);
-
-    if (get_string (&t, &m->data))
-    {
-        map_str_destroy (&m);
-        return 1;
-    }
-    eat_spaces (&t);
-
-    *text = t;
-    *ma = m;
-    return 0;
-}
-
-/*
-    returns 0 on success,
-    returns 1 otherwise,
-*/
-static int get_error (const byte **text, error **er, map_str *maps)
-{
-    const byte *t = *text;
-    byte *temp = NULL;
-
-    if (*t != '.')
-        return 0;
-
-    t++;
-    if (get_identifier (&t, &temp))
-        return 1;
-    eat_spaces (&t);
-
-    if (!str_equal ((byte *) "error", temp))
-    {
-        mem_free ((void **) (void *) &temp);
-        return 0;
-    }
-
-    mem_free ((void **) (void *) &temp);
-
-    error_create (er);
-    if (*er == NULL)
-        return 1;
-
-    if (*t == '\"')
-    {
-        if (get_string (&t, &(**er).m_text))
-        {
-            error_destroy (er);
-            return 1;
-        }
-        eat_spaces (&t);
-    }
-    else
-    {
-        if (get_identifier (&t, &temp))
-        {
-            error_destroy (er);
-            return 1;
-        }
-        eat_spaces (&t);
-
-        if (map_str_find (&maps, temp, &(**er).m_text))
-        {
-            mem_free ((void **) (void *) &temp);
-            error_destroy (er);
-            return 1;
-        }
-
-        mem_free ((void **) (void *) &temp);
-    }
-
-    /* try to extract "token" from "...$token$..." */
-    {
-        byte *processed = NULL;
-        unsigned int len = 0;
-      int i = 0;
-
-        if (string_grow (&processed, &len, '\0'))
-        {
-            error_destroy (er);
-            return 1;
-        }
-
-        while (i < str_length ((**er).m_text))
-        {
-            /* check if the dollar sign is repeated - if so skip it */
-            if ((**er).m_text[i] == '$' && (**er).m_text[i + 1] == '$')
-            {
-                if (string_grow (&processed, &len, '$'))
-                {
-                    mem_free ((void **) (void *) &processed);
-                    error_destroy (er);
-                    return 1;
-                }
-
-                i += 2;
-            }
-            else if ((**er).m_text[i] != '$')
-            {
-                if (string_grow (&processed, &len, (**er).m_text[i]))
-                {
-                    mem_free ((void **) (void *) &processed);
-                    error_destroy (er);
-                    return 1;
-                }
-
-                i++;
-            }
-            else
-            {
-                if (string_grow (&processed, &len, '$'))
-                {
-                    mem_free ((void **) (void *) &processed);
-                    error_destroy (er);
-                    return 1;
-                }
-
-                {
-                    /* length of token being extracted */
-                    unsigned int tlen = 0;
-
-                    if (string_grow (&(**er).m_token_name, &tlen, '\0'))
-                    {
-                        mem_free ((void **) (void *) &processed);
-                        error_destroy (er);
-                        return 1;
-                    }
-
-                    /* skip the dollar sign */
-                    i++;
-
-                    while ((**er).m_text[i] != '$')
-                    {
-                        if (string_grow (&(**er).m_token_name, &tlen, (**er).m_text[i]))
-                        {
-                            mem_free ((void **) (void *) &processed);
-                            error_destroy (er);
-                            return 1;
-                        }
-
-                        i++;
-                    }
-
-                    /* skip the dollar sign */
-                    i++;
-                }
-            }
-        }
-
-        mem_free ((void **) &(**er).m_text);
-        (**er).m_text = processed;
-    }
-
-    *text = t;
-    return 0;
-}
-
-/*
-    returns 0 on success,
-    returns 1 otherwise,
-*/
-static int get_emits (const byte **text, emit **em, map_byte *mapb)
-{
-    const byte *t = *text;
-    byte *temp = NULL;
-    emit *e = NULL;
-    emit_dest dest;
-
-    if (*t != '.')
-        return 0;
-
-    t++;
-    if (get_identifier (&t, &temp))
-        return 1;
-    eat_spaces (&t);
-
-    /* .emit */
-    if (str_equal ((byte *) "emit", temp))
-        dest = ed_output;
-    /* .load */
-    else if (str_equal ((byte *) "load", temp))
-        dest = ed_regbyte;
-    else
-    {
-        mem_free ((void **) (void *) &temp);
-        return 0;
-    }
-
-    mem_free ((void **) (void *) &temp);
-
-    emit_create (&e);
-    if (e == NULL)
-        return 1;
-
-    e->m_emit_dest = dest;
-
-    if (dest == ed_regbyte)
-    {
-        if (get_identifier (&t, &e->m_regname))
-        {
-            emit_destroy (&e);
-            return 1;
-        }
-        eat_spaces (&t);
-    }
-
-    /* 0xNN */
-    if (*t == '0' && (t[1] == 'x' || t[1] == 'X'))
-    {
-        t += 2;
-        e->m_byte = (byte) hex_convert (&t);
-
-        e->m_emit_type = et_byte;
-    }
-    /* NNN */
-    else if (*t >= '0' && *t <= '9')
-    {
-        e->m_byte = (byte) dec_convert (&t);
-
-        e->m_emit_type = et_byte;
-    }
-    /* * */
-    else if (*t == '*')
-    {
-        t++;
-
-        e->m_emit_type = et_stream;
-    }
-    /* $ */
-    else if (*t == '$')
-    {
-        t++;
-
-        e->m_emit_type = et_position;
-    }
-    /* 'c' */
-    else if (*t == '\'')
-    {
-        if (get_string (&t, &temp))
-        {
-            emit_destroy (&e);
-            return 1;
-        }
-        e->m_byte = (byte) temp[0];
-
-        mem_free ((void **) (void *) &temp);
-
-        e->m_emit_type = et_byte;
-    }
-    else
-    {
-        if (get_identifier (&t, &temp))
-        {
-            emit_destroy (&e);
-            return 1;
-        }
-
-        if (map_byte_find (&mapb, temp, &e->m_byte))
-        {
-            mem_free ((void **) (void *) &temp);
-            emit_destroy (&e);
-            return 1;
-        }
-
-        mem_free ((void **) (void *) &temp);
-
-        e->m_emit_type = et_byte;
-    }
-
-    eat_spaces (&t);
-
-    if (get_emits (&t, &e->m_next, mapb))
-    {
-        emit_destroy (&e);
-        return 1;
-    }
-
-    *text = t;
-    *em = e;
-    return 0;
-}
-
-/*
-    returns 0 on success,
-    returns 1 otherwise,
-*/
-static int get_spec (const byte **text, spec **sp, map_str *maps, map_byte *mapb)
-{
-    const byte *t = *text;
-    spec *s = NULL;
-
-    spec_create (&s);
-    if (s == NULL)
-        return 1;
-
-    /* first - read optional .if statement */
-    if (*t == '.')
-    {
-        const byte *u = t;
-        byte *keyword = NULL;
-
-        /* skip the dot */
-        u++;
-
-        if (get_identifier (&u, &keyword))
-        {
-            spec_destroy (&s);
-            return 1;
-        }
-
-        /* .if */
-        if (str_equal ((byte *) "if", keyword))
-        {
-            cond_create (&s->m_cond);
-            if (s->m_cond == NULL)
-            {
-                spec_destroy (&s);
-                return 1;
-            }
-
-            /* skip the left paren */
-            eat_spaces (&u);
-            u++;
-
-            /* get the left operand */
-            eat_spaces (&u);
-            if (get_identifier (&u, &s->m_cond->m_operands[0].m_regname))
-            {
-                spec_destroy (&s);
-                return 1;
-            }
-            s->m_cond->m_operands[0].m_type = cot_regbyte;
-
-            /* get the operator (!= or ==) */
-            eat_spaces (&u);
-            if (*u == '!')
-                s->m_cond->m_type = ct_not_equal;
-            else
-                s->m_cond->m_type = ct_equal;
-            u += 2;
-            eat_spaces (&u);
-
-            if (u[0] == '0' && (u[1] == 'x' || u[1] == 'X'))
-            {
-                /* skip the 0x prefix */
-                u += 2;
-
-                /* get the right operand */
-                s->m_cond->m_operands[1].m_byte = hex_convert (&u);
-                s->m_cond->m_operands[1].m_type = cot_byte;
-            }
-            else /*if (*u >= '0' && *u <= '9')*/
-            {
-                /* get the right operand */
-                s->m_cond->m_operands[1].m_byte = dec_convert (&u);
-                s->m_cond->m_operands[1].m_type = cot_byte;
-            }
-
-            /* skip the right paren */
-            eat_spaces (&u);
-            u++;
-
-            eat_spaces (&u);
-
-            t = u;
-        }
-
-        mem_free ((void **) (void *) &keyword);
-    }
-
-    if (*t == '\'')
-    {
-        byte *temp = NULL;
-
-        if (get_string (&t, &temp))
-        {
-            spec_destroy (&s);
-            return 1;
-        }
-        eat_spaces (&t);
-
-        if (*t == '-')
-        {
-            byte *temp2 = NULL;
-
-            /* skip the '-' character */
-            t++;
-            eat_spaces (&t);
-
-            if (get_string (&t, &temp2))
-            {
-                mem_free ((void **) (void *) &temp);
-                spec_destroy (&s);
-                return 1;
-            }
-            eat_spaces (&t);
-
-            s->m_spec_type = st_byte_range;
-            s->m_byte[0] = *temp;
-            s->m_byte[1] = *temp2;
-
-            mem_free ((void **) (void *) &temp2);
-        }
-        else
-        {
-            s->m_spec_type = st_byte;
-            *s->m_byte = *temp;
-        }
-
-        mem_free ((void **) (void *) &temp);
-    }
-    else if (*t == '"')
-    {
-        if (get_string (&t, &s->m_string))
-        {
-            spec_destroy (&s);
-            return 1;
-        }
-        eat_spaces (&t);
-
-        s->m_spec_type = st_string;
-    }
-    else if (*t == '.')
-    {
-        byte *keyword = NULL;
-
-        /* skip the dot */
-        t++;
-
-        if (get_identifier (&t, &keyword))
-        {
-            spec_destroy (&s);
-            return 1;
-        }
-        eat_spaces (&t);
-
-        /* .true */
-        if (str_equal ((byte *) "true", keyword))
-        {
-            s->m_spec_type = st_true;
-        }
-        /* .false */
-        else if (str_equal ((byte *) "false", keyword))
-        {
-            s->m_spec_type = st_false;
-        }
-        /* .debug */
-        else if (str_equal ((byte *) "debug", keyword))
-        {
-            s->m_spec_type = st_debug;
-        }
-        /* .loop */
-        else if (str_equal ((byte *) "loop", keyword))
-        {
-            if (get_identifier (&t, &s->m_string))
-            {
-                mem_free ((void **) (void *) &keyword);
-                spec_destroy (&s);
-                return 1;
-            }
-            eat_spaces (&t);
-
-            s->m_spec_type = st_identifier_loop;
-        }
-        mem_free ((void **) (void *) &keyword);
-    }
-    else
-    {
-        if (get_identifier (&t, &s->m_string))
-        {
-            spec_destroy (&s);
-            return 1;
-        }
-        eat_spaces (&t);
-
-        s->m_spec_type = st_identifier;
-    }
-
-    if (get_error (&t, &s->m_errtext, maps))
-    {
-        spec_destroy (&s);
-        return 1;
-    }
-
-    if (get_emits (&t, &s->m_emits, mapb))
-    {
-        spec_destroy (&s);
-        return 1;
-    }
-
-    *text = t;
-    *sp = s;
-    return 0;
-}
-
-/*
-    returns 0 on success,
-    returns 1 otherwise,
-*/
-static int get_rule (const byte **text, rule **ru, map_str *maps, map_byte *mapb)
-{
-    const byte *t = *text;
-    rule *r = NULL;
-
-    rule_create (&r);
-    if (r == NULL)
-        return 1;
-
-    if (get_spec (&t, &r->m_specs, maps, mapb))
-    {
-        rule_destroy (&r);
-        return 1;
-    }
-
-    while (*t != ';')
-    {
-        byte *op = NULL;
-        spec *sp = NULL;
-
-        /* skip the dot that precedes "and" or "or" */
-        t++;
-
-        /* read "and" or "or" keyword */
-        if (get_identifier (&t, &op))
-        {
-            rule_destroy (&r);
-            return 1;
-        }
-        eat_spaces (&t);
-
-        if (r->m_oper == op_none)
-        {
-            /* .and */
-            if (str_equal ((byte *) "and", op))
-                r->m_oper = op_and;
-            /* .or */
-            else
-                r->m_oper = op_or;
-        }
-
-        mem_free ((void **) (void *) &op);
-
-        if (get_spec (&t, &sp, maps, mapb))
-        {
-            rule_destroy (&r);
-            return 1;
-        }
-
-        spec_append (&r->m_specs, sp);
-    }
-
-    /* skip the semicolon */
-    t++;
-    eat_spaces (&t);
-
-    *text = t;
-    *ru = r;
-    return 0;
-}
-
-/*
-    returns 0 on success,
-    returns 1 otherwise,
-*/
-static int update_dependency (map_rule *mapr, byte *symbol, rule **ru)
-{
-    if (map_rule_find (&mapr, symbol, ru))
-        return 1;
-
-    (**ru).m_referenced = 1;
-
-    return 0;
-}
-
-/*
-    returns 0 on success,
-    returns 1 otherwise,
-*/
-static int update_dependencies (dict *di, map_rule *mapr, byte **syntax_symbol,
-    byte **string_symbol, map_byte *regbytes)
-{
-    rule *rulez = di->m_rulez;
-
-    /* update dependecies for the root and lexer symbols */
-    if (update_dependency (mapr, *syntax_symbol, &di->m_syntax) ||
-        (*string_symbol != NULL && update_dependency (mapr, *string_symbol, &di->m_string)))
-        return 1;
-
-    mem_free ((void **) syntax_symbol);
-    mem_free ((void **) string_symbol);
-
-    /* update dependecies for the rest of the rules */
-    while (rulez)
-    {
-        spec *sp = rulez->m_specs;
-
-        /* iterate through all the specifiers */
-        while (sp)
-        {
-            /* update dependency for identifier */
-            if (sp->m_spec_type == st_identifier || sp->m_spec_type == st_identifier_loop)
-            {
-                if (update_dependency (mapr, sp->m_string, &sp->m_rule))
-                    return 1;
-
-                mem_free ((void **) &sp->m_string);
-            }
-
-            /* some errtexts reference to a rule */
-            if (sp->m_errtext && sp->m_errtext->m_token_name)
-            {
-                if (update_dependency (mapr, sp->m_errtext->m_token_name, &sp->m_errtext->m_token))
-                    return 1;
-
-                mem_free ((void **) &sp->m_errtext->m_token_name);
-            }
-
-            /* update dependency for condition */
-            if (sp->m_cond)
-            {
-                int i;
-                for (i = 0; i < 2; i++)
-                    if (sp->m_cond->m_operands[i].m_type == cot_regbyte)
-                    {
-                        sp->m_cond->m_operands[i].m_regbyte = map_byte_locate (&regbytes,
-                            sp->m_cond->m_operands[i].m_regname);
-
-                        if (sp->m_cond->m_operands[i].m_regbyte == NULL)
-                            return 1;
-
-                        mem_free ((void **) &sp->m_cond->m_operands[i].m_regname);
-                    }
-            }
-
-            /* update dependency for all .load instructions */
-            if (sp->m_emits)
-            {
-                emit *em = sp->m_emits;
-                while (em != NULL)
-                {
-                    if (em->m_emit_dest == ed_regbyte)
-                    {
-                        em->m_regbyte = map_byte_locate (&regbytes, em->m_regname);
-
-                        if (em->m_regbyte == NULL)
-                            return 1;
-
-                        mem_free ((void **) &em->m_regname);
-                    }
-
-                    em = em->m_next;
-                }
-            }
-
-            sp = sp->next;
-        }
-
-        rulez = rulez->next;
-    }
-
-    /* check for unreferenced symbols */
-    rulez = di->m_rulez;
-    while (rulez != NULL)
-    {
-        if (!rulez->m_referenced)
-        {
-            map_rule *ma = mapr;
-            while (ma)
-            {
-                if (ma->data == rulez)
-                {
-                    set_last_error (UNREFERENCED_IDENTIFIER, str_duplicate (ma->key), -1);
-                    return 1;
-                }
-                ma = ma->next;
-            }
-        }
-        rulez = rulez->next;
-    }
-
-    return 0;
-}
-
-static int satisfies_condition (cond *co, regbyte_ctx *ctx)
-{
-    byte values[2];
-    int i;
-
-    if (co == NULL)
-        return 1;
-
-    for (i = 0; i < 2; i++)
-        switch (co->m_operands[i].m_type)
-        {
-        case cot_byte:
-            values[i] = co->m_operands[i].m_byte;
-            break;
-        case cot_regbyte:
-            values[i] = regbyte_ctx_extract (&ctx, co->m_operands[i].m_regbyte);
-            break;
-        }
-
-    switch (co->m_type)
-    {
-    case ct_equal:
-        return values[0] == values[1];
-    case ct_not_equal:
-        return values[0] != values[1];
-    }
-
-    return 0;
-}
-
-static void free_regbyte_ctx_stack (regbyte_ctx *top, regbyte_ctx *limit)
-{
-    while (top != limit)
-    {
-        regbyte_ctx *rbc = top->m_prev;
-        regbyte_ctx_destroy (&top);
-        top = rbc;
-    }
-}
-
-typedef enum match_result_
-{
-    mr_not_matched,     /* the examined string does not match */
-    mr_matched,         /* the examined string matches */
-    mr_error_raised,    /* mr_not_matched + error has been raised */
-    mr_dont_emit,       /* used by identifier loops only */
-    mr_internal_error   /* an internal error has occured such as out of memory */
-} match_result;
-
-/*
- * This function does the main job. It parses the text and generates output data.
- */
-static match_result
-match (dict *di, const byte *text, int *index, rule *ru, barray **ba, int filtering_string,
-       regbyte_ctx **rbc)
-{
-   int ind = *index;
-    match_result status = mr_not_matched;
-    spec *sp = ru->m_specs;
-    regbyte_ctx *ctx = *rbc;
-
-    /* for every specifier in the rule */
-    while (sp)
-    {
-      int i, len, save_ind = ind;
-        barray *array = NULL;
-
-        if (satisfies_condition (sp->m_cond, ctx))
-        {
-            switch (sp->m_spec_type)
-            {
-            case st_identifier:
-                barray_create (&array);
-                if (array == NULL)
-                {
-                    free_regbyte_ctx_stack (ctx, *rbc);
-                    return mr_internal_error;
-                }
-
-                status = match (di, text, &ind, sp->m_rule, &array, filtering_string, &ctx);
-
-                if (status == mr_internal_error)
-                {
-                    free_regbyte_ctx_stack (ctx, *rbc);
-                    barray_destroy (&array);
-                    return mr_internal_error;
-                }
-                break;
-            case st_string:
-                len = str_length (sp->m_string);
-
-                /* prefilter the stream */
-                if (!filtering_string && di->m_string)
-                {
-                    barray *ba;
-               int filter_index = 0;
-                    match_result result;
-                    regbyte_ctx *null_ctx = NULL;
-
-                    barray_create (&ba);
-                    if (ba == NULL)
-                    {
-                        free_regbyte_ctx_stack (ctx, *rbc);
-                        return mr_internal_error;
-                    }
-
-                    result = match (di, text + ind, &filter_index, di->m_string, &ba, 1, &null_ctx);
-
-                    if (result == mr_internal_error)
-                    {
-                        free_regbyte_ctx_stack (ctx, *rbc);
-                        barray_destroy (&ba);
-                        return mr_internal_error;
-                    }
-
-                    if (result != mr_matched)
-                    {
-                        barray_destroy (&ba);
-                        status = mr_not_matched;
-                        break;
-                    }
-
-                    barray_destroy (&ba);
-
-                    if (filter_index != len || !str_equal_n (sp->m_string, text + ind, len))
-                    {
-                        status = mr_not_matched;
-                        break;
-                    }
-
-                    status = mr_matched;
-                    ind += len;
-                }
-                else
-                {
-                    status = mr_matched;
-                    for (i = 0; status == mr_matched && i < len; i++)
-                        if (text[ind + i] != sp->m_string[i])
-                            status = mr_not_matched;
-
-                    if (status == mr_matched)
-                        ind += len;
-                }
-                break;
-            case st_byte:
-                status = text[ind] == *sp->m_byte ? mr_matched : mr_not_matched;
-                if (status == mr_matched)
-                    ind++;
-                break;
-            case st_byte_range:
-                status = (text[ind] >= sp->m_byte[0] && text[ind] <= sp->m_byte[1]) ?
-                    mr_matched : mr_not_matched;
-                if (status == mr_matched)
-                    ind++;
-                break;
-            case st_true:
-                status = mr_matched;
-                break;
-            case st_false:
-                status = mr_not_matched;
-                break;
-            case st_debug:
-                status = ru->m_oper == op_and ? mr_matched : mr_not_matched;
-                break;
-            case st_identifier_loop:
-                barray_create (&array);
-                if (array == NULL)
-                {
-                    free_regbyte_ctx_stack (ctx, *rbc);
-                    return mr_internal_error;
-                }
-
-                status = mr_dont_emit;
-                for (;;)
-                {
-                    match_result result;
-
-                    save_ind = ind;
-                    result = match (di, text, &ind, sp->m_rule, &array, filtering_string, &ctx);
-
-                    if (result == mr_error_raised)
-                    {
-                        status = result;
-                        break;
-                    }
-                    else if (result == mr_matched)
-                    {
-                        if (barray_push (ba, sp->m_emits, text[ind - 1], save_ind, &ctx) ||
-                            barray_append (ba, &array))
-                        {
-                            free_regbyte_ctx_stack (ctx, *rbc);
-                            barray_destroy (&array);
-                            return mr_internal_error;
-                        }
-                        barray_destroy (&array);
-                        barray_create (&array);
-                        if (array == NULL)
-                        {
-                            free_regbyte_ctx_stack (ctx, *rbc);
-                            return mr_internal_error;
-                        }
-                    }
-                    else if (result == mr_internal_error)
-                    {
-                        free_regbyte_ctx_stack (ctx, *rbc);
-                        barray_destroy (&array);
-                        return mr_internal_error;
-                    }
-                    else
-                        break;
-                }
-                break;
-            }
-        }
-        else
-        {
-            status = mr_not_matched;
-        }
-
-        if (status == mr_error_raised)
-        {
-            free_regbyte_ctx_stack (ctx, *rbc);
-            barray_destroy (&array);
-
-            return mr_error_raised;
-        }
-
-        if (ru->m_oper == op_and && status != mr_matched && status != mr_dont_emit)
-        {
-            free_regbyte_ctx_stack (ctx, *rbc);
-            barray_destroy (&array);
-
-            if (sp->m_errtext)
-            {
-                set_last_error (sp->m_errtext->m_text, error_get_token (sp->m_errtext, di, text,
-                    ind), ind);
-
-                return mr_error_raised;
-            }
-
-            return mr_not_matched;
-        }
-
-        if (status == mr_matched)
-        {
-            if (sp->m_emits)
-                if (barray_push (ba, sp->m_emits, text[ind - 1], save_ind, &ctx))
-                {
-                    free_regbyte_ctx_stack (ctx, *rbc);
-                    barray_destroy (&array);
-                    return mr_internal_error;
-                }
-
-            if (array)
-                if (barray_append (ba, &array))
-                {
-                    free_regbyte_ctx_stack (ctx, *rbc);
-                    barray_destroy (&array);
-                    return mr_internal_error;
-                }
-        }
-
-        barray_destroy (&array);
-
-        /* if the rule operator is a logical or, we pick up the first matching specifier */
-        if (ru->m_oper == op_or && (status == mr_matched || status == mr_dont_emit))
-        {
-            *index = ind;
-            *rbc = ctx;
-            return mr_matched;
-        }
-
-        sp = sp->next;
-    }
-
-    /* everything went fine - all specifiers match up */
-    if (ru->m_oper == op_and && (status == mr_matched || status == mr_dont_emit))
-    {
-        *index = ind;
-        *rbc = ctx;
-        return mr_matched;
-    }
-
-    free_regbyte_ctx_stack (ctx, *rbc);
-    return mr_not_matched;
-}
-
-static match_result
-fast_match (dict *di, const byte *text, int *index, rule *ru, int *_PP, bytepool *_BP,
-            int filtering_string, regbyte_ctx **rbc)
-{
-   int ind = *index;
-    int _P = filtering_string ? 0 : *_PP;
-    int _P2;
-    match_result status = mr_not_matched;
-    spec *sp = ru->m_specs;
-    regbyte_ctx *ctx = *rbc;
-
-    /* for every specifier in the rule */
-    while (sp)
-    {
-      int i, len, save_ind = ind;
-
-        _P2 = _P + (sp->m_emits ? emit_size (sp->m_emits) : 0);
-        if (bytepool_reserve (_BP, _P2))
-        {
-            free_regbyte_ctx_stack (ctx, *rbc);
-            return mr_internal_error;
-        }
-
-        if (satisfies_condition (sp->m_cond, ctx))
-        {
-            switch (sp->m_spec_type)
-            {
-            case st_identifier:
-                status = fast_match (di, text, &ind, sp->m_rule, &_P2, _BP, filtering_string, &ctx);
-
-                if (status == mr_internal_error)
-                {
-                    free_regbyte_ctx_stack (ctx, *rbc);
-                    return mr_internal_error;
-                }
-                break;
-            case st_string:
-                len = str_length (sp->m_string);
-
-                /* prefilter the stream */
-                if (!filtering_string && di->m_string)
-                {
-               int filter_index = 0;
-                    match_result result;
-                    regbyte_ctx *null_ctx = NULL;
-
-                    result = fast_match (di, text + ind, &filter_index, di->m_string, NULL, _BP, 1, &null_ctx);
-
-                    if (result == mr_internal_error)
-                    {
-                        free_regbyte_ctx_stack (ctx, *rbc);
-                        return mr_internal_error;
-                    }
-
-                    if (result != mr_matched)
-                    {
-                        status = mr_not_matched;
-                        break;
-                    }
-
-                    if (filter_index != len || !str_equal_n (sp->m_string, text + ind, len))
-                    {
-                        status = mr_not_matched;
-                        break;
-                    }
-
-                    status = mr_matched;
-                    ind += len;
-                }
-                else
-                {
-                    status = mr_matched;
-                    for (i = 0; status == mr_matched && i < len; i++)
-                        if (text[ind + i] != sp->m_string[i])
-                            status = mr_not_matched;
-
-                    if (status == mr_matched)
-                        ind += len;
-                }
-                break;
-            case st_byte:
-                status = text[ind] == *sp->m_byte ? mr_matched : mr_not_matched;
-                if (status == mr_matched)
-                    ind++;
-                break;
-            case st_byte_range:
-                status = (text[ind] >= sp->m_byte[0] && text[ind] <= sp->m_byte[1]) ?
-                    mr_matched : mr_not_matched;
-                if (status == mr_matched)
-                    ind++;
-                break;
-            case st_true:
-                status = mr_matched;
-                break;
-            case st_false:
-                status = mr_not_matched;
-                break;
-            case st_debug:
-                status = ru->m_oper == op_and ? mr_matched : mr_not_matched;
-                break;
-            case st_identifier_loop:
-                status = mr_dont_emit;
-                for (;;)
-                {
-                    match_result result;
-
-                    save_ind = ind;
-                    result = fast_match (di, text, &ind, sp->m_rule, &_P2, _BP, filtering_string, &ctx);
-
-                    if (result == mr_error_raised)
-                    {
-                        status = result;
-                        break;
-                    }
-                    else if (result == mr_matched)
-                    {
-                        if (!filtering_string)
-                        {
-                            if (sp->m_emits != NULL)
-                            {
-                                if (emit_push (sp->m_emits, _BP->_F + _P, text[ind - 1], save_ind, &ctx))
-                                {
-                                    free_regbyte_ctx_stack (ctx, *rbc);
-                                    return mr_internal_error;
-                                }
-                            }
-
-                            _P = _P2;
-                            _P2 += sp->m_emits ? emit_size (sp->m_emits) : 0;
-                            if (bytepool_reserve (_BP, _P2))
-                            {
-                                free_regbyte_ctx_stack (ctx, *rbc);
-                                return mr_internal_error;
-                            }
-                        }
-                    }
-                    else if (result == mr_internal_error)
-                    {
-                        free_regbyte_ctx_stack (ctx, *rbc);
-                        return mr_internal_error;
-                    }
-                    else
-                        break;
-                }
-                break;
-            }
-        }
-        else
-        {
-            status = mr_not_matched;
-        }
-
-        if (status == mr_error_raised)
-        {
-            free_regbyte_ctx_stack (ctx, *rbc);
-
-            return mr_error_raised;
-        }
-
-        if (ru->m_oper == op_and && status != mr_matched && status != mr_dont_emit)
-        {
-            free_regbyte_ctx_stack (ctx, *rbc);
-
-            if (sp->m_errtext)
-            {
-                set_last_error (sp->m_errtext->m_text, error_get_token (sp->m_errtext, di, text,
-                    ind), ind);
-
-                return mr_error_raised;
-            }
-
-            return mr_not_matched;
-        }
-
-        if (status == mr_matched)
-        {
-            if (sp->m_emits != NULL) {
-                const byte ch = (ind <= 0) ? 0 : text[ind - 1];
-                if (emit_push (sp->m_emits, _BP->_F + _P, ch, save_ind, &ctx))
-                {
-                    free_regbyte_ctx_stack (ctx, *rbc);
-                    return mr_internal_error;
-                }
-
-           }
-           _P = _P2;
-        }
-
-        /* if the rule operator is a logical or, we pick up the first matching specifier */
-        if (ru->m_oper == op_or && (status == mr_matched || status == mr_dont_emit))
-        {
-            *index = ind;
-            *rbc = ctx;
-            if (!filtering_string)
-                *_PP = _P;
-            return mr_matched;
-        }
-
-        sp = sp->next;
-    }
-
-    /* everything went fine - all specifiers match up */
-    if (ru->m_oper == op_and && (status == mr_matched || status == mr_dont_emit))
-    {
-        *index = ind;
-        *rbc = ctx;
-        if (!filtering_string)
-            *_PP = _P;
-        return mr_matched;
-    }
-
-    free_regbyte_ctx_stack (ctx, *rbc);
-    return mr_not_matched;
-}
-
-static byte *
-error_get_token (error *er, dict *di, const byte *text, int ind)
-{
-    byte *str = NULL;
-
-    if (er->m_token)
-    {
-        barray *ba;
-      int filter_index = 0;
-        regbyte_ctx *ctx = NULL;
-
-        barray_create (&ba);
-        if (ba != NULL)
-        {
-            if (match (di, text + ind, &filter_index, er->m_token, &ba, 0, &ctx) == mr_matched &&
-                filter_index)
-            {
-                str = (byte *) mem_alloc (filter_index + 1);
-                if (str != NULL)
-                {
-                    str_copy_n (str, text + ind, filter_index);
-                    str[filter_index] = '\0';
-                }
-            }
-            barray_destroy (&ba);
-        }
-    }
-
-    return str;
-}
-
-typedef struct grammar_load_state_
-{
-    dict *di;
-    byte *syntax_symbol;
-    byte *string_symbol;
-    map_str *maps;
-    map_byte *mapb;
-    map_rule *mapr;
-} grammar_load_state;
-
-static void grammar_load_state_create (grammar_load_state **gr)
-{
-    *gr = (grammar_load_state *) mem_alloc (sizeof (grammar_load_state));
-    if (*gr)
-    {
-        (**gr).di = NULL;
-        (**gr).syntax_symbol = NULL;
-        (**gr).string_symbol = NULL;
-        (**gr).maps = NULL;
-        (**gr).mapb = NULL;
-        (**gr).mapr = NULL;
-    }
-}
-
-static void grammar_load_state_destroy (grammar_load_state **gr)
-{
-    if (*gr)
-    {
-        dict_destroy (&(**gr).di);
-        mem_free ((void **) &(**gr).syntax_symbol);
-        mem_free ((void **) &(**gr).string_symbol);
-        map_str_destroy (&(**gr).maps);
-        map_byte_destroy (&(**gr).mapb);
-        map_rule_destroy (&(**gr).mapr);
-        mem_free ((void **) gr);
-    }
-}
-
-
-static void error_msg(int line, const char *msg)
-{
-   fprintf(stderr, "Error in grammar_load_from_text() at line %d: %s\n", line, msg);
-}
-
-
-/*
-    the API
-*/
-grammar grammar_load_from_text (const byte *text)
-{
-    grammar_load_state *g = NULL;
-    grammar id = 0;
-
-    clear_last_error ();
-
-    grammar_load_state_create (&g);
-    if (g == NULL) {
-        error_msg(__LINE__, "");
-        return 0;
-    }
-
-    dict_create (&g->di);
-    if (g->di == NULL)
-    {
-        grammar_load_state_destroy (&g);
-        error_msg(__LINE__, "");
-        return 0;
-    }
-
-    eat_spaces (&text);
-
-    /* skip ".syntax" keyword */
-    text += 7;
-    eat_spaces (&text);
-
-    /* retrieve root symbol */
-    if (get_identifier (&text, &g->syntax_symbol))
-    {
-        grammar_load_state_destroy (&g);
-        error_msg(__LINE__, "");
-        return 0;
-    }
-    eat_spaces (&text);
-
-    /* skip semicolon */
-    text++;
-    eat_spaces (&text);
-
-    while (*text)
-    {
-        byte *symbol = NULL;
-        int is_dot = *text == '.';
-
-        if (is_dot)
-            text++;
-
-        if (get_identifier (&text, &symbol))
-        {
-            grammar_load_state_destroy (&g);
-            error_msg(__LINE__, "");
-            return 0;
-        }
-        eat_spaces (&text);
-
-        /* .emtcode */
-        if (is_dot && str_equal (symbol, (byte *) "emtcode"))
-        {
-            map_byte *ma = NULL;
-
-            mem_free ((void **) (void *) &symbol);
-
-            if (get_emtcode (&text, &ma))
-            {
-                grammar_load_state_destroy (&g);
-                error_msg(__LINE__, "");
-                return 0;
-            }
-
-            map_byte_append (&g->mapb, ma);
-        }
-        /* .regbyte */
-        else if (is_dot && str_equal (symbol, (byte *) "regbyte"))
-        {
-            map_byte *ma = NULL;
-
-            mem_free ((void **) (void *) &symbol);
-
-            if (get_regbyte (&text, &ma))
-            {
-                grammar_load_state_destroy (&g);
-                error_msg(__LINE__, "");
-                return 0;
-            }
-
-            map_byte_append (&g->di->m_regbytes, ma);
-        }
-        /* .errtext */
-        else if (is_dot && str_equal (symbol, (byte *) "errtext"))
-        {
-            map_str *ma = NULL;
-
-            mem_free ((void **) (void *) &symbol);
-
-            if (get_errtext (&text, &ma))
-            {
-                grammar_load_state_destroy (&g);
-                error_msg(__LINE__, "");
-                return 0;
-            }
-
-            map_str_append (&g->maps, ma);
-        }
-        /* .string */
-        else if (is_dot && str_equal (symbol, (byte *) "string"))
-        {
-            mem_free ((void **) (void *) &symbol);
-
-            if (g->di->m_string != NULL)
-            {
-                grammar_load_state_destroy (&g);
-                error_msg(__LINE__, "");
-                return 0;
-            }
-
-            if (get_identifier (&text, &g->string_symbol))
-            {
-                grammar_load_state_destroy (&g);
-                error_msg(__LINE__, "");
-                return 0;
-            }
-
-            /* skip semicolon */
-            eat_spaces (&text);
-            text++;
-            eat_spaces (&text);
-        }
-        else
-        {
-            rule *ru = NULL;
-            map_rule *ma = NULL;
-
-            if (get_rule (&text, &ru, g->maps, g->mapb))
-            {
-                grammar_load_state_destroy (&g);
-                error_msg(__LINE__, "");
-                return 0;
-            }
-
-            rule_append (&g->di->m_rulez, ru);
-
-            /* if a rule consist of only one specifier, give it an ".and" operator */
-            if (ru->m_oper == op_none)
-                ru->m_oper = op_and;
-
-            map_rule_create (&ma);
-            if (ma == NULL)
-            {
-                grammar_load_state_destroy (&g);
-                error_msg(__LINE__, "");
-                return 0;
-            }
-
-            ma->key = symbol;
-            ma->data = ru;
-            map_rule_append (&g->mapr, ma);
-        }
-    }
-
-    if (update_dependencies (g->di, g->mapr, &g->syntax_symbol, &g->string_symbol,
-        g->di->m_regbytes))
-    {
-        grammar_load_state_destroy (&g);
-        error_msg(__LINE__, "update_dependencies() failed");
-        return 0;
-    }
-
-    dict_append (&g_dicts, g->di);
-    id = g->di->m_id;
-    g->di = NULL;
-
-    grammar_load_state_destroy (&g);
-
-    return id;
-}
-
-int grammar_set_reg8 (grammar id, const byte *name, byte value)
-{
-    dict *di = NULL;
-    map_byte *reg = NULL;
-
-    clear_last_error ();
-
-    dict_find (&g_dicts, id, &di);
-    if (di == NULL)
-    {
-        set_last_error (INVALID_GRAMMAR_ID, NULL, -1);
-        return 0;
-    }
-
-    reg = map_byte_locate (&di->m_regbytes, name);
-    if (reg == NULL)
-    {
-        set_last_error (INVALID_REGISTER_NAME, str_duplicate (name), -1);
-        return 0;
-    }
-
-    reg->data = value;
-    return 1;
-}
-
-/*
-    internal checking function used by both grammar_check and grammar_fast_check functions
-*/
-static int _grammar_check (grammar id, const byte *text, byte **prod, unsigned int *size,
-    unsigned int estimate_prod_size, int use_fast_path)
-{
-    dict *di = NULL;
-   int index = 0;
-
-    clear_last_error ();
-
-    dict_find (&g_dicts, id, &di);
-    if (di == NULL)
-    {
-        set_last_error (INVALID_GRAMMAR_ID, NULL, -1);
-        return 0;
-    }
-
-    *prod = NULL;
-    *size = 0;
-
-    if (use_fast_path)
-    {
-        regbyte_ctx *rbc = NULL;
-        bytepool *bp = NULL;
-        int _P = 0;
-
-        bytepool_create (&bp, estimate_prod_size);
-        if (bp == NULL)
-            return 0;
-
-        if (fast_match (di, text, &index, di->m_syntax, &_P, bp, 0, &rbc) != mr_matched)
-        {
-            bytepool_destroy (&bp);
-            free_regbyte_ctx_stack (rbc, NULL);
-            return 0;
-        }
-
-        free_regbyte_ctx_stack (rbc, NULL);
-
-        *prod = bp->_F;
-        *size = _P;
-        bp->_F = NULL;
-        bytepool_destroy (&bp);
-    }
-    else
-    {
-        regbyte_ctx *rbc = NULL;
-        barray *ba = NULL;
-
-        barray_create (&ba);
-        if (ba == NULL)
-            return 0;
-
-        if (match (di, text, &index, di->m_syntax, &ba, 0, &rbc) != mr_matched)
-        {
-            barray_destroy (&ba);
-            free_regbyte_ctx_stack (rbc, NULL);
-            return 0;
-        }
-
-        free_regbyte_ctx_stack (rbc, NULL);
-
-        *prod = (byte *) mem_alloc (ba->len * sizeof (byte));
-        if (*prod == NULL)
-        {
-            barray_destroy (&ba);
-            return 0;
-        }
-
-        mem_copy (*prod, ba->data, ba->len * sizeof (byte));
-        *size = ba->len;
-        barray_destroy (&ba);
-    }
-
-    return 1;
-}
-
-int grammar_check (grammar id, const byte *text, byte **prod, unsigned int *size)
-{
-    return _grammar_check (id, text, prod, size, 0, 0);
-}
-
-int grammar_fast_check (grammar id, const byte *text, byte **prod, unsigned int *size,
-    unsigned int estimate_prod_size)
-{
-    return _grammar_check (id, text, prod, size, estimate_prod_size, 1);
-}
-
-int grammar_destroy (grammar id)
-{
-    dict **di = &g_dicts;
-
-    clear_last_error ();
-
-    while (*di != NULL)
-    {
-        if ((**di).m_id == id)
-        {
-            dict *tmp = *di;
-            *di = (**di).next;
-            dict_destroy (&tmp);
-            return 1;
-        }
-
-        di = &(**di).next;
-    }
-
-    set_last_error (INVALID_GRAMMAR_ID, NULL, -1);
-    return 0;
-}
-
-static void append_character (const char x, byte *text, int *dots_made, int *len, int size)
-{
-    if (*dots_made == 0)
-    {
-        if (*len < size - 1)
-        {
-            text[(*len)++] = x;
-            text[*len] = '\0';
-        }
-        else
-        {
-            int i;
-            for (i = 0; i < 3; i++)
-                if (--(*len) >= 0)
-                    text[*len] = '.';
-            *dots_made = 1;
-        }
-    }
-}
-
-void grammar_get_last_error (byte *text, unsigned int size, int *pos)
-{
-    int len = 0, dots_made = 0;
-    const byte *p = error_message;
-
-    *text = '\0';
-
-    if (p)
-    {
-        while (*p)
-        {
-            if (*p == '$')
-            {
-                const byte *r = error_param;
-
-                while (*r)
-                {
-                    append_character (*r++, text, &dots_made, &len, (int) size);
-                }
-
-                p++;
-            }
-            else
-            {
-                append_character (*p++, text, &dots_made, &len, size);
-            }
-        }
-    }
-
-    *pos = error_position;
-}
diff --git a/src/mesa/shader/grammar/grammar.h b/src/mesa/shader/grammar/grammar.h
deleted file mode 100644 (file)
index 591e38a..0000000
+++ /dev/null
@@ -1,103 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  6.2
- *
- * Copyright (C) 1999-2004  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-#ifndef GRAMMAR_H
-#define GRAMMAR_H
-
-
-#ifndef GRAMMAR_PORT_INCLUDE
-#error Do not include this file directly, include your grammar_XXX.h instead
-#endif
-
-
-#ifdef __cplusplus
-extern "C" {
-#endif
-
-void grammar_alloc_free (void *);
-void *grammar_alloc_malloc (size_t);
-void *grammar_alloc_realloc (void *, size_t, size_t);
-void *grammar_memory_copy (void *, const void *, size_t);
-int grammar_string_compare (const byte *, const byte *);
-int grammar_string_compare_n (const byte *, const byte *, size_t);
-byte *grammar_string_copy (byte *, const byte *);
-byte *grammar_string_copy_n (byte *, const byte *, size_t);
-byte *grammar_string_duplicate (const byte *);
-unsigned int grammar_string_length (const byte *);
-
-/*
-    loads grammar script from null-terminated ASCII <text>
-    returns unique grammar id to grammar object
-    returns 0 if an error occurs (call grammar_get_last_error to retrieve the error text)
-*/
-grammar grammar_load_from_text (const byte *text);
-
-/*
-    sets a new <value> to a register <name> for grammar <id>
-    returns 0 on error (call grammar_get_last_error to retrieve the error text)
-    returns 1 on success
-*/
-int grammar_set_reg8 (grammar id, const byte *name, byte value);
-
-/*
-    this function is obsolete, use only for debugging purposes
-
-    checks if a null-terminated <text> matches given grammar <id>
-    returns 0 on error (call grammar_get_last_error to retrieve the error text)
-    returns 1 on success, the <prod> points to newly allocated buffer with production and <size>
-    is filled with the production size
-    call grammar_alloc_free to free the memory block pointed by <prod>
-*/
-int grammar_check (grammar id, const byte *text, byte **prod, unsigned int *size);
-
-/*
-    does the same what grammar_check does but much more (approx. 4 times) faster
-    use this function instead of grammar_check
-    <estimate_prod_size> is a hint - the initial production buffer size will be of this size,
-    but if more room is needed it will be safely resized; set it to 0x1000 or so
-*/
-int grammar_fast_check (grammar id, const byte *text, byte **prod, unsigned int *size,
-    unsigned int estimate_prod_size);
-
-/*
-    destroys grammar object identified by <id>
-    returns 0 on error (call grammar_get_last_error to retrieve the error text)
-    returns 1 on success
-*/
-int grammar_destroy (grammar id);
-
-/*
-    retrieves last grammar error reported either by grammar_load_from_text, grammar_check
-    or grammar_destroy
-    the user allocated <text> buffer receives error description, <pos> points to error position,
-    <size> is the size of the text buffer to fill in - it must be at least 4 bytes long,
-*/
-void grammar_get_last_error (byte *text, unsigned int size, int *pos);
-
-#ifdef __cplusplus
-}
-#endif
-
-#endif
-
diff --git a/src/mesa/shader/grammar/grammar.syn b/src/mesa/shader/grammar/grammar.syn
deleted file mode 100644 (file)
index 5d99f65..0000000
+++ /dev/null
@@ -1,567 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  6.2
- *
- * Copyright (C) 1999-2004  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
- /**
- * \file grammar.syn
- * syntax of .syn script - used to validate other syntax files
- * \author Michal Krol
- */
-.syntax grammar;
-
-/* declaration */
-.emtcode DECLARATION_END                            0
-.emtcode DECLARATION_EMITCODE                       1
-.emtcode DECLARATION_ERRORTEXT                      2
-.emtcode DECLARATION_REGBYTE                        3
-.emtcode DECLARATION_LEXER                          4
-.emtcode DECLARATION_RULE                           5
-
-/* specifier */
-.emtcode SPECIFIER_END                              0
-.emtcode SPECIFIER_AND_TAG                          1
-.emtcode SPECIFIER_OR_TAG                           2
-.emtcode SPECIFIER_CHARACTER_RANGE                  3
-.emtcode SPECIFIER_CHARACTER                        4
-.emtcode SPECIFIER_STRING                           5
-.emtcode SPECIFIER_IDENTIFIER                       6
-.emtcode SPECIFIER_TRUE                             7
-.emtcode SPECIFIER_FALSE                            8
-.emtcode SPECIFIER_DEBUG                            9
-
-/* identifier */
-.emtcode IDENTIFIER_SIMPLE                          0
-.emtcode IDENTIFIER_LOOP                            1
-
-/* error */
-.emtcode ERROR_NOT_PRESENT                          0
-.emtcode ERROR_PRESENT                              1
-
-/* emit */
-.emtcode EMIT_NULL                                  0
-.emtcode EMIT_INTEGER                               1
-.emtcode EMIT_IDENTIFIER                            2
-.emtcode EMIT_CHARACTER                             3
-.emtcode EMIT_LAST_CHARACTER                        4
-.emtcode EMIT_CURRENT_POSITION                      5
-
-.errtext INVALID_GRAMMAR                            "internal error 2001: invalid grammar script"
-.errtext SYNTAX_EXPECTED                            "internal error 2002: '.syntax' keyword expected"
-.errtext IDENTIFIER_EXPECTED                        "internal error 2003: identifier expected"
-.errtext MISSING_SEMICOLON                          "internal error 2004: missing ';'"
-.errtext INTEGER_EXPECTED                           "internal error 2005: integer value expected"
-.errtext STRING_EXPECTED                            "internal error 2006: string expected"
-
-/*
-    <grammar>                           ::= ".syntax" <identifier> ";" <declaration_list>
-*/
-grammar
-    grammar_1 .error INVALID_GRAMMAR;
-grammar_1
-    optional_space .and ".syntax" .error SYNTAX_EXPECTED .and space .and identifier .and
-    semicolon .and declaration_list .and optional_space .and '\0' .emit DECLARATION_END;
-
-/*
-    <optional_space>                    ::= <space>
-                                          | ""
-*/
-optional_space
-    space .or .true;
-
-/*
-    <space>                             ::= <single_space> <single_space>*
-*/
-space
-    single_space .and .loop single_space;
-
-/*
-    <single_space>                      ::= <white_char>
-                                          | <comment_block>
-*/
-single_space
-    white_char .or comment_block;
-
-/*
-    <white_char>                        ::= " " 
-                                          | "\t"
-                                          | "\n"
-                                          | "\r"
-*/
-white_char
-    ' ' .or '\t' .or '\n' .or '\r';
-
-/*
-    <comment_block>                     ::= "/" "*" <comment_rest>
-*/
-comment_block
-    '/' .and '*' .and comment_rest;
-
-/*
-    <comment_rest>                      ::= <comment_char_no_star>* <comment_end>
-                                          | <comment_char_no_star>* "*" <comment_rest>
-*/
-comment_rest
-    .loop comment_char_no_star .and comment_rest_1;
-comment_rest_1
-    comment_end .or comment_rest_2;
-comment_rest_2
-    '*' .and comment_rest;
-
-/*
-    <comment_char_no_star>              ::= All ASCII characters except "*" and "\0"
-*/
-comment_char_no_star
-    '\x2B'-'\xFF' .or '\x01'-'\x29';
-
-/*
-    <comment_end>                       ::= "*" "/"
-*/
-comment_end
-    '*' .and '/';
-
-/*
-    <identifier>                        ::= <identifier>
-*/
-identifier
-    identifier_ne .error IDENTIFIER_EXPECTED;
-
-/*
-    <identifier_ne>                     ::= <first_idchar> <follow_idchar>*
-*/
-identifier_ne
-    first_idchar .emit * .and .loop follow_idchar .emit * .and .true .emit '\0';
-
-/*
-    <first_idchar>                      ::= "a"-"z"
-                                          | "A"-"Z"
-                                          | "_"
-*/
-first_idchar
-    'a'-'z' .or 'A'-'Z' .or '_';
-
-/*
-    <follow_idchar>                     ::= <first_idchar>
-                                          | <digit_dec>
-*/
-follow_idchar
-    first_idchar .or digit_dec;
-
-/*
-    <digit_dec>                         ::= "0"-"9"
-*/
-digit_dec
-    '0'-'9';
-
-/*
-    <semicolon>                         ::= ";"
-*/
-semicolon
-    optional_space .and ';' .error MISSING_SEMICOLON .and optional_space;
-
-/*
-    <declaration_list>                  ::= <declaration>
-                                          | <declaration_list> <declaration>
-*/
-declaration_list
-    declaration .and .loop declaration;
-
-/*
-    <declaration>                       ::= <emitcode_definition>
-                                          | <errortext_definition>
-                                          | <lexer_definition>
-                                          | <rule_definition>
-*/
-declaration
-    emitcode_definition .emit DECLARATION_EMITCODE .or
-    errortext_definition .emit DECLARATION_ERRORTEXT .or
-    regbyte_definition .emit DECLARATION_REGBYTE .or
-    lexer_definition .emit DECLARATION_LEXER .or
-    rule_definition .emit DECLARATION_RULE;
-
-/*
-    <emitcode_definition>               ::= ".emtcode" <identifier> <integer>
-*/
-emitcode_definition
-    ".emtcode" .and space .and identifier .and space .and integer .and space_or_null;
-
-/*
-    <integer>                           ::= <integer_ne>
-*/
-integer
-    integer_ne .error INTEGER_EXPECTED;
-
-/*
-    <integer_ne>                        ::= <hex_integer>
-                                          | <dec_integer>
-*/
-integer_ne
-    hex_integer .emit 0x10 .or dec_integer .emit 10;
-
-/*
-    <hex_integer>                       :: <hex_prefix> <digit_hex> <digit_hex>*
-*/
-hex_integer
-    hex_prefix .and digit_hex .emit * .and .loop digit_hex .emit * .and .true .emit '\0';
-
-/*
-    <hex_prefix>                        ::= "0x"
-                                          | "0X"
-*/
-hex_prefix
-    '0' .and hex_prefix_1;
-hex_prefix_1
-    'x' .or 'X';
-
-/*
-    <digit_hex>                         ::= "0"-"9"
-                                          | "a"-"f"
-                                          | "A"-"F"
-*/
-digit_hex
-    '0'-'9' .or 'a'-'f' .or 'A'-'F';
-
-/*
-    <dec_integer>                       :: <digit_dec> <digit_dec>*
-*/
-dec_integer
-    digit_dec .emit * .and .loop digit_dec .emit * .and .true .emit '\0';
-
-/*
-    <space_or_null>                     ::= <space>
-                                          | "\0"
-*/
-space_or_null
-    space .or '\0';
-
-/*
-    <errortext_definition>              ::= ".errtext" <identifier> <string>
-*/
-errortext_definition
-    ".errtext" .and space .and identifier .and space .and string .and space_or_null;
-
-/*
-    <string>                            ::= <string_ne>
-*/
-string
-    string_ne .error STRING_EXPECTED;
-
-/*
-    <string_ne>                         ::= "\"" <string_char_double_quotes> "\""
-*/
-string_ne
-    '"' .and .loop string_char_double_quotes .and '"' .emit '\0';
-
-/*
-    <string_char_double_quotes>         ::= <escape_sequence>
-                                          | <string_char>
-                                          | "\'"
-*/
-string_char_double_quotes
-    escape_sequence .or string_char .emit * .or '\'' .emit *;
-
-/*
-    <string_char>                       ::= All ASCII characters except "\'", "\"", "\n", "\r",
-                                            "\0" and "\\"
-*/
-string_char
-    '\x5D'-'\xFF' .or '\x28'-'\x5B' .or '\x23'-'\x26' .or '\x0E'-'\x21' .or '\x0B'-'\x0C' .or
-    '\x01'-'\x09';
-
-/*
-    <escape_sequence>                   ::= "\\" <escape_code>
-*/
-escape_sequence
-    '\\' .emit * .and escape_code;
-
-/*
-    <escape_code>                       ::= <simple_escape_code>
-                                          | <hex_escape_code>
-                                          | <oct_escape_code>
-*/
-escape_code
-    simple_escape_code .emit * .or hex_escape_code .or oct_escape_code;
-
-/*
-    <simple_escape_code>                ::= "\'"
-                                          | "\""
-                                          | "?"
-                                          | "\\"
-                                          | "a"
-                                          | "b"
-                                          | "f"
-                                          | "n"
-                                          | "r"
-                                          | "t"
-                                          | "v"
-*/
-simple_escape_code
-    '\'' .or '"' .or '?' .or '\\' .or 'a' .or 'b' .or 'f' .or 'n' .or 'r' .or 't' .or 'v';
-
-/*
-    <hex_escape_code>                   ::= "x" <digit_hex> <digit_hex>*
-*/
-hex_escape_code
-    'x' .emit * .and digit_hex .emit * .and .loop digit_hex .emit *;
-
-/*
-    <oct_escape_code>                   ::= <digit_oct> <optional_digit_oct> <optional_digit_oct>
-*/
-oct_escape_code
-    digit_oct .emit * .and optional_digit_oct .and optional_digit_oct;
-
-/*
-    <digit_oct>                         ::= "0"-"7"
-*/
-digit_oct
-    '0'-'7';
-
-/*
-    <optional_digit_oct>                ::= <digit_oct>
-                                          | ""
-*/
-optional_digit_oct
-    digit_oct .emit * .or .true;
-
-/*
-    <regbyte_definition>                ::= ".regbyte" <identifier> <integer>
-*/
-regbyte_definition
-    ".regbyte" .and space .and identifier .and space .and integer .and space_or_null;
-
-/*
-    <lexer_definition>                  ::= ".string" <identifier> ";"
-*/
-lexer_definition
-    ".string" .and space .and identifier .and semicolon;
-
-/*
-    <rule_definition>                   ::= <identifier_ne> <definition>
-*/
-rule_definition
-    identifier_ne .and space .and definition;
-
-/*
-    <definition>                        ::= <specifier> <optional_specifiers_and_or> ";"
-*/
-definition
-    specifier .and optional_specifiers_and_or .and semicolon .emit SPECIFIER_END;
-
-/*
-    <optional_specifiers_and_or>        ::= <and_specifiers>
-                                          | <or_specifiers>
-                                          | ""
-*/
-optional_specifiers_and_or
-    and_specifiers .emit SPECIFIER_AND_TAG .or or_specifiers .emit SPECIFIER_OR_TAG .or .true;
-
-/*
-    <specifier>                         ::= <specifier_condition> <specifier_rule>
-*/
-specifier
-    specifier_condition .and optional_space .and specifier_rule;
-
-/*
-    <specifier_condition>               ::= ".if" "(" <left_operand> <operator> <right_operand> ")"
-*/
-specifier_condition
-    specifier_condition_1 .or .true;
-specifier_condition_1
-    ".if" .and optional_space .and '(' .and optional_space .and left_operand .and operator .and
-    right_operand .and optional_space .and ')';
-
-/*
-    <left_operand>                      ::= <identifier>
-*/
-left_operand
-    identifier;
-
-/*
-    <operator>                          ::= "!="
-                                          | "=="
-*/
-operator
-    operator_1 .or operator_2;
-operator_1
-    optional_space .and '!' .and '=' .and optional_space;
-operator_2
-    optional_space .and '=' .and '=' .and optional_space;
-
-/*
-    <right_operand>                     ::= <integer>
-*/
-right_operand
-    integer;
-
-/*
-    <specifier_rule>                    ::= <character_range> <optional_error> <emit>*
-                                          | <character> <optional_error> <emit>*
-                                          | <string> <optional_error> <emit>*
-                                          | ".true" <optional_error> <emit>*
-                                          | ".false" <optional_error> <emit>*
-                                          | ".debug" <optional_error> <emit>*
-                                          | <advanced_identifier> <optional_error> <emit>*
-*/
-specifier_rule
-    specifier_rule_1 .and optional_error .and .loop emit .and .true .emit EMIT_NULL;
-specifier_rule_1
-    character_range .emit SPECIFIER_CHARACTER_RANGE .or
-    character .emit SPECIFIER_CHARACTER .or
-    string_ne .emit SPECIFIER_STRING .or
-    ".true" .emit SPECIFIER_TRUE .or
-    ".false" .emit SPECIFIER_FALSE .or
-    ".debug" .emit SPECIFIER_DEBUG .or
-    advanced_identifier .emit SPECIFIER_IDENTIFIER;
-
-/*
-    <character>                         ::= "\'" <string_char_single_quotes "\'"
-*/
-character
-    '\'' .and string_char_single_quotes .and '\'' .emit '\0';
-
-/*
-    <string_char_single_quotes>         ::= <escape_sequence>
-                                          | <string_char>
-                                          | "\""
-*/
-string_char_single_quotes
-    escape_sequence .or string_char .emit * .or '"' .emit *;
-
-/*
-    <character_range>                   ::= <character> "-" <character>
-*/
-character_range
-    character .and optional_space .and '-' .and optional_space .and character;
-
-/*
-    <advanced_identifier>               ::= <optional_loop> <identifier>
-*/
-advanced_identifier
-    optional_loop .and identifier;
-
-/*
-    <optional_loop>                     ::= ".loop"
-                                          | ""
-*/
-optional_loop
-    optional_loop_1 .emit IDENTIFIER_LOOP .or .true .emit IDENTIFIER_SIMPLE;
-optional_loop_1
-    ".loop" .and space;
-
-/*
-    <optional_error>                    ::= <error>
-                                          | ""
-*/
-optional_error
-    error .emit ERROR_PRESENT .or .true .emit ERROR_NOT_PRESENT;
-
-/*
-    <error>                             :: ".error" <identifier>
-*/
-error
-    space .and ".error" .and space .and identifier;
-
-/*
-    <emit>                              ::= <emit_output>
-                                          | <emit_regbyte>
-*/
-emit
-    emit_output .or emit_regbyte;
-
-/*
-    <emit_output>                       ::= ".emit" <emit_param>
-*/
-emit_output
-    space .and ".emit" .and space .and emit_param;
-
-/*
-    <emit_param>                        ::= <integer>
-                                          | <identifier>
-                                          | <character>
-                                          | "*"
-                                          | "$"
-*/
-emit_param
-    integer_ne .emit EMIT_INTEGER .or
-    identifier_ne .emit EMIT_IDENTIFIER .or
-    character .emit EMIT_CHARACTER .or
-    '*' .emit EMIT_LAST_CHARACTER .or
-    '$' .emit EMIT_CURRENT_POSITION;
-
-/*
-    <emit_regbyte>                      ::= ".load" <identifier> <emit_param>
-*/
-emit_regbyte
-    space .and ".load" .and space .and identifier .and space .and emit_param;
-
-/*
-    <and_specifiers>                    ::= <and_specifier> <and_specifier>*
-*/
-and_specifiers
-    and_specifier .and .loop and_specifier;
-
-/*
-    <or_specifiers>                     ::= <or_specifier> <or_specifier>*
-*/
-or_specifiers
-    or_specifier .and .loop or_specifier;
-
-/*
-    <and_specifier>                     ::= ".and" <specifier>
-*/
-and_specifier
-    space .and ".and" .and space .and specifier;
-
-/*
-    <or_specifier>                      ::= ".or" <specifier>
-*/
-or_specifier
-    space .and ".or" .and space .and specifier;
-
-
-.string __string_filter;
-
-/*
-    <__string_filter>                   ::= <__first_identifier_char> <__next_identifier_char>*
-*/
-__string_filter
-    __first_identifier_char .and .loop __next_identifier_char;
-
-/*
-    <__first_identifier_char>           ::= "a"-"z"
-                                          | "A"-"Z"
-                                          | "_"
-                                          | "."
-*/
-__first_identifier_char
-    'a'-'z' .or 'A'-'Z' .or '_' .or '.';
-
-/*
-    <__next_identifier_char>            ::= "a"-"z"
-                                          | "A"-"Z"
-                                          | "_"
-                                          | "0"-"9"
-*/
-__next_identifier_char
-    'a'-'z' .or 'A'-'Z' .or '_' .or '0'-'9';
-
diff --git a/src/mesa/shader/grammar/grammar_crt.c b/src/mesa/shader/grammar/grammar_crt.c
deleted file mode 100644 (file)
index d2c95d1..0000000
+++ /dev/null
@@ -1,64 +0,0 @@
-#include "grammar_crt.h"
-
-#define GRAMMAR_PORT_BUILD 1
-#include "grammar.c"
-#undef GRAMMAR_PORT_BUILD
-
-
-void grammar_alloc_free (void *ptr)
-{
-    free (ptr);
-}
-
-void *grammar_alloc_malloc (size_t size)
-{
-    return malloc (size);
-}
-
-void *grammar_alloc_realloc (void *ptr, size_t old_size, size_t size)
-{
-    return realloc (ptr, size);
-}
-
-void *grammar_memory_copy (void *dst, const void * src, size_t size)
-{
-    return memcpy (dst, src, size);
-}
-
-int grammar_string_compare (const byte *str1, const byte *str2)
-{
-    return strcmp ((const char *) str1, (const char *) str2);
-}
-
-int grammar_string_compare_n (const byte *str1, const byte *str2, size_t n)
-{
-    return strncmp ((const char *) str1, (const char *) str2, n);
-}
-
-byte *grammar_string_copy (byte *dst, const byte *src)
-{
-    return (byte *) strcpy ((char *) dst, (const char *) src);
-}
-
-byte *grammar_string_copy_n (byte *dst, const byte *src, size_t n)
-{
-    return (byte *) strncpy ((char *) dst, (const char *) src, n);
-}
-
-unsigned int grammar_string_length (const byte *str)
-{
-    return strlen ((const char *) str);
-}
-
-byte *grammar_string_duplicate (const byte *src)
-{
-    const unsigned int size = grammar_string_length (src);
-    byte *str = grammar_alloc_malloc (size + 1);
-    if (str != NULL)
-    {
-        grammar_memory_copy (str, src, size);
-        str[size] = '\0';
-    }
-    return str;
-}
-
diff --git a/src/mesa/shader/grammar/grammar_crt.h b/src/mesa/shader/grammar/grammar_crt.h
deleted file mode 100644 (file)
index 492711e..0000000
+++ /dev/null
@@ -1,20 +0,0 @@
-#ifndef GRAMMAR_CRT_H
-#define GRAMMAR_CRT_H
-
-
-#include <stdlib.h>
-#include <malloc.h>
-#include <string.h>
-
-
-typedef unsigned long grammar;
-typedef unsigned char byte;
-
-
-#define GRAMMAR_PORT_INCLUDE 1
-#include "grammar.h"
-#undef GRAMMAR_PORT_INCLUDE
-
-
-#endif
-
diff --git a/src/mesa/shader/grammar/grammar_mesa.c b/src/mesa/shader/grammar/grammar_mesa.c
deleted file mode 100644 (file)
index eb96250..0000000
+++ /dev/null
@@ -1,87 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  6.1
- *
- * Copyright (C) 1999-2004  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/**
- * \file grammar_mesa.c
- * mesa3d port to syntax parsing engine
- * \author Michal Krol
- */
-
-#include "grammar_mesa.h"
-
-#define GRAMMAR_PORT_BUILD 1
-#include "grammar.c"
-#undef GRAMMAR_PORT_BUILD
-
-
-void grammar_alloc_free (void *ptr)
-{
-    _mesa_free (ptr);
-}
-
-void *grammar_alloc_malloc (size_t size)
-{
-    return _mesa_malloc (size);
-}
-
-void *grammar_alloc_realloc (void *ptr, size_t old_size, size_t size)
-{
-    return _mesa_realloc (ptr, old_size, size);
-}
-
-void *grammar_memory_copy (void *dst, const void * src, size_t size)
-{
-    return _mesa_memcpy (dst, src, size);
-}
-
-int grammar_string_compare (const byte *str1, const byte *str2)
-{
-    return _mesa_strcmp ((const char *) str1, (const char *) str2);
-}
-
-int grammar_string_compare_n (const byte *str1, const byte *str2, size_t n)
-{
-    return _mesa_strncmp ((const char *) str1, (const char *) str2, n);
-}
-
-byte *grammar_string_copy (byte *dst, const byte *src)
-{
-    return (byte *) _mesa_strcpy ((char *) dst, (const char *) src);
-}
-
-byte *grammar_string_copy_n (byte *dst, const byte *src, size_t n)
-{
-    return (byte *) _mesa_strncpy ((char *) dst, (const char *) src, n);
-}
-
-byte *grammar_string_duplicate (const byte *src)
-{
-    return (byte *) _mesa_strdup ((const char *) src);
-}
-
-unsigned int grammar_string_length (const byte *str)
-{
-    return (unsigned int)_mesa_strlen ((const char *) str);
-}
-
diff --git a/src/mesa/shader/grammar/grammar_mesa.h b/src/mesa/shader/grammar/grammar_mesa.h
deleted file mode 100644 (file)
index 6c92c58..0000000
+++ /dev/null
@@ -1,43 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  6.1
- *
- * Copyright (C) 1999-2004  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-#ifndef GRAMMAR_MESA_H
-#define GRAMMAR_MESA_H
-
-
-#include "main/imports.h"
-/* NOTE: include Mesa 3-D specific headers here */
-
-
-typedef GLuint grammar;
-typedef GLubyte byte;
-
-
-#define GRAMMAR_PORT_INCLUDE 1
-#include "grammar.h"
-#undef GRAMMAR_PORT_INCLUDE
-
-
-#endif
-
diff --git a/src/mesa/shader/grammar/grammar_syn.h b/src/mesa/shader/grammar/grammar_syn.h
deleted file mode 100644 (file)
index 840a1ab..0000000
+++ /dev/null
@@ -1,202 +0,0 @@
-".syntax grammar;\n"
-".emtcode DECLARATION_END 0\n"
-".emtcode DECLARATION_EMITCODE 1\n"
-".emtcode DECLARATION_ERRORTEXT 2\n"
-".emtcode DECLARATION_REGBYTE 3\n"
-".emtcode DECLARATION_LEXER 4\n"
-".emtcode DECLARATION_RULE 5\n"
-".emtcode SPECIFIER_END 0\n"
-".emtcode SPECIFIER_AND_TAG 1\n"
-".emtcode SPECIFIER_OR_TAG 2\n"
-".emtcode SPECIFIER_CHARACTER_RANGE 3\n"
-".emtcode SPECIFIER_CHARACTER 4\n"
-".emtcode SPECIFIER_STRING 5\n"
-".emtcode SPECIFIER_IDENTIFIER 6\n"
-".emtcode SPECIFIER_TRUE 7\n"
-".emtcode SPECIFIER_FALSE 8\n"
-".emtcode SPECIFIER_DEBUG 9\n"
-".emtcode IDENTIFIER_SIMPLE 0\n"
-".emtcode IDENTIFIER_LOOP 1\n"
-".emtcode ERROR_NOT_PRESENT 0\n"
-".emtcode ERROR_PRESENT 1\n"
-".emtcode EMIT_NULL 0\n"
-".emtcode EMIT_INTEGER 1\n"
-".emtcode EMIT_IDENTIFIER 2\n"
-".emtcode EMIT_CHARACTER 3\n"
-".emtcode EMIT_LAST_CHARACTER 4\n"
-".emtcode EMIT_CURRENT_POSITION 5\n"
-".errtext INVALID_GRAMMAR \"internal error 2001: invalid grammar script\"\n"
-".errtext SYNTAX_EXPECTED \"internal error 2002: '.syntax' keyword expected\"\n"
-".errtext IDENTIFIER_EXPECTED \"internal error 2003: identifier expected\"\n"
-".errtext MISSING_SEMICOLON \"internal error 2004: missing ';'\"\n"
-".errtext INTEGER_EXPECTED \"internal error 2005: integer value expected\"\n"
-".errtext STRING_EXPECTED \"internal error 2006: string expected\"\n"
-"grammar\n"
-" grammar_1 .error INVALID_GRAMMAR;\n"
-"grammar_1\n"
-" optional_space .and \".syntax\" .error SYNTAX_EXPECTED .and space .and identifier .and\n"
-" semicolon .and declaration_list .and optional_space .and '\\0' .emit DECLARATION_END;\n"
-"optional_space\n"
-" space .or .true;\n"
-"space\n"
-" single_space .and .loop single_space;\n"
-"single_space\n"
-" white_char .or comment_block;\n"
-"white_char\n"
-" ' ' .or '\\t' .or '\\n' .or '\\r';\n"
-"comment_block\n"
-" '/' .and '*' .and comment_rest;\n"
-"comment_rest\n"
-" .loop comment_char_no_star .and comment_rest_1;\n"
-"comment_rest_1\n"
-" comment_end .or comment_rest_2;\n"
-"comment_rest_2\n"
-" '*' .and comment_rest;\n"
-"comment_char_no_star\n"
-" '\\x2B'-'\\xFF' .or '\\x01'-'\\x29';\n"
-"comment_end\n"
-" '*' .and '/';\n"
-"identifier\n"
-" identifier_ne .error IDENTIFIER_EXPECTED;\n"
-"identifier_ne\n"
-" first_idchar .emit * .and .loop follow_idchar .emit * .and .true .emit '\\0';\n"
-"first_idchar\n"
-" 'a'-'z' .or 'A'-'Z' .or '_';\n"
-"follow_idchar\n"
-" first_idchar .or digit_dec;\n"
-"digit_dec\n"
-" '0'-'9';\n"
-"semicolon\n"
-" optional_space .and ';' .error MISSING_SEMICOLON .and optional_space;\n"
-"declaration_list\n"
-" declaration .and .loop declaration;\n"
-"declaration\n"
-" emitcode_definition .emit DECLARATION_EMITCODE .or\n"
-" errortext_definition .emit DECLARATION_ERRORTEXT .or\n"
-" regbyte_definition .emit DECLARATION_REGBYTE .or\n"
-" lexer_definition .emit DECLARATION_LEXER .or\n"
-" rule_definition .emit DECLARATION_RULE;\n"
-"emitcode_definition\n"
-" \".emtcode\" .and space .and identifier .and space .and integer .and space_or_null;\n"
-"integer\n"
-" integer_ne .error INTEGER_EXPECTED;\n"
-"integer_ne\n"
-" hex_integer .emit 0x10 .or dec_integer .emit 10;\n"
-"hex_integer\n"
-" hex_prefix .and digit_hex .emit * .and .loop digit_hex .emit * .and .true .emit '\\0';\n"
-"hex_prefix\n"
-" '0' .and hex_prefix_1;\n"
-"hex_prefix_1\n"
-" 'x' .or 'X';\n"
-"digit_hex\n"
-" '0'-'9' .or 'a'-'f' .or 'A'-'F';\n"
-"dec_integer\n"
-" digit_dec .emit * .and .loop digit_dec .emit * .and .true .emit '\\0';\n"
-"space_or_null\n"
-" space .or '\\0';\n"
-"errortext_definition\n"
-" \".errtext\" .and space .and identifier .and space .and string .and space_or_null;\n"
-"string\n"
-" string_ne .error STRING_EXPECTED;\n"
-"string_ne\n"
-" '\"' .and .loop string_char_double_quotes .and '\"' .emit '\\0';\n"
-"string_char_double_quotes\n"
-" escape_sequence .or string_char .emit * .or '\\'' .emit *;\n"
-"string_char\n"
-" '\\x5D'-'\\xFF' .or '\\x28'-'\\x5B' .or '\\x23'-'\\x26' .or '\\x0E'-'\\x21' .or '\\x0B'-'\\x0C' .or\n"
-" '\\x01'-'\\x09';\n"
-"escape_sequence\n"
-" '\\\\' .emit * .and escape_code;\n"
-"escape_code\n"
-" simple_escape_code .emit * .or hex_escape_code .or oct_escape_code;\n"
-"simple_escape_code\n"
-" '\\'' .or '\"' .or '?' .or '\\\\' .or 'a' .or 'b' .or 'f' .or 'n' .or 'r' .or 't' .or 'v';\n"
-"hex_escape_code\n"
-" 'x' .emit * .and digit_hex .emit * .and .loop digit_hex .emit *;\n"
-"oct_escape_code\n"
-" digit_oct .emit * .and optional_digit_oct .and optional_digit_oct;\n"
-"digit_oct\n"
-" '0'-'7';\n"
-"optional_digit_oct\n"
-" digit_oct .emit * .or .true;\n"
-"regbyte_definition\n"
-" \".regbyte\" .and space .and identifier .and space .and integer .and space_or_null;\n"
-"lexer_definition\n"
-" \".string\" .and space .and identifier .and semicolon;\n"
-"rule_definition\n"
-" identifier_ne .and space .and definition;\n"
-"definition\n"
-" specifier .and optional_specifiers_and_or .and semicolon .emit SPECIFIER_END;\n"
-"optional_specifiers_and_or\n"
-" and_specifiers .emit SPECIFIER_AND_TAG .or or_specifiers .emit SPECIFIER_OR_TAG .or .true;\n"
-"specifier\n"
-" specifier_condition .and optional_space .and specifier_rule;\n"
-"specifier_condition\n"
-" specifier_condition_1 .or .true;\n"
-"specifier_condition_1\n"
-" \".if\" .and optional_space .and '(' .and optional_space .and left_operand .and operator .and\n"
-" right_operand .and optional_space .and ')';\n"
-"left_operand\n"
-" identifier;\n"
-"operator\n"
-" operator_1 .or operator_2;\n"
-"operator_1\n"
-" optional_space .and '!' .and '=' .and optional_space;\n"
-"operator_2\n"
-" optional_space .and '=' .and '=' .and optional_space;\n"
-"right_operand\n"
-" integer;\n"
-"specifier_rule\n"
-" specifier_rule_1 .and optional_error .and .loop emit .and .true .emit EMIT_NULL;\n"
-"specifier_rule_1\n"
-" character_range .emit SPECIFIER_CHARACTER_RANGE .or\n"
-" character .emit SPECIFIER_CHARACTER .or\n"
-" string_ne .emit SPECIFIER_STRING .or\n"
-" \".true\" .emit SPECIFIER_TRUE .or\n"
-" \".false\" .emit SPECIFIER_FALSE .or\n"
-" \".debug\" .emit SPECIFIER_DEBUG .or\n"
-" advanced_identifier .emit SPECIFIER_IDENTIFIER;\n"
-"character\n"
-" '\\'' .and string_char_single_quotes .and '\\'' .emit '\\0';\n"
-"string_char_single_quotes\n"
-" escape_sequence .or string_char .emit * .or '\"' .emit *;\n"
-"character_range\n"
-" character .and optional_space .and '-' .and optional_space .and character;\n"
-"advanced_identifier\n"
-" optional_loop .and identifier;\n"
-"optional_loop\n"
-" optional_loop_1 .emit IDENTIFIER_LOOP .or .true .emit IDENTIFIER_SIMPLE;\n"
-"optional_loop_1\n"
-" \".loop\" .and space;\n"
-"optional_error\n"
-" error .emit ERROR_PRESENT .or .true .emit ERROR_NOT_PRESENT;\n"
-"error\n"
-" space .and \".error\" .and space .and identifier;\n"
-"emit\n"
-" emit_output .or emit_regbyte;\n"
-"emit_output\n"
-" space .and \".emit\" .and space .and emit_param;\n"
-"emit_param\n"
-" integer_ne .emit EMIT_INTEGER .or\n"
-" identifier_ne .emit EMIT_IDENTIFIER .or\n"
-" character .emit EMIT_CHARACTER .or\n"
-" '*' .emit EMIT_LAST_CHARACTER .or\n"
-" '$' .emit EMIT_CURRENT_POSITION;\n"
-"emit_regbyte\n"
-" space .and \".load\" .and space .and identifier .and space .and emit_param;\n"
-"and_specifiers\n"
-" and_specifier .and .loop and_specifier;\n"
-"or_specifiers\n"
-" or_specifier .and .loop or_specifier;\n"
-"and_specifier\n"
-" space .and \".and\" .and space .and specifier;\n"
-"or_specifier\n"
-" space .and \".or\" .and space .and specifier;\n"
-".string __string_filter;\n"
-"__string_filter\n"
-" __first_identifier_char .and .loop __next_identifier_char;\n"
-"__first_identifier_char\n"
-" 'a'-'z' .or 'A'-'Z' .or '_' .or '.';\n"
-"__next_identifier_char\n"
-" 'a'-'z' .or 'A'-'Z' .or '_' .or '0'-'9';\n"
-""
index 192d39aed1ee40ac7f3e4bf8aa8c8b68f6e5fb84..7f034520cd474ff967af35748107074fdfc3f15c 100644 (file)
  * Set x to positive or negative infinity.
  */
 #if defined(USE_IEEE) || defined(_WIN32)
-#define SET_POS_INFINITY(x)  ( *((GLuint *) (void *)&x) = 0x7F800000 )
-#define SET_NEG_INFINITY(x)  ( *((GLuint *) (void *)&x) = 0xFF800000 )
+#define SET_POS_INFINITY(x)                  \
+   do {                                      \
+         fi_type fi;                         \
+         fi.i = 0x7F800000;                  \
+         x = fi.f;                           \
+   } while (0)
+#define SET_NEG_INFINITY(x)                  \
+   do {                                      \
+         fi_type fi;                         \
+         fi.i = 0xFF800000;                  \
+         x = fi.f;                           \
+   } while (0)
 #elif defined(VMS)
 #define SET_POS_INFINITY(x)  x = __MAXFLOAT
 #define SET_NEG_INFINITY(x)  x = -__MAXFLOAT
@@ -674,9 +684,13 @@ _mesa_execute_program(GLcontext * ctx,
          break;
       case OPCODE_BGNLOOP:
          /* no-op */
+         ASSERT(program->Instructions[inst->BranchTarget].Opcode
+                == OPCODE_ENDLOOP);
          break;
       case OPCODE_ENDLOOP:
          /* subtract 1 here since pc is incremented by for(pc) loop */
+         ASSERT(program->Instructions[inst->BranchTarget].Opcode
+                == OPCODE_BGNLOOP);
          pc = inst->BranchTarget - 1;   /* go to matching BNGLOOP */
          break;
       case OPCODE_BGNSUB:      /* begin subroutine */
@@ -684,12 +698,26 @@ _mesa_execute_program(GLcontext * ctx,
       case OPCODE_ENDSUB:      /* end subroutine */
          break;
       case OPCODE_BRA:         /* branch (conditional) */
-         /* fall-through */
+         if (eval_condition(machine, inst)) {
+            /* take branch */
+            /* Subtract 1 here since we'll do pc++ below */
+            pc = inst->BranchTarget - 1;
+         }
+         break;
       case OPCODE_BRK:         /* break out of loop (conditional) */
-         /* fall-through */
+         ASSERT(program->Instructions[inst->BranchTarget].Opcode
+                == OPCODE_ENDLOOP);
+         if (eval_condition(machine, inst)) {
+            /* break out of loop */
+            /* pc++ at end of for-loop will put us after the ENDLOOP inst */
+            pc = inst->BranchTarget;
+         }
+         break;
       case OPCODE_CONT:        /* continue loop (conditional) */
+         ASSERT(program->Instructions[inst->BranchTarget].Opcode
+                == OPCODE_ENDLOOP);
          if (eval_condition(machine, inst)) {
-            /* take branch */
+            /* continue at ENDLOOP */
             /* Subtract 1 here since we'll do pc++ at end of for-loop */
             pc = inst->BranchTarget - 1;
          }
@@ -882,6 +910,10 @@ _mesa_execute_program(GLcontext * ctx,
       case OPCODE_IF:
          {
             GLboolean cond;
+            ASSERT(program->Instructions[inst->BranchTarget].Opcode
+                   == OPCODE_ELSE ||
+                   program->Instructions[inst->BranchTarget].Opcode
+                   == OPCODE_ENDIF);
             /* eval condition */
             if (inst->SrcReg[0].File != PROGRAM_UNDEFINED) {
                GLfloat a[4];
@@ -901,14 +933,16 @@ _mesa_execute_program(GLcontext * ctx,
             else {
                /* go to the instruction after ELSE or ENDIF */
                assert(inst->BranchTarget >= 0);
-               pc = inst->BranchTarget - 1;
+               pc = inst->BranchTarget;
             }
          }
          break;
       case OPCODE_ELSE:
          /* goto ENDIF */
+         ASSERT(program->Instructions[inst->BranchTarget].Opcode
+                == OPCODE_ENDIF);
          assert(inst->BranchTarget >= 0);
-         pc = inst->BranchTarget - 1;
+         pc = inst->BranchTarget;
          break;
       case OPCODE_ENDIF:
          /* nothing */
@@ -1635,11 +1669,12 @@ _mesa_execute_program(GLcontext * ctx,
       case OPCODE_UP2H:        /* unpack two 16-bit floats */
          {
             GLfloat a[4], result[4];
-            const GLuint *rawBits = (const GLuint *) a;
+            fi_type fi;
             GLhalfNV hx, hy;
             fetch_vector1(&inst->SrcReg[0], machine, a);
-            hx = rawBits[0] & 0xffff;
-            hy = rawBits[0] >> 16;
+            fi.f = a[0];
+            hx = fi.i & 0xffff;
+            hy = fi.i >> 16;
             result[0] = result[2] = _mesa_half_to_float(hx);
             result[1] = result[3] = _mesa_half_to_float(hy);
             store_vector4(inst, machine, result);
@@ -1648,11 +1683,12 @@ _mesa_execute_program(GLcontext * ctx,
       case OPCODE_UP2US:       /* unpack two GLushorts */
          {
             GLfloat a[4], result[4];
-            const GLuint *rawBits = (const GLuint *) a;
+            fi_type fi;
             GLushort usx, usy;
             fetch_vector1(&inst->SrcReg[0], machine, a);
-            usx = rawBits[0] & 0xffff;
-            usy = rawBits[0] >> 16;
+            fi.f = a[0];
+            usx = fi.i & 0xffff;
+            usy = fi.i >> 16;
             result[0] = result[2] = usx * (1.0f / 65535.0f);
             result[1] = result[3] = usy * (1.0f / 65535.0f);
             store_vector4(inst, machine, result);
@@ -1661,24 +1697,26 @@ _mesa_execute_program(GLcontext * ctx,
       case OPCODE_UP4B:        /* unpack four GLbytes */
          {
             GLfloat a[4], result[4];
-            const GLuint *rawBits = (const GLuint *) a;
+            fi_type fi;
             fetch_vector1(&inst->SrcReg[0], machine, a);
-            result[0] = (((rawBits[0] >> 0) & 0xff) - 128) / 127.0F;
-            result[1] = (((rawBits[0] >> 8) & 0xff) - 128) / 127.0F;
-            result[2] = (((rawBits[0] >> 16) & 0xff) - 128) / 127.0F;
-            result[3] = (((rawBits[0] >> 24) & 0xff) - 128) / 127.0F;
+            fi.f = a[0];
+            result[0] = (((fi.i >> 0) & 0xff) - 128) / 127.0F;
+            result[1] = (((fi.i >> 8) & 0xff) - 128) / 127.0F;
+            result[2] = (((fi.i >> 16) & 0xff) - 128) / 127.0F;
+            result[3] = (((fi.i >> 24) & 0xff) - 128) / 127.0F;
             store_vector4(inst, machine, result);
          }
          break;
       case OPCODE_UP4UB:       /* unpack four GLubytes */
          {
             GLfloat a[4], result[4];
-            const GLuint *rawBits = (const GLuint *) a;
+            fi_type fi;
             fetch_vector1(&inst->SrcReg[0], machine, a);
-            result[0] = ((rawBits[0] >> 0) & 0xff) / 255.0F;
-            result[1] = ((rawBits[0] >> 8) & 0xff) / 255.0F;
-            result[2] = ((rawBits[0] >> 16) & 0xff) / 255.0F;
-            result[3] = ((rawBits[0] >> 24) & 0xff) / 255.0F;
+            fi.f = a[0];
+            result[0] = ((fi.i >> 0) & 0xff) / 255.0F;
+            result[1] = ((fi.i >> 8) & 0xff) / 255.0F;
+            result[2] = ((fi.i >> 16) & 0xff) / 255.0F;
+            result[3] = ((fi.i >> 24) & 0xff) / 255.0F;
             store_vector4(inst, machine, result);
          }
          break;
index 1c687bc16cfffce591f59979df31f0262b6242c2..224350caac6c33efcecae83cb57a77a9d044e792 100644 (file)
@@ -312,7 +312,6 @@ struct prog_dst_register
     */
    GLuint CondSrc:1;
    /*@}*/
-   GLuint pad:28;
 };
 
 
index 6eefbcf5bdd73d426c5eaef312de376d61614f6e..674b786ac086c10389666aabb289bf74e5ef6692 100644 (file)
 
 VPATH = RCS
 
-INCDIR = [----.include],[--.main],[--.glapi],[-.slang],[-.grammar],[-]
+INCDIR = [----.include],[--.main],[--.glapi],[-.slang],[-]
 LIBDIR = [----.lib]
 CFLAGS = /include=($(INCDIR),[])/define=(PTHREADS=1)/name=(as_is,short)/float=ieee/ieee=denorm
 
 SOURCES = \
-       slang_compile.c,slang_preprocess.c
+       slang_compile.c
 
 OBJECTS = slang_builtin.obj,slang_codegen.obj,slang_compile.obj,\
        slang_compile_function.obj,slang_compile_operation.obj,\
@@ -59,7 +59,6 @@ slang_library_noise.obj : slang_library_noise.c
 slang_link.obj : slang_link.c
 slang_log.obj : slang_log.c
 slang_mem.obj : slang_mem.c
-slang_preprocess.obj : slang_preprocess.c
 slang_print.obj : slang_print.c
 slang_simplify.obj : slang_simplify.c
 slang_storage.obj : slang_storage.c
diff --git a/src/mesa/shader/slang/library/.gitignore b/src/mesa/shader/slang/library/.gitignore
new file mode 100644 (file)
index 0000000..02a89fc
--- /dev/null
@@ -0,0 +1 @@
+*_gc.h
index 0e03fac2ee176ad8c296acf752785b1c3bac8869..c6964512bfec548cf1f14e5d82a42429e0a3d231 100644 (file)
@@ -4,9 +4,7 @@ TOP = ../../../../..
 
 include $(TOP)/configs/current
 
-INCDIR = $(TOP)/include
-
-LIB_DEP = $(TOP)/$(LIB_DIR)/$(GL_LIB_NAME)
+GLSL_CL = $(TOP)/src/glsl/apps/compile
 
 #
 # targets
@@ -14,41 +12,13 @@ LIB_DEP = $(TOP)/$(LIB_DIR)/$(GL_LIB_NAME)
 
 .PHONY: default clean
 
-default: syntax builtin
+default: builtin
 
 clean:
-       -rm -f syn_to_c gc_to_bin *_syn.h *_gc.h
-
-syntax: slang_pp_directives_syn.h slang_pp_expression_syn.h slang_shader_syn.h slang_pp_version_syn.h
+       -rm -f *_gc.h
 
 builtin: builtin_110 builtin_120
 
-#
-# executables
-#
-
-syn_to_c: syn_to_c.c
-       $(CC) syn_to_c.c -o syn_to_c
-
-gc_to_bin: gc_to_bin.c slang_shader_syn.h
-       $(CC) gc_to_bin.c -o gc_to_bin
-
-#
-# syntax scripts
-#
-
-slang_pp_directives_syn.h: syn_to_c slang_pp_directives.syn
-       ./syn_to_c slang_pp_directives.syn > slang_pp_directives_syn.h
-
-slang_pp_expression_syn.h: syn_to_c slang_pp_expression.syn
-       ./syn_to_c slang_pp_expression.syn > slang_pp_expression_syn.h
-
-slang_shader_syn.h: syn_to_c slang_shader.syn
-       ./syn_to_c slang_shader.syn > slang_shader_syn.h
-
-slang_pp_version_syn.h: syn_to_c slang_pp_version.syn
-       ./syn_to_c slang_pp_version.syn > slang_pp_version_syn.h
-
 #
 # builtin library sources
 #
@@ -58,24 +28,24 @@ builtin_110: slang_common_builtin_gc.h slang_core_gc.h slang_fragment_builtin_gc
 builtin_120: slang_120_core_gc.h slang_builtin_120_common_gc.h slang_builtin_120_fragment_gc.h
 
 
-slang_120_core_gc.h: gc_to_bin slang_120_core.gc
-       ./gc_to_bin 1 slang_120_core.gc slang_120_core_gc.h
+slang_120_core_gc.h: slang_120_core.gc
+       $(GLSL_CL) fragment slang_120_core.gc slang_120_core_gc.h
 
-slang_builtin_120_common_gc.h: gc_to_bin slang_builtin_120_common.gc
-       ./gc_to_bin 1 slang_builtin_120_common.gc slang_builtin_120_common_gc.h
+slang_builtin_120_common_gc.h: slang_builtin_120_common.gc
+       $(GLSL_CL) fragment slang_builtin_120_common.gc slang_builtin_120_common_gc.h
 
-slang_builtin_120_fragment_gc.h: gc_to_bin slang_builtin_120_fragment.gc
-       ./gc_to_bin 1 slang_builtin_120_fragment.gc slang_builtin_120_fragment_gc.h
+slang_builtin_120_fragment_gc.h: slang_builtin_120_fragment.gc
+       $(GLSL_CL) fragment slang_builtin_120_fragment.gc slang_builtin_120_fragment_gc.h
 
-slang_common_builtin_gc.h: gc_to_bin slang_common_builtin.gc
-       ./gc_to_bin 1 slang_common_builtin.gc slang_common_builtin_gc.h
+slang_common_builtin_gc.h: slang_common_builtin.gc
+       $(GLSL_CL) fragment slang_common_builtin.gc slang_common_builtin_gc.h
 
-slang_core_gc.h: gc_to_bin slang_core.gc
-       ./gc_to_bin 1 slang_core.gc slang_core_gc.h
+slang_core_gc.h: slang_core.gc
+       $(GLSL_CL) fragment slang_core.gc slang_core_gc.h
 
-slang_fragment_builtin_gc.h: gc_to_bin slang_fragment_builtin.gc
-       ./gc_to_bin 1 slang_fragment_builtin.gc slang_fragment_builtin_gc.h
+slang_fragment_builtin_gc.h: slang_fragment_builtin.gc
+       $(GLSL_CL) fragment slang_fragment_builtin.gc slang_fragment_builtin_gc.h
 
-slang_vertex_builtin_gc.h: gc_to_bin slang_vertex_builtin.gc
-       ./gc_to_bin 2 slang_vertex_builtin.gc slang_vertex_builtin_gc.h
+slang_vertex_builtin_gc.h: slang_vertex_builtin.gc
+       $(GLSL_CL) vertex slang_vertex_builtin.gc slang_vertex_builtin_gc.h
 
diff --git a/src/mesa/shader/slang/library/SConscript b/src/mesa/shader/slang/library/SConscript
new file mode 100644 (file)
index 0000000..ef13114
--- /dev/null
@@ -0,0 +1,52 @@
+#######################################################################
+# SConscript for GLSL builtin library
+
+Import('*')
+
+env = env.Clone()
+
+# See also http://www.scons.org/wiki/UsingCodeGenerators
+
+def glsl_compile_emitter(target, source, env):
+       env.Depends(target, glsl_compile)
+       return (target, source)
+bld_frag = Builder(
+       action = glsl_compile[0].abspath + ' fragment $SOURCE $TARGET',
+       emitter = glsl_compile_emitter,
+       suffix = '.gc',
+       src_suffix = '_gc.h')
+       
+bld_vert = Builder(
+       action = glsl_compile[0].abspath + ' vertex $SOURCE $TARGET',
+       emitter = glsl_compile_emitter,
+       suffix = '.gc',
+       src_suffix = '_gc.h')
+
+env['BUILDERS']['bld_frag'] = bld_frag
+env['BUILDERS']['bld_vert'] = bld_vert
+
+# Generate GLSL builtin library binaries
+env.bld_frag(
+       '#src/mesa/shader/slang/library/slang_core_gc.h',
+       '#src/mesa/shader/slang/library/slang_core.gc')
+env.bld_frag(
+       '#src/mesa/shader/slang/library/slang_common_builtin_gc.h',
+       '#src/mesa/shader/slang/library/slang_common_builtin.gc')
+env.bld_frag(
+       '#src/mesa/shader/slang/library/slang_fragment_builtin_gc.h',
+       '#src/mesa/shader/slang/library/slang_fragment_builtin.gc')
+env.bld_vert(
+       '#src/mesa/shader/slang/library/slang_vertex_builtin_gc.h',
+       '#src/mesa/shader/slang/library/slang_vertex_builtin.gc')
+
+# Generate GLSL 1.20 builtin library binaries
+env.bld_frag(
+       '#src/mesa/shader/slang/library/slang_120_core_gc.h',
+       '#src/mesa/shader/slang/library/slang_120_core.gc')
+env.bld_frag(
+       '#src/mesa/shader/slang/library/slang_builtin_120_common_gc.h',
+       '#src/mesa/shader/slang/library/slang_builtin_120_common.gc')
+env.bld_frag(
+       '#src/mesa/shader/slang/library/slang_builtin_120_fragment_gc.h',
+       '#src/mesa/shader/slang/library/slang_builtin_120_fragment.gc')
diff --git a/src/mesa/shader/slang/library/gc_to_bin.c b/src/mesa/shader/slang/library/gc_to_bin.c
deleted file mode 100644 (file)
index 8aef7b5..0000000
+++ /dev/null
@@ -1,85 +0,0 @@
-#include "../../grammar/grammar_crt.h"
-#include "../../grammar/grammar_crt.c"
-#include <stdlib.h>
-#include <stdio.h>
-
-static const char *slang_shader_syn =
-#include "slang_shader_syn.h"
-;
-
-static int gc_to_bin (grammar id, const char *in, const char *out)
-{
-       FILE *f;
-       byte *source, *prod;
-       unsigned int size, i, line = 0;
-
-       printf ("Precompiling %s\n", in);
-
-       f = fopen (in, "r");
-       if (f == NULL)
-               return 1;
-       fseek (f, 0, SEEK_END);
-       size = ftell (f);
-       fseek (f, 0, SEEK_SET);
-       source = (byte *) grammar_alloc_malloc (size + 1);
-       source[fread (source, 1, size, f)] = '\0';
-       fclose (f);
-
-       if (!grammar_fast_check (id, source, &prod, &size, 65536))
-       {
-               grammar_alloc_free (source);
-               return 1;
-       }
-
-       f = fopen (out, "w");
-       fprintf (f, "\n");
-       fprintf (f, "/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE FOLLOWING FILE: */\n");
-       fprintf (f, "/* %s */\n", in);
-       fprintf (f, "\n");
-       for (i = 0; i < size; i++)
-       {
-               unsigned int a;
-               if (prod[i] < 10)
-                       a = 1;
-               else if (prod[i] < 100)
-                       a = 2;
-               else
-                       a = 3;
-               if (i < size - 1)
-                       a++;
-               if (line + a >= 100)
-               {
-                       fprintf (f, "\n");
-                       line = 0;
-               }
-               line += a;
-               fprintf (f, "%d", prod[i]);
-               if (i < size - 1)
-                       fprintf (f, ",");
-       }
-       fprintf (f, "\n");
-       fclose (f);
-       grammar_alloc_free (prod);
-   return 0;
-}
-
-int main (int argc, char *argv[])
-{
-   grammar id;
-
-   id = grammar_load_from_text ((const byte *) slang_shader_syn);
-   if (id == 0) {
-      fprintf(stderr, "Error loading grammar from text\n");
-      return 1;
-   }
-   grammar_set_reg8 (id, (const byte *) "parsing_builtin", 1);
-   grammar_set_reg8 (id, (const byte *) "shader_type", atoi (argv[1]));
-   if (gc_to_bin (id, argv[2], argv[3])) {
-      fprintf(stderr, "Error in gc_to_bin %s %s\n", argv[2], argv[3]);
-      grammar_destroy (id);
-      return 1;
-   }
-   grammar_destroy (id);
-   return 0;
-}
-
diff --git a/src/mesa/shader/slang/library/slang_120_core_gc.h b/src/mesa/shader/slang/library/slang_120_core_gc.h
deleted file mode 100644 (file)
index 1fdbddf..0000000
+++ /dev/null
@@ -1,764 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE FOLLOWING FILE: */
-/* slang_120_core.gc */
-
-5,1,90,95,0,0,26,0,1,1,1,0,0,9,0,102,48,48,0,0,1,1,0,0,9,0,102,49,48,0,0,1,1,0,0,9,0,102,50,48,0,0,
-1,1,0,0,9,0,102,48,49,0,0,1,1,0,0,9,0,102,49,49,0,0,1,1,0,0,9,0,102,50,49,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,16,8,48,0,57,59,120,0,18,102,48,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,
-16,8,48,0,57,59,121,0,18,102,49,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,122,0,
-18,102,50,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,120,0,18,102,48,49,0,20,0,
-9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,121,0,18,102,49,49,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,49,0,57,59,122,0,18,102,50,49,0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,9,0,102,0,
-0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,120,51,0,0,18,102,0,0,17,48,0,48,0,0,0,
-17,48,0,48,0,0,0,17,48,0,48,0,0,0,18,102,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,
-5,0,105,0,0,0,1,3,2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,105,0,0,0,0,0,9,18,95,95,
-114,101,116,86,97,108,0,58,109,97,116,50,120,51,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,
-1,0,98,0,0,0,1,3,2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,98,0,0,0,0,0,9,18,95,95,
-114,101,116,86,97,108,0,58,109,97,116,50,120,51,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,
-11,0,99,48,0,0,1,1,0,0,11,0,99,49,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,99,48,
-0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,99,49,0,20,0,0,1,90,95,0,0,28,0,1,1,1,0,
-0,9,0,102,48,48,0,0,1,1,0,0,9,0,102,49,48,0,0,1,1,0,0,9,0,102,50,48,0,0,1,1,0,0,9,0,102,51,48,0,0,
-1,1,0,0,9,0,102,48,49,0,0,1,1,0,0,9,0,102,49,49,0,0,1,1,0,0,9,0,102,50,49,0,0,1,1,0,0,9,0,102,51,
-49,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,120,0,18,102,48,48,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,8,48,0,57,59,121,0,18,102,49,48,0,20,0,9,18,95,95,114,101,116,86,97,108,
-0,16,8,48,0,57,59,122,0,18,102,50,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,119,
-0,18,102,51,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,120,0,18,102,48,49,0,20,
-0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,121,0,18,102,49,49,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,49,0,57,59,122,0,18,102,50,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,
-10,49,0,57,59,119,0,18,102,51,49,0,20,0,0,1,90,95,0,0,28,0,1,1,1,0,0,9,0,102,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,58,109,97,116,50,120,52,0,0,18,102,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,
-17,48,0,48,0,0,0,17,48,0,48,0,0,0,18,102,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,
-0,28,0,1,1,1,0,0,5,0,105,0,0,0,1,3,2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,105,0,0,
-0,0,0,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,120,52,0,0,18,102,0,0,0,20,0,0,1,90,95,0,
-0,28,0,1,1,1,0,0,1,0,98,0,0,0,1,3,2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,98,0,0,0,
-0,0,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,120,52,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,
-28,0,1,1,1,0,0,12,0,99,48,0,0,1,1,0,0,12,0,99,49,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,
-48,0,57,18,99,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,99,49,0,20,0,0,1,90,95,
-0,0,27,0,1,1,1,0,0,9,0,102,48,48,0,0,1,1,0,0,9,0,102,49,48,0,0,1,1,0,0,9,0,102,48,49,0,0,1,1,0,0,9,
-0,102,49,49,0,0,1,1,0,0,9,0,102,48,50,0,0,1,1,0,0,9,0,102,49,50,0,0,0,1,9,18,95,95,114,101,116,86,
-97,108,0,16,8,48,0,57,59,120,0,18,102,48,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,
-59,121,0,18,102,49,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,120,0,18,102,48,
-49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,121,0,18,102,49,49,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,50,0,57,59,120,0,18,102,48,50,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,50,0,57,59,121,0,18,102,49,50,0,20,0,0,1,90,95,0,0,27,0,1,1,1,0,0,9,0,102,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,58,109,97,116,51,120,50,0,0,18,102,0,0,17,48,0,48,0,0,0,17,48,0,48,0,
-0,0,18,102,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,0,27,0,1,1,1,0,0,5,0,105,0,0,0,
-1,3,2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,105,0,0,0,0,0,9,18,95,95,114,101,116,86,
-97,108,0,58,109,97,116,51,120,50,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,27,0,1,1,1,0,0,1,0,98,0,0,0,1,
-3,2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,98,0,0,0,0,0,9,18,95,95,114,101,116,86,97,
-108,0,58,109,97,116,51,120,50,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,27,0,1,1,1,0,0,10,0,99,48,0,0,1,
-1,0,0,10,0,99,49,0,0,1,1,0,0,10,0,99,50,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-99,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,99,49,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,50,0,57,18,99,50,0,20,0,0,1,90,95,0,0,30,0,1,1,1,0,0,9,0,102,48,48,0,0,1,1,0,
-0,9,0,102,49,48,0,0,1,1,0,0,9,0,102,50,48,0,0,1,1,0,0,9,0,102,51,48,0,0,1,1,0,0,9,0,102,48,49,0,0,
-1,1,0,0,9,0,102,49,49,0,0,1,1,0,0,9,0,102,50,49,0,0,1,1,0,0,9,0,102,51,49,0,0,1,1,0,0,9,0,102,48,
-50,0,0,1,1,0,0,9,0,102,49,50,0,0,1,1,0,0,9,0,102,50,50,0,0,1,1,0,0,9,0,102,51,50,0,0,0,1,9,18,95,
-95,114,101,116,86,97,108,0,16,8,48,0,57,59,120,0,18,102,48,48,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,8,48,0,57,59,121,0,18,102,49,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,
-122,0,18,102,50,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,119,0,18,102,51,48,0,
-20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,120,0,18,102,48,49,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,16,10,49,0,57,59,121,0,18,102,49,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,
-16,10,49,0,57,59,122,0,18,102,50,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,119,
-0,18,102,51,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,59,120,0,18,102,48,50,0,20,
-0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,59,121,0,18,102,49,50,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,50,0,57,59,122,0,18,102,50,50,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,
-10,50,0,57,59,119,0,18,102,51,50,0,20,0,0,1,90,95,0,0,30,0,1,1,1,0,0,9,0,102,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,58,109,97,116,51,120,52,0,0,18,102,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,
-17,48,0,48,0,0,0,17,48,0,48,0,0,0,18,102,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,17,
-48,0,48,0,0,0,18,102,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,0,30,0,1,1,1,0,0,5,0,105,0,0,0,1,3,2,
-90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,105,0,0,0,0,0,9,18,95,95,114,101,116,86,97,
-108,0,58,109,97,116,51,120,52,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,30,0,1,1,1,0,0,1,0,98,0,0,0,1,3,
-2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,98,0,0,0,0,0,9,18,95,95,114,101,116,86,97,
-108,0,58,109,97,116,51,120,52,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,30,0,1,1,1,0,0,12,0,99,48,0,0,1,
-1,0,0,12,0,99,49,0,0,1,1,0,0,12,0,99,50,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-99,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,99,49,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,50,0,57,18,99,50,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,9,0,102,48,48,0,0,1,1,0,
-0,9,0,102,49,48,0,0,1,1,0,0,9,0,102,48,49,0,0,1,1,0,0,9,0,102,49,49,0,0,1,1,0,0,9,0,102,48,50,0,0,
-1,1,0,0,9,0,102,49,50,0,0,1,1,0,0,9,0,102,48,51,0,0,1,1,0,0,9,0,102,49,51,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,16,8,48,0,57,59,120,0,18,102,48,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,
-16,8,48,0,57,59,121,0,18,102,49,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,120,
-0,18,102,48,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,121,0,18,102,49,49,0,20,
-0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,59,120,0,18,102,48,50,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,50,0,57,59,121,0,18,102,49,50,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,
-10,51,0,57,59,120,0,18,102,48,51,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,59,121,0,
-18,102,49,51,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,9,0,102,0,0,0,1,9,18,95,95,114,101,116,86,97,108,
-0,58,109,97,116,52,120,50,0,0,18,102,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,16,10,52,0,0,17,48,0,48,
-0,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,5,0,
-105,0,0,0,1,3,2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,105,0,0,0,0,0,9,18,95,95,114,
-101,116,86,97,108,0,58,109,97,116,52,120,50,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,1,0,
-98,0,0,0,1,3,2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,98,0,0,0,0,0,9,18,95,95,114,
-101,116,86,97,108,0,58,109,97,116,52,120,50,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,10,
-0,99,48,0,0,1,1,0,0,10,0,99,49,0,0,1,1,0,0,10,0,99,50,0,0,1,1,0,0,10,0,99,51,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,16,8,48,0,57,18,99,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,
-57,18,99,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,99,50,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,16,10,51,0,57,18,99,51,0,20,0,0,1,90,95,0,0,31,0,1,1,1,0,0,9,0,102,48,48,0,0,1,
-1,0,0,9,0,102,49,48,0,0,1,1,0,0,9,0,102,50,48,0,0,1,1,0,0,9,0,102,48,49,0,0,1,1,0,0,9,0,102,49,49,
-0,0,1,1,0,0,9,0,102,50,49,0,0,1,1,0,0,9,0,102,48,50,0,0,1,1,0,0,9,0,102,49,50,0,0,1,1,0,0,9,0,102,
-50,50,0,0,1,1,0,0,9,0,102,48,51,0,0,1,1,0,0,9,0,102,49,51,0,0,1,1,0,0,9,0,102,50,51,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,120,0,18,102,48,48,0,20,0,9,18,95,95,114,101,116,86,
-97,108,0,16,8,48,0,57,59,121,0,18,102,49,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,
-59,122,0,18,102,50,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,120,0,18,102,48,
-49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,121,0,18,102,49,49,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,49,0,57,59,122,0,18,102,50,49,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,50,0,57,59,120,0,18,102,48,50,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,
-59,121,0,18,102,49,50,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,59,122,0,18,102,50,
-50,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,59,120,0,18,102,48,51,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,51,0,57,59,121,0,18,102,49,51,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,51,0,57,59,122,0,18,102,50,51,0,20,0,0,1,90,95,0,0,31,0,1,1,1,0,0,9,0,102,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,51,0,0,18,102,0,0,17,48,0,48,0,0,0,17,48,0,48,0,
-0,0,17,48,0,48,0,0,0,18,102,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,18,102,0,0,17,
-48,0,48,0,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,0,31,0,1,1,1,0,0,5,0,105,0,0,0,
-1,3,2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,105,0,0,0,0,0,9,18,95,95,114,101,116,86,
-97,108,0,58,109,97,116,52,120,51,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,31,0,1,1,1,0,0,1,0,98,0,0,0,1,
-3,2,90,95,1,0,9,0,1,102,0,2,58,102,108,111,97,116,0,0,18,98,0,0,0,0,0,9,18,95,95,114,101,116,86,97,
-108,0,58,109,97,116,52,120,51,0,0,18,102,0,0,0,20,0,0,1,90,95,0,0,31,0,1,1,1,0,0,11,0,99,48,0,0,1,
-1,0,0,11,0,99,49,0,0,1,1,0,0,11,0,99,50,0,0,1,1,0,0,11,0,99,51,0,0,0,1,9,18,95,95,114,101,116,86,
-97,108,0,16,8,48,0,57,18,99,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,99,49,0,
-20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,99,50,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,51,0,57,18,99,51,0,20,0,0,1,90,95,0,0,13,0,1,1,1,0,0,13,0,109,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,13,0,1,1,1,0,0,27,0,109,0,0,0,1,9,18,95,95,114,101,
-116,86,97,108,0,58,109,97,116,50,0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,10,49,0,57,0,0,20,0,0,1,
-90,95,0,0,13,0,1,1,1,0,0,29,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,0,0,
-18,109,0,16,8,48,0,57,0,18,109,0,16,10,49,0,57,0,0,20,0,0,1,90,95,0,0,13,0,1,1,1,0,0,26,0,109,0,0,
-0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,0,0,18,109,0,16,8,48,0,57,59,120,121,0,0,
-18,109,0,16,10,49,0,57,59,120,121,0,0,0,20,0,0,1,90,95,0,0,13,0,1,1,1,0,0,28,0,109,0,0,0,1,9,18,95,
-95,114,101,116,86,97,108,0,58,109,97,116,50,0,0,18,109,0,16,8,48,0,57,59,120,121,0,0,18,109,0,16,
-10,49,0,57,59,120,121,0,0,0,20,0,0,1,90,95,0,0,13,0,1,1,1,0,0,14,0,109,0,0,0,1,9,18,95,95,114,101,
-116,86,97,108,0,58,109,97,116,50,0,0,18,109,0,16,8,48,0,57,59,120,121,0,0,18,109,0,16,10,49,0,57,
-59,120,121,0,0,0,20,0,0,1,90,95,0,0,13,0,1,1,1,0,0,30,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,58,109,97,116,50,0,0,18,109,0,16,8,48,0,57,59,120,121,0,0,18,109,0,16,10,49,0,57,59,120,121,
-0,0,0,20,0,0,1,90,95,0,0,13,0,1,1,1,0,0,31,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,
-97,116,50,0,0,18,109,0,16,8,48,0,57,59,120,121,0,0,18,109,0,16,10,49,0,57,59,120,121,0,0,0,20,0,0,
-1,90,95,0,0,13,0,1,1,1,0,0,15,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,0,
-0,18,109,0,16,8,48,0,57,59,120,121,0,0,18,109,0,16,10,49,0,57,59,120,121,0,0,0,20,0,0,1,90,95,0,0,
-26,0,1,1,1,0,0,26,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,26,
-0,1,1,1,0,0,14,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,120,51,0,0,18,109,
-0,16,8,48,0,57,0,18,109,0,16,10,49,0,57,0,0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,31,0,109,0,0,0,1,9,
-18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,
-10,49,0,57,0,0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,28,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,
-0,58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,59,120,121,122,0,0,18,109,0,16,10,49,0,57,59,
-120,121,122,0,0,0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,30,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,59,120,121,122,0,0,18,109,0,16,10,49,0,57,
-59,120,121,122,0,0,0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,15,0,109,0,0,0,1,9,18,95,95,114,101,116,86,
-97,108,0,58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,59,120,121,122,0,0,18,109,0,16,10,49,0,
-57,59,120,121,122,0,0,0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,13,0,109,0,0,0,1,9,18,95,95,114,101,116,
-86,97,108,0,58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,
-121,0,0,17,48,0,48,0,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,121,0,0,17,48,
-0,48,0,0,0,0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,27,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
-58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,17,
-48,0,48,0,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,121,0,0,17,48,0,48,0,0,0,
-0,20,0,0,1,90,95,0,0,26,0,1,1,1,0,0,29,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,
-116,50,120,51,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,17,48,0,48,0,0,
-0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,121,0,0,17,48,0,48,0,0,0,0,20,0,0,1,
-90,95,0,0,28,0,1,1,1,0,0,28,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,
-95,0,0,28,0,1,1,1,0,0,30,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,120,52,
-0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,10,49,0,57,0,0,20,0,0,1,90,95,0,0,28,0,1,1,1,0,0,15,0,109,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,120,52,0,0,18,109,0,16,8,48,0,57,0,18,
-109,0,16,10,49,0,57,0,0,20,0,0,1,90,95,0,0,28,0,1,1,1,0,0,26,0,109,0,0,0,1,9,18,95,95,114,101,116,
-86,97,108,0,58,109,97,116,50,120,52,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,
-121,0,0,18,109,0,16,8,48,0,57,59,122,0,0,17,48,0,48,0,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,
-0,16,10,49,0,57,59,121,0,0,18,109,0,16,10,49,0,57,59,122,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,0,
-28,0,1,1,1,0,0,14,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,120,52,0,0,18,
-109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,0,16,8,48,0,57,59,122,0,0,17,
-48,0,48,0,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,121,0,0,18,109,0,16,10,
-49,0,57,59,122,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,0,28,0,1,1,1,0,0,31,0,109,0,0,0,1,9,18,95,
-95,114,101,116,86,97,108,0,58,109,97,116,50,120,52,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,
-16,8,48,0,57,59,121,0,0,18,109,0,16,8,48,0,57,59,122,0,0,17,48,0,48,0,0,0,18,109,0,16,10,49,0,57,
-59,120,0,0,18,109,0,16,10,49,0,57,59,121,0,0,18,109,0,16,10,49,0,57,59,122,0,0,17,48,0,48,0,0,0,0,
-20,0,0,1,90,95,0,0,28,0,1,1,1,0,0,13,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,
-116,50,120,52,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,121,0,0,17,48,0,48,0,
-0,0,17,48,0,48,0,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,121,0,0,17,48,0,
-48,0,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,0,28,0,1,1,1,0,0,27,0,109,0,0,0,1,9,18,95,95,114,101,
-116,86,97,108,0,58,109,97,116,50,120,52,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,8,48,0,57,
-59,121,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,49,0,
-57,59,121,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,0,28,0,1,1,1,0,0,29,0,109,0,0,0,
-1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,50,120,52,0,0,18,109,0,16,8,48,0,57,59,120,0,0,
-18,109,0,16,8,48,0,57,59,121,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,18,109,0,16,10,49,0,57,59,120,0,
-0,18,109,0,16,10,49,0,57,59,121,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,0,0,27,0,1,
-1,1,0,0,27,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,27,0,1,1,1,
-0,0,29,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,120,50,0,0,18,109,0,16,8,
-48,0,57,0,18,109,0,16,10,49,0,57,0,18,109,0,16,10,50,0,57,0,0,20,0,0,1,90,95,0,0,27,0,1,1,1,0,0,14,
-0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,120,50,0,0,18,109,0,16,8,48,0,57,
-0,18,109,0,16,10,49,0,57,0,18,109,0,16,10,50,0,57,0,0,20,0,0,1,90,95,0,0,27,0,1,1,1,0,0,30,0,109,0,
-0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,120,50,0,0,18,109,0,16,8,48,0,57,59,120,
-0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,
-121,0,0,18,109,0,16,10,50,0,57,59,120,0,0,18,109,0,16,10,50,0,57,59,121,0,0,0,20,0,0,1,90,95,0,0,
-27,0,1,1,1,0,0,31,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,120,50,0,0,18,
-109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,0,16,10,49,0,57,59,120,0,0,
-18,109,0,16,10,49,0,57,59,121,0,0,18,109,0,16,10,50,0,57,59,120,0,0,18,109,0,16,10,50,0,57,59,121,
-0,0,0,20,0,0,1,90,95,0,0,27,0,1,1,1,0,0,15,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,
-97,116,51,120,50,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,0,16,
-10,49,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,121,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,0,20,0,0,
-1,90,95,0,0,27,0,1,1,1,0,0,13,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,
-120,50,0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,10,49,0,57,0,58,118,101,99,50,0,0,17,48,0,48,0,0,0,
-0,0,0,20,0,0,1,90,95,0,0,27,0,1,1,1,0,0,26,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,
-97,116,51,120,50,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,0,16,
-10,49,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,121,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,0,20,0,0,
-1,90,95,0,0,27,0,1,1,1,0,0,28,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,
-120,50,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,0,16,10,49,0,
-57,59,120,0,0,18,109,0,16,10,49,0,57,59,121,0,0,17,48,0,48,0,0,0,17,48,0,48,0,0,0,0,20,0,0,1,90,95,
-0,0,14,0,1,1,1,0,0,14,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,
-14,0,1,1,1,0,0,31,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,0,0,18,109,0,
-16,8,48,0,57,0,18,109,0,16,10,49,0,57,0,18,109,0,16,10,50,0,57,0,0,20,0,0,1,90,95,0,0,14,0,1,1,1,0,
-0,30,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,0,0,18,109,0,16,8,48,0,57,
-59,120,121,122,0,0,18,109,0,16,10,49,0,57,59,120,121,122,0,0,18,109,0,16,10,50,0,57,59,120,121,122,
-0,0,0,20,0,0,1,90,95,0,0,14,0,1,1,1,0,0,15,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,
-97,116,51,0,0,18,109,0,16,8,48,0,57,59,120,121,122,0,0,18,109,0,16,10,49,0,57,59,120,121,122,0,0,
-18,109,0,16,10,50,0,57,59,120,121,122,0,0,0,20,0,0,1,90,95,0,0,14,0,1,1,1,0,0,26,0,109,0,0,0,1,9,
-18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,10,49,0,
-57,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,0,20,0,0,1,90,95,0,0,14,0,1,1,1,0,0,28,0,109,0,0,0,
-1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,0,0,18,109,0,16,8,48,0,57,59,120,121,122,0,0,
-18,109,0,16,10,49,0,57,59,120,121,122,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,0,20,0,0,1,90,
-95,0,0,14,0,1,1,1,0,0,27,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,0,0,18,
-109,0,16,8,48,0,57,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,18,109,0,16,10,50,0,57,0,
-17,49,0,0,0,0,0,20,0,0,1,90,95,0,0,14,0,1,1,1,0,0,29,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,58,109,97,116,51,0,0,18,109,0,16,8,48,0,57,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,
-0,0,0,18,109,0,16,10,50,0,57,0,17,49,0,0,0,0,0,20,0,0,1,90,95,0,0,14,0,1,1,1,0,0,13,0,109,0,0,0,1,
-9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,0,0,18,109,0,16,8,48,0,57,0,17,48,0,0,0,0,18,
-109,0,16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,0,20,0,0,1,90,95,0,0,
-30,0,1,1,1,0,0,30,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,30,
-0,1,1,1,0,0,15,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,120,52,0,0,18,109,
-0,16,8,48,0,57,0,18,109,0,16,10,49,0,57,0,18,109,0,16,10,50,0,57,0,0,20,0,0,1,90,95,0,0,30,0,1,1,1,
-0,0,14,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,120,52,0,0,18,109,0,16,8,
-48,0,57,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,18,109,0,16,10,50,0,57,0,17,48,0,0,
-0,0,0,20,0,0,1,90,95,0,0,30,0,1,1,1,0,0,31,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,
-97,116,51,120,52,0,0,18,109,0,16,8,48,0,57,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,
-18,109,0,16,10,50,0,57,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,30,0,1,1,1,0,0,28,0,109,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,58,109,97,116,51,120,52,0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,10,
-49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,30,0,1,1,1,
-0,0,26,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,120,52,0,0,18,109,0,16,8,
-48,0,57,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,
-0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,30,0,1,1,1,0,0,27,0,109,0,0,0,1,9,18,95,95,114,101,116,86,
-97,108,0,58,109,97,116,51,120,52,0,0,18,109,0,16,8,48,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,18,109,0,
-16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,18,109,0,16,10,50,0,57,0,17,49,0,0,0,0,17,48,0,0,0,0,0,
-20,0,0,1,90,95,0,0,30,0,1,1,1,0,0,29,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,
-116,51,120,52,0,0,18,109,0,16,8,48,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,
-48,0,0,0,0,17,48,0,0,0,0,18,109,0,16,10,50,0,57,0,17,49,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,
-30,0,1,1,1,0,0,13,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,51,120,52,0,0,18,
-109,0,16,8,48,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,
-0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,29,
-0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,31,0,
-109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,50,0,0,18,109,0,16,8,48,0,57,
-59,120,121,0,0,18,109,0,16,10,49,0,57,59,120,121,0,0,18,109,0,16,10,50,0,57,59,120,121,0,0,18,109,
-0,16,10,51,0,57,59,120,121,0,0,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,15,0,109,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,58,109,97,116,52,120,50,0,0,18,109,0,16,8,48,0,57,59,120,121,0,0,18,109,0,16,
-10,49,0,57,59,120,121,0,0,18,109,0,16,10,50,0,57,59,120,121,0,0,18,109,0,16,10,51,0,57,59,120,121,
-0,0,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,27,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,
-97,116,52,120,50,0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,
-0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,14,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,
-116,52,120,50,0,0,18,109,0,16,8,48,0,57,59,120,121,0,0,18,109,0,16,10,49,0,57,59,120,121,0,0,18,
-109,0,16,10,50,0,57,59,120,121,0,0,17,48,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,
-30,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,50,0,0,18,109,0,16,8,48,0,
-57,59,120,121,0,0,18,109,0,16,10,49,0,57,59,120,121,0,0,18,109,0,16,10,50,0,57,59,120,121,0,0,17,
-48,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,13,0,109,0,0,0,1,9,18,95,95,114,101,
-116,86,97,108,0,58,109,97,116,52,120,50,0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,10,49,0,57,0,17,48,
-0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,26,0,109,0,0,
-0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,50,0,0,18,109,0,16,8,48,0,57,59,120,
-121,0,0,18,109,0,16,10,49,0,57,59,120,121,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,
-0,0,0,20,0,0,1,90,95,0,0,29,0,1,1,1,0,0,28,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,
-97,116,52,120,50,0,0,18,109,0,16,8,48,0,57,59,120,121,0,0,18,109,0,16,10,49,0,57,59,120,121,0,0,17,
-48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,31,0,1,1,1,0,0,31,0,109,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,31,0,1,1,1,0,0,15,0,109,0,0,
-0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,51,0,0,18,109,0,16,8,48,0,57,59,120,
-121,122,0,0,18,109,0,16,10,49,0,57,59,120,121,122,0,0,18,109,0,16,10,50,0,57,59,120,121,122,0,0,18,
-109,0,16,10,51,0,57,59,120,121,122,0,0,0,20,0,0,1,90,95,0,0,31,0,1,1,1,0,0,14,0,109,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,51,0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,10,
-49,0,57,0,18,109,0,16,10,50,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,
-31,0,1,1,1,0,0,30,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,51,0,0,18,
-109,0,16,8,48,0,57,59,120,121,122,0,0,18,109,0,16,10,49,0,57,59,120,121,122,0,0,18,109,0,16,10,50,
-0,57,59,120,121,122,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,31,0,1,1,1,
-0,0,29,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,51,0,0,18,109,0,16,8,
-48,0,57,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,18,109,0,16,10,50,0,57,0,17,49,0,0,
-0,0,18,109,0,16,10,51,0,57,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,31,0,1,1,1,0,0,26,0,109,0,0,0,1,9,
-18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,51,0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,
-10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,0,
-20,0,0,1,90,95,0,0,31,0,1,1,1,0,0,27,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,
-116,52,120,51,0,0,18,109,0,16,8,48,0,57,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,18,
-109,0,16,10,50,0,57,0,17,49,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,
-31,0,1,1,1,0,0,28,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,51,0,0,18,
-109,0,16,8,48,0,57,59,120,121,122,0,0,18,109,0,16,10,49,0,57,59,120,121,122,0,0,17,48,0,0,0,0,17,
-48,0,0,0,0,17,49,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,31,0,1,1,1,
-0,0,13,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,120,51,0,0,18,109,0,16,8,
-48,0,57,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,
-0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,0,20,0,0,1,90,95,0,0,15,0,1,1,1,0,0,15,0,109,0,0,0,
-1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,15,0,1,1,1,0,0,30,0,109,0,0,0,1,9,
-18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,0,0,18,109,0,16,8,48,0,57,0,18,109,0,16,10,49,0,
-57,0,18,109,0,16,10,50,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,0,20,0,0,1,
-90,95,0,0,15,0,1,1,1,0,0,31,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,0,0,
-18,109,0,16,8,48,0,57,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,18,109,0,16,10,50,0,
-57,0,17,48,0,0,0,0,18,109,0,16,10,51,0,57,0,17,49,0,0,0,0,0,20,0,0,1,90,95,0,0,15,0,1,1,1,0,0,28,0,
-109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,0,0,18,109,0,16,8,48,0,57,0,18,109,
-0,16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,
-0,0,17,48,0,0,0,0,17,49,0,0,0,0,0,20,0,0,1,90,95,0,0,15,0,1,1,1,0,0,29,0,109,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,58,109,97,116,52,0,0,18,109,0,16,8,48,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,
-18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,18,109,0,16,10,50,0,57,0,17,49,0,0,0,0,17,48,
-0,0,0,0,18,109,0,16,10,51,0,57,0,17,48,0,0,0,0,17,49,0,0,0,0,0,20,0,0,1,90,95,0,0,15,0,1,1,1,0,0,
-14,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,0,0,18,109,0,16,8,48,0,57,0,
-17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,18,109,0,16,10,50,0,57,0,17,48,0,0,0,0,17,48,
-0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,0,20,0,0,1,90,95,0,0,15,0,1,1,1,0,0,26,0,109,0,0,
-0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,0,0,18,109,0,16,8,48,0,57,0,17,48,0,0,0,0,
-18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,17,48,0,0,0,0,17,
-48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,0,20,0,0,1,90,95,0,0,15,0,1,1,1,0,0,27,0,109,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,0,0,18,109,0,16,8,48,0,57,0,17,48,0,0,
-0,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,18,109,0,16,10,50,0,57,0,17,
-49,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,0,20,0,0,1,90,95,
-0,0,15,0,1,1,1,0,0,13,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,109,97,116,52,0,0,18,109,
-0,16,8,48,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,18,109,0,16,10,49,0,57,0,17,48,0,0,0,0,17,48,0,0,0,0,
-17,48,0,0,0,0,17,48,0,0,0,0,17,49,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,17,48,0,0,0,0,
-17,49,0,0,0,0,0,20,0,0,1,90,95,0,0,0,0,2,1,1,0,2,0,26,0,109,0,0,1,1,0,0,26,0,110,0,0,0,1,9,18,109,
-0,16,8,48,0,57,18,110,0,16,8,48,0,57,21,0,9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,21,0,0,1,
-90,95,0,0,0,0,2,1,1,0,2,0,28,0,109,0,0,1,1,0,0,28,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,
-16,8,48,0,57,21,0,9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,21,0,0,1,90,95,0,0,0,0,2,1,1,0,2,
-0,27,0,109,0,0,1,1,0,0,27,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,21,0,9,18,
-109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,21,0,9,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,21,
-0,0,1,90,95,0,0,0,0,2,1,1,0,2,0,30,0,109,0,0,1,1,0,0,30,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,
-110,0,16,8,48,0,57,21,0,9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,21,0,9,18,109,0,16,10,50,0,
-57,18,110,0,16,10,50,0,57,21,0,0,1,90,95,0,0,0,0,2,1,1,0,2,0,29,0,109,0,0,1,1,0,0,29,0,110,0,0,0,1,
-9,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,21,0,9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,
-21,0,9,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,21,0,9,18,109,0,16,10,51,0,57,18,110,0,16,10,
-51,0,57,21,0,0,1,90,95,0,0,0,0,2,1,1,0,2,0,31,0,109,0,0,1,1,0,0,31,0,110,0,0,0,1,9,18,109,0,16,8,
-48,0,57,18,110,0,16,8,48,0,57,21,0,9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,21,0,9,18,109,0,
-16,10,50,0,57,18,110,0,16,10,50,0,57,21,0,9,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,21,0,0,1,
-90,95,0,0,0,0,2,2,1,0,2,0,26,0,109,0,0,1,1,0,0,26,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,
-16,8,48,0,57,22,0,9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,22,0,0,1,90,95,0,0,0,0,2,2,1,0,2,
-0,28,0,109,0,0,1,1,0,0,28,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,22,0,9,18,
-109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,22,0,0,1,90,95,0,0,0,0,2,2,1,0,2,0,27,0,109,0,0,1,1,0,0,
-27,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,22,0,9,18,109,0,16,10,49,0,57,18,
-110,0,16,10,49,0,57,22,0,9,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,22,0,0,1,90,95,0,0,0,0,2,
-2,1,0,2,0,30,0,109,0,0,1,1,0,0,30,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,22,0,
-9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,22,0,9,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,
-57,22,0,0,1,90,95,0,0,0,0,2,2,1,0,2,0,29,0,109,0,0,1,1,0,0,29,0,110,0,0,0,1,9,18,109,0,16,8,48,0,
-57,18,110,0,16,8,48,0,57,22,0,9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,22,0,9,18,109,0,16,
-10,50,0,57,18,110,0,16,10,50,0,57,22,0,9,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,22,0,0,1,90,
-95,0,0,0,0,2,2,1,0,2,0,31,0,109,0,0,1,1,0,0,31,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,16,8,
-48,0,57,22,0,9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,22,0,9,18,109,0,16,10,50,0,57,18,110,
-0,16,10,50,0,57,22,0,9,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,22,0,0,1,90,95,0,0,0,0,2,4,1,
-0,2,0,26,0,109,0,0,1,1,0,0,26,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,24,0,9,
-18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,24,0,0,1,90,95,0,0,0,0,2,4,1,0,2,0,28,0,109,0,0,1,1,
-0,0,28,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,24,0,9,18,109,0,16,10,49,0,57,
-18,110,0,16,10,49,0,57,24,0,0,1,90,95,0,0,0,0,2,4,1,0,2,0,27,0,109,0,0,1,1,0,0,27,0,110,0,0,0,1,9,
-18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,24,0,9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,
-24,0,9,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,24,0,0,1,90,95,0,0,0,0,2,4,1,0,2,0,30,0,109,0,
-0,1,1,0,0,30,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,24,0,9,18,109,0,16,10,49,
-0,57,18,110,0,16,10,49,0,57,24,0,9,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,24,0,0,1,90,95,0,
-0,0,0,2,4,1,0,2,0,29,0,109,0,0,1,1,0,0,29,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,
-57,24,0,9,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,24,0,9,18,109,0,16,10,50,0,57,18,110,0,16,
-10,50,0,57,24,0,9,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,24,0,0,1,90,95,0,0,0,0,2,4,1,0,2,0,
-31,0,109,0,0,1,1,0,0,31,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,24,0,9,18,109,
-0,16,10,49,0,57,18,110,0,16,10,49,0,57,24,0,9,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,24,0,9,
-18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,24,0,0,1,90,95,0,0,11,0,2,21,1,1,0,0,26,0,109,0,0,1,
-1,0,0,10,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,118,0,59,120,0,18,109,0,16,8,
-48,0,57,59,120,0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,120,0,48,46,20,0,9,18,95,95,114,
-101,116,86,97,108,0,59,121,0,18,118,0,59,120,0,18,109,0,16,8,48,0,57,59,121,0,48,18,118,0,59,121,0,
-18,109,0,16,10,49,0,57,59,121,0,48,46,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,118,0,59,
-120,0,18,109,0,16,8,48,0,57,59,122,0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,122,0,48,46,20,
-0,0,1,90,95,0,0,12,0,2,21,1,1,0,0,28,0,109,0,0,1,1,0,0,10,0,118,0,0,0,1,9,18,95,95,114,101,116,86,
-97,108,0,59,120,0,18,118,0,59,120,0,18,109,0,16,8,48,0,57,59,120,0,48,18,118,0,59,121,0,18,109,0,
-16,10,49,0,57,59,120,0,48,46,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,118,0,59,120,0,18,
-109,0,16,8,48,0,57,59,121,0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,121,0,48,46,20,0,9,18,
-95,95,114,101,116,86,97,108,0,59,122,0,18,118,0,59,120,0,18,109,0,16,8,48,0,57,59,122,0,48,18,118,
-0,59,121,0,18,109,0,16,10,49,0,57,59,122,0,48,46,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,
-18,118,0,59,120,0,18,109,0,16,8,48,0,57,59,119,0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,
-119,0,48,46,20,0,0,1,90,95,0,0,10,0,2,21,1,1,0,0,27,0,109,0,0,1,1,0,0,11,0,118,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,59,120,0,18,118,0,59,120,0,18,109,0,16,8,48,0,57,59,120,0,48,18,118,0,59,
-121,0,18,109,0,16,10,49,0,57,59,120,0,48,46,18,118,0,59,122,0,18,109,0,16,10,50,0,57,59,120,0,48,
-46,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,118,0,59,120,0,18,109,0,16,8,48,0,57,59,121,
-0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,121,0,48,46,18,118,0,59,122,0,18,109,0,16,10,50,0,
-57,59,121,0,48,46,20,0,0,1,90,95,0,0,12,0,2,21,1,1,0,0,30,0,109,0,0,1,1,0,0,11,0,118,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,59,120,0,18,118,0,59,120,0,18,109,0,16,8,48,0,57,59,120,0,48,18,118,
-0,59,121,0,18,109,0,16,10,49,0,57,59,120,0,48,46,18,118,0,59,122,0,18,109,0,16,10,50,0,57,59,120,0,
-48,46,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,118,0,59,120,0,18,109,0,16,8,48,0,57,59,
-121,0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,121,0,48,46,18,118,0,59,122,0,18,109,0,16,10,
-50,0,57,59,121,0,48,46,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,118,0,59,120,0,18,109,0,
-16,8,48,0,57,59,122,0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,122,0,48,46,18,118,0,59,122,0,
-18,109,0,16,10,50,0,57,59,122,0,48,46,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,18,118,0,59,
-120,0,18,109,0,16,8,48,0,57,59,119,0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,119,0,48,46,18,
-118,0,59,122,0,18,109,0,16,10,50,0,57,59,119,0,48,46,20,0,0,1,90,95,0,0,10,0,2,21,1,1,0,0,29,0,109,
-0,0,1,1,0,0,12,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,118,0,59,120,0,18,109,
-0,16,8,48,0,57,59,120,0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,120,0,48,46,18,118,0,59,122,
-0,18,109,0,16,10,50,0,57,59,120,0,48,46,18,118,0,59,119,0,18,109,0,16,10,51,0,57,59,120,0,48,46,20,
-0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,118,0,59,120,0,18,109,0,16,8,48,0,57,59,121,0,48,
-18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,121,0,48,46,18,118,0,59,122,0,18,109,0,16,10,50,0,57,
-59,121,0,48,46,18,118,0,59,119,0,18,109,0,16,10,51,0,57,59,121,0,48,46,20,0,0,1,90,95,0,0,11,0,2,
-21,1,1,0,0,31,0,109,0,0,1,1,0,0,12,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,
-118,0,59,120,0,18,109,0,16,8,48,0,57,59,120,0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,120,0,
-48,46,18,118,0,59,122,0,18,109,0,16,10,50,0,57,59,120,0,48,46,18,118,0,59,119,0,18,109,0,16,10,51,
-0,57,59,120,0,48,46,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,118,0,59,120,0,18,109,0,16,
-8,48,0,57,59,121,0,48,18,118,0,59,121,0,18,109,0,16,10,49,0,57,59,121,0,48,46,18,118,0,59,122,0,18,
-109,0,16,10,50,0,57,59,121,0,48,46,18,118,0,59,119,0,18,109,0,16,10,51,0,57,59,121,0,48,46,20,0,9,
-18,95,95,114,101,116,86,97,108,0,59,122,0,18,118,0,59,120,0,18,109,0,16,8,48,0,57,59,122,0,48,18,
-118,0,59,121,0,18,109,0,16,10,49,0,57,59,122,0,48,46,18,118,0,59,122,0,18,109,0,16,10,50,0,57,59,
-122,0,48,46,18,118,0,59,119,0,18,109,0,16,10,51,0,57,59,122,0,48,46,20,0,0,1,90,95,0,0,27,0,2,21,1,
-1,0,0,13,0,109,0,0,1,1,0,0,27,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,
-0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,
-16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,18,110,0,16,10,50,
-0,57,48,20,0,0,1,90,95,0,0,29,0,2,21,1,1,0,0,13,0,109,0,0,1,1,0,0,29,0,110,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,
-97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,
-16,10,50,0,57,18,109,0,18,110,0,16,10,50,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,
-0,57,18,109,0,18,110,0,16,10,51,0,57,48,20,0,0,1,90,95,0,0,26,0,2,21,1,1,0,0,26,0,109,0,0,1,1,0,0,
-13,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,
-20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,0,1,
-90,95,0,0,14,0,2,21,1,1,0,0,26,0,109,0,0,1,1,0,0,27,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,
-0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,
-0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,
-109,0,18,110,0,16,10,50,0,57,48,20,0,0,1,90,95,0,0,31,0,2,21,1,1,0,0,26,0,109,0,0,1,1,0,0,29,0,110,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,
-18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,18,110,0,16,10,50,0,57,48,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,51,0,57,18,109,0,18,110,0,16,10,51,0,57,48,20,0,0,1,90,95,0,0,28,0,2,21,1,1,
-0,0,28,0,109,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,
-18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,
-16,10,49,0,57,48,20,0,0,1,90,95,0,0,30,0,2,21,1,1,0,0,28,0,109,0,0,1,1,0,0,27,0,110,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,
-101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,
-86,97,108,0,16,10,50,0,57,18,109,0,18,110,0,16,10,50,0,57,48,20,0,0,1,90,95,0,0,15,0,2,21,1,1,0,0,
-28,0,109,0,0,1,1,0,0,29,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,
-110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,
-10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,18,110,0,16,10,50,0,
-57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,18,110,0,16,10,51,0,57,48,20,
-0,0,1,90,95,0,0,13,0,2,21,1,1,0,0,27,0,109,0,0,1,1,0,0,26,0,110,0,0,0,1,9,18,95,95,114,101,116,86,
-97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,
-10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,0,1,90,95,0,0,27,0,2,21,1,1,0,0,27,0,109,0,0,1,
-1,0,0,14,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,
-57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,
-0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,18,110,0,16,10,50,0,57,48,20,0,0,1,90,
-95,0,0,29,0,2,21,1,1,0,0,27,0,109,0,0,1,1,0,0,31,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
-16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,
-57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,
-0,18,110,0,16,10,50,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,18,110,
-0,16,10,51,0,57,48,20,0,0,1,90,95,0,0,26,0,2,21,1,1,0,0,14,0,109,0,0,1,1,0,0,26,0,110,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,
-101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,0,1,90,95,0,0,31,0,2,21,
-1,1,0,0,14,0,109,0,0,1,1,0,0,31,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,
-110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,18,110,0,16,
-10,50,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,18,110,0,16,10,51,0,
-57,48,20,0,0,1,90,95,0,0,28,0,2,21,1,1,0,0,30,0,109,0,0,1,1,0,0,26,0,110,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,
-97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,0,1,90,95,0,0,30,0,2,21,1,1,0,0,30,
-0,109,0,0,1,1,0,0,14,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,
-0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,
-0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,18,110,0,16,10,50,0,57,48,
-20,0,0,1,90,95,0,0,15,0,2,21,1,1,0,0,30,0,109,0,0,1,1,0,0,31,0,110,0,0,0,1,9,18,95,95,114,101,116,
-86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,
-16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,
-0,57,18,109,0,18,110,0,16,10,50,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,
-109,0,18,110,0,16,10,51,0,57,48,20,0,0,1,90,95,0,0,13,0,2,21,1,1,0,0,29,0,109,0,0,1,1,0,0,28,0,110,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,
-18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,0,1,90,95,0,
-0,27,0,2,21,1,1,0,0,29,0,109,0,0,1,1,0,0,30,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,
-48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,
-109,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,18,
-110,0,16,10,50,0,57,48,20,0,0,1,90,95,0,0,29,0,2,21,1,1,0,0,29,0,109,0,0,1,1,0,0,15,0,110,0,0,0,1,
-9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,50,0,57,18,109,0,18,110,0,16,10,50,0,57,48,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,51,0,57,18,109,0,18,110,0,16,10,51,0,57,48,20,0,0,1,90,95,0,0,26,0,2,21,1,1,0,0,31,0,
-109,0,0,1,1,0,0,28,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,
-16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,
-57,48,20,0,0,1,90,95,0,0,14,0,2,21,1,1,0,0,31,0,109,0,0,1,1,0,0,30,0,110,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,
-97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,
-16,10,50,0,57,18,109,0,18,110,0,16,10,50,0,57,48,20,0,0,1,90,95,0,0,31,0,2,21,1,1,0,0,31,0,109,0,0,
-1,1,0,0,15,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,
-0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,
-20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,18,110,0,16,10,50,0,57,48,20,0,9,18,
-95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,18,110,0,16,10,51,0,57,48,20,0,0,1,90,95,0,0,
-28,0,2,21,1,1,0,0,15,0,109,0,0,1,1,0,0,28,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,
-0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,
-109,0,18,110,0,16,10,49,0,57,48,20,0,0,1,90,95,0,0,30,0,2,21,1,1,0,0,15,0,109,0,0,1,1,0,0,30,0,110,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,18,110,0,16,8,48,0,57,48,20,0,9,
-18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,18,110,0,16,10,50,0,57,48,20,0,0,1,90,95,0,0,0,0,2,
-3,1,0,2,0,26,0,109,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,109,0,18,109,0,18,110,0,48,20,0,0,1,90,95,0,0,
-0,0,2,3,1,0,2,0,28,0,109,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,109,0,18,109,0,18,110,0,48,20,0,0,1,90,
-95,0,0,0,0,2,3,1,0,2,0,27,0,109,0,0,1,1,0,0,14,0,110,0,0,0,1,9,18,109,0,18,109,0,18,110,0,48,20,0,
-0,1,90,95,0,0,0,0,2,3,1,0,2,0,30,0,109,0,0,1,1,0,0,14,0,110,0,0,0,1,9,18,109,0,18,109,0,18,110,0,
-48,20,0,0,1,90,95,0,0,0,0,2,3,1,0,2,0,29,0,109,0,0,1,1,0,0,15,0,110,0,0,0,1,9,18,109,0,18,109,0,18,
-110,0,48,20,0,0,1,90,95,0,0,0,0,2,3,1,0,2,0,31,0,109,0,0,1,1,0,0,15,0,110,0,0,0,1,9,18,109,0,18,
-109,0,18,110,0,48,20,0,0,1,90,95,0,0,11,0,2,21,1,1,0,0,10,0,118,0,0,1,1,0,0,27,0,109,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,59,120,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,8,48,0,57,0,0,20,
-0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,49,0,57,
-0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,
-50,0,57,0,0,20,0,0,1,90,95,0,0,12,0,2,21,1,1,0,0,10,0,118,0,0,1,1,0,0,29,0,109,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,59,120,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,8,48,0,57,0,0,20,0,9,18,
-95,95,114,101,116,86,97,108,0,59,121,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,49,0,57,0,0,20,
-0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,50,0,57,
-0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,
-51,0,57,0,0,20,0,0,1,90,95,0,0,10,0,2,21,1,1,0,0,11,0,118,0,0,1,1,0,0,26,0,109,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,59,120,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,8,48,0,57,0,0,20,0,9,18,
-95,95,114,101,116,86,97,108,0,59,121,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,49,0,57,0,0,20,
-0,0,1,90,95,0,0,12,0,2,21,1,1,0,0,11,0,118,0,0,1,1,0,0,31,0,109,0,0,0,1,9,18,95,95,114,101,116,86,
-97,108,0,59,120,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,8,48,0,57,0,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,59,121,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,49,0,57,0,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,59,122,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,50,0,57,0,0,20,0,9,
-18,95,95,114,101,116,86,97,108,0,59,119,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,51,0,57,0,0,
-20,0,0,1,90,95,0,0,10,0,2,21,1,1,0,0,12,0,118,0,0,1,1,0,0,28,0,109,0,0,0,1,9,18,95,95,114,101,116,
-86,97,108,0,59,120,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,8,48,0,57,0,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,59,121,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,49,0,57,0,0,20,0,0,1,90,
-95,0,0,11,0,2,21,1,1,0,0,12,0,118,0,0,1,1,0,0,30,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
-59,120,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,8,48,0,57,0,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,59,121,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,49,0,57,0,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,59,122,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,50,0,57,0,0,20,0,0,1,90,95,0,
-0,0,0,2,1,1,0,2,0,26,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,21,0,9,18,
-109,0,16,10,49,0,57,18,97,0,21,0,0,1,90,95,0,0,0,0,2,1,1,0,2,0,28,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,
-9,18,109,0,16,8,48,0,57,18,97,0,21,0,9,18,109,0,16,10,49,0,57,18,97,0,21,0,0,1,90,95,0,0,0,0,2,1,1,
-0,2,0,27,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,21,0,9,18,109,0,16,10,49,
-0,57,18,97,0,21,0,9,18,109,0,16,10,50,0,57,18,97,0,21,0,0,1,90,95,0,0,0,0,2,1,1,0,2,0,30,0,109,0,0,
-1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,21,0,9,18,109,0,16,10,49,0,57,18,97,0,21,0,
-9,18,109,0,16,10,50,0,57,18,97,0,21,0,0,1,90,95,0,0,0,0,2,1,1,0,2,0,29,0,109,0,0,1,1,0,0,9,0,97,0,
-0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,21,0,9,18,109,0,16,10,49,0,57,18,97,0,21,0,9,18,109,0,16,10,
-50,0,57,18,97,0,21,0,9,18,109,0,16,10,51,0,57,18,97,0,21,0,0,1,90,95,0,0,0,0,2,1,1,0,2,0,31,0,109,
-0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,21,0,9,18,109,0,16,10,49,0,57,18,97,0,
-21,0,9,18,109,0,16,10,50,0,57,18,97,0,21,0,9,18,109,0,16,10,51,0,57,18,97,0,21,0,0,1,90,95,0,0,0,0,
-2,2,1,0,2,0,26,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,22,0,9,18,109,0,16,
-10,49,0,57,18,97,0,22,0,0,1,90,95,0,0,0,0,2,2,1,0,2,0,28,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,
-0,16,8,48,0,57,18,97,0,22,0,9,18,109,0,16,10,49,0,57,18,97,0,22,0,0,1,90,95,0,0,0,0,2,2,1,0,2,0,27,
-0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,22,0,9,18,109,0,16,10,49,0,57,18,
-97,0,22,0,9,18,109,0,16,10,50,0,57,18,97,0,22,0,0,1,90,95,0,0,0,0,2,2,1,0,2,0,30,0,109,0,0,1,1,0,0,
-9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,22,0,9,18,109,0,16,10,49,0,57,18,97,0,22,0,9,18,109,
-0,16,10,50,0,57,18,97,0,22,0,0,1,90,95,0,0,0,0,2,2,1,0,2,0,29,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,
-18,109,0,16,8,48,0,57,18,97,0,22,0,9,18,109,0,16,10,49,0,57,18,97,0,22,0,9,18,109,0,16,10,50,0,57,
-18,97,0,22,0,9,18,109,0,16,10,51,0,57,18,97,0,22,0,0,1,90,95,0,0,0,0,2,2,1,0,2,0,31,0,109,0,0,1,1,
-0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,22,0,9,18,109,0,16,10,49,0,57,18,97,0,22,0,9,18,
-109,0,16,10,50,0,57,18,97,0,22,0,9,18,109,0,16,10,51,0,57,18,97,0,22,0,0,1,90,95,0,0,0,0,2,3,1,0,2,
-0,26,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,23,0,9,18,109,0,16,10,49,0,
-57,18,97,0,23,0,0,1,90,95,0,0,0,0,2,3,1,0,2,0,28,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,
-48,0,57,18,97,0,23,0,9,18,109,0,16,10,49,0,57,18,97,0,23,0,0,1,90,95,0,0,0,0,2,3,1,0,2,0,27,0,109,
-0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,23,0,9,18,109,0,16,10,49,0,57,18,97,0,
-23,0,9,18,109,0,16,10,50,0,57,18,97,0,23,0,0,1,90,95,0,0,0,0,2,3,1,0,2,0,30,0,109,0,0,1,1,0,0,9,0,
-97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,23,0,9,18,109,0,16,10,49,0,57,18,97,0,23,0,9,18,109,0,
-16,10,50,0,57,18,97,0,23,0,0,1,90,95,0,0,0,0,2,3,1,0,2,0,29,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,
-109,0,16,8,48,0,57,18,97,0,23,0,9,18,109,0,16,10,49,0,57,18,97,0,23,0,9,18,109,0,16,10,50,0,57,18,
-97,0,23,0,9,18,109,0,16,10,51,0,57,18,97,0,23,0,0,1,90,95,0,0,0,0,2,3,1,0,2,0,31,0,109,0,0,1,1,0,0,
-9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,23,0,9,18,109,0,16,10,49,0,57,18,97,0,23,0,9,18,109,
-0,16,10,50,0,57,18,97,0,23,0,9,18,109,0,16,10,51,0,57,18,97,0,23,0,0,1,90,95,0,0,0,0,2,4,1,0,2,0,
-26,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,24,0,9,18,109,0,16,10,49,0,57,
-18,97,0,24,0,0,1,90,95,0,0,0,0,2,4,1,0,2,0,28,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,
-0,57,18,97,0,24,0,9,18,109,0,16,10,49,0,57,18,97,0,24,0,0,1,90,95,0,0,0,0,2,4,1,0,2,0,27,0,109,0,0,
-1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,24,0,9,18,109,0,16,10,49,0,57,18,97,0,24,0,
-9,18,109,0,16,10,50,0,57,18,97,0,24,0,0,1,90,95,0,0,0,0,2,4,1,0,2,0,30,0,109,0,0,1,1,0,0,9,0,97,0,
-0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,24,0,9,18,109,0,16,10,49,0,57,18,97,0,24,0,9,18,109,0,16,10,
-50,0,57,18,97,0,24,0,0,1,90,95,0,0,0,0,2,4,1,0,2,0,29,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,109,0,
-16,8,48,0,57,18,97,0,24,0,9,18,109,0,16,10,49,0,57,18,97,0,24,0,9,18,109,0,16,10,50,0,57,18,97,0,
-24,0,9,18,109,0,16,10,51,0,57,18,97,0,24,0,0,1,90,95,0,0,0,0,2,4,1,0,2,0,31,0,109,0,0,1,1,0,0,9,0,
-97,0,0,0,1,9,18,109,0,16,8,48,0,57,18,97,0,24,0,9,18,109,0,16,10,49,0,57,18,97,0,24,0,9,18,109,0,
-16,10,50,0,57,18,97,0,24,0,9,18,109,0,16,10,51,0,57,18,97,0,24,0,0,1,90,95,0,0,26,0,2,26,1,1,0,0,
-26,0,109,0,0,1,1,0,0,26,0,110,0,0,0,1,8,58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,18,110,0,
-16,8,48,0,57,46,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,46,0,0,0,0,1,90,95,0,0,28,0,2,26,1,
-1,0,0,28,0,109,0,0,1,1,0,0,28,0,110,0,0,0,1,8,58,109,97,116,50,120,52,0,0,18,109,0,16,8,48,0,57,18,
-110,0,16,8,48,0,57,46,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,46,0,0,0,0,1,90,95,0,0,27,0,
-2,26,1,1,0,0,27,0,109,0,0,1,1,0,0,27,0,110,0,0,0,1,8,58,109,97,116,51,120,50,0,0,18,109,0,16,8,48,
-0,57,18,110,0,16,8,48,0,57,46,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,46,0,18,109,0,16,10,
-50,0,57,18,110,0,16,10,50,0,57,46,0,0,0,0,1,90,95,0,0,30,0,2,26,1,1,0,0,30,0,109,0,0,1,1,0,0,30,0,
-110,0,0,0,1,8,58,109,97,116,51,120,52,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,46,0,18,109,
-0,16,10,49,0,57,18,110,0,16,10,49,0,57,46,0,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,46,0,0,0,
-0,1,90,95,0,0,29,0,2,26,1,1,0,0,29,0,109,0,0,1,1,0,0,29,0,110,0,0,0,1,8,58,109,97,116,52,120,50,0,
-0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,46,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,
-46,0,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,46,0,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,
-57,46,0,0,0,0,1,90,95,0,0,31,0,2,26,1,1,0,0,31,0,109,0,0,1,1,0,0,31,0,110,0,0,0,1,8,58,109,97,116,
-52,120,51,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,46,0,18,109,0,16,10,49,0,57,18,110,0,16,
-10,49,0,57,46,0,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,46,0,18,109,0,16,10,51,0,57,18,110,0,
-16,10,51,0,57,46,0,0,0,0,1,90,95,0,0,26,0,2,27,1,1,0,0,26,0,109,0,0,1,1,0,0,26,0,110,0,0,0,1,8,58,
-109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,47,0,18,109,0,16,10,49,0,57,
-18,110,0,16,10,49,0,57,47,0,0,0,0,1,90,95,0,0,28,0,2,27,1,1,0,0,28,0,109,0,0,1,1,0,0,28,0,110,0,0,
-0,1,8,58,109,97,116,50,120,52,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,47,0,18,109,0,16,10,
-49,0,57,18,110,0,16,10,49,0,57,47,0,0,0,0,1,90,95,0,0,27,0,2,27,1,1,0,0,27,0,109,0,0,1,1,0,0,27,0,
-110,0,0,0,1,8,58,109,97,116,51,120,50,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,47,0,18,109,
-0,16,10,49,0,57,18,110,0,16,10,49,0,57,47,0,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,47,0,0,0,
-0,1,90,95,0,0,30,0,2,27,1,1,0,0,30,0,109,0,0,1,1,0,0,30,0,110,0,0,0,1,8,58,109,97,116,51,120,52,0,
-0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,47,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,
-47,0,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,47,0,0,0,0,1,90,95,0,0,29,0,2,27,1,1,0,0,29,0,
-109,0,0,1,1,0,0,29,0,110,0,0,0,1,8,58,109,97,116,52,120,50,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,
-48,0,57,47,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,47,0,18,109,0,16,10,50,0,57,18,110,0,16,
-10,50,0,57,47,0,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,47,0,0,0,0,1,90,95,0,0,31,0,2,27,1,1,
-0,0,31,0,109,0,0,1,1,0,0,31,0,110,0,0,0,1,8,58,109,97,116,52,120,51,0,0,18,109,0,16,8,48,0,57,18,
-110,0,16,8,48,0,57,47,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,47,0,18,109,0,16,10,50,0,57,
-18,110,0,16,10,50,0,57,47,0,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,47,0,0,0,0,1,90,95,0,0,
-26,0,2,22,1,1,0,0,26,0,109,0,0,1,1,0,0,26,0,110,0,0,0,1,8,58,109,97,116,50,120,51,0,0,18,109,0,16,
-8,48,0,57,18,110,0,16,8,48,0,57,49,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,49,0,0,0,0,1,90,
-95,0,0,28,0,2,22,1,1,0,0,28,0,109,0,0,1,1,0,0,28,0,110,0,0,0,1,8,58,109,97,116,50,120,52,0,0,18,
-109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,49,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,49,0,0,
-0,0,1,90,95,0,0,27,0,2,22,1,1,0,0,27,0,109,0,0,1,1,0,0,27,0,110,0,0,0,1,8,58,109,97,116,51,120,50,
-0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,49,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,
-49,0,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,49,0,0,0,0,1,90,95,0,0,30,0,2,22,1,1,0,0,30,0,
-109,0,0,1,1,0,0,30,0,110,0,0,0,1,8,58,109,97,116,51,120,52,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,
-48,0,57,49,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,49,0,18,109,0,16,10,50,0,57,18,110,0,16,
-10,50,0,57,49,0,0,0,0,1,90,95,0,0,29,0,2,22,1,1,0,0,29,0,109,0,0,1,1,0,0,29,0,110,0,0,0,1,8,58,109,
-97,116,52,120,50,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,49,0,18,109,0,16,10,49,0,57,18,
-110,0,16,10,49,0,57,49,0,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,49,0,18,109,0,16,10,51,0,57,
-18,110,0,16,10,51,0,57,49,0,0,0,0,1,90,95,0,0,31,0,2,22,1,1,0,0,31,0,109,0,0,1,1,0,0,31,0,110,0,0,
-0,1,8,58,109,97,116,52,120,51,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,49,0,18,109,0,16,10,
-49,0,57,18,110,0,16,10,49,0,57,49,0,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,49,0,18,109,0,16,
-10,51,0,57,18,110,0,16,10,51,0,57,49,0,0,0,0,1,90,95,0,0,26,0,2,26,1,1,0,0,9,0,97,0,0,1,1,0,0,26,0,
-110,0,0,0,1,8,58,109,97,116,50,120,51,0,0,18,97,0,18,110,0,16,8,48,0,57,46,0,18,97,0,18,110,0,16,
-10,49,0,57,46,0,0,0,0,1,90,95,0,0,26,0,2,26,1,1,0,0,26,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,8,58,109,
-97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,18,98,0,46,0,18,109,0,16,10,49,0,57,18,98,0,46,0,0,0,0,
-1,90,95,0,0,28,0,2,26,1,1,0,0,9,0,97,0,0,1,1,0,0,28,0,110,0,0,0,1,8,58,109,97,116,50,120,52,0,0,18,
-97,0,18,110,0,16,8,48,0,57,46,0,18,97,0,18,110,0,16,10,49,0,57,46,0,0,0,0,1,90,95,0,0,28,0,2,26,1,
-1,0,0,28,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,8,58,109,97,116,50,120,52,0,0,18,109,0,16,8,48,0,57,18,
-98,0,46,0,18,109,0,16,10,49,0,57,18,98,0,46,0,0,0,0,1,90,95,0,0,27,0,2,26,1,1,0,0,9,0,97,0,0,1,1,0,
-0,27,0,110,0,0,0,1,8,58,109,97,116,51,120,50,0,0,18,97,0,18,110,0,16,8,48,0,57,46,0,18,97,0,18,110,
-0,16,10,49,0,57,46,0,18,97,0,18,110,0,16,10,50,0,57,46,0,0,0,0,1,90,95,0,0,27,0,2,26,1,1,0,0,27,0,
-109,0,0,1,1,0,0,9,0,98,0,0,0,1,8,58,109,97,116,51,120,50,0,0,18,109,0,16,8,48,0,57,18,98,0,46,0,18,
-109,0,16,10,49,0,57,18,98,0,46,0,18,109,0,16,10,50,0,57,18,98,0,46,0,0,0,0,1,90,95,0,0,30,0,2,26,1,
-1,0,0,9,0,97,0,0,1,1,0,0,30,0,110,0,0,0,1,8,58,109,97,116,51,120,52,0,0,18,97,0,18,110,0,16,8,48,0,
-57,46,0,18,97,0,18,110,0,16,10,49,0,57,46,0,18,97,0,18,110,0,16,10,50,0,57,46,0,0,0,0,1,90,95,0,0,
-30,0,2,26,1,1,0,0,30,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,8,58,109,97,116,51,120,52,0,0,18,109,0,16,8,
-48,0,57,18,98,0,46,0,18,109,0,16,10,49,0,57,18,98,0,46,0,18,109,0,16,10,50,0,57,18,98,0,46,0,0,0,0,
-1,90,95,0,0,29,0,2,26,1,1,0,0,29,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,8,58,109,97,116,52,120,50,0,0,18,
-109,0,16,8,48,0,57,18,98,0,46,0,18,109,0,16,10,49,0,57,18,98,0,46,0,18,109,0,16,10,50,0,57,18,98,0,
-46,0,18,109,0,16,10,51,0,57,18,98,0,46,0,0,0,0,1,90,95,0,0,29,0,2,26,1,1,0,0,9,0,97,0,0,1,1,0,0,29,
-0,110,0,0,0,1,8,58,109,97,116,52,120,50,0,0,18,97,0,18,110,0,16,8,48,0,57,46,0,18,97,0,18,110,0,16,
-10,49,0,57,46,0,18,97,0,18,110,0,16,10,50,0,57,46,0,18,97,0,18,110,0,16,10,51,0,57,46,0,0,0,0,1,90,
-95,0,0,31,0,2,26,1,1,0,0,31,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,8,58,109,97,116,52,120,51,0,0,18,109,
-0,16,8,48,0,57,18,98,0,46,0,18,109,0,16,10,49,0,57,18,98,0,46,0,18,109,0,16,10,50,0,57,18,98,0,46,
-0,18,109,0,16,10,51,0,57,18,98,0,46,0,0,0,0,1,90,95,0,0,31,0,2,26,1,1,0,0,9,0,97,0,0,1,1,0,0,31,0,
-110,0,0,0,1,8,58,109,97,116,52,120,51,0,0,18,97,0,18,110,0,16,8,48,0,57,46,0,18,97,0,18,110,0,16,
-10,49,0,57,46,0,18,97,0,18,110,0,16,10,50,0,57,46,0,18,97,0,18,110,0,16,10,51,0,57,46,0,0,0,0,1,90,
-95,0,0,26,0,2,27,1,1,0,0,9,0,97,0,0,1,1,0,0,26,0,110,0,0,0,1,8,58,109,97,116,50,120,51,0,0,18,97,0,
-18,110,0,16,8,48,0,57,47,0,18,97,0,18,110,0,16,10,49,0,57,47,0,0,0,0,1,90,95,0,0,26,0,2,27,1,1,0,0,
-26,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,8,58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,18,98,0,47,
-0,18,109,0,16,10,49,0,57,18,98,0,47,0,0,0,0,1,90,95,0,0,28,0,2,27,1,1,0,0,9,0,97,0,0,1,1,0,0,28,0,
-110,0,0,0,1,8,58,109,97,116,50,120,52,0,0,18,97,0,18,110,0,16,8,48,0,57,47,0,18,97,0,18,110,0,16,
-10,49,0,57,47,0,0,0,0,1,90,95,0,0,28,0,2,27,1,1,0,0,28,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,8,58,109,
-97,116,50,120,52,0,0,18,109,0,16,8,48,0,57,18,98,0,47,0,18,109,0,16,10,49,0,57,18,98,0,47,0,0,0,0,
-1,90,95,0,0,27,0,2,27,1,1,0,0,9,0,97,0,0,1,1,0,0,27,0,110,0,0,0,1,8,58,109,97,116,51,120,50,0,0,18,
-97,0,18,110,0,16,8,48,0,57,47,0,18,97,0,18,110,0,16,10,49,0,57,47,0,18,97,0,18,110,0,16,10,50,0,57,
-47,0,0,0,0,1,90,95,0,0,27,0,2,27,1,1,0,0,27,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,8,58,109,97,116,51,
-120,50,0,0,18,109,0,16,8,48,0,57,18,98,0,47,0,18,109,0,16,10,49,0,57,18,98,0,47,0,18,109,0,16,10,
-50,0,57,18,98,0,47,0,0,0,0,1,90,95,0,0,30,0,2,27,1,1,0,0,9,0,97,0,0,1,1,0,0,30,0,110,0,0,0,1,8,58,
-109,97,116,51,120,52,0,0,18,97,0,18,110,0,16,8,48,0,57,47,0,18,97,0,18,110,0,16,10,49,0,57,47,0,18,
-97,0,18,110,0,16,10,50,0,57,47,0,0,0,0,1,90,95,0,0,30,0,2,27,1,1,0,0,30,0,109,0,0,1,1,0,0,9,0,98,0,
-0,0,1,8,58,109,97,116,51,120,52,0,0,18,109,0,16,8,48,0,57,18,98,0,47,0,18,109,0,16,10,49,0,57,18,
-98,0,47,0,18,109,0,16,10,50,0,57,18,98,0,47,0,0,0,0,1,90,95,0,0,29,0,2,27,1,1,0,0,29,0,109,0,0,1,1,
-0,0,9,0,98,0,0,0,1,8,58,109,97,116,52,120,50,0,0,18,109,0,16,8,48,0,57,18,98,0,47,0,18,109,0,16,10,
-49,0,57,18,98,0,47,0,18,109,0,16,10,50,0,57,18,98,0,47,0,18,109,0,16,10,51,0,57,18,98,0,47,0,0,0,0,
-1,90,95,0,0,29,0,2,27,1,1,0,0,9,0,97,0,0,1,1,0,0,29,0,110,0,0,0,1,8,58,109,97,116,52,120,50,0,0,18,
-97,0,18,110,0,16,8,48,0,57,47,0,18,97,0,18,110,0,16,10,49,0,57,47,0,18,97,0,18,110,0,16,10,50,0,57,
-47,0,18,97,0,18,110,0,16,10,51,0,57,47,0,0,0,0,1,90,95,0,0,31,0,2,27,1,1,0,0,31,0,109,0,0,1,1,0,0,
-9,0,98,0,0,0,1,8,58,109,97,116,52,120,51,0,0,18,109,0,16,8,48,0,57,18,98,0,47,0,18,109,0,16,10,49,
-0,57,18,98,0,47,0,18,109,0,16,10,50,0,57,18,98,0,47,0,18,109,0,16,10,51,0,57,18,98,0,47,0,0,0,0,1,
-90,95,0,0,31,0,2,27,1,1,0,0,9,0,97,0,0,1,1,0,0,31,0,110,0,0,0,1,8,58,109,97,116,52,120,51,0,0,18,
-97,0,18,110,0,16,8,48,0,57,47,0,18,97,0,18,110,0,16,10,49,0,57,47,0,18,97,0,18,110,0,16,10,50,0,57,
-47,0,18,97,0,18,110,0,16,10,51,0,57,47,0,0,0,0,1,90,95,0,0,26,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,0,26,
-0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,48,20,
-0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,0,16,10,49,0,57,48,20,0,0,1,90,
-95,0,0,26,0,2,21,1,1,0,0,26,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,
-8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,
-18,109,0,16,10,49,0,57,18,98,0,48,20,0,0,1,90,95,0,0,28,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,0,28,0,110,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,48,20,0,9,18,
-95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,0,16,10,49,0,57,48,20,0,0,1,90,95,0,0,
-28,0,2,21,1,1,0,0,28,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,
-57,18,109,0,16,8,48,0,57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,
-16,10,49,0,57,18,98,0,48,20,0,0,1,90,95,0,0,27,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,0,27,0,110,0,0,0,1,
-9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,50,0,57,18,97,0,18,110,0,16,10,50,0,57,48,20,0,0,1,90,95,0,0,27,0,2,21,1,1,0,
-0,27,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,
-8,48,0,57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,
-18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,98,0,48,
-20,0,0,1,90,95,0,0,30,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,0,30,0,110,0,0,0,1,9,18,95,95,114,101,116,86,
-97,108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,
-10,49,0,57,18,97,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,
-18,97,0,18,110,0,16,10,50,0,57,48,20,0,0,1,90,95,0,0,30,0,2,21,1,1,0,0,30,0,109,0,0,1,1,0,0,9,0,98,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,48,20,0,9,18,
-95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,48,20,0,9,18,95,95,114,
-101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,98,0,48,20,0,0,1,90,95,0,0,29,0,2,21,1,
-1,0,0,29,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,
-16,8,48,0,57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,
-57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,98,0,
-48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,16,10,51,0,57,18,98,0,48,20,0,0,
-1,90,95,0,0,29,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,0,29,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,
-0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,
-57,18,97,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,97,0,
-18,110,0,16,10,50,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,97,0,18,110,0,
-16,10,51,0,57,48,20,0,0,1,90,95,0,0,31,0,2,21,1,1,0,0,31,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,
-95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,48,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,
-10,51,0,57,18,109,0,16,10,51,0,57,18,98,0,48,20,0,0,1,90,95,0,0,31,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,
-0,31,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,
-48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,0,16,10,49,0,57,48,20,0,9,
-18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,97,0,18,110,0,16,10,50,0,57,48,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,51,0,57,18,97,0,18,110,0,16,10,51,0,57,48,20,0,0,1,90,95,0,0,26,0,2,
-22,1,1,0,0,9,0,97,0,0,1,1,0,0,26,0,110,0,0,0,1,3,2,90,95,1,0,9,0,1,105,110,118,0,2,17,49,0,48,0,0,
-18,97,0,49,0,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,105,110,118,0,18,110,0,16,8,48,0,
-57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,105,110,118,0,18,110,0,16,10,49,0,
-57,48,20,0,0,1,90,95,0,0,26,0,2,22,1,1,0,0,26,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,1,0,9,0,1,
-105,110,118,0,2,17,49,0,48,0,0,18,98,0,49,0,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-109,0,16,8,48,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,
-109,0,16,10,49,0,57,18,105,110,118,0,48,20,0,0,1,90,95,0,0,28,0,2,22,1,1,0,0,9,0,97,0,0,1,1,0,0,28,
-0,110,0,0,0,1,3,2,90,95,1,0,9,0,1,105,110,118,0,2,17,49,0,48,0,0,18,97,0,49,0,0,9,18,95,95,114,101,
-116,86,97,108,0,16,8,48,0,57,18,105,110,118,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,
-86,97,108,0,16,10,49,0,57,18,105,110,118,0,18,110,0,16,10,49,0,57,48,20,0,0,1,90,95,0,0,28,0,2,22,
-1,1,0,0,28,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,1,0,9,0,1,105,110,118,0,2,17,49,0,48,0,0,18,
-98,0,49,0,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,105,110,118,0,
-48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,105,110,118,0,
-48,20,0,0,1,90,95,0,0,27,0,2,22,1,1,0,0,9,0,97,0,0,1,1,0,0,27,0,110,0,0,0,1,3,2,90,95,1,0,9,0,1,
-105,110,118,0,2,17,49,0,48,0,0,18,97,0,49,0,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-105,110,118,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,
-105,110,118,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,
-105,110,118,0,18,110,0,16,10,50,0,57,48,20,0,0,1,90,95,0,0,27,0,2,22,1,1,0,0,27,0,109,0,0,1,1,0,0,
-9,0,98,0,0,0,1,3,2,90,95,1,0,9,0,1,105,110,118,0,2,17,49,0,48,0,0,18,98,0,49,0,0,9,18,95,95,114,
-101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,105,110,118,0,48,20,0,0,1,90,95,0,0,30,0,2,
-22,1,1,0,0,9,0,97,0,0,1,1,0,0,30,0,110,0,0,0,1,3,2,90,95,1,0,9,0,1,105,110,118,0,2,17,49,0,48,0,0,
-18,97,0,49,0,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,105,110,118,0,18,110,0,16,8,48,0,
-57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,105,110,118,0,18,110,0,16,10,49,0,
-57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,105,110,118,0,18,110,0,16,10,50,0,
-57,48,20,0,0,1,90,95,0,0,30,0,2,22,1,1,0,0,30,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,1,0,9,0,1,
-105,110,118,0,2,17,49,0,48,0,0,18,98,0,49,0,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-109,0,16,8,48,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,
-109,0,16,10,49,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,
-109,0,16,10,50,0,57,18,105,110,118,0,48,20,0,0,1,90,95,0,0,29,0,2,22,1,1,0,0,29,0,109,0,0,1,1,0,0,
-9,0,98,0,0,0,1,3,2,90,95,1,0,9,0,1,105,110,118,0,2,17,49,0,48,0,0,18,98,0,49,0,0,9,18,95,95,114,
-101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,51,0,57,18,109,0,16,10,51,0,57,18,105,110,118,0,48,20,0,0,1,90,95,0,0,29,0,2,
-22,1,1,0,0,9,0,97,0,0,1,1,0,0,29,0,110,0,0,0,1,3,2,90,95,1,0,9,0,1,105,110,118,0,2,17,49,0,48,0,0,
-18,97,0,49,0,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,105,110,118,0,18,110,0,16,8,48,0,
-57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,105,110,118,0,18,110,0,16,10,49,0,
-57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,105,110,118,0,18,110,0,16,10,50,0,
-57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,105,110,118,0,18,110,0,16,10,51,0,
-57,48,20,0,0,1,90,95,0,0,31,0,2,22,1,1,0,0,31,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,1,0,9,0,1,
-105,110,118,0,2,17,49,0,48,0,0,18,98,0,49,0,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-109,0,16,8,48,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,
-109,0,16,10,49,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,
-109,0,16,10,50,0,57,18,105,110,118,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,
-109,0,16,10,51,0,57,18,105,110,118,0,48,20,0,0,1,90,95,0,0,31,0,2,22,1,1,0,0,9,0,97,0,0,1,1,0,0,31,
-0,110,0,0,0,1,3,2,90,95,1,0,9,0,1,105,110,118,0,2,17,49,0,48,0,0,18,97,0,49,0,0,9,18,95,95,114,101,
-116,86,97,108,0,16,8,48,0,57,18,105,110,118,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,
-86,97,108,0,16,10,49,0,57,18,105,110,118,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,
-86,97,108,0,16,10,50,0,57,18,105,110,118,0,18,110,0,16,10,50,0,57,48,20,0,9,18,95,95,114,101,116,
-86,97,108,0,16,10,51,0,57,18,105,110,118,0,18,110,0,16,10,51,0,57,48,20,0,0,1,90,95,0,0,26,0,2,27,
-1,1,0,0,26,0,109,0,0,0,1,8,58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,54,0,18,109,0,16,10,
-49,0,57,54,0,0,0,0,1,90,95,0,0,28,0,2,27,1,1,0,0,28,0,109,0,0,0,1,8,58,109,97,116,50,120,52,0,0,18,
-109,0,16,8,48,0,57,54,0,18,109,0,16,10,49,0,57,54,0,0,0,0,1,90,95,0,0,27,0,2,27,1,1,0,0,27,0,109,0,
-0,0,1,8,58,109,97,116,51,120,50,0,0,18,109,0,16,8,48,0,57,54,0,18,109,0,16,10,49,0,57,54,0,18,109,
-0,16,10,50,0,57,54,0,0,0,0,1,90,95,0,0,30,0,2,27,1,1,0,0,30,0,109,0,0,0,1,8,58,109,97,116,51,120,
-52,0,0,18,109,0,16,8,48,0,57,54,0,18,109,0,16,10,49,0,57,54,0,18,109,0,16,10,50,0,57,54,0,0,0,0,1,
-90,95,0,0,29,0,2,27,1,1,0,0,29,0,109,0,0,0,1,8,58,109,97,116,52,120,50,0,0,18,109,0,16,8,48,0,57,
-54,0,18,109,0,16,10,49,0,57,54,0,18,109,0,16,10,50,0,57,54,0,18,109,0,16,10,51,0,57,54,0,0,0,0,1,
-90,95,0,0,31,0,2,27,1,1,0,0,31,0,109,0,0,0,1,8,58,109,97,116,52,120,51,0,0,18,109,0,16,8,48,0,57,
-54,0,18,109,0,16,10,49,0,57,54,0,18,109,0,16,10,50,0,57,54,0,18,109,0,16,10,51,0,57,54,0,0,0,0,1,
-90,95,0,0,0,0,2,25,1,0,2,0,26,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,52,0,9,18,109,0,16,10,49,0,57,
-52,0,0,1,90,95,0,0,0,0,2,25,1,0,2,0,28,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,52,0,9,18,109,0,16,10,
-49,0,57,52,0,0,1,90,95,0,0,0,0,2,25,1,0,2,0,27,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,52,0,9,18,109,
-0,16,10,49,0,57,52,0,9,18,109,0,16,10,50,0,57,52,0,0,1,90,95,0,0,0,0,2,25,1,0,2,0,30,0,109,0,0,0,1,
-9,18,109,0,16,8,48,0,57,52,0,9,18,109,0,16,10,49,0,57,52,0,9,18,109,0,16,10,50,0,57,52,0,0,1,90,95,
-0,0,0,0,2,25,1,0,2,0,29,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,52,0,9,18,109,0,16,10,49,0,57,52,0,9,
-18,109,0,16,10,50,0,57,52,0,9,18,109,0,16,10,51,0,57,52,0,0,1,90,95,0,0,0,0,2,25,1,0,2,0,31,0,109,
-0,0,0,1,9,18,109,0,16,8,48,0,57,52,0,9,18,109,0,16,10,49,0,57,52,0,9,18,109,0,16,10,50,0,57,52,0,9,
-18,109,0,16,10,51,0,57,52,0,0,1,90,95,0,0,0,0,2,24,1,0,2,0,26,0,109,0,0,0,1,9,18,109,0,16,8,48,0,
-57,51,0,9,18,109,0,16,10,49,0,57,51,0,0,1,90,95,0,0,0,0,2,24,1,0,2,0,28,0,109,0,0,0,1,9,18,109,0,
-16,8,48,0,57,51,0,9,18,109,0,16,10,49,0,57,51,0,0,1,90,95,0,0,0,0,2,24,1,0,2,0,27,0,109,0,0,0,1,9,
-18,109,0,16,8,48,0,57,51,0,9,18,109,0,16,10,49,0,57,51,0,9,18,109,0,16,10,50,0,57,51,0,0,1,90,95,0,
-0,0,0,2,24,1,0,2,0,30,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,51,0,9,18,109,0,16,10,49,0,57,51,0,9,
-18,109,0,16,10,50,0,57,51,0,0,1,90,95,0,0,0,0,2,24,1,0,2,0,29,0,109,0,0,0,1,9,18,109,0,16,8,48,0,
-57,51,0,9,18,109,0,16,10,49,0,57,51,0,9,18,109,0,16,10,50,0,57,51,0,9,18,109,0,16,10,51,0,57,51,0,
-0,1,90,95,0,0,0,0,2,24,1,0,2,0,31,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,51,0,9,18,109,0,16,10,49,0,
-57,51,0,9,18,109,0,16,10,50,0,57,51,0,9,18,109,0,16,10,51,0,57,51,0,0,0
diff --git a/src/mesa/shader/slang/library/slang_builtin_120_common_gc.h b/src/mesa/shader/slang/library/slang_builtin_120_common_gc.h
deleted file mode 100644 (file)
index c397b9f..0000000
+++ /dev/null
@@ -1,108 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE FOLLOWING FILE: */
-/* slang_builtin_120_common.gc */
-
-5,1,90,95,0,0,26,0,0,109,97,116,114,105,120,67,111,109,112,77,117,108,116,0,1,0,0,0,26,0,109,0,0,1,
-0,0,0,26,0,110,0,0,0,1,8,58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,
-48,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,48,0,0,0,0,1,90,95,0,0,28,0,0,109,97,116,114,
-105,120,67,111,109,112,77,117,108,116,0,1,0,0,0,28,0,109,0,0,1,0,0,0,28,0,110,0,0,0,1,8,58,109,97,
-116,50,120,52,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,48,0,18,109,0,16,10,49,0,57,18,110,0,
-16,10,49,0,57,48,0,0,0,0,1,90,95,0,0,27,0,0,109,97,116,114,105,120,67,111,109,112,77,117,108,116,0,
-1,0,0,0,27,0,109,0,0,1,0,0,0,27,0,110,0,0,0,1,8,58,109,97,116,51,120,50,0,0,18,109,0,16,8,48,0,57,
-18,110,0,16,8,48,0,57,48,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,48,0,18,109,0,16,10,50,0,
-57,18,110,0,16,10,50,0,57,48,0,0,0,0,1,90,95,0,0,30,0,0,109,97,116,114,105,120,67,111,109,112,77,
-117,108,116,0,1,0,0,0,30,0,109,0,0,1,0,0,0,30,0,110,0,0,0,1,8,58,109,97,116,51,120,52,0,0,18,109,0,
-16,8,48,0,57,18,110,0,16,8,48,0,57,48,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,48,0,18,109,
-0,16,10,50,0,57,18,110,0,16,10,50,0,57,48,0,0,0,0,1,90,95,0,0,29,0,0,109,97,116,114,105,120,67,111,
-109,112,77,117,108,116,0,1,0,0,0,29,0,109,0,0,1,0,0,0,29,0,110,0,0,0,1,8,58,109,97,116,52,120,50,0,
-0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,48,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,
-48,0,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,48,0,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,
-57,48,0,0,0,0,1,90,95,0,0,31,0,0,109,97,116,114,105,120,67,111,109,112,77,117,108,116,0,1,0,0,0,31,
-0,109,0,0,1,0,0,0,31,0,110,0,0,0,1,8,58,109,97,116,52,120,51,0,0,18,109,0,16,8,48,0,57,18,110,0,16,
-8,48,0,57,48,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,48,0,18,109,0,16,10,50,0,57,18,110,0,
-16,10,50,0,57,48,0,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,48,0,0,0,0,1,90,95,0,0,13,0,0,111,
-117,116,101,114,80,114,111,100,117,99,116,0,1,0,0,0,10,0,99,0,0,1,0,0,0,10,0,114,0,0,0,1,8,58,109,
-97,116,50,0,0,18,99,0,59,120,0,18,114,0,59,120,0,48,0,18,99,0,59,121,0,18,114,0,59,120,0,48,0,18,
-99,0,59,120,0,18,114,0,59,121,0,48,0,18,99,0,59,121,0,18,114,0,59,121,0,48,0,0,0,0,1,90,95,0,0,14,
-0,0,111,117,116,101,114,80,114,111,100,117,99,116,0,1,0,0,0,11,0,99,0,0,1,0,0,0,11,0,114,0,0,0,1,8,
-58,109,97,116,51,0,0,18,99,0,59,120,0,18,114,0,59,120,0,48,0,18,99,0,59,121,0,18,114,0,59,120,0,48,
-0,18,99,0,59,122,0,18,114,0,59,120,0,48,0,18,99,0,59,120,0,18,114,0,59,121,0,48,0,18,99,0,59,121,0,
-18,114,0,59,121,0,48,0,18,99,0,59,122,0,18,114,0,59,121,0,48,0,18,99,0,59,120,0,18,114,0,59,122,0,
-48,0,18,99,0,59,121,0,18,114,0,59,122,0,48,0,18,99,0,59,122,0,18,114,0,59,122,0,48,0,0,0,0,1,90,95,
-0,0,15,0,0,111,117,116,101,114,80,114,111,100,117,99,116,0,1,0,0,0,12,0,99,0,0,1,0,0,0,12,0,114,0,
-0,0,1,8,58,109,97,116,52,0,0,18,99,0,59,120,0,18,114,0,59,120,0,48,0,18,99,0,59,121,0,18,114,0,59,
-120,0,48,0,18,99,0,59,122,0,18,114,0,59,120,0,48,0,18,99,0,59,119,0,18,114,0,59,120,0,48,0,18,99,0,
-59,120,0,18,114,0,59,121,0,48,0,18,99,0,59,121,0,18,114,0,59,121,0,48,0,18,99,0,59,122,0,18,114,0,
-59,121,0,48,0,18,99,0,59,119,0,18,114,0,59,121,0,48,0,18,99,0,59,120,0,18,114,0,59,122,0,48,0,18,
-99,0,59,121,0,18,114,0,59,122,0,48,0,18,99,0,59,122,0,18,114,0,59,122,0,48,0,18,99,0,59,119,0,18,
-114,0,59,122,0,48,0,18,99,0,59,120,0,18,114,0,59,119,0,48,0,18,99,0,59,121,0,18,114,0,59,119,0,48,
-0,18,99,0,59,122,0,18,114,0,59,119,0,48,0,18,99,0,59,119,0,18,114,0,59,119,0,48,0,0,0,0,1,90,95,0,
-0,26,0,0,111,117,116,101,114,80,114,111,100,117,99,116,0,1,0,0,0,11,0,99,0,0,1,0,0,0,10,0,114,0,0,
-0,1,8,58,109,97,116,50,120,51,0,0,18,99,0,59,120,0,18,114,0,59,120,0,48,0,18,99,0,59,121,0,18,114,
-0,59,120,0,48,0,18,99,0,59,122,0,18,114,0,59,120,0,48,0,18,99,0,59,120,0,18,114,0,59,121,0,48,0,18,
-99,0,59,121,0,18,114,0,59,121,0,48,0,18,99,0,59,122,0,18,114,0,59,121,0,48,0,0,0,0,1,90,95,0,0,27,
-0,0,111,117,116,101,114,80,114,111,100,117,99,116,0,1,0,0,0,10,0,99,0,0,1,0,0,0,11,0,114,0,0,0,1,8,
-58,109,97,116,51,120,50,0,0,18,99,0,59,120,0,18,114,0,59,120,0,48,0,18,99,0,59,121,0,18,114,0,59,
-120,0,48,0,18,99,0,59,120,0,18,114,0,59,121,0,48,0,18,99,0,59,121,0,18,114,0,59,121,0,48,0,18,99,0,
-59,120,0,18,114,0,59,122,0,48,0,18,99,0,59,121,0,18,114,0,59,122,0,48,0,0,0,0,1,90,95,0,0,28,0,0,
-111,117,116,101,114,80,114,111,100,117,99,116,0,1,0,0,0,12,0,99,0,0,1,0,0,0,10,0,114,0,0,0,1,8,58,
-109,97,116,50,120,52,0,0,18,99,0,59,120,0,18,114,0,59,120,0,48,0,18,99,0,59,121,0,18,114,0,59,120,
-0,48,0,18,99,0,59,122,0,18,114,0,59,120,0,48,0,18,99,0,59,119,0,18,114,0,59,120,0,48,0,18,99,0,59,
-120,0,18,114,0,59,121,0,48,0,18,99,0,59,121,0,18,114,0,59,121,0,48,0,18,99,0,59,122,0,18,114,0,59,
-121,0,48,0,18,99,0,59,119,0,18,114,0,59,121,0,48,0,0,0,0,1,90,95,0,0,29,0,0,111,117,116,101,114,80,
-114,111,100,117,99,116,0,1,0,0,0,10,0,99,0,0,1,0,0,0,12,0,114,0,0,0,1,8,58,109,97,116,52,120,50,0,
-0,18,99,0,59,120,0,18,114,0,59,120,0,48,0,18,99,0,59,121,0,18,114,0,59,120,0,48,0,18,99,0,59,120,0,
-18,114,0,59,121,0,48,0,18,99,0,59,121,0,18,114,0,59,121,0,48,0,18,99,0,59,120,0,18,114,0,59,122,0,
-48,0,18,99,0,59,121,0,18,114,0,59,122,0,48,0,18,99,0,59,120,0,18,114,0,59,119,0,48,0,18,99,0,59,
-121,0,18,114,0,59,119,0,48,0,0,0,0,1,90,95,0,0,30,0,0,111,117,116,101,114,80,114,111,100,117,99,
-116,0,1,0,0,0,12,0,99,0,0,1,0,0,0,11,0,114,0,0,0,1,8,58,109,97,116,51,120,52,0,0,18,99,0,59,120,0,
-18,114,0,59,120,0,48,0,18,99,0,59,121,0,18,114,0,59,120,0,48,0,18,99,0,59,122,0,18,114,0,59,120,0,
-48,0,18,99,0,59,119,0,18,114,0,59,120,0,48,0,18,99,0,59,120,0,18,114,0,59,121,0,48,0,18,99,0,59,
-121,0,18,114,0,59,121,0,48,0,18,99,0,59,122,0,18,114,0,59,121,0,48,0,18,99,0,59,119,0,18,114,0,59,
-121,0,48,0,18,99,0,59,120,0,18,114,0,59,122,0,48,0,18,99,0,59,121,0,18,114,0,59,122,0,48,0,18,99,0,
-59,122,0,18,114,0,59,122,0,48,0,18,99,0,59,119,0,18,114,0,59,122,0,48,0,0,0,0,1,90,95,0,0,31,0,0,
-111,117,116,101,114,80,114,111,100,117,99,116,0,1,0,0,0,11,0,99,0,0,1,0,0,0,12,0,114,0,0,0,1,8,58,
-109,97,116,52,120,51,0,0,18,99,0,59,120,0,18,114,0,59,120,0,48,0,18,99,0,59,121,0,18,114,0,59,120,
-0,48,0,18,99,0,59,122,0,18,114,0,59,120,0,48,0,18,99,0,59,120,0,18,114,0,59,121,0,48,0,18,99,0,59,
-121,0,18,114,0,59,121,0,48,0,18,99,0,59,122,0,18,114,0,59,121,0,48,0,18,99,0,59,120,0,18,114,0,59,
-122,0,48,0,18,99,0,59,121,0,18,114,0,59,122,0,48,0,18,99,0,59,122,0,18,114,0,59,122,0,48,0,18,99,0,
-59,120,0,18,114,0,59,119,0,48,0,18,99,0,59,121,0,18,114,0,59,119,0,48,0,18,99,0,59,122,0,18,114,0,
-59,119,0,48,0,0,0,0,1,90,95,0,0,13,0,0,116,114,97,110,115,112,111,115,101,0,1,0,0,0,13,0,109,0,0,0,
-1,8,58,109,97,116,50,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,
-0,16,8,48,0,57,59,121,0,0,18,109,0,16,10,49,0,57,59,121,0,0,0,0,0,1,90,95,0,0,14,0,0,116,114,97,
-110,115,112,111,115,101,0,1,0,0,0,14,0,109,0,0,0,1,8,58,109,97,116,51,0,0,18,109,0,16,8,48,0,57,59,
-120,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,50,0,57,59,120,0,0,18,109,0,16,8,48,0,57,
-59,121,0,0,18,109,0,16,10,49,0,57,59,121,0,0,18,109,0,16,10,50,0,57,59,121,0,0,18,109,0,16,8,48,0,
-57,59,122,0,0,18,109,0,16,10,49,0,57,59,122,0,0,18,109,0,16,10,50,0,57,59,122,0,0,0,0,0,1,90,95,0,
-0,15,0,0,116,114,97,110,115,112,111,115,101,0,1,0,0,0,15,0,109,0,0,0,1,8,58,109,97,116,52,0,0,18,
-109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,50,0,57,59,120,0,0,
-18,109,0,16,10,51,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,0,16,10,49,0,57,59,121,0,
-0,18,109,0,16,10,50,0,57,59,121,0,0,18,109,0,16,10,51,0,57,59,121,0,0,18,109,0,16,8,48,0,57,59,122,
-0,0,18,109,0,16,10,49,0,57,59,122,0,0,18,109,0,16,10,50,0,57,59,122,0,0,18,109,0,16,10,51,0,57,59,
-122,0,0,18,109,0,16,8,48,0,57,59,119,0,0,18,109,0,16,10,49,0,57,59,119,0,0,18,109,0,16,10,50,0,57,
-59,119,0,0,18,109,0,16,10,51,0,57,59,119,0,0,0,0,0,1,90,95,0,0,26,0,0,116,114,97,110,115,112,111,
-115,101,0,1,0,0,0,27,0,109,0,0,0,1,8,58,109,97,116,50,120,51,0,0,18,109,0,16,8,48,0,57,59,120,0,0,
-18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,50,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,
-0,18,109,0,16,10,49,0,57,59,121,0,0,18,109,0,16,10,50,0,57,59,121,0,0,0,0,0,1,90,95,0,0,27,0,0,116,
-114,97,110,115,112,111,115,101,0,1,0,0,0,26,0,109,0,0,0,1,8,58,109,97,116,51,120,50,0,0,18,109,0,
-16,8,48,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,
-0,16,10,49,0,57,59,121,0,0,18,109,0,16,8,48,0,57,59,122,0,0,18,109,0,16,10,49,0,57,59,122,0,0,0,0,
-0,1,90,95,0,0,28,0,0,116,114,97,110,115,112,111,115,101,0,1,0,0,0,29,0,109,0,0,0,1,8,58,109,97,116,
-50,120,52,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,50,
-0,57,59,120,0,0,18,109,0,16,10,51,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,0,16,10,
-49,0,57,59,121,0,0,18,109,0,16,10,50,0,57,59,121,0,0,18,109,0,16,10,51,0,57,59,121,0,0,0,0,0,1,90,
-95,0,0,29,0,0,116,114,97,110,115,112,111,115,101,0,1,0,0,0,28,0,109,0,0,0,1,8,58,109,97,116,52,120,
-50,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,
-121,0,0,18,109,0,16,10,49,0,57,59,121,0,0,18,109,0,16,8,48,0,57,59,122,0,0,18,109,0,16,10,49,0,57,
-59,122,0,0,18,109,0,16,8,48,0,57,59,119,0,0,18,109,0,16,10,49,0,57,59,119,0,0,0,0,0,1,90,95,0,0,30,
-0,0,116,114,97,110,115,112,111,115,101,0,1,0,0,0,31,0,109,0,0,0,1,8,58,109,97,116,51,120,52,0,0,18,
-109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,109,0,16,10,50,0,57,59,120,0,0,
-18,109,0,16,10,51,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,0,16,10,49,0,57,59,121,0,
-0,18,109,0,16,10,50,0,57,59,121,0,0,18,109,0,16,10,51,0,57,59,121,0,0,18,109,0,16,8,48,0,57,59,122,
-0,0,18,109,0,16,10,49,0,57,59,122,0,0,18,109,0,16,10,50,0,57,59,122,0,0,18,109,0,16,10,51,0,57,59,
-122,0,0,0,0,0,1,90,95,0,0,31,0,0,116,114,97,110,115,112,111,115,101,0,1,0,0,0,30,0,109,0,0,0,1,8,
-58,109,97,116,52,120,51,0,0,18,109,0,16,8,48,0,57,59,120,0,0,18,109,0,16,10,49,0,57,59,120,0,0,18,
-109,0,16,10,50,0,57,59,120,0,0,18,109,0,16,8,48,0,57,59,121,0,0,18,109,0,16,10,49,0,57,59,121,0,0,
-18,109,0,16,10,50,0,57,59,121,0,0,18,109,0,16,8,48,0,57,59,122,0,0,18,109,0,16,10,49,0,57,59,122,0,
-0,18,109,0,16,10,50,0,57,59,122,0,0,18,109,0,16,8,48,0,57,59,119,0,0,18,109,0,16,10,49,0,57,59,119,
-0,0,18,109,0,16,10,50,0,57,59,119,0,0,0,0,0,0
diff --git a/src/mesa/shader/slang/library/slang_builtin_120_fragment_gc.h b/src/mesa/shader/slang/library/slang_builtin_120_fragment_gc.h
deleted file mode 100644 (file)
index add3b5a..0000000
+++ /dev/null
@@ -1,5 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE FOLLOWING FILE: */
-/* slang_builtin_120_fragment.gc */
-
-5,2,2,90,95,3,0,10,0,1,103,108,95,80,111,105,110,116,67,111,111,114,100,0,0,0,0
diff --git a/src/mesa/shader/slang/library/slang_common_builtin_gc.h b/src/mesa/shader/slang/library/slang_common_builtin_gc.h
deleted file mode 100644 (file)
index 3c3666e..0000000
+++ /dev/null
@@ -1,880 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE FOLLOWING FILE: */
-/* slang_common_builtin.gc */
-
-5,2,2,90,95,1,0,5,0,1,103,108,95,77,97,120,76,105,103,104,116,115,0,2,16,10,56,0,0,0,2,2,90,95,1,0,
-5,0,1,103,108,95,77,97,120,67,108,105,112,80,108,97,110,101,115,0,2,16,10,54,0,0,0,2,2,90,95,1,0,5,
-0,1,103,108,95,77,97,120,84,101,120,116,117,114,101,85,110,105,116,115,0,2,16,10,56,0,0,0,2,2,90,
-95,1,0,5,0,1,103,108,95,77,97,120,84,101,120,116,117,114,101,67,111,111,114,100,115,0,2,16,10,56,0,
-0,0,2,2,90,95,1,0,5,0,1,103,108,95,77,97,120,86,101,114,116,101,120,65,116,116,114,105,98,115,0,2,
-16,10,49,54,0,0,0,2,2,90,95,1,0,5,0,1,103,108,95,77,97,120,86,101,114,116,101,120,85,110,105,102,
-111,114,109,67,111,109,112,111,110,101,110,116,115,0,2,16,10,53,49,50,0,0,0,2,2,90,95,1,0,5,0,1,
-103,108,95,77,97,120,86,97,114,121,105,110,103,70,108,111,97,116,115,0,2,16,10,51,50,0,0,0,2,2,90,
-95,1,0,5,0,1,103,108,95,77,97,120,86,101,114,116,101,120,84,101,120,116,117,114,101,73,109,97,103,
-101,85,110,105,116,115,0,2,16,8,48,0,0,0,2,2,90,95,1,0,5,0,1,103,108,95,77,97,120,67,111,109,98,
-105,110,101,100,84,101,120,116,117,114,101,73,109,97,103,101,85,110,105,116,115,0,2,16,10,50,0,0,0,
-2,2,90,95,1,0,5,0,1,103,108,95,77,97,120,84,101,120,116,117,114,101,73,109,97,103,101,85,110,105,
-116,115,0,2,16,10,50,0,0,0,2,2,90,95,1,0,5,0,1,103,108,95,77,97,120,70,114,97,103,109,101,110,116,
-85,110,105,102,111,114,109,67,111,109,112,111,110,101,110,116,115,0,2,16,10,54,52,0,0,0,2,2,90,95,
-1,0,5,0,1,103,108,95,77,97,120,68,114,97,119,66,117,102,102,101,114,115,0,2,16,10,49,0,0,0,2,2,90,
-95,4,0,15,0,1,103,108,95,77,111,100,101,108,86,105,101,119,77,97,116,114,105,120,0,0,0,2,2,90,95,4,
-0,15,0,1,103,108,95,80,114,111,106,101,99,116,105,111,110,77,97,116,114,105,120,0,0,0,2,2,90,95,4,
-0,15,0,1,103,108,95,77,111,100,101,108,86,105,101,119,80,114,111,106,101,99,116,105,111,110,77,97,
-116,114,105,120,0,0,0,2,2,90,95,4,0,15,0,1,103,108,95,84,101,120,116,117,114,101,77,97,116,114,105,
-120,0,3,18,103,108,95,77,97,120,84,101,120,116,117,114,101,67,111,111,114,100,115,0,0,0,2,2,90,95,
-4,0,14,0,1,103,108,95,78,111,114,109,97,108,77,97,116,114,105,120,0,0,0,2,2,90,95,4,0,15,0,1,103,
-108,95,77,111,100,101,108,86,105,101,119,77,97,116,114,105,120,73,110,118,101,114,115,101,0,0,0,2,
-2,90,95,4,0,15,0,1,103,108,95,80,114,111,106,101,99,116,105,111,110,77,97,116,114,105,120,73,110,
-118,101,114,115,101,0,0,0,2,2,90,95,4,0,15,0,1,103,108,95,77,111,100,101,108,86,105,101,119,80,114,
-111,106,101,99,116,105,111,110,77,97,116,114,105,120,73,110,118,101,114,115,101,0,0,0,2,2,90,95,4,
-0,15,0,1,103,108,95,84,101,120,116,117,114,101,77,97,116,114,105,120,73,110,118,101,114,115,101,0,
-3,18,103,108,95,77,97,120,84,101,120,116,117,114,101,67,111,111,114,100,115,0,0,0,2,2,90,95,4,0,15,
-0,1,103,108,95,77,111,100,101,108,86,105,101,119,77,97,116,114,105,120,84,114,97,110,115,112,111,
-115,101,0,0,0,2,2,90,95,4,0,15,0,1,103,108,95,80,114,111,106,101,99,116,105,111,110,77,97,116,114,
-105,120,84,114,97,110,115,112,111,115,101,0,0,0,2,2,90,95,4,0,15,0,1,103,108,95,77,111,100,101,108,
-86,105,101,119,80,114,111,106,101,99,116,105,111,110,77,97,116,114,105,120,84,114,97,110,115,112,
-111,115,101,0,0,0,2,2,90,95,4,0,15,0,1,103,108,95,84,101,120,116,117,114,101,77,97,116,114,105,120,
-84,114,97,110,115,112,111,115,101,0,3,18,103,108,95,77,97,120,84,101,120,116,117,114,101,67,111,
-111,114,100,115,0,0,0,2,2,90,95,4,0,15,0,1,103,108,95,77,111,100,101,108,86,105,101,119,77,97,116,
-114,105,120,73,110,118,101,114,115,101,84,114,97,110,115,112,111,115,101,0,0,0,2,2,90,95,4,0,15,0,
-1,103,108,95,80,114,111,106,101,99,116,105,111,110,77,97,116,114,105,120,73,110,118,101,114,115,
-101,84,114,97,110,115,112,111,115,101,0,0,0,2,2,90,95,4,0,15,0,1,103,108,95,77,111,100,101,108,86,
-105,101,119,80,114,111,106,101,99,116,105,111,110,77,97,116,114,105,120,73,110,118,101,114,115,101,
-84,114,97,110,115,112,111,115,101,0,0,0,2,2,90,95,4,0,15,0,1,103,108,95,84,101,120,116,117,114,101,
-77,97,116,114,105,120,73,110,118,101,114,115,101,84,114,97,110,115,112,111,115,101,0,3,18,103,108,
-95,77,97,120,84,101,120,116,117,114,101,67,111,111,114,100,115,0,0,0,2,2,90,95,4,0,9,0,1,103,108,
-95,78,111,114,109,97,108,83,99,97,108,101,0,0,0,2,2,90,95,0,0,24,103,108,95,68,101,112,116,104,82,
-97,110,103,101,80,97,114,97,109,101,116,101,114,115,0,9,0,110,101,97,114,0,0,0,1,9,0,102,97,114,0,
-0,0,1,9,0,100,105,102,102,0,0,0,0,0,0,0,2,2,90,95,4,0,25,103,108,95,68,101,112,116,104,82,97,110,
-103,101,80,97,114,97,109,101,116,101,114,115,0,0,1,103,108,95,68,101,112,116,104,82,97,110,103,101,
-0,0,0,2,2,90,95,4,0,12,0,1,103,108,95,67,108,105,112,80,108,97,110,101,0,3,18,103,108,95,77,97,120,
-67,108,105,112,80,108,97,110,101,115,0,0,0,2,2,90,95,0,0,24,103,108,95,80,111,105,110,116,80,97,
-114,97,109,101,116,101,114,115,0,9,0,115,105,122,101,0,0,0,1,9,0,115,105,122,101,77,105,110,0,0,0,
-1,9,0,115,105,122,101,77,97,120,0,0,0,1,9,0,102,97,100,101,84,104,114,101,115,104,111,108,100,83,
-105,122,101,0,0,0,1,9,0,100,105,115,116,97,110,99,101,67,111,110,115,116,97,110,116,65,116,116,101,
-110,117,97,116,105,111,110,0,0,0,1,9,0,100,105,115,116,97,110,99,101,76,105,110,101,97,114,65,116,
-116,101,110,117,97,116,105,111,110,0,0,0,1,9,0,100,105,115,116,97,110,99,101,81,117,97,100,114,97,
-116,105,99,65,116,116,101,110,117,97,116,105,111,110,0,0,0,0,0,0,0,2,2,90,95,4,0,25,103,108,95,80,
-111,105,110,116,80,97,114,97,109,101,116,101,114,115,0,0,1,103,108,95,80,111,105,110,116,0,0,0,2,2,
-90,95,0,0,24,103,108,95,77,97,116,101,114,105,97,108,80,97,114,97,109,101,116,101,114,115,0,12,0,
-101,109,105,115,115,105,111,110,0,0,0,1,12,0,97,109,98,105,101,110,116,0,0,0,1,12,0,100,105,102,
-102,117,115,101,0,0,0,1,12,0,115,112,101,99,117,108,97,114,0,0,0,1,9,0,115,104,105,110,105,110,101,
-115,115,0,0,0,0,0,0,0,2,2,90,95,4,0,25,103,108,95,77,97,116,101,114,105,97,108,80,97,114,97,109,
-101,116,101,114,115,0,0,1,103,108,95,70,114,111,110,116,77,97,116,101,114,105,97,108,0,0,0,2,2,90,
-95,4,0,25,103,108,95,77,97,116,101,114,105,97,108,80,97,114,97,109,101,116,101,114,115,0,0,1,103,
-108,95,66,97,99,107,77,97,116,101,114,105,97,108,0,0,0,2,2,90,95,0,0,24,103,108,95,76,105,103,104,
-116,83,111,117,114,99,101,80,97,114,97,109,101,116,101,114,115,0,12,0,97,109,98,105,101,110,116,0,
-0,0,1,12,0,100,105,102,102,117,115,101,0,0,0,1,12,0,115,112,101,99,117,108,97,114,0,0,0,1,12,0,112,
-111,115,105,116,105,111,110,0,0,0,1,12,0,104,97,108,102,86,101,99,116,111,114,0,0,0,1,11,0,115,112,
-111,116,68,105,114,101,99,116,105,111,110,0,0,0,1,9,0,115,112,111,116,67,111,115,67,117,116,111,
-102,102,0,0,0,1,9,0,99,111,110,115,116,97,110,116,65,116,116,101,110,117,97,116,105,111,110,0,0,0,
-1,9,0,108,105,110,101,97,114,65,116,116,101,110,117,97,116,105,111,110,0,0,0,1,9,0,113,117,97,100,
-114,97,116,105,99,65,116,116,101,110,117,97,116,105,111,110,0,0,0,1,9,0,115,112,111,116,69,120,112,
-111,110,101,110,116,0,0,0,1,9,0,115,112,111,116,67,117,116,111,102,102,0,0,0,0,0,0,0,2,2,90,95,4,0,
-25,103,108,95,76,105,103,104,116,83,111,117,114,99,101,80,97,114,97,109,101,116,101,114,115,0,0,1,
-103,108,95,76,105,103,104,116,83,111,117,114,99,101,0,3,18,103,108,95,77,97,120,76,105,103,104,116,
-115,0,0,0,2,2,90,95,0,0,24,103,108,95,76,105,103,104,116,77,111,100,101,108,80,97,114,97,109,101,
-116,101,114,115,0,12,0,97,109,98,105,101,110,116,0,0,0,0,0,0,0,2,2,90,95,4,0,25,103,108,95,76,105,
-103,104,116,77,111,100,101,108,80,97,114,97,109,101,116,101,114,115,0,0,1,103,108,95,76,105,103,
-104,116,77,111,100,101,108,0,0,0,2,2,90,95,0,0,24,103,108,95,76,105,103,104,116,77,111,100,101,108,
-80,114,111,100,117,99,116,115,0,12,0,115,99,101,110,101,67,111,108,111,114,0,0,0,0,0,0,0,2,2,90,95,
-4,0,25,103,108,95,76,105,103,104,116,77,111,100,101,108,80,114,111,100,117,99,116,115,0,0,1,103,
-108,95,70,114,111,110,116,76,105,103,104,116,77,111,100,101,108,80,114,111,100,117,99,116,0,0,0,2,
-2,90,95,4,0,25,103,108,95,76,105,103,104,116,77,111,100,101,108,80,114,111,100,117,99,116,115,0,0,
-1,103,108,95,66,97,99,107,76,105,103,104,116,77,111,100,101,108,80,114,111,100,117,99,116,0,0,0,2,
-2,90,95,0,0,24,103,108,95,76,105,103,104,116,80,114,111,100,117,99,116,115,0,12,0,97,109,98,105,
-101,110,116,0,0,0,1,12,0,100,105,102,102,117,115,101,0,0,0,1,12,0,115,112,101,99,117,108,97,114,0,
-0,0,0,0,0,0,2,2,90,95,4,0,25,103,108,95,76,105,103,104,116,80,114,111,100,117,99,116,115,0,0,1,103,
-108,95,70,114,111,110,116,76,105,103,104,116,80,114,111,100,117,99,116,0,3,18,103,108,95,77,97,120,
-76,105,103,104,116,115,0,0,0,2,2,90,95,4,0,25,103,108,95,76,105,103,104,116,80,114,111,100,117,99,
-116,115,0,0,1,103,108,95,66,97,99,107,76,105,103,104,116,80,114,111,100,117,99,116,0,3,18,103,108,
-95,77,97,120,76,105,103,104,116,115,0,0,0,2,2,90,95,4,0,12,0,1,103,108,95,84,101,120,116,117,114,
-101,69,110,118,67,111,108,111,114,0,3,18,103,108,95,77,97,120,84,101,120,116,117,114,101,73,109,97,
-103,101,85,110,105,116,115,0,0,0,2,2,90,95,4,0,12,0,1,103,108,95,69,121,101,80,108,97,110,101,83,0,
-3,18,103,108,95,77,97,120,84,101,120,116,117,114,101,67,111,111,114,100,115,0,0,0,2,2,90,95,4,0,12,
-0,1,103,108,95,69,121,101,80,108,97,110,101,84,0,3,18,103,108,95,77,97,120,84,101,120,116,117,114,
-101,67,111,111,114,100,115,0,0,0,2,2,90,95,4,0,12,0,1,103,108,95,69,121,101,80,108,97,110,101,82,0,
-3,18,103,108,95,77,97,120,84,101,120,116,117,114,101,67,111,111,114,100,115,0,0,0,2,2,90,95,4,0,12,
-0,1,103,108,95,69,121,101,80,108,97,110,101,81,0,3,18,103,108,95,77,97,120,84,101,120,116,117,114,
-101,67,111,111,114,100,115,0,0,0,2,2,90,95,4,0,12,0,1,103,108,95,79,98,106,101,99,116,80,108,97,
-110,101,83,0,3,18,103,108,95,77,97,120,84,101,120,116,117,114,101,67,111,111,114,100,115,0,0,0,2,2,
-90,95,4,0,12,0,1,103,108,95,79,98,106,101,99,116,80,108,97,110,101,84,0,3,18,103,108,95,77,97,120,
-84,101,120,116,117,114,101,67,111,111,114,100,115,0,0,0,2,2,90,95,4,0,12,0,1,103,108,95,79,98,106,
-101,99,116,80,108,97,110,101,82,0,3,18,103,108,95,77,97,120,84,101,120,116,117,114,101,67,111,111,
-114,100,115,0,0,0,2,2,90,95,4,0,12,0,1,103,108,95,79,98,106,101,99,116,80,108,97,110,101,81,0,3,18,
-103,108,95,77,97,120,84,101,120,116,117,114,101,67,111,111,114,100,115,0,0,0,2,2,90,95,0,0,24,103,
-108,95,70,111,103,80,97,114,97,109,101,116,101,114,115,0,12,0,99,111,108,111,114,0,0,0,1,9,0,100,
-101,110,115,105,116,121,0,0,0,1,9,0,115,116,97,114,116,0,0,0,1,9,0,101,110,100,0,0,0,1,9,0,115,99,
-97,108,101,0,0,0,0,0,0,0,2,2,90,95,4,0,25,103,108,95,70,111,103,80,97,114,97,109,101,116,101,114,
-115,0,0,1,103,108,95,70,111,103,0,0,0,1,90,95,0,0,9,0,0,114,97,100,105,97,110,115,0,1,1,0,0,9,0,
-100,101,103,0,0,0,1,3,2,90,95,1,0,9,0,1,99,0,2,17,51,0,49,52,49,53,57,50,54,0,0,17,49,56,48,0,48,0,
-0,49,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,
-18,100,101,103,0,0,18,99,0,0,0,0,1,90,95,0,0,10,0,0,114,97,100,105,97,110,115,0,1,1,0,0,10,0,100,
-101,103,0,0,0,1,3,2,90,95,1,0,9,0,1,99,0,2,17,51,0,49,52,49,53,57,50,54,0,0,17,49,56,48,0,48,0,0,
-49,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,
-120,121,0,0,18,100,101,103,0,59,120,121,0,0,18,99,0,59,120,120,0,0,0,0,1,90,95,0,0,11,0,0,114,97,
-100,105,97,110,115,0,1,1,0,0,11,0,100,101,103,0,0,0,1,3,2,90,95,1,0,9,0,1,99,0,2,17,51,0,49,52,49,
-53,57,50,54,0,0,17,49,56,48,0,48,0,0,49,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,
-18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,100,101,103,0,59,120,121,122,0,0,18,99,0,59,
-120,120,120,0,0,0,0,1,90,95,0,0,12,0,0,114,97,100,105,97,110,115,0,1,1,0,0,12,0,100,101,103,0,0,0,
-1,3,2,90,95,1,0,9,0,1,99,0,2,17,51,0,49,52,49,53,57,50,54,0,0,17,49,56,48,0,48,0,0,49,0,0,4,118,
-101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,18,100,101,103,0,
-0,18,99,0,59,120,120,120,120,0,0,0,0,1,90,95,0,0,9,0,0,100,101,103,114,101,101,115,0,1,1,0,0,9,0,
-114,97,100,0,0,0,1,3,2,90,95,1,0,9,0,1,99,0,2,17,49,56,48,0,48,0,0,17,51,0,49,52,49,53,57,50,54,0,
-0,49,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,
-18,114,97,100,0,0,18,99,0,0,0,0,1,90,95,0,0,10,0,0,100,101,103,114,101,101,115,0,1,1,0,0,10,0,114,
-97,100,0,0,0,1,3,2,90,95,1,0,9,0,1,99,0,2,17,49,56,48,0,48,0,0,17,51,0,49,52,49,53,57,50,54,0,0,49,
-0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,120,
-121,0,0,18,114,97,100,0,59,120,121,0,0,18,99,0,59,120,120,0,0,0,0,1,90,95,0,0,11,0,0,100,101,103,
-114,101,101,115,0,1,1,0,0,11,0,114,97,100,0,0,0,1,3,2,90,95,1,0,9,0,1,99,0,2,17,49,56,48,0,48,0,0,
-17,51,0,49,52,49,53,57,50,54,0,0,49,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,
-95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,114,97,100,0,59,120,121,122,0,0,18,99,0,59,120,
-120,120,0,0,0,0,1,90,95,0,0,12,0,0,100,101,103,114,101,101,115,0,1,1,0,0,12,0,114,97,100,0,0,0,1,3,
-2,90,95,1,0,9,0,1,99,0,2,17,49,56,48,0,48,0,0,17,51,0,49,52,49,53,57,50,54,0,0,49,0,0,4,118,101,99,
-52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,18,114,97,100,0,0,18,99,
-0,59,120,120,120,120,0,0,0,0,1,90,95,0,0,9,0,0,115,105,110,0,1,1,0,0,9,0,114,97,100,105,97,110,115,
-0,0,0,1,4,102,108,111,97,116,95,115,105,110,101,0,18,95,95,114,101,116,86,97,108,0,0,18,114,97,100,
-105,97,110,115,0,0,0,0,1,90,95,0,0,10,0,0,115,105,110,0,1,1,0,0,10,0,114,97,100,105,97,110,115,0,0,
-0,1,4,102,108,111,97,116,95,115,105,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,114,
-97,100,105,97,110,115,0,59,120,0,0,0,4,102,108,111,97,116,95,115,105,110,101,0,18,95,95,114,101,
-116,86,97,108,0,59,121,0,0,18,114,97,100,105,97,110,115,0,59,121,0,0,0,0,1,90,95,0,0,11,0,0,115,
-105,110,0,1,1,0,0,11,0,114,97,100,105,97,110,115,0,0,0,1,4,102,108,111,97,116,95,115,105,110,101,0,
-18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,114,97,100,105,97,110,115,0,59,120,0,0,0,4,102,108,
-111,97,116,95,115,105,110,101,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,114,97,100,105,97,
-110,115,0,59,121,0,0,0,4,102,108,111,97,116,95,115,105,110,101,0,18,95,95,114,101,116,86,97,108,0,
-59,122,0,0,18,114,97,100,105,97,110,115,0,59,122,0,0,0,0,1,90,95,0,0,12,0,0,115,105,110,0,1,1,0,0,
-12,0,114,97,100,105,97,110,115,0,0,0,1,4,102,108,111,97,116,95,115,105,110,101,0,18,95,95,114,101,
-116,86,97,108,0,59,120,0,0,18,114,97,100,105,97,110,115,0,59,120,0,0,0,4,102,108,111,97,116,95,115,
-105,110,101,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,114,97,100,105,97,110,115,0,59,121,0,
-0,0,4,102,108,111,97,116,95,115,105,110,101,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,18,114,
-97,100,105,97,110,115,0,59,122,0,0,0,4,102,108,111,97,116,95,115,105,110,101,0,18,95,95,114,101,
-116,86,97,108,0,59,119,0,0,18,114,97,100,105,97,110,115,0,59,119,0,0,0,0,1,90,95,0,0,9,0,0,99,111,
-115,0,1,1,0,0,9,0,114,97,100,105,97,110,115,0,0,0,1,4,102,108,111,97,116,95,99,111,115,105,110,101,
-0,18,95,95,114,101,116,86,97,108,0,0,18,114,97,100,105,97,110,115,0,0,0,0,1,90,95,0,0,10,0,0,99,
-111,115,0,1,1,0,0,10,0,114,97,100,105,97,110,115,0,0,0,1,4,102,108,111,97,116,95,99,111,115,105,
-110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,114,97,100,105,97,110,115,0,59,120,0,0,0,
-4,102,108,111,97,116,95,99,111,115,105,110,101,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,
-114,97,100,105,97,110,115,0,59,121,0,0,0,0,1,90,95,0,0,11,0,0,99,111,115,0,1,1,0,0,11,0,114,97,100,
-105,97,110,115,0,0,0,1,4,102,108,111,97,116,95,99,111,115,105,110,101,0,18,95,95,114,101,116,86,97,
-108,0,59,120,0,0,18,114,97,100,105,97,110,115,0,59,120,0,0,0,4,102,108,111,97,116,95,99,111,115,
-105,110,101,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,114,97,100,105,97,110,115,0,59,121,0,
-0,0,4,102,108,111,97,116,95,99,111,115,105,110,101,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,
-18,114,97,100,105,97,110,115,0,59,122,0,0,0,0,1,90,95,0,0,12,0,0,99,111,115,0,1,1,0,0,12,0,114,97,
-100,105,97,110,115,0,0,0,1,4,102,108,111,97,116,95,99,111,115,105,110,101,0,18,95,95,114,101,116,
-86,97,108,0,59,120,0,0,18,114,97,100,105,97,110,115,0,59,120,0,0,0,4,102,108,111,97,116,95,99,111,
-115,105,110,101,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,114,97,100,105,97,110,115,0,59,
-121,0,0,0,4,102,108,111,97,116,95,99,111,115,105,110,101,0,18,95,95,114,101,116,86,97,108,0,59,122,
-0,0,18,114,97,100,105,97,110,115,0,59,122,0,0,0,4,102,108,111,97,116,95,99,111,115,105,110,101,0,
-18,95,95,114,101,116,86,97,108,0,59,119,0,0,18,114,97,100,105,97,110,115,0,59,119,0,0,0,0,1,90,95,
-0,0,9,0,0,116,97,110,0,1,1,0,0,9,0,97,110,103,108,101,0,0,0,1,3,2,90,95,1,0,9,0,1,115,0,2,58,115,
-105,110,0,0,18,97,110,103,108,101,0,0,0,0,0,3,2,90,95,1,0,9,0,1,99,0,2,58,99,111,115,0,0,18,97,110,
-103,108,101,0,0,0,0,0,8,18,115,0,18,99,0,49,0,0,1,90,95,0,0,10,0,0,116,97,110,0,1,1,0,0,10,0,97,
-110,103,108,101,0,0,0,1,3,2,90,95,1,0,10,0,1,115,0,2,58,115,105,110,0,0,18,97,110,103,108,101,0,0,
-0,0,0,3,2,90,95,1,0,10,0,1,99,0,2,58,99,111,115,0,0,18,97,110,103,108,101,0,0,0,0,0,8,18,115,0,18,
-99,0,49,0,0,1,90,95,0,0,11,0,0,116,97,110,0,1,1,0,0,11,0,97,110,103,108,101,0,0,0,1,3,2,90,95,1,0,
-11,0,1,115,0,2,58,115,105,110,0,0,18,97,110,103,108,101,0,0,0,0,0,3,2,90,95,1,0,11,0,1,99,0,2,58,
-99,111,115,0,0,18,97,110,103,108,101,0,0,0,0,0,8,18,115,0,18,99,0,49,0,0,1,90,95,0,0,12,0,0,116,97,
-110,0,1,1,0,0,12,0,97,110,103,108,101,0,0,0,1,3,2,90,95,1,0,12,0,1,115,0,2,58,115,105,110,0,0,18,
-97,110,103,108,101,0,0,0,0,0,3,2,90,95,1,0,12,0,1,99,0,2,58,99,111,115,0,0,18,97,110,103,108,101,0,
-0,0,0,0,8,18,115,0,18,99,0,49,0,0,1,90,95,0,0,9,0,0,97,115,105,110,0,1,1,0,0,9,0,120,0,0,0,1,3,2,
-90,95,1,0,9,0,1,97,48,0,2,17,49,0,53,55,48,55,50,56,56,0,0,0,0,3,2,90,95,1,0,9,0,1,97,49,0,2,17,48,
-0,50,49,50,49,49,52,52,0,0,54,0,0,3,2,90,95,1,0,9,0,1,97,50,0,2,17,48,0,48,55,52,50,54,49,48,0,0,0,
-0,3,2,90,95,1,0,9,0,1,104,97,108,102,80,105,0,2,17,51,0,49,52,49,53,57,50,54,0,0,17,48,0,53,0,0,48,
-0,0,3,2,90,95,1,0,9,0,1,121,0,2,58,97,98,115,0,0,18,120,0,0,0,0,0,9,18,95,95,114,101,116,86,97,108,
-0,18,104,97,108,102,80,105,0,58,115,113,114,116,0,0,17,49,0,48,0,0,18,121,0,47,0,0,18,97,48,0,18,
-121,0,18,97,49,0,18,97,50,0,18,121,0,48,46,48,46,48,47,58,115,105,103,110,0,0,18,120,0,0,0,48,20,0,
-0,1,90,95,0,0,10,0,0,97,115,105,110,0,1,1,0,0,10,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
-59,120,0,58,97,115,105,110,0,0,18,118,0,59,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,
-121,0,58,97,115,105,110,0,0,18,118,0,59,121,0,0,0,20,0,0,1,90,95,0,0,11,0,0,97,115,105,110,0,1,1,0,
-0,11,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,97,115,105,110,0,0,18,118,0,59,
-120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,97,115,105,110,0,0,18,118,0,59,121,0,
-0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,97,115,105,110,0,0,18,118,0,59,122,0,0,0,
-20,0,0,1,90,95,0,0,12,0,0,97,115,105,110,0,1,1,0,0,12,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,59,120,0,58,97,115,105,110,0,0,18,118,0,59,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,
-59,121,0,58,97,115,105,110,0,0,18,118,0,59,121,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,
-122,0,58,97,115,105,110,0,0,18,118,0,59,122,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,
-58,97,115,105,110,0,0,18,118,0,59,119,0,0,0,20,0,0,1,90,95,0,0,9,0,0,97,99,111,115,0,1,1,0,0,9,0,
-120,0,0,0,1,3,2,90,95,1,0,9,0,1,104,97,108,102,80,105,0,2,17,51,0,49,52,49,53,57,50,54,0,0,17,48,0,
-53,0,0,48,0,0,9,18,95,95,114,101,116,86,97,108,0,18,104,97,108,102,80,105,0,58,97,115,105,110,0,0,
-18,120,0,0,0,47,20,0,0,1,90,95,0,0,10,0,0,97,99,111,115,0,1,1,0,0,10,0,118,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,59,120,0,58,97,99,111,115,0,0,18,118,0,59,120,0,0,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,59,121,0,58,97,99,111,115,0,0,18,118,0,59,121,0,0,0,20,0,0,1,90,95,0,0,11,0,0,97,
-99,111,115,0,1,1,0,0,11,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,97,99,111,115,
-0,0,18,118,0,59,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,97,99,111,115,0,0,18,
-118,0,59,121,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,97,99,111,115,0,0,18,118,0,
-59,122,0,0,0,20,0,0,1,90,95,0,0,12,0,0,97,99,111,115,0,1,1,0,0,12,0,118,0,0,0,1,9,18,95,95,114,101,
-116,86,97,108,0,59,120,0,58,97,99,111,115,0,0,18,118,0,59,120,0,0,0,20,0,9,18,95,95,114,101,116,86,
-97,108,0,59,121,0,58,97,99,111,115,0,0,18,118,0,59,121,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,
-0,59,122,0,58,97,99,111,115,0,0,18,118,0,59,122,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,
-119,0,58,97,99,111,115,0,0,18,118,0,59,119,0,0,0,20,0,0,1,90,95,0,0,9,0,0,97,116,97,110,0,1,1,0,0,
-9,0,120,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,97,115,105,110,0,0,18,120,0,58,105,110,118,
-101,114,115,101,115,113,114,116,0,0,18,120,0,18,120,0,48,17,49,0,48,0,0,46,0,0,48,0,0,20,0,0,1,90,
-95,0,0,10,0,0,97,116,97,110,0,1,1,0,0,10,0,121,95,111,118,101,114,95,120,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,59,120,0,58,97,116,97,110,0,0,18,121,95,111,118,101,114,95,120,0,59,120,0,0,0,
-20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,97,116,97,110,0,0,18,121,95,111,118,101,114,95,
-120,0,59,121,0,0,0,20,0,0,1,90,95,0,0,11,0,0,97,116,97,110,0,1,1,0,0,11,0,121,95,111,118,101,114,
-95,120,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,97,116,97,110,0,0,18,121,95,111,118,
-101,114,95,120,0,59,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,97,116,97,110,0,
-0,18,121,95,111,118,101,114,95,120,0,59,121,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,
-58,97,116,97,110,0,0,18,121,95,111,118,101,114,95,120,0,59,122,0,0,0,20,0,0,1,90,95,0,0,12,0,0,97,
-116,97,110,0,1,1,0,0,12,0,121,95,111,118,101,114,95,120,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
-59,120,0,58,97,116,97,110,0,0,18,121,95,111,118,101,114,95,120,0,59,120,0,0,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,59,121,0,58,97,116,97,110,0,0,18,121,95,111,118,101,114,95,120,0,59,121,0,0,0,
-20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,97,116,97,110,0,0,18,121,95,111,118,101,114,95,
-120,0,59,122,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,58,97,116,97,110,0,0,18,121,95,
-111,118,101,114,95,120,0,59,119,0,0,0,20,0,0,1,90,95,0,0,9,0,0,97,116,97,110,0,1,1,0,0,9,0,121,0,0,
-1,1,0,0,9,0,120,0,0,0,1,3,2,90,95,0,0,9,0,1,114,0,0,0,10,58,97,98,115,0,0,18,120,0,0,0,17,49,0,48,
-0,45,52,0,41,0,2,9,18,114,0,58,97,116,97,110,0,0,18,121,0,18,120,0,49,0,0,20,0,10,18,120,0,17,48,0,
-48,0,0,40,0,2,9,18,114,0,18,114,0,58,115,105,103,110,0,0,18,121,0,0,0,17,51,0,49,52,49,53,57,51,0,
-0,48,46,20,0,0,9,14,0,0,2,9,18,114,0,58,115,105,103,110,0,0,18,121,0,0,0,17,49,0,53,55,48,55,57,54,
-53,0,0,48,20,0,0,8,18,114,0,0,0,1,90,95,0,0,10,0,0,97,116,97,110,0,1,1,0,0,10,0,117,0,0,1,1,0,0,10,
-0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,97,116,97,110,0,0,18,117,0,59,120,0,0,
-18,118,0,59,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,97,116,97,110,0,0,18,117,
-0,59,121,0,0,18,118,0,59,121,0,0,0,20,0,0,1,90,95,0,0,11,0,0,97,116,97,110,0,1,1,0,0,11,0,117,0,0,
-1,1,0,0,11,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,97,116,97,110,0,0,18,117,0,
-59,120,0,0,18,118,0,59,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,97,116,97,110,
-0,0,18,117,0,59,121,0,0,18,118,0,59,121,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,
-97,116,97,110,0,0,18,117,0,59,122,0,0,18,118,0,59,122,0,0,0,20,0,0,1,90,95,0,0,12,0,0,97,116,97,
-110,0,1,1,0,0,12,0,117,0,0,1,1,0,0,12,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,
-97,116,97,110,0,0,18,117,0,59,120,0,0,18,118,0,59,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,
-0,59,121,0,58,97,116,97,110,0,0,18,117,0,59,121,0,0,18,118,0,59,121,0,0,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,59,122,0,58,97,116,97,110,0,0,18,117,0,59,122,0,0,18,118,0,59,122,0,0,0,20,0,9,18,
-95,95,114,101,116,86,97,108,0,59,119,0,58,97,116,97,110,0,0,18,117,0,59,119,0,0,18,118,0,59,119,0,
-0,0,20,0,0,1,90,95,0,0,9,0,0,112,111,119,0,1,1,0,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,102,108,111,
-97,116,95,112,111,119,101,114,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,
-0,0,10,0,0,112,111,119,0,1,1,0,0,10,0,97,0,0,1,1,0,0,10,0,98,0,0,0,1,4,102,108,111,97,116,95,112,
-111,119,101,114,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,97,0,59,120,0,0,18,98,0,59,120,0,
-0,0,4,102,108,111,97,116,95,112,111,119,101,114,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,
-97,0,59,121,0,0,18,98,0,59,121,0,0,0,0,1,90,95,0,0,11,0,0,112,111,119,0,1,1,0,0,11,0,97,0,0,1,1,0,
-0,11,0,98,0,0,0,1,4,102,108,111,97,116,95,112,111,119,101,114,0,18,95,95,114,101,116,86,97,108,0,
-59,120,0,0,18,97,0,59,120,0,0,18,98,0,59,120,0,0,0,4,102,108,111,97,116,95,112,111,119,101,114,0,
-18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,97,0,59,121,0,0,18,98,0,59,121,0,0,0,4,102,108,111,
-97,116,95,112,111,119,101,114,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,18,97,0,59,122,0,0,18,
-98,0,59,122,0,0,0,0,1,90,95,0,0,12,0,0,112,111,119,0,1,1,0,0,12,0,97,0,0,1,1,0,0,12,0,98,0,0,0,1,4,
-102,108,111,97,116,95,112,111,119,101,114,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,97,0,59,
-120,0,0,18,98,0,59,120,0,0,0,4,102,108,111,97,116,95,112,111,119,101,114,0,18,95,95,114,101,116,86,
-97,108,0,59,121,0,0,18,97,0,59,121,0,0,18,98,0,59,121,0,0,0,4,102,108,111,97,116,95,112,111,119,
-101,114,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,18,97,0,59,122,0,0,18,98,0,59,122,0,0,0,4,
-102,108,111,97,116,95,112,111,119,101,114,0,18,95,95,114,101,116,86,97,108,0,59,119,0,0,18,97,0,59,
-119,0,0,18,98,0,59,119,0,0,0,0,1,90,95,0,0,9,0,0,101,120,112,0,1,1,0,0,9,0,97,0,0,0,1,3,2,90,95,0,
-0,9,0,1,116,0,2,18,97,0,17,49,0,52,52,50,54,57,53,48,50,0,0,48,0,0,4,102,108,111,97,116,95,101,120,
-112,50,0,18,95,95,114,101,116,86,97,108,0,0,18,116,0,0,0,0,1,90,95,0,0,10,0,0,101,120,112,0,1,1,0,
-0,10,0,97,0,0,0,1,3,2,90,95,0,0,10,0,1,116,0,2,18,97,0,17,49,0,52,52,50,54,57,53,48,50,0,0,48,0,0,
-4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,116,0,59,
-120,0,0,0,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,
-116,0,59,121,0,0,0,0,1,90,95,0,0,11,0,0,101,120,112,0,1,1,0,0,11,0,97,0,0,0,1,3,2,90,95,0,0,11,0,1,
-116,0,2,18,97,0,17,49,0,52,52,50,54,57,53,48,50,0,0,48,0,0,4,102,108,111,97,116,95,101,120,112,50,
-0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,116,0,59,120,0,0,0,4,102,108,111,97,116,95,101,
-120,112,50,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,116,0,59,121,0,0,0,4,102,108,111,97,
-116,95,101,120,112,50,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,18,116,0,59,122,0,0,0,0,1,90,
-95,0,0,12,0,0,101,120,112,0,1,1,0,0,12,0,97,0,0,0,1,3,2,90,95,0,0,12,0,1,116,0,2,18,97,0,17,49,0,
-52,52,50,54,57,53,48,50,0,0,48,0,0,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,116,
-86,97,108,0,59,120,0,0,18,116,0,59,120,0,0,0,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,
-101,116,86,97,108,0,59,121,0,0,18,116,0,59,121,0,0,0,4,102,108,111,97,116,95,101,120,112,50,0,18,
-95,95,114,101,116,86,97,108,0,59,122,0,0,18,116,0,59,122,0,0,0,4,102,108,111,97,116,95,101,120,112,
-50,0,18,95,95,114,101,116,86,97,108,0,59,119,0,0,18,116,0,59,119,0,0,0,0,1,90,95,0,0,9,0,0,108,111,
-103,50,0,1,1,0,0,9,0,120,0,0,0,1,4,102,108,111,97,116,95,108,111,103,50,0,18,95,95,114,101,116,86,
-97,108,0,0,18,120,0,0,0,0,1,90,95,0,0,10,0,0,108,111,103,50,0,1,1,0,0,10,0,118,0,0,0,1,4,102,108,
-111,97,116,95,108,111,103,50,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,118,0,59,120,0,0,0,4,
-102,108,111,97,116,95,108,111,103,50,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,118,0,59,121,
-0,0,0,0,1,90,95,0,0,11,0,0,108,111,103,50,0,1,1,0,0,11,0,118,0,0,0,1,4,102,108,111,97,116,95,108,
-111,103,50,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,118,0,59,120,0,0,0,4,102,108,111,97,
-116,95,108,111,103,50,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,118,0,59,121,0,0,0,4,102,
-108,111,97,116,95,108,111,103,50,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,18,118,0,59,122,0,0,
-0,0,1,90,95,0,0,12,0,0,108,111,103,50,0,1,1,0,0,12,0,118,0,0,0,1,4,102,108,111,97,116,95,108,111,
-103,50,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,118,0,59,120,0,0,0,4,102,108,111,97,116,95,
-108,111,103,50,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,118,0,59,121,0,0,0,4,102,108,111,
-97,116,95,108,111,103,50,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,18,118,0,59,122,0,0,0,4,102,
-108,111,97,116,95,108,111,103,50,0,18,95,95,114,101,116,86,97,108,0,59,119,0,0,18,118,0,59,119,0,0,
-0,0,1,90,95,0,0,9,0,0,108,111,103,0,1,1,0,0,9,0,120,0,0,0,1,3,2,90,95,1,0,9,0,1,99,0,2,17,48,0,54,
-57,51,49,52,55,49,56,49,0,0,0,0,8,58,108,111,103,50,0,0,18,120,0,0,0,18,99,0,48,0,0,1,90,95,0,0,10,
-0,0,108,111,103,0,1,1,0,0,10,0,118,0,0,0,1,3,2,90,95,1,0,9,0,1,99,0,2,17,48,0,54,57,51,49,52,55,49,
-56,49,0,0,0,0,8,58,108,111,103,50,0,0,18,118,0,0,0,18,99,0,48,0,0,1,90,95,0,0,11,0,0,108,111,103,0,
-1,1,0,0,11,0,118,0,0,0,1,3,2,90,95,1,0,9,0,1,99,0,2,17,48,0,54,57,51,49,52,55,49,56,49,0,0,0,0,8,
-58,108,111,103,50,0,0,18,118,0,0,0,18,99,0,48,0,0,1,90,95,0,0,12,0,0,108,111,103,0,1,1,0,0,12,0,
-118,0,0,0,1,3,2,90,95,1,0,9,0,1,99,0,2,17,48,0,54,57,51,49,52,55,49,56,49,0,0,0,0,8,58,108,111,103,
-50,0,0,18,118,0,0,0,18,99,0,48,0,0,1,90,95,0,0,9,0,0,101,120,112,50,0,1,1,0,0,9,0,97,0,0,0,1,4,102,
-108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,0,0,1,90,95,0,0,10,
-0,0,101,120,112,50,0,1,1,0,0,10,0,97,0,0,0,1,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,
-101,116,86,97,108,0,59,120,0,0,18,97,0,59,120,0,0,0,4,102,108,111,97,116,95,101,120,112,50,0,18,95,
-95,114,101,116,86,97,108,0,59,121,0,0,18,97,0,59,121,0,0,0,0,1,90,95,0,0,11,0,0,101,120,112,50,0,1,
-1,0,0,11,0,97,0,0,0,1,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,116,86,97,108,0,59,
-120,0,0,18,97,0,59,120,0,0,0,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,116,86,97,
-108,0,59,121,0,0,18,97,0,59,121,0,0,0,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,
-116,86,97,108,0,59,122,0,0,18,97,0,59,122,0,0,0,0,1,90,95,0,0,12,0,0,101,120,112,50,0,1,1,0,0,12,0,
-97,0,0,0,1,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,
-97,0,59,120,0,0,0,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,116,86,97,108,0,59,121,
-0,0,18,97,0,59,121,0,0,0,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,116,86,97,108,0,
-59,122,0,0,18,97,0,59,122,0,0,0,4,102,108,111,97,116,95,101,120,112,50,0,18,95,95,114,101,116,86,
-97,108,0,59,119,0,0,18,97,0,59,119,0,0,0,0,1,90,95,0,0,9,0,0,115,113,114,116,0,1,1,0,0,9,0,120,0,0,
-0,1,3,2,90,95,1,0,9,0,1,110,120,0,2,18,120,0,54,0,0,3,2,90,95,0,0,9,0,1,114,0,0,0,4,102,108,111,97,
-116,95,114,115,113,0,18,114,0,0,18,120,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,114,0,0,18,
-114,0,0,0,4,118,101,99,52,95,99,109,112,0,18,95,95,114,101,116,86,97,108,0,0,18,110,120,0,0,18,114,
-0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,10,0,0,115,113,114,116,0,1,1,0,0,10,0,120,0,0,0,1,3,2,90,95,1,
-0,10,0,1,110,120,0,2,18,120,0,54,0,1,1,122,101,114,111,0,2,58,118,101,99,50,0,0,17,48,0,48,0,0,0,0,
-0,0,3,2,90,95,0,0,10,0,1,114,0,0,0,4,102,108,111,97,116,95,114,115,113,0,18,114,0,59,120,0,0,18,
-120,0,59,120,0,0,0,4,102,108,111,97,116,95,114,115,113,0,18,114,0,59,121,0,0,18,120,0,59,121,0,0,0,
-4,102,108,111,97,116,95,114,99,112,0,18,114,0,59,120,0,0,18,114,0,59,120,0,0,0,4,102,108,111,97,
-116,95,114,99,112,0,18,114,0,59,121,0,0,18,114,0,59,121,0,0,0,4,118,101,99,52,95,99,109,112,0,18,
-95,95,114,101,116,86,97,108,0,0,18,110,120,0,0,18,114,0,0,18,122,101,114,111,0,0,0,0,1,90,95,0,0,
-11,0,0,115,113,114,116,0,1,1,0,0,11,0,120,0,0,0,1,3,2,90,95,1,0,11,0,1,110,120,0,2,18,120,0,54,0,1,
-1,122,101,114,111,0,2,58,118,101,99,51,0,0,17,48,0,48,0,0,0,0,0,0,3,2,90,95,0,0,11,0,1,114,0,0,0,4,
-102,108,111,97,116,95,114,115,113,0,18,114,0,59,120,0,0,18,120,0,59,120,0,0,0,4,102,108,111,97,116,
-95,114,115,113,0,18,114,0,59,121,0,0,18,120,0,59,121,0,0,0,4,102,108,111,97,116,95,114,115,113,0,
-18,114,0,59,122,0,0,18,120,0,59,122,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,114,0,59,120,0,0,
-18,114,0,59,120,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,114,0,59,121,0,0,18,114,0,59,121,0,0,
-0,4,102,108,111,97,116,95,114,99,112,0,18,114,0,59,122,0,0,18,114,0,59,122,0,0,0,4,118,101,99,52,
-95,99,109,112,0,18,95,95,114,101,116,86,97,108,0,0,18,110,120,0,0,18,114,0,0,18,122,101,114,111,0,
-0,0,0,1,90,95,0,0,12,0,0,115,113,114,116,0,1,1,0,0,12,0,120,0,0,0,1,3,2,90,95,1,0,12,0,1,110,120,0,
-2,18,120,0,54,0,1,1,122,101,114,111,0,2,58,118,101,99,52,0,0,17,48,0,48,0,0,0,0,0,0,3,2,90,95,0,0,
-12,0,1,114,0,0,0,4,102,108,111,97,116,95,114,115,113,0,18,114,0,59,120,0,0,18,120,0,59,120,0,0,0,4,
-102,108,111,97,116,95,114,115,113,0,18,114,0,59,121,0,0,18,120,0,59,121,0,0,0,4,102,108,111,97,116,
-95,114,115,113,0,18,114,0,59,122,0,0,18,120,0,59,122,0,0,0,4,102,108,111,97,116,95,114,115,113,0,
-18,114,0,59,119,0,0,18,120,0,59,119,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,114,0,59,120,0,0,
-18,114,0,59,120,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,114,0,59,121,0,0,18,114,0,59,121,0,0,
-0,4,102,108,111,97,116,95,114,99,112,0,18,114,0,59,122,0,0,18,114,0,59,122,0,0,0,4,102,108,111,97,
-116,95,114,99,112,0,18,114,0,59,119,0,0,18,114,0,59,119,0,0,0,4,118,101,99,52,95,99,109,112,0,18,
-95,95,114,101,116,86,97,108,0,0,18,110,120,0,0,18,114,0,0,18,122,101,114,111,0,0,0,0,1,90,95,0,0,9,
-0,0,105,110,118,101,114,115,101,115,113,114,116,0,1,1,0,0,9,0,120,0,0,0,1,4,102,108,111,97,116,95,
-114,115,113,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,120,0,0,0,0,1,90,95,0,0,10,0,0,105,
-110,118,101,114,115,101,115,113,114,116,0,1,1,0,0,10,0,118,0,0,0,1,4,102,108,111,97,116,95,114,115,
-113,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,118,0,59,120,0,0,0,4,102,108,111,97,116,95,
-114,115,113,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,118,0,59,121,0,0,0,0,1,90,95,0,0,11,0,
-0,105,110,118,101,114,115,101,115,113,114,116,0,1,1,0,0,11,0,118,0,0,0,1,4,102,108,111,97,116,95,
-114,115,113,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,118,0,59,120,0,0,0,4,102,108,111,97,
-116,95,114,115,113,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,118,0,59,121,0,0,0,4,102,108,
-111,97,116,95,114,115,113,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,18,118,0,59,122,0,0,0,0,1,
-90,95,0,0,12,0,0,105,110,118,101,114,115,101,115,113,114,116,0,1,1,0,0,12,0,118,0,0,0,1,4,102,108,
-111,97,116,95,114,115,113,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,118,0,59,120,0,0,0,4,
-102,108,111,97,116,95,114,115,113,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,118,0,59,121,0,
-0,0,4,102,108,111,97,116,95,114,115,113,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,18,118,0,59,
-122,0,0,0,4,102,108,111,97,116,95,114,115,113,0,18,95,95,114,101,116,86,97,108,0,59,119,0,0,18,118,
-0,59,119,0,0,0,0,1,90,95,0,0,9,0,0,110,111,114,109,97,108,105,122,101,0,1,1,0,0,9,0,120,0,0,0,1,9,
-18,95,95,114,101,116,86,97,108,0,17,49,0,48,0,0,20,0,0,1,90,95,0,0,10,0,0,110,111,114,109,97,108,
-105,122,101,0,1,1,0,0,10,0,118,0,0,0,1,3,2,90,95,1,0,9,0,1,115,0,2,58,105,110,118,101,114,115,101,
-115,113,114,116,0,0,58,100,111,116,0,0,18,118,0,0,18,118,0,0,0,0,0,0,0,4,118,101,99,52,95,109,117,
-108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,118,0,0,18,115,0,0,0,
-0,1,90,95,0,0,11,0,0,110,111,114,109,97,108,105,122,101,0,1,1,0,0,11,0,118,0,0,0,1,3,2,90,95,0,0,9,
-0,1,116,109,112,0,0,0,4,118,101,99,51,95,100,111,116,0,18,116,109,112,0,0,18,118,0,0,18,118,0,0,0,
-4,102,108,111,97,116,95,114,115,113,0,18,116,109,112,0,0,18,116,109,112,0,0,0,4,118,101,99,52,95,
-109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,118,0,0,
-18,116,109,112,0,0,0,0,1,90,95,0,0,12,0,0,110,111,114,109,97,108,105,122,101,0,1,1,0,0,12,0,118,0,
-0,0,1,3,2,90,95,0,0,9,0,1,116,109,112,0,0,0,4,118,101,99,52,95,100,111,116,0,18,116,109,112,0,0,18,
-118,0,0,18,118,0,0,0,4,102,108,111,97,116,95,114,115,113,0,18,116,109,112,0,0,18,116,109,112,0,0,0,
-4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,120,121,
-122,0,0,18,118,0,0,18,116,109,112,0,0,0,0,1,90,95,0,0,9,0,0,97,98,115,0,1,1,0,0,9,0,97,0,0,0,1,4,
-118,101,99,52,95,97,98,115,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,0,0,1,90,95,0,0,10,0,0,
-97,98,115,0,1,1,0,0,10,0,97,0,0,0,1,4,118,101,99,52,95,97,98,115,0,18,95,95,114,101,116,86,97,108,
-0,59,120,121,0,0,18,97,0,0,0,0,1,90,95,0,0,11,0,0,97,98,115,0,1,1,0,0,11,0,97,0,0,0,1,4,118,101,99,
-52,95,97,98,115,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,97,0,0,0,0,1,90,95,0,0,12,
-0,0,97,98,115,0,1,1,0,0,12,0,97,0,0,0,1,4,118,101,99,52,95,97,98,115,0,18,95,95,114,101,116,86,97,
-108,0,0,18,97,0,0,0,0,1,90,95,0,0,9,0,0,115,105,103,110,0,1,1,0,0,9,0,120,0,0,0,1,3,2,90,95,0,0,9,
-0,1,112,0,0,1,1,110,0,0,0,4,118,101,99,52,95,115,103,116,0,18,112,0,0,18,120,0,0,17,48,0,48,0,0,0,
-0,4,118,101,99,52,95,115,103,116,0,18,110,0,0,17,48,0,48,0,0,0,18,120,0,0,0,4,118,101,99,52,95,115,
-117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,0,18,112,0,0,18,110,0,0,0,0,1,90,95,0,
-0,10,0,0,115,105,103,110,0,1,1,0,0,10,0,118,0,0,0,1,3,2,90,95,0,0,10,0,1,112,0,0,1,1,110,0,0,0,4,
-118,101,99,52,95,115,103,116,0,18,112,0,59,120,121,0,0,18,118,0,0,17,48,0,48,0,0,0,0,4,118,101,99,
-52,95,115,103,116,0,18,110,0,59,120,121,0,0,17,48,0,48,0,0,0,18,118,0,0,0,4,118,101,99,52,95,115,
-117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,112,0,0,18,110,0,0,0,
-0,1,90,95,0,0,11,0,0,115,105,103,110,0,1,1,0,0,11,0,118,0,0,0,1,3,2,90,95,0,0,11,0,1,112,0,0,1,1,
-110,0,0,0,4,118,101,99,52,95,115,103,116,0,18,112,0,59,120,121,122,0,0,18,118,0,0,17,48,0,48,0,0,0,
-0,4,118,101,99,52,95,115,103,116,0,18,110,0,59,120,121,122,0,0,17,48,0,48,0,0,0,18,118,0,0,0,4,118,
-101,99,52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,
-112,0,0,18,110,0,0,0,0,1,90,95,0,0,12,0,0,115,105,103,110,0,1,1,0,0,12,0,118,0,0,0,1,3,2,90,95,0,0,
-12,0,1,112,0,0,1,1,110,0,0,0,4,118,101,99,52,95,115,103,116,0,18,112,0,0,18,118,0,0,17,48,0,48,0,0,
-0,0,4,118,101,99,52,95,115,103,116,0,18,110,0,0,17,48,0,48,0,0,0,18,118,0,0,0,4,118,101,99,52,95,
-115,117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,0,18,112,0,0,18,110,0,0,0,0,1,90,
-95,0,0,9,0,0,102,108,111,111,114,0,1,1,0,0,9,0,97,0,0,0,1,4,118,101,99,52,95,102,108,111,111,114,0,
-18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,0,0,1,90,95,0,0,10,0,0,102,108,111,111,114,0,1,1,0,0,
-10,0,97,0,0,0,1,4,118,101,99,52,95,102,108,111,111,114,0,18,95,95,114,101,116,86,97,108,0,59,120,
-121,0,0,18,97,0,0,0,0,1,90,95,0,0,11,0,0,102,108,111,111,114,0,1,1,0,0,11,0,97,0,0,0,1,4,118,101,
-99,52,95,102,108,111,111,114,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,97,0,0,0,0,1,
-90,95,0,0,12,0,0,102,108,111,111,114,0,1,1,0,0,12,0,97,0,0,0,1,4,118,101,99,52,95,102,108,111,111,
-114,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,0,0,1,90,95,0,0,9,0,0,99,101,105,108,0,1,1,0,0,
-9,0,97,0,0,0,1,3,2,90,95,0,0,9,0,1,98,0,2,18,97,0,54,0,0,4,118,101,99,52,95,102,108,111,111,114,0,
-18,98,0,0,18,98,0,0,0,9,18,95,95,114,101,116,86,97,108,0,18,98,0,54,20,0,0,1,90,95,0,0,10,0,0,99,
-101,105,108,0,1,1,0,0,10,0,97,0,0,0,1,3,2,90,95,0,0,10,0,1,98,0,2,18,97,0,54,0,0,4,118,101,99,52,
-95,102,108,111,111,114,0,18,98,0,0,18,98,0,0,0,9,18,95,95,114,101,116,86,97,108,0,59,120,121,0,18,
-98,0,54,20,0,0,1,90,95,0,0,11,0,0,99,101,105,108,0,1,1,0,0,11,0,97,0,0,0,1,3,2,90,95,0,0,11,0,1,98,
-0,2,18,97,0,54,0,0,4,118,101,99,52,95,102,108,111,111,114,0,18,98,0,0,18,98,0,0,0,9,18,95,95,114,
-101,116,86,97,108,0,59,120,121,122,0,18,98,0,54,20,0,0,1,90,95,0,0,12,0,0,99,101,105,108,0,1,1,0,0,
-12,0,97,0,0,0,1,3,2,90,95,0,0,12,0,1,98,0,2,18,97,0,54,0,0,4,118,101,99,52,95,102,108,111,111,114,
-0,18,98,0,0,18,98,0,0,0,9,18,95,95,114,101,116,86,97,108,0,18,98,0,54,20,0,0,1,90,95,0,0,9,0,0,102,
-114,97,99,116,0,1,1,0,0,9,0,97,0,0,0,1,4,118,101,99,52,95,102,114,97,99,0,18,95,95,114,101,116,86,
-97,108,0,0,18,97,0,0,0,0,1,90,95,0,0,10,0,0,102,114,97,99,116,0,1,1,0,0,10,0,97,0,0,0,1,4,118,101,
-99,52,95,102,114,97,99,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,97,0,0,0,0,1,90,95,0,0,
-11,0,0,102,114,97,99,116,0,1,1,0,0,11,0,97,0,0,0,1,4,118,101,99,52,95,102,114,97,99,0,18,95,95,114,
-101,116,86,97,108,0,59,120,121,122,0,0,18,97,0,0,0,0,1,90,95,0,0,12,0,0,102,114,97,99,116,0,1,1,0,
-0,12,0,97,0,0,0,1,4,118,101,99,52,95,102,114,97,99,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,
-0,0,1,90,95,0,0,9,0,0,109,111,100,0,1,1,0,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,0,0,9,0,1,
-111,110,101,79,118,101,114,66,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,118,101,
-114,66,0,0,18,98,0,0,0,9,18,95,95,114,101,116,86,97,108,0,18,97,0,18,98,0,58,102,108,111,111,114,0,
-0,18,97,0,18,111,110,101,79,118,101,114,66,0,48,0,0,48,47,20,0,0,1,90,95,0,0,10,0,0,109,111,100,0,
-1,1,0,0,10,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,0,0,9,0,1,111,110,101,79,118,101,114,66,0,0,0,
-4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,118,101,114,66,0,0,18,98,0,0,0,9,18,95,95,
-114,101,116,86,97,108,0,59,120,121,0,18,97,0,18,98,0,58,102,108,111,111,114,0,0,18,97,0,18,111,110,
-101,79,118,101,114,66,0,48,0,0,48,47,20,0,0,1,90,95,0,0,11,0,0,109,111,100,0,1,1,0,0,11,0,97,0,0,1,
-1,0,0,9,0,98,0,0,0,1,3,2,90,95,0,0,9,0,1,111,110,101,79,118,101,114,66,0,0,0,4,102,108,111,97,116,
-95,114,99,112,0,18,111,110,101,79,118,101,114,66,0,0,18,98,0,0,0,9,18,95,95,114,101,116,86,97,108,
-0,59,120,121,122,0,18,97,0,18,98,0,58,102,108,111,111,114,0,0,18,97,0,18,111,110,101,79,118,101,
-114,66,0,48,0,0,48,47,20,0,0,1,90,95,0,0,12,0,0,109,111,100,0,1,1,0,0,12,0,97,0,0,1,1,0,0,9,0,98,0,
-0,0,1,3,2,90,95,0,0,9,0,1,111,110,101,79,118,101,114,66,0,0,0,4,102,108,111,97,116,95,114,99,112,0,
-18,111,110,101,79,118,101,114,66,0,0,18,98,0,0,0,9,18,95,95,114,101,116,86,97,108,0,18,97,0,18,98,
-0,58,102,108,111,111,114,0,0,18,97,0,18,111,110,101,79,118,101,114,66,0,48,0,0,48,47,20,0,0,1,90,
-95,0,0,10,0,0,109,111,100,0,1,1,0,0,10,0,97,0,0,1,1,0,0,10,0,98,0,0,0,1,3,2,90,95,0,0,10,0,1,111,
-110,101,79,118,101,114,66,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,118,101,114,
-66,0,59,120,0,0,18,98,0,59,120,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,118,
-101,114,66,0,59,121,0,0,18,98,0,59,121,0,0,0,9,18,95,95,114,101,116,86,97,108,0,18,97,0,18,98,0,58,
-102,108,111,111,114,0,0,18,97,0,18,111,110,101,79,118,101,114,66,0,48,0,0,48,47,20,0,0,1,90,95,0,0,
-11,0,0,109,111,100,0,1,1,0,0,11,0,97,0,0,1,1,0,0,11,0,98,0,0,0,1,3,2,90,95,0,0,11,0,1,111,110,101,
-79,118,101,114,66,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,118,101,114,66,0,59,
-120,0,0,18,98,0,59,120,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,118,101,114,66,
-0,59,121,0,0,18,98,0,59,121,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,118,101,
-114,66,0,59,122,0,0,18,98,0,59,122,0,0,0,9,18,95,95,114,101,116,86,97,108,0,18,97,0,18,98,0,58,102,
-108,111,111,114,0,0,18,97,0,18,111,110,101,79,118,101,114,66,0,48,0,0,48,47,20,0,0,1,90,95,0,0,12,
-0,0,109,111,100,0,1,1,0,0,12,0,97,0,0,1,1,0,0,12,0,98,0,0,0,1,3,2,90,95,0,0,12,0,1,111,110,101,79,
-118,101,114,66,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,118,101,114,66,0,59,
-120,0,0,18,98,0,59,120,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,118,101,114,66,
-0,59,121,0,0,18,98,0,59,121,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,118,101,
-114,66,0,59,122,0,0,18,98,0,59,122,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,111,110,101,79,
-118,101,114,66,0,59,119,0,0,18,98,0,59,119,0,0,0,9,18,95,95,114,101,116,86,97,108,0,18,97,0,18,98,
-0,58,102,108,111,111,114,0,0,18,97,0,18,111,110,101,79,118,101,114,66,0,48,0,0,48,47,20,0,0,1,90,
-95,0,0,9,0,0,109,105,110,0,1,1,0,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,109,105,
-110,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,10,0,0,109,105,110,0,
-1,1,0,0,10,0,97,0,0,1,1,0,0,10,0,98,0,0,0,1,4,118,101,99,52,95,109,105,110,0,18,95,95,114,101,116,
-86,97,108,0,59,120,121,0,0,18,97,0,59,120,121,0,0,18,98,0,59,120,121,0,0,0,0,1,90,95,0,0,11,0,0,
-109,105,110,0,1,1,0,0,11,0,97,0,0,1,1,0,0,11,0,98,0,0,0,1,4,118,101,99,52,95,109,105,110,0,18,95,
-95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,97,0,59,120,121,122,0,0,18,98,0,59,120,121,122,0,
-0,0,0,1,90,95,0,0,12,0,0,109,105,110,0,1,1,0,0,12,0,97,0,0,1,1,0,0,12,0,98,0,0,0,1,4,118,101,99,52,
-95,109,105,110,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,10,0,0,109,
-105,110,0,1,1,0,0,10,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,109,105,110,0,18,95,95,114,
-101,116,86,97,108,0,0,18,97,0,59,120,121,0,0,18,98,0,0,0,0,1,90,95,0,0,11,0,0,109,105,110,0,1,1,0,
-0,11,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,109,105,110,0,18,95,95,114,101,116,86,97,
-108,0,0,18,97,0,59,120,121,122,0,0,18,98,0,0,0,0,1,90,95,0,0,12,0,0,109,105,110,0,1,1,0,0,12,0,97,
-0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,109,105,110,0,18,95,95,114,101,116,86,97,108,0,0,18,
-97,0,0,18,98,0,0,0,0,1,90,95,0,0,9,0,0,109,97,120,0,1,1,0,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,
-118,101,99,52,95,109,97,120,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,
-0,10,0,0,109,97,120,0,1,1,0,0,10,0,97,0,0,1,1,0,0,10,0,98,0,0,0,1,4,118,101,99,52,95,109,97,120,0,
-18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,97,0,59,120,121,0,0,18,98,0,59,120,121,0,0,0,0,
-1,90,95,0,0,11,0,0,109,97,120,0,1,1,0,0,11,0,97,0,0,1,1,0,0,11,0,98,0,0,0,1,4,118,101,99,52,95,109,
-97,120,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,97,0,59,120,121,122,0,0,18,98,0,59,
-120,121,122,0,0,0,0,1,90,95,0,0,12,0,0,109,97,120,0,1,1,0,0,12,0,97,0,0,1,1,0,0,12,0,98,0,0,0,1,4,
-118,101,99,52,95,109,97,120,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,
-0,10,0,0,109,97,120,0,1,1,0,0,10,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,109,97,120,0,
-18,95,95,114,101,116,86,97,108,0,0,18,97,0,59,120,121,0,0,18,98,0,0,0,0,1,90,95,0,0,11,0,0,109,97,
-120,0,1,1,0,0,11,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,109,97,120,0,18,95,95,114,101,
-116,86,97,108,0,0,18,97,0,59,120,121,122,0,0,18,98,0,0,0,0,1,90,95,0,0,12,0,0,109,97,120,0,1,1,0,0,
-12,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,109,97,120,0,18,95,95,114,101,116,86,97,108,
-0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,9,0,0,99,108,97,109,112,0,1,1,0,0,9,0,118,97,108,0,0,1,1,0,
-0,9,0,109,105,110,86,97,108,0,0,1,1,0,0,9,0,109,97,120,86,97,108,0,0,0,1,4,118,101,99,52,95,99,108,
-97,109,112,0,18,95,95,114,101,116,86,97,108,0,0,18,118,97,108,0,0,18,109,105,110,86,97,108,0,0,18,
-109,97,120,86,97,108,0,0,0,0,1,90,95,0,0,10,0,0,99,108,97,109,112,0,1,1,0,0,10,0,118,97,108,0,0,1,
-1,0,0,9,0,109,105,110,86,97,108,0,0,1,1,0,0,9,0,109,97,120,86,97,108,0,0,0,1,4,118,101,99,52,95,99,
-108,97,109,112,0,18,95,95,114,101,116,86,97,108,0,0,18,118,97,108,0,0,18,109,105,110,86,97,108,0,0,
-18,109,97,120,86,97,108,0,0,0,0,1,90,95,0,0,11,0,0,99,108,97,109,112,0,1,1,0,0,11,0,118,97,108,0,0,
-1,1,0,0,9,0,109,105,110,86,97,108,0,0,1,1,0,0,9,0,109,97,120,86,97,108,0,0,0,1,4,118,101,99,52,95,
-99,108,97,109,112,0,18,95,95,114,101,116,86,97,108,0,0,18,118,97,108,0,0,18,109,105,110,86,97,108,
-0,0,18,109,97,120,86,97,108,0,0,0,0,1,90,95,0,0,12,0,0,99,108,97,109,112,0,1,1,0,0,12,0,118,97,108,
-0,0,1,1,0,0,9,0,109,105,110,86,97,108,0,0,1,1,0,0,9,0,109,97,120,86,97,108,0,0,0,1,4,118,101,99,52,
-95,99,108,97,109,112,0,18,95,95,114,101,116,86,97,108,0,0,18,118,97,108,0,0,18,109,105,110,86,97,
-108,0,0,18,109,97,120,86,97,108,0,0,0,0,1,90,95,0,0,10,0,0,99,108,97,109,112,0,1,1,0,0,10,0,118,97,
-108,0,0,1,1,0,0,10,0,109,105,110,86,97,108,0,0,1,1,0,0,10,0,109,97,120,86,97,108,0,0,0,1,4,118,101,
-99,52,95,99,108,97,109,112,0,18,95,95,114,101,116,86,97,108,0,0,18,118,97,108,0,0,18,109,105,110,
-86,97,108,0,0,18,109,97,120,86,97,108,0,0,0,0,1,90,95,0,0,11,0,0,99,108,97,109,112,0,1,1,0,0,11,0,
-118,97,108,0,0,1,1,0,0,11,0,109,105,110,86,97,108,0,0,1,1,0,0,11,0,109,97,120,86,97,108,0,0,0,1,4,
-118,101,99,52,95,99,108,97,109,112,0,18,95,95,114,101,116,86,97,108,0,0,18,118,97,108,0,0,18,109,
-105,110,86,97,108,0,0,18,109,97,120,86,97,108,0,0,0,0,1,90,95,0,0,12,0,0,99,108,97,109,112,0,1,1,0,
-0,12,0,118,97,108,0,0,1,1,0,0,12,0,109,105,110,86,97,108,0,0,1,1,0,0,12,0,109,97,120,86,97,108,0,0,
-0,1,4,118,101,99,52,95,99,108,97,109,112,0,18,95,95,114,101,116,86,97,108,0,0,18,118,97,108,0,0,18,
-109,105,110,86,97,108,0,0,18,109,97,120,86,97,108,0,0,0,0,1,90,95,0,0,9,0,0,109,105,120,0,1,1,0,0,
-9,0,120,0,0,1,1,0,0,9,0,121,0,0,1,1,0,0,9,0,97,0,0,0,1,4,118,101,99,52,95,108,114,112,0,18,95,95,
-114,101,116,86,97,108,0,0,18,97,0,0,18,121,0,0,18,120,0,0,0,0,1,90,95,0,0,10,0,0,109,105,120,0,1,1,
-0,0,10,0,120,0,0,1,1,0,0,10,0,121,0,0,1,1,0,0,9,0,97,0,0,0,1,4,118,101,99,52,95,108,114,112,0,18,
-95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,121,0,0,18,120,0,0,0,0,1,90,95,0,0,11,0,0,109,105,120,
-0,1,1,0,0,11,0,120,0,0,1,1,0,0,11,0,121,0,0,1,1,0,0,9,0,97,0,0,0,1,4,118,101,99,52,95,108,114,112,
-0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,121,0,0,18,120,0,0,0,0,1,90,95,0,0,12,0,0,109,
-105,120,0,1,1,0,0,12,0,120,0,0,1,1,0,0,12,0,121,0,0,1,1,0,0,9,0,97,0,0,0,1,4,118,101,99,52,95,108,
-114,112,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,121,0,0,18,120,0,0,0,0,1,90,95,0,0,10,0,
-0,109,105,120,0,1,1,0,0,10,0,120,0,0,1,1,0,0,10,0,121,0,0,1,1,0,0,10,0,97,0,0,0,1,4,118,101,99,52,
-95,108,114,112,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,121,0,0,18,120,0,0,0,0,1,90,95,0,
-0,11,0,0,109,105,120,0,1,1,0,0,11,0,120,0,0,1,1,0,0,11,0,121,0,0,1,1,0,0,11,0,97,0,0,0,1,4,118,101,
-99,52,95,108,114,112,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,121,0,0,18,120,0,0,0,0,1,90,
-95,0,0,12,0,0,109,105,120,0,1,1,0,0,12,0,120,0,0,1,1,0,0,12,0,121,0,0,1,1,0,0,12,0,97,0,0,0,1,4,
-118,101,99,52,95,108,114,112,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,121,0,0,18,120,0,0,
-0,0,1,90,95,0,0,9,0,0,115,116,101,112,0,1,1,0,0,9,0,101,100,103,101,0,0,1,1,0,0,9,0,120,0,0,0,1,4,
-118,101,99,52,95,115,103,101,0,18,95,95,114,101,116,86,97,108,0,0,18,120,0,0,18,101,100,103,101,0,
-0,0,0,1,90,95,0,0,10,0,0,115,116,101,112,0,1,1,0,0,10,0,101,100,103,101,0,0,1,1,0,0,10,0,120,0,0,0,
-1,4,118,101,99,52,95,115,103,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,120,0,0,18,
-101,100,103,101,0,0,0,0,1,90,95,0,0,11,0,0,115,116,101,112,0,1,1,0,0,11,0,101,100,103,101,0,0,1,1,
-0,0,11,0,120,0,0,0,1,4,118,101,99,52,95,115,103,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,
-122,0,0,18,120,0,0,18,101,100,103,101,0,0,0,0,1,90,95,0,0,12,0,0,115,116,101,112,0,1,1,0,0,12,0,
-101,100,103,101,0,0,1,1,0,0,12,0,120,0,0,0,1,4,118,101,99,52,95,115,103,101,0,18,95,95,114,101,116,
-86,97,108,0,0,18,120,0,0,18,101,100,103,101,0,0,0,0,1,90,95,0,0,10,0,0,115,116,101,112,0,1,1,0,0,9,
-0,101,100,103,101,0,0,1,1,0,0,10,0,118,0,0,0,1,4,118,101,99,52,95,115,103,101,0,18,95,95,114,101,
-116,86,97,108,0,59,120,121,0,0,18,118,0,0,18,101,100,103,101,0,0,0,0,1,90,95,0,0,11,0,0,115,116,
-101,112,0,1,1,0,0,9,0,101,100,103,101,0,0,1,1,0,0,11,0,118,0,0,0,1,4,118,101,99,52,95,115,103,101,
-0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,118,0,0,18,101,100,103,101,0,0,0,0,1,90,
-95,0,0,12,0,0,115,116,101,112,0,1,1,0,0,9,0,101,100,103,101,0,0,1,1,0,0,12,0,118,0,0,0,1,4,118,101,
-99,52,95,115,103,101,0,18,95,95,114,101,116,86,97,108,0,0,18,118,0,0,18,101,100,103,101,0,0,0,0,1,
-90,95,0,0,9,0,0,115,109,111,111,116,104,115,116,101,112,0,1,1,0,0,9,0,101,100,103,101,48,0,0,1,1,0,
-0,9,0,101,100,103,101,49,0,0,1,1,0,0,9,0,120,0,0,0,1,3,2,90,95,0,0,9,0,1,116,0,2,58,99,108,97,109,
-112,0,0,18,120,0,18,101,100,103,101,48,0,47,18,101,100,103,101,49,0,18,101,100,103,101,48,0,47,49,
-0,17,48,0,48,0,0,0,17,49,0,48,0,0,0,0,0,0,8,18,116,0,18,116,0,48,17,51,0,48,0,0,17,50,0,48,0,0,18,
-116,0,48,47,48,0,0,1,90,95,0,0,10,0,0,115,109,111,111,116,104,115,116,101,112,0,1,1,0,0,10,0,101,
-100,103,101,48,0,0,1,1,0,0,10,0,101,100,103,101,49,0,0,1,1,0,0,10,0,118,0,0,0,1,3,2,90,95,0,0,10,0,
-1,116,0,2,58,99,108,97,109,112,0,0,18,118,0,18,101,100,103,101,48,0,47,18,101,100,103,101,49,0,18,
-101,100,103,101,48,0,47,49,0,17,48,0,48,0,0,0,17,49,0,48,0,0,0,0,0,0,8,18,116,0,18,116,0,48,17,51,
-0,48,0,0,17,50,0,48,0,0,18,116,0,48,47,48,0,0,1,90,95,0,0,11,0,0,115,109,111,111,116,104,115,116,
-101,112,0,1,1,0,0,11,0,101,100,103,101,48,0,0,1,1,0,0,11,0,101,100,103,101,49,0,0,1,1,0,0,11,0,118,
-0,0,0,1,3,2,90,95,0,0,11,0,1,116,0,2,58,99,108,97,109,112,0,0,18,118,0,18,101,100,103,101,48,0,47,
-18,101,100,103,101,49,0,18,101,100,103,101,48,0,47,49,0,17,48,0,48,0,0,0,17,49,0,48,0,0,0,0,0,0,8,
-18,116,0,18,116,0,48,17,51,0,48,0,0,17,50,0,48,0,0,18,116,0,48,47,48,0,0,1,90,95,0,0,12,0,0,115,
-109,111,111,116,104,115,116,101,112,0,1,1,0,0,12,0,101,100,103,101,48,0,0,1,1,0,0,12,0,101,100,103,
-101,49,0,0,1,1,0,0,12,0,118,0,0,0,1,3,2,90,95,0,0,12,0,1,116,0,2,58,99,108,97,109,112,0,0,18,118,0,
-18,101,100,103,101,48,0,47,18,101,100,103,101,49,0,18,101,100,103,101,48,0,47,49,0,17,48,0,48,0,0,
-0,17,49,0,48,0,0,0,0,0,0,8,18,116,0,18,116,0,48,17,51,0,48,0,0,17,50,0,48,0,0,18,116,0,48,47,48,0,
-0,1,90,95,0,0,10,0,0,115,109,111,111,116,104,115,116,101,112,0,1,1,0,0,9,0,101,100,103,101,48,0,0,
-1,1,0,0,9,0,101,100,103,101,49,0,0,1,1,0,0,10,0,118,0,0,0,1,3,2,90,95,0,0,10,0,1,116,0,2,58,99,108,
-97,109,112,0,0,18,118,0,18,101,100,103,101,48,0,47,18,101,100,103,101,49,0,18,101,100,103,101,48,0,
-47,49,0,17,48,0,48,0,0,0,17,49,0,48,0,0,0,0,0,0,8,18,116,0,18,116,0,48,17,51,0,48,0,0,17,50,0,48,0,
-0,18,116,0,48,47,48,0,0,1,90,95,0,0,11,0,0,115,109,111,111,116,104,115,116,101,112,0,1,1,0,0,9,0,
-101,100,103,101,48,0,0,1,1,0,0,9,0,101,100,103,101,49,0,0,1,1,0,0,11,0,118,0,0,0,1,3,2,90,95,0,0,
-11,0,1,116,0,2,58,99,108,97,109,112,0,0,18,118,0,18,101,100,103,101,48,0,47,18,101,100,103,101,49,
-0,18,101,100,103,101,48,0,47,49,0,17,48,0,48,0,0,0,17,49,0,48,0,0,0,0,0,0,8,18,116,0,18,116,0,48,
-17,51,0,48,0,0,17,50,0,48,0,0,18,116,0,48,47,48,0,0,1,90,95,0,0,12,0,0,115,109,111,111,116,104,115,
-116,101,112,0,1,1,0,0,9,0,101,100,103,101,48,0,0,1,1,0,0,9,0,101,100,103,101,49,0,0,1,1,0,0,12,0,
-118,0,0,0,1,3,2,90,95,0,0,12,0,1,116,0,2,58,99,108,97,109,112,0,0,18,118,0,18,101,100,103,101,48,0,
-47,18,101,100,103,101,49,0,18,101,100,103,101,48,0,47,49,0,17,48,0,48,0,0,0,17,49,0,48,0,0,0,0,0,0,
-8,18,116,0,18,116,0,48,17,51,0,48,0,0,17,50,0,48,0,0,18,116,0,48,47,48,0,0,1,90,95,0,0,9,0,0,108,
-101,110,103,116,104,0,1,1,0,0,9,0,120,0,0,0,1,8,58,97,98,115,0,0,18,120,0,0,0,0,0,1,90,95,0,0,9,0,
-0,108,101,110,103,116,104,0,1,1,0,0,10,0,118,0,0,0,1,3,2,90,95,0,0,9,0,1,114,0,0,0,3,2,90,95,1,0,9,
-0,1,112,0,2,58,100,111,116,0,0,18,118,0,0,18,118,0,0,0,0,0,4,102,108,111,97,116,95,114,115,113,0,
-18,114,0,0,18,112,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,95,95,114,101,116,86,97,108,0,59,
-120,0,0,18,114,0,0,0,0,1,90,95,0,0,9,0,0,108,101,110,103,116,104,0,1,1,0,0,11,0,118,0,0,0,1,3,2,90,
-95,0,0,9,0,1,114,0,0,0,3,2,90,95,1,0,9,0,1,112,0,2,58,100,111,116,0,0,18,118,0,0,18,118,0,0,0,0,0,
-4,102,108,111,97,116,95,114,115,113,0,18,114,0,0,18,112,0,0,0,4,102,108,111,97,116,95,114,99,112,0,
-18,95,95,114,101,116,86,97,108,0,0,18,114,0,0,0,0,1,90,95,0,0,9,0,0,108,101,110,103,116,104,0,1,1,
-0,0,12,0,118,0,0,0,1,3,2,90,95,0,0,9,0,1,114,0,0,0,3,2,90,95,1,0,9,0,1,112,0,2,58,100,111,116,0,0,
-18,118,0,0,18,118,0,0,0,0,0,4,102,108,111,97,116,95,114,115,113,0,18,114,0,0,18,112,0,0,0,4,102,
-108,111,97,116,95,114,99,112,0,18,95,95,114,101,116,86,97,108,0,0,18,114,0,0,0,0,1,90,95,0,0,9,0,0,
-100,105,115,116,97,110,99,101,0,1,1,0,0,9,0,120,0,0,1,1,0,0,9,0,121,0,0,0,1,3,2,90,95,1,0,9,0,1,
-100,0,2,18,120,0,18,121,0,47,0,0,9,18,95,95,114,101,116,86,97,108,0,58,108,101,110,103,116,104,0,0,
-18,100,0,0,0,20,0,0,1,90,95,0,0,9,0,0,100,105,115,116,97,110,99,101,0,1,1,0,0,10,0,118,0,0,1,1,0,0,
-10,0,117,0,0,0,1,3,2,90,95,1,0,10,0,1,100,50,0,2,18,118,0,18,117,0,47,0,0,9,18,95,95,114,101,116,
-86,97,108,0,58,108,101,110,103,116,104,0,0,18,100,50,0,0,0,20,0,0,1,90,95,0,0,9,0,0,100,105,115,
-116,97,110,99,101,0,1,1,0,0,11,0,118,0,0,1,1,0,0,11,0,117,0,0,0,1,3,2,90,95,1,0,11,0,1,100,51,0,2,
-18,118,0,18,117,0,47,0,0,9,18,95,95,114,101,116,86,97,108,0,58,108,101,110,103,116,104,0,0,18,100,
-51,0,0,0,20,0,0,1,90,95,0,0,9,0,0,100,105,115,116,97,110,99,101,0,1,1,0,0,12,0,118,0,0,1,1,0,0,12,
-0,117,0,0,0,1,3,2,90,95,1,0,12,0,1,100,52,0,2,18,118,0,18,117,0,47,0,0,9,18,95,95,114,101,116,86,
-97,108,0,58,108,101,110,103,116,104,0,0,18,100,52,0,0,0,20,0,0,1,90,95,0,0,11,0,0,99,114,111,115,
-115,0,1,1,0,0,11,0,118,0,0,1,1,0,0,11,0,117,0,0,0,1,4,118,101,99,51,95,99,114,111,115,115,0,18,95,
-95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,118,0,0,18,117,0,0,0,0,1,90,95,0,0,9,0,0,102,97,
-99,101,102,111,114,119,97,114,100,0,1,1,0,0,9,0,78,0,0,1,1,0,0,9,0,73,0,0,1,1,0,0,9,0,78,114,101,
-102,0,0,0,1,3,2,90,95,1,0,9,0,1,100,0,2,58,100,111,116,0,0,18,78,114,101,102,0,0,18,73,0,0,0,0,0,3,
-2,90,95,0,0,9,0,1,115,0,0,0,4,118,101,99,52,95,115,103,116,0,18,115,0,0,17,48,0,48,0,0,0,18,100,0,
-0,0,8,58,109,105,120,0,0,18,78,0,54,0,18,78,0,0,18,115,0,0,0,0,0,1,90,95,0,0,10,0,0,102,97,99,101,
-102,111,114,119,97,114,100,0,1,1,0,0,10,0,78,0,0,1,1,0,0,10,0,73,0,0,1,1,0,0,10,0,78,114,101,102,0,
-0,0,1,3,2,90,95,1,0,9,0,1,100,0,2,58,100,111,116,0,0,18,78,114,101,102,0,0,18,73,0,0,0,0,0,3,2,90,
-95,0,0,9,0,1,115,0,0,0,4,118,101,99,52,95,115,103,116,0,18,115,0,0,17,48,0,48,0,0,0,18,100,0,0,0,8,
-58,109,105,120,0,0,18,78,0,54,0,18,78,0,0,18,115,0,0,0,0,0,1,90,95,0,0,11,0,0,102,97,99,101,102,
-111,114,119,97,114,100,0,1,1,0,0,11,0,78,0,0,1,1,0,0,11,0,73,0,0,1,1,0,0,11,0,78,114,101,102,0,0,0,
-1,3,2,90,95,1,0,9,0,1,100,0,2,58,100,111,116,0,0,18,78,114,101,102,0,0,18,73,0,0,0,0,0,3,2,90,95,0,
-0,9,0,1,115,0,0,0,4,118,101,99,52,95,115,103,116,0,18,115,0,0,17,48,0,48,0,0,0,18,100,0,0,0,8,58,
-109,105,120,0,0,18,78,0,54,0,18,78,0,0,18,115,0,0,0,0,0,1,90,95,0,0,12,0,0,102,97,99,101,102,111,
-114,119,97,114,100,0,1,1,0,0,12,0,78,0,0,1,1,0,0,12,0,73,0,0,1,1,0,0,12,0,78,114,101,102,0,0,0,1,3,
-2,90,95,1,0,9,0,1,100,0,2,58,100,111,116,0,0,18,78,114,101,102,0,0,18,73,0,0,0,0,0,3,2,90,95,0,0,9,
-0,1,115,0,0,0,4,118,101,99,52,95,115,103,116,0,18,115,0,0,17,48,0,48,0,0,0,18,100,0,0,0,8,58,109,
-105,120,0,0,18,78,0,54,0,18,78,0,0,18,115,0,0,0,0,0,1,90,95,0,0,9,0,0,114,101,102,108,101,99,116,0,
-1,1,0,0,9,0,73,0,0,1,1,0,0,9,0,78,0,0,0,1,8,18,73,0,17,50,0,48,0,0,58,100,111,116,0,0,18,78,0,0,18,
-73,0,0,0,48,18,78,0,48,47,0,0,1,90,95,0,0,10,0,0,114,101,102,108,101,99,116,0,1,1,0,0,10,0,73,0,0,
-1,1,0,0,10,0,78,0,0,0,1,8,18,73,0,17,50,0,48,0,0,58,100,111,116,0,0,18,78,0,0,18,73,0,0,0,48,18,78,
-0,48,47,0,0,1,90,95,0,0,11,0,0,114,101,102,108,101,99,116,0,1,1,0,0,11,0,73,0,0,1,1,0,0,11,0,78,0,
-0,0,1,8,18,73,0,17,50,0,48,0,0,58,100,111,116,0,0,18,78,0,0,18,73,0,0,0,48,18,78,0,48,47,0,0,1,90,
-95,0,0,12,0,0,114,101,102,108,101,99,116,0,1,1,0,0,12,0,73,0,0,1,1,0,0,12,0,78,0,0,0,1,8,18,73,0,
-17,50,0,48,0,0,58,100,111,116,0,0,18,78,0,0,18,73,0,0,0,48,18,78,0,48,47,0,0,1,90,95,0,0,9,0,0,114,
-101,102,114,97,99,116,0,1,1,0,0,9,0,73,0,0,1,1,0,0,9,0,78,0,0,1,1,0,0,9,0,101,116,97,0,0,0,1,3,2,
-90,95,0,0,9,0,1,110,95,100,111,116,95,105,0,2,58,100,111,116,0,0,18,78,0,0,18,73,0,0,0,0,0,3,2,90,
-95,0,0,9,0,1,107,0,2,17,49,0,48,0,0,18,101,116,97,0,18,101,116,97,0,48,17,49,0,48,0,0,18,110,95,
-100,111,116,95,105,0,18,110,95,100,111,116,95,105,0,48,47,48,47,0,0,3,2,90,95,0,0,9,0,1,114,101,
-116,118,97,108,0,0,0,10,18,107,0,17,48,0,48,0,0,40,0,9,18,114,101,116,118,97,108,0,17,48,0,48,0,0,
-20,0,9,18,114,101,116,118,97,108,0,18,101,116,97,0,18,73,0,48,18,101,116,97,0,18,110,95,100,111,
-116,95,105,0,48,58,115,113,114,116,0,0,18,107,0,0,0,46,18,78,0,48,47,20,0,8,18,114,101,116,118,97,
-108,0,0,0,1,90,95,0,0,10,0,0,114,101,102,114,97,99,116,0,1,1,0,0,10,0,73,0,0,1,1,0,0,10,0,78,0,0,1,
-1,0,0,9,0,101,116,97,0,0,0,1,3,2,90,95,0,0,9,0,1,110,95,100,111,116,95,105,0,2,58,100,111,116,0,0,
-18,78,0,0,18,73,0,0,0,0,0,3,2,90,95,0,0,9,0,1,107,0,2,17,49,0,48,0,0,18,101,116,97,0,18,101,116,97,
-0,48,17,49,0,48,0,0,18,110,95,100,111,116,95,105,0,18,110,95,100,111,116,95,105,0,48,47,48,47,0,0,
-3,2,90,95,0,0,10,0,1,114,101,116,118,97,108,0,0,0,10,18,107,0,17,48,0,48,0,0,40,0,9,18,114,101,116,
-118,97,108,0,58,118,101,99,50,0,0,17,48,0,48,0,0,0,0,20,0,9,18,114,101,116,118,97,108,0,18,101,116,
-97,0,18,73,0,48,18,101,116,97,0,18,110,95,100,111,116,95,105,0,48,58,115,113,114,116,0,0,18,107,0,
-0,0,46,18,78,0,48,47,20,0,8,18,114,101,116,118,97,108,0,0,0,1,90,95,0,0,11,0,0,114,101,102,114,97,
-99,116,0,1,1,0,0,11,0,73,0,0,1,1,0,0,11,0,78,0,0,1,1,0,0,9,0,101,116,97,0,0,0,1,3,2,90,95,0,0,9,0,
-1,110,95,100,111,116,95,105,0,2,58,100,111,116,0,0,18,78,0,0,18,73,0,0,0,0,0,3,2,90,95,0,0,9,0,1,
-107,0,2,17,49,0,48,0,0,18,101,116,97,0,18,101,116,97,0,48,17,49,0,48,0,0,18,110,95,100,111,116,95,
-105,0,18,110,95,100,111,116,95,105,0,48,47,48,47,0,0,3,2,90,95,0,0,11,0,1,114,101,116,118,97,108,0,
-0,0,10,18,107,0,17,48,0,48,0,0,40,0,9,18,114,101,116,118,97,108,0,58,118,101,99,51,0,0,17,48,0,48,
-0,0,0,0,20,0,9,18,114,101,116,118,97,108,0,18,101,116,97,0,18,73,0,48,18,101,116,97,0,18,110,95,
-100,111,116,95,105,0,48,58,115,113,114,116,0,0,18,107,0,0,0,46,18,78,0,48,47,20,0,8,18,114,101,116,
-118,97,108,0,0,0,1,90,95,0,0,12,0,0,114,101,102,114,97,99,116,0,1,1,0,0,12,0,73,0,0,1,1,0,0,12,0,
-78,0,0,1,1,0,0,9,0,101,116,97,0,0,0,1,3,2,90,95,0,0,9,0,1,110,95,100,111,116,95,105,0,2,58,100,111,
-116,0,0,18,78,0,0,18,73,0,0,0,0,0,3,2,90,95,0,0,9,0,1,107,0,2,17,49,0,48,0,0,18,101,116,97,0,18,
-101,116,97,0,48,17,49,0,48,0,0,18,110,95,100,111,116,95,105,0,18,110,95,100,111,116,95,105,0,48,47,
-48,47,0,0,3,2,90,95,0,0,12,0,1,114,101,116,118,97,108,0,0,0,10,18,107,0,17,48,0,48,0,0,40,0,9,18,
-114,101,116,118,97,108,0,58,118,101,99,52,0,0,17,48,0,48,0,0,0,0,20,0,9,18,114,101,116,118,97,108,
-0,18,101,116,97,0,18,73,0,48,18,101,116,97,0,18,110,95,100,111,116,95,105,0,48,58,115,113,114,116,
-0,0,18,107,0,0,0,46,18,78,0,48,47,20,0,8,18,114,101,116,118,97,108,0,0,0,1,90,95,0,0,13,0,0,109,97,
-116,114,105,120,67,111,109,112,77,117,108,116,0,1,0,0,0,13,0,109,0,0,1,0,0,0,13,0,110,0,0,0,1,8,58,
-109,97,116,50,0,0,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,48,0,18,109,0,16,10,49,0,57,18,110,0,
-16,10,49,0,57,48,0,0,0,0,1,90,95,0,0,14,0,0,109,97,116,114,105,120,67,111,109,112,77,117,108,116,0,
-1,0,0,0,14,0,109,0,0,1,0,0,0,14,0,110,0,0,0,1,8,58,109,97,116,51,0,0,18,109,0,16,8,48,0,57,18,110,
-0,16,8,48,0,57,48,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,48,0,18,109,0,16,10,50,0,57,18,
-110,0,16,10,50,0,57,48,0,0,0,0,1,90,95,0,0,15,0,0,109,97,116,114,105,120,67,111,109,112,77,117,108,
-116,0,1,0,0,0,15,0,109,0,0,1,0,0,0,15,0,110,0,0,0,1,8,58,109,97,116,52,0,0,18,109,0,16,8,48,0,57,
-18,110,0,16,8,48,0,57,48,0,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,48,0,18,109,0,16,10,50,0,
-57,18,110,0,16,10,50,0,57,48,0,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,48,0,0,0,0,1,90,95,0,
-0,2,0,0,108,101,115,115,84,104,97,110,0,1,1,0,0,10,0,117,0,0,1,1,0,0,10,0,118,0,0,0,1,4,118,101,99,
-52,95,115,108,116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,117,0,0,18,118,0,0,0,0,1,90,
-95,0,0,3,0,0,108,101,115,115,84,104,97,110,0,1,1,0,0,11,0,117,0,0,1,1,0,0,11,0,118,0,0,0,1,4,118,
-101,99,52,95,115,108,116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,117,0,0,18,118,0,
-0,0,0,1,90,95,0,0,4,0,0,108,101,115,115,84,104,97,110,0,1,1,0,0,12,0,117,0,0,1,1,0,0,12,0,118,0,0,
-0,1,4,118,101,99,52,95,115,108,116,0,18,95,95,114,101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,0,0,
-1,90,95,0,0,2,0,0,108,101,115,115,84,104,97,110,0,1,1,0,0,6,0,117,0,0,1,1,0,0,6,0,118,0,0,0,1,4,
-118,101,99,52,95,115,108,116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,117,0,0,18,118,0,
-0,0,0,1,90,95,0,0,3,0,0,108,101,115,115,84,104,97,110,0,1,1,0,0,7,0,117,0,0,1,1,0,0,7,0,118,0,0,0,
-1,4,118,101,99,52,95,115,108,116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,117,0,0,
-18,118,0,0,0,0,1,90,95,0,0,4,0,0,108,101,115,115,84,104,97,110,0,1,1,0,0,8,0,117,0,0,1,1,0,0,8,0,
-118,0,0,0,1,4,118,101,99,52,95,115,108,116,0,18,95,95,114,101,116,86,97,108,0,0,18,117,0,0,18,118,
-0,0,0,0,1,90,95,0,0,2,0,0,108,101,115,115,84,104,97,110,69,113,117,97,108,0,1,1,0,0,10,0,117,0,0,1,
-1,0,0,10,0,118,0,0,0,1,4,118,101,99,52,95,115,108,101,0,18,95,95,114,101,116,86,97,108,0,59,120,
-121,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,3,0,0,108,101,115,115,84,104,97,110,69,113,117,97,
-108,0,1,1,0,0,11,0,117,0,0,1,1,0,0,11,0,118,0,0,0,1,4,118,101,99,52,95,115,108,101,0,18,95,95,114,
-101,116,86,97,108,0,59,120,121,122,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,4,0,0,108,101,115,115,
-84,104,97,110,69,113,117,97,108,0,1,1,0,0,12,0,117,0,0,1,1,0,0,12,0,118,0,0,0,1,4,118,101,99,52,95,
-115,108,101,0,18,95,95,114,101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,2,0,0,108,
-101,115,115,84,104,97,110,69,113,117,97,108,0,1,1,0,0,6,0,117,0,0,1,1,0,0,6,0,118,0,0,0,1,4,118,
-101,99,52,95,115,108,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,117,0,0,18,118,0,0,0,
-0,1,90,95,0,0,3,0,0,108,101,115,115,84,104,97,110,69,113,117,97,108,0,1,1,0,0,7,0,117,0,0,1,1,0,0,
-7,0,118,0,0,0,1,4,118,101,99,52,95,115,108,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,
-0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,4,0,0,108,101,115,115,84,104,97,110,69,113,117,97,108,0,1,
-1,0,0,8,0,117,0,0,1,1,0,0,8,0,118,0,0,0,1,4,118,101,99,52,95,115,108,101,0,18,95,95,114,101,116,86,
-97,108,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,2,0,0,103,114,101,97,116,101,114,84,104,97,110,0,
-1,1,0,0,10,0,117,0,0,1,1,0,0,10,0,118,0,0,0,1,4,118,101,99,52,95,115,103,116,0,18,95,95,114,101,
-116,86,97,108,0,59,120,121,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,3,0,0,103,114,101,97,116,101,
-114,84,104,97,110,0,1,1,0,0,11,0,117,0,0,1,1,0,0,11,0,118,0,0,0,1,4,118,101,99,52,95,115,103,116,0,
-18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,4,0,0,
-103,114,101,97,116,101,114,84,104,97,110,0,1,1,0,0,12,0,117,0,0,1,1,0,0,12,0,118,0,0,0,1,4,118,101,
-99,52,95,115,103,116,0,18,95,95,114,101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,2,
-0,0,103,114,101,97,116,101,114,84,104,97,110,0,1,1,0,0,6,0,117,0,0,1,1,0,0,6,0,118,0,0,0,1,4,118,
-101,99,52,95,115,103,116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,117,0,59,120,121,0,0,
-18,118,0,59,120,121,0,0,0,0,1,90,95,0,0,3,0,0,103,114,101,97,116,101,114,84,104,97,110,0,1,1,0,0,7,
-0,117,0,0,1,1,0,0,7,0,118,0,0,0,1,4,118,101,99,52,95,115,103,116,0,18,95,95,114,101,116,86,97,108,
-0,59,120,121,122,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,4,0,0,103,114,101,97,116,101,114,84,104,
-97,110,0,1,1,0,0,8,0,117,0,0,1,1,0,0,8,0,118,0,0,0,1,4,118,101,99,52,95,115,103,116,0,18,95,95,114,
-101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,2,0,0,103,114,101,97,116,101,114,84,
-104,97,110,69,113,117,97,108,0,1,1,0,0,10,0,117,0,0,1,1,0,0,10,0,118,0,0,0,1,4,118,101,99,52,95,
-115,103,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,
-0,3,0,0,103,114,101,97,116,101,114,84,104,97,110,69,113,117,97,108,0,1,1,0,0,11,0,117,0,0,1,1,0,0,
-11,0,118,0,0,0,1,4,118,101,99,52,95,115,103,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,
-0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,4,0,0,103,114,101,97,116,101,114,84,104,97,110,69,113,
-117,97,108,0,1,1,0,0,12,0,117,0,0,1,1,0,0,12,0,118,0,0,0,1,4,118,101,99,52,95,115,103,101,0,18,95,
-95,114,101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,2,0,0,103,114,101,97,116,101,
-114,84,104,97,110,69,113,117,97,108,0,1,1,0,0,6,0,117,0,0,1,1,0,0,6,0,118,0,0,0,1,4,118,101,99,52,
-95,115,103,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,
-0,0,3,0,0,103,114,101,97,116,101,114,84,104,97,110,69,113,117,97,108,0,1,1,0,0,7,0,117,0,0,1,1,0,0,
-7,0,118,0,0,0,1,4,118,101,99,52,95,115,103,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,
-0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,4,0,0,103,114,101,97,116,101,114,84,104,97,110,69,113,117,
-97,108,0,1,1,0,0,8,0,117,0,0,1,1,0,0,8,0,118,0,0,0,1,4,118,101,99,52,95,115,103,101,0,18,95,95,114,
-101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,2,0,0,101,113,117,97,108,0,1,1,0,0,10,
-0,117,0,0,1,1,0,0,10,0,118,0,0,0,1,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,
-0,59,120,121,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,3,0,0,101,113,117,97,108,0,1,1,0,0,11,0,117,
-0,0,1,1,0,0,11,0,118,0,0,0,1,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,
-120,121,122,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,4,0,0,101,113,117,97,108,0,1,1,0,0,12,0,117,
-0,0,1,1,0,0,12,0,118,0,0,0,1,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,0,
-18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,2,0,0,101,113,117,97,108,0,1,1,0,0,6,0,117,0,0,1,1,0,0,6,0,
-118,0,0,0,1,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,
-117,0,0,18,118,0,0,0,0,1,90,95,0,0,3,0,0,101,113,117,97,108,0,1,1,0,0,7,0,117,0,0,1,1,0,0,7,0,118,
-0,0,0,1,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,
-117,0,0,18,118,0,0,0,0,1,90,95,0,0,4,0,0,101,113,117,97,108,0,1,1,0,0,8,0,117,0,0,1,1,0,0,8,0,118,
-0,0,0,1,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,
-0,0,1,90,95,0,0,2,0,0,101,113,117,97,108,0,1,1,0,0,2,0,117,0,0,1,1,0,0,2,0,118,0,0,0,1,4,118,101,
-99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,117,0,0,18,118,0,0,0,0,1,
-90,95,0,0,3,0,0,101,113,117,97,108,0,1,1,0,0,3,0,117,0,0,1,1,0,0,3,0,118,0,0,0,1,4,118,101,99,52,
-95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,117,0,0,18,118,0,0,0,0,1,
-90,95,0,0,4,0,0,101,113,117,97,108,0,1,1,0,0,4,0,117,0,0,1,1,0,0,4,0,118,0,0,0,1,4,118,101,99,52,
-95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,0,2,0,0,
-110,111,116,69,113,117,97,108,0,1,1,0,0,10,0,117,0,0,1,1,0,0,10,0,118,0,0,0,1,4,118,101,99,52,95,
-115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,117,0,0,18,118,0,0,0,0,1,90,95,0,
-0,3,0,0,110,111,116,69,113,117,97,108,0,1,1,0,0,11,0,117,0,0,1,1,0,0,11,0,118,0,0,0,1,4,118,101,99,
-52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,117,0,0,18,118,0,0,0,0,
-1,90,95,0,0,4,0,0,110,111,116,69,113,117,97,108,0,1,1,0,0,12,0,117,0,0,1,1,0,0,12,0,118,0,0,0,1,4,
-118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,0,0,1,90,
-95,0,0,2,0,0,110,111,116,69,113,117,97,108,0,1,1,0,0,6,0,117,0,0,1,1,0,0,6,0,118,0,0,0,1,4,118,101,
-99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,117,0,0,18,118,0,0,0,0,1,
-90,95,0,0,3,0,0,110,111,116,69,113,117,97,108,0,1,1,0,0,7,0,117,0,0,1,1,0,0,7,0,118,0,0,0,1,4,118,
-101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,117,0,0,18,118,0,
-0,0,0,1,90,95,0,0,4,0,0,110,111,116,69,113,117,97,108,0,1,1,0,0,8,0,117,0,0,1,1,0,0,8,0,118,0,0,0,
-1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,0,0,1,
-90,95,0,0,2,0,0,110,111,116,69,113,117,97,108,0,1,1,0,0,2,0,117,0,0,1,1,0,0,2,0,118,0,0,0,1,4,118,
-101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,117,0,0,18,118,0,0,0,
-0,1,90,95,0,0,3,0,0,110,111,116,69,113,117,97,108,0,1,1,0,0,3,0,117,0,0,1,1,0,0,3,0,118,0,0,0,1,4,
-118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,117,0,0,18,
-118,0,0,0,0,1,90,95,0,0,4,0,0,110,111,116,69,113,117,97,108,0,1,1,0,0,4,0,117,0,0,1,1,0,0,4,0,118,
-0,0,0,1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,0,18,117,0,0,18,118,0,0,
-0,0,1,90,95,0,0,1,0,0,97,110,121,0,1,1,0,0,2,0,118,0,0,0,1,3,2,90,95,0,0,9,0,1,115,117,109,0,0,0,4,
-118,101,99,52,95,97,100,100,0,18,115,117,109,0,59,120,0,0,18,118,0,59,120,0,0,18,118,0,59,121,0,0,
-0,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,115,117,109,0,59,
-120,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,1,0,0,97,110,121,0,1,1,0,0,3,0,118,0,0,0,1,3,2,90,95,0,0,
-9,0,1,115,117,109,0,0,0,4,118,101,99,52,95,97,100,100,0,18,115,117,109,0,59,120,0,0,18,118,0,59,
-120,0,0,18,118,0,59,121,0,0,0,4,118,101,99,52,95,97,100,100,0,18,115,117,109,0,59,120,0,0,18,115,
-117,109,0,59,120,0,0,18,118,0,59,122,0,0,0,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,
-86,97,108,0,59,120,0,0,18,115,117,109,0,59,120,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,1,0,0,97,110,
-121,0,1,1,0,0,4,0,118,0,0,0,1,3,2,90,95,0,0,9,0,1,115,117,109,0,0,0,4,118,101,99,52,95,97,100,100,
-0,18,115,117,109,0,59,120,0,0,18,118,0,59,120,0,0,18,118,0,59,121,0,0,0,4,118,101,99,52,95,97,100,
-100,0,18,115,117,109,0,59,120,0,0,18,115,117,109,0,59,120,0,0,18,118,0,59,122,0,0,0,4,118,101,99,
-52,95,97,100,100,0,18,115,117,109,0,59,120,0,0,18,115,117,109,0,59,120,0,0,18,118,0,59,119,0,0,0,4,
-118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,115,117,109,0,59,120,
-0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,1,0,0,97,108,108,0,1,1,0,0,2,0,118,0,0,0,1,3,2,90,95,0,0,9,0,
-1,112,114,111,100,0,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,112,114,111,100,0,
-0,18,118,0,59,120,0,0,18,118,0,59,121,0,0,0,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,
-86,97,108,0,0,18,112,114,111,100,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,1,0,0,97,108,108,0,1,1,0,0,3,
-0,118,0,0,0,1,3,2,90,95,0,0,9,0,1,112,114,111,100,0,0,0,4,118,101,99,52,95,109,117,108,116,105,112,
-108,121,0,18,112,114,111,100,0,0,18,118,0,59,120,0,0,18,118,0,59,121,0,0,0,4,118,101,99,52,95,109,
-117,108,116,105,112,108,121,0,18,112,114,111,100,0,0,18,112,114,111,100,0,0,18,118,0,59,122,0,0,0,
-4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,0,18,112,114,111,100,0,0,17,48,0,
-48,0,0,0,0,0,1,90,95,0,0,1,0,0,97,108,108,0,1,1,0,0,4,0,118,0,0,0,1,3,2,90,95,0,0,9,0,1,112,114,
-111,100,0,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,112,114,111,100,0,0,18,118,0,
-59,120,0,0,18,118,0,59,121,0,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,112,114,
-111,100,0,0,18,112,114,111,100,0,0,18,118,0,59,122,0,0,0,4,118,101,99,52,95,109,117,108,116,105,
-112,108,121,0,18,112,114,111,100,0,0,18,112,114,111,100,0,0,18,118,0,59,119,0,0,0,4,118,101,99,52,
-95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,0,18,112,114,111,100,0,0,17,48,0,48,0,0,0,0,0,1,
-90,95,0,0,2,0,0,110,111,116,0,1,1,0,0,2,0,118,0,0,0,1,4,118,101,99,52,95,115,101,113,0,18,95,95,
-114,101,116,86,97,108,0,59,120,121,0,0,18,118,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,3,0,0,110,111,
-116,0,1,1,0,0,3,0,118,0,0,0,1,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,59,
-120,121,122,0,0,18,118,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,4,0,0,110,111,116,0,1,1,0,0,4,0,118,0,
-0,0,1,4,118,101,99,52,95,115,101,113,0,18,95,95,114,101,116,86,97,108,0,0,18,118,0,0,17,48,0,48,0,
-0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,49,68,0,1,1,0,0,16,0,115,97,109,112,108,101,
-114,0,0,1,1,0,0,9,0,99,111,111,114,100,0,0,0,1,4,118,101,99,52,95,116,101,120,95,49,100,0,18,95,95,
-114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,0,0,0,0,1,90,95,
-0,0,12,0,0,116,101,120,116,117,114,101,49,68,80,114,111,106,0,1,1,0,0,16,0,115,97,109,112,108,101,
-114,0,0,1,1,0,0,10,0,99,111,111,114,100,0,0,0,1,4,118,101,99,52,95,116,101,120,95,49,100,95,112,
-114,111,106,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,
-114,100,0,59,120,121,121,121,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,49,68,80,114,
-111,106,0,1,1,0,0,16,0,115,97,109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,0,1,4,
-118,101,99,52,95,116,101,120,95,49,100,95,112,114,111,106,0,18,95,95,114,101,116,86,97,108,0,0,18,
-115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,
-117,114,101,50,68,0,1,1,0,0,17,0,115,97,109,112,108,101,114,0,0,1,1,0,0,10,0,99,111,111,114,100,0,
-0,0,1,4,118,101,99,52,95,116,101,120,95,50,100,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,
-112,108,101,114,0,0,18,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,
-50,68,80,114,111,106,0,1,1,0,0,17,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,
-0,0,0,1,4,118,101,99,52,95,116,101,120,95,50,100,95,112,114,111,106,0,18,95,95,114,101,116,86,97,
-108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,0,59,120,121,122,122,0,0,0,0,1,90,
-95,0,0,12,0,0,116,101,120,116,117,114,101,50,68,80,114,111,106,0,1,1,0,0,17,0,115,97,109,112,108,
-101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,0,1,4,118,101,99,52,95,116,101,120,95,50,100,95,
-112,114,111,106,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,
-111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,51,68,0,1,1,0,0,18,0,115,97,109,
-112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,0,0,0,1,4,118,101,99,52,95,116,101,120,95,51,
-100,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,0,
-0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,51,68,80,114,111,106,0,1,1,0,0,18,0,115,97,
-109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,0,1,4,118,101,99,52,95,116,101,120,95,
-51,100,95,112,114,111,106,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,
-18,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,67,117,98,101,0,1,1,0,
-0,19,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,0,0,0,1,4,118,101,99,52,95,
-116,101,120,95,99,117,98,101,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,
-0,18,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,115,104,97,100,111,119,49,68,0,1,1,0,0,20,0,115,
-97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,0,0,0,1,4,118,101,99,52,95,116,101,120,
-95,49,100,95,115,104,97,100,111,119,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,
-114,0,0,18,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,115,104,97,100,111,119,49,68,80,114,111,
-106,0,1,1,0,0,20,0,115,97,109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,0,1,4,118,
-101,99,52,95,116,101,120,95,49,100,95,112,114,111,106,95,115,104,97,100,111,119,0,18,95,95,114,101,
-116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,
-0,115,104,97,100,111,119,50,68,0,1,1,0,0,21,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,
-111,114,100,0,0,0,1,4,118,101,99,52,95,116,101,120,95,50,100,95,115,104,97,100,111,119,0,18,95,95,
-114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,0,0,0,0,1,90,95,
-0,0,12,0,0,115,104,97,100,111,119,50,68,80,114,111,106,0,1,1,0,0,21,0,115,97,109,112,108,101,114,0,
-0,1,1,0,0,12,0,99,111,111,114,100,0,0,0,1,4,118,101,99,52,95,116,101,120,95,50,100,95,112,114,111,
-106,95,115,104,97,100,111,119,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,
-0,18,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,50,68,82,101,99,116,
-0,1,1,0,0,22,0,115,97,109,112,108,101,114,0,0,1,1,0,0,10,0,99,111,111,114,100,0,0,0,1,4,118,101,99,
-52,95,116,101,120,95,114,101,99,116,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,
-114,0,0,18,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,50,68,82,101,
-99,116,80,114,111,106,0,1,1,0,0,22,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,
-100,0,0,0,1,4,118,101,99,52,95,116,101,120,95,114,101,99,116,95,112,114,111,106,0,18,95,95,114,101,
-116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,0,59,120,121,122,122,0,0,
-0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,50,68,82,101,99,116,80,114,111,106,0,1,1,0,0,22,
-0,115,97,109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,0,1,4,118,101,99,52,95,116,
-101,120,95,114,101,99,116,95,112,114,111,106,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,
-112,108,101,114,0,0,18,99,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,115,104,97,100,111,119,50,
-68,82,101,99,116,0,1,1,0,0,23,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,0,0,
-0,1,4,118,101,99,52,95,116,101,120,95,114,101,99,116,95,115,104,97,100,111,119,0,18,95,95,114,101,
-116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,
-0,115,104,97,100,111,119,50,68,82,101,99,116,80,114,111,106,0,1,1,0,0,23,0,115,97,109,112,108,101,
-114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,0,1,4,118,101,99,52,95,116,101,120,95,114,101,99,116,
-95,112,114,111,106,95,115,104,97,100,111,119,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,
-112,108,101,114,0,0,18,99,111,111,114,100,0,0,0,0,1,90,95,0,0,9,0,0,110,111,105,115,101,49,0,1,1,0,
-0,9,0,120,0,0,0,1,4,102,108,111,97,116,95,110,111,105,115,101,49,0,18,95,95,114,101,116,86,97,108,
-0,0,18,120,0,0,0,0,1,90,95,0,0,9,0,0,110,111,105,115,101,49,0,1,1,0,0,10,0,120,0,0,0,1,4,102,108,
-111,97,116,95,110,111,105,115,101,50,0,18,95,95,114,101,116,86,97,108,0,0,18,120,0,0,0,0,1,90,95,0,
-0,9,0,0,110,111,105,115,101,49,0,1,1,0,0,11,0,120,0,0,0,1,4,102,108,111,97,116,95,110,111,105,115,
-101,51,0,18,95,95,114,101,116,86,97,108,0,0,18,120,0,0,0,0,1,90,95,0,0,9,0,0,110,111,105,115,101,
-49,0,1,1,0,0,12,0,120,0,0,0,1,4,102,108,111,97,116,95,110,111,105,115,101,52,0,18,95,95,114,101,
-116,86,97,108,0,0,18,120,0,0,0,0,1,90,95,0,0,10,0,0,110,111,105,115,101,50,0,1,1,0,0,9,0,120,0,0,0,
-1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,110,111,105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,
-95,95,114,101,116,86,97,108,0,59,121,0,58,110,111,105,115,101,49,0,0,18,120,0,17,49,57,0,51,52,0,0,
-46,0,0,20,0,0,1,90,95,0,0,10,0,0,110,111,105,115,101,50,0,1,1,0,0,10,0,120,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,59,120,0,58,110,111,105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,59,121,0,58,110,111,105,115,101,49,0,0,18,120,0,58,118,101,99,50,0,0,17,49,57,0,51,
-52,0,0,0,17,55,0,54,54,0,0,0,0,46,0,0,20,0,0,1,90,95,0,0,10,0,0,110,111,105,115,101,50,0,1,1,0,0,
-11,0,120,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,110,111,105,115,101,49,0,0,18,120,
-0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,110,111,105,115,101,49,0,0,18,120,0,58,
-118,101,99,51,0,0,17,49,57,0,51,52,0,0,0,17,55,0,54,54,0,0,0,17,51,0,50,51,0,0,0,0,46,0,0,20,0,0,1,
-90,95,0,0,10,0,0,110,111,105,115,101,50,0,1,1,0,0,12,0,120,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,59,120,0,58,110,111,105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,
-59,121,0,58,110,111,105,115,101,49,0,0,18,120,0,58,118,101,99,52,0,0,17,49,57,0,51,52,0,0,0,17,55,
-0,54,54,0,0,0,17,51,0,50,51,0,0,0,17,50,0,55,55,0,0,0,0,46,0,0,20,0,0,1,90,95,0,0,11,0,0,110,111,
-105,115,101,51,0,1,1,0,0,9,0,120,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,110,111,
-105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,110,111,105,
-115,101,49,0,0,18,120,0,17,49,57,0,51,52,0,0,46,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,
-0,58,110,111,105,115,101,49,0,0,18,120,0,17,53,0,52,55,0,0,46,0,0,20,0,0,1,90,95,0,0,11,0,0,110,
-111,105,115,101,51,0,1,1,0,0,10,0,120,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,110,
-111,105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,110,111,
-105,115,101,49,0,0,18,120,0,58,118,101,99,50,0,0,17,49,57,0,51,52,0,0,0,17,55,0,54,54,0,0,0,0,46,0,
-0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,110,111,105,115,101,49,0,0,18,120,0,58,118,
-101,99,50,0,0,17,53,0,52,55,0,0,0,17,49,55,0,56,53,0,0,0,0,46,0,0,20,0,0,1,90,95,0,0,11,0,0,110,
-111,105,115,101,51,0,1,1,0,0,11,0,120,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,110,
-111,105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,110,111,
-105,115,101,49,0,0,18,120,0,58,118,101,99,51,0,0,17,49,57,0,51,52,0,0,0,17,55,0,54,54,0,0,0,17,51,
-0,50,51,0,0,0,0,46,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,110,111,105,115,101,49,
-0,0,18,120,0,58,118,101,99,51,0,0,17,53,0,52,55,0,0,0,17,49,55,0,56,53,0,0,0,17,49,49,0,48,52,0,0,
-0,0,46,0,0,20,0,0,1,90,95,0,0,11,0,0,110,111,105,115,101,51,0,1,1,0,0,12,0,120,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,59,120,0,58,110,111,105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,59,121,0,58,110,111,105,115,101,49,0,0,18,120,0,58,118,101,99,52,0,0,17,49,57,
-0,51,52,0,0,0,17,55,0,54,54,0,0,0,17,51,0,50,51,0,0,0,17,50,0,55,55,0,0,0,0,46,0,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,59,122,0,58,110,111,105,115,101,49,0,0,18,120,0,58,118,101,99,52,0,0,17,53,
-0,52,55,0,0,0,17,49,55,0,56,53,0,0,0,17,49,49,0,48,52,0,0,0,17,49,51,0,49,57,0,0,0,0,46,0,0,20,0,0,
-1,90,95,0,0,12,0,0,110,111,105,115,101,52,0,1,1,0,0,9,0,120,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,59,120,0,58,110,111,105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,
-59,121,0,58,110,111,105,115,101,49,0,0,18,120,0,17,49,57,0,51,52,0,0,46,0,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,59,122,0,58,110,111,105,115,101,49,0,0,18,120,0,17,53,0,52,55,0,0,46,0,0,20,0,
-9,18,95,95,114,101,116,86,97,108,0,59,119,0,58,110,111,105,115,101,49,0,0,18,120,0,17,50,51,0,53,
-52,0,0,46,0,0,20,0,0,1,90,95,0,0,12,0,0,110,111,105,115,101,52,0,1,1,0,0,10,0,120,0,0,0,1,9,18,95,
-95,114,101,116,86,97,108,0,59,120,0,58,110,111,105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,59,121,0,58,110,111,105,115,101,49,0,0,18,120,0,58,118,101,99,50,0,0,17,49,57,
-0,51,52,0,0,0,17,55,0,54,54,0,0,0,0,46,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,110,
-111,105,115,101,49,0,0,18,120,0,58,118,101,99,50,0,0,17,53,0,52,55,0,0,0,17,49,55,0,56,53,0,0,0,0,
-46,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,58,110,111,105,115,101,49,0,0,18,120,0,58,
-118,101,99,50,0,0,17,50,51,0,53,52,0,0,0,17,50,57,0,49,49,0,0,0,0,46,0,0,20,0,0,1,90,95,0,0,12,0,0,
-110,111,105,115,101,52,0,1,1,0,0,11,0,120,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,
-110,111,105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,110,
-111,105,115,101,49,0,0,18,120,0,58,118,101,99,51,0,0,17,49,57,0,51,52,0,0,0,17,55,0,54,54,0,0,0,17,
-51,0,50,51,0,0,0,0,46,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,110,111,105,115,101,
-49,0,0,18,120,0,58,118,101,99,51,0,0,17,53,0,52,55,0,0,0,17,49,55,0,56,53,0,0,0,17,49,49,0,48,52,0,
-0,0,0,46,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,58,110,111,105,115,101,49,0,0,18,120,
-0,58,118,101,99,51,0,0,17,50,51,0,53,52,0,0,0,17,50,57,0,49,49,0,0,0,17,51,49,0,57,49,0,0,0,0,46,0,
-0,20,0,0,1,90,95,0,0,12,0,0,110,111,105,115,101,52,0,1,1,0,0,12,0,120,0,0,0,1,9,18,95,95,114,101,
-116,86,97,108,0,59,120,0,58,110,111,105,115,101,49,0,0,18,120,0,0,0,20,0,9,18,95,95,114,101,116,86,
-97,108,0,59,121,0,58,110,111,105,115,101,49,0,0,18,120,0,58,118,101,99,52,0,0,17,49,57,0,51,52,0,0,
-0,17,55,0,54,54,0,0,0,17,51,0,50,51,0,0,0,17,50,0,55,55,0,0,0,0,46,0,0,20,0,9,18,95,95,114,101,116,
-86,97,108,0,59,122,0,58,110,111,105,115,101,49,0,0,18,120,0,58,118,101,99,52,0,0,17,53,0,52,55,0,0,
-0,17,49,55,0,56,53,0,0,0,17,49,49,0,48,52,0,0,0,17,49,51,0,49,57,0,0,0,0,46,0,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,59,119,0,58,110,111,105,115,101,49,0,0,18,120,0,58,118,101,99,52,0,0,17,50,
-51,0,53,52,0,0,0,17,50,57,0,49,49,0,0,0,17,51,49,0,57,49,0,0,0,17,51,55,0,52,56,0,0,0,0,46,0,0,20,
-0,0,0
diff --git a/src/mesa/shader/slang/library/slang_core_gc.h b/src/mesa/shader/slang/library/slang_core_gc.h
deleted file mode 100644 (file)
index b3d3e87..0000000
+++ /dev/null
@@ -1,869 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE FOLLOWING FILE: */
-/* slang_core.gc */
-
-5,1,90,95,0,0,5,0,1,1,1,0,0,9,0,102,0,0,0,1,4,118,101,99,52,95,116,111,95,105,118,101,99,52,0,18,
-95,95,114,101,116,86,97,108,0,0,18,102,0,0,0,0,1,90,95,0,0,5,0,1,1,1,0,0,1,0,98,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,18,98,0,20,0,0,1,90,95,0,0,5,0,1,1,1,0,0,5,0,105,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,18,105,0,20,0,0,1,90,95,0,0,1,0,1,1,1,0,0,5,0,105,0,0,0,1,4,118,101,99,52,95,
-115,110,101,0,18,95,95,114,101,116,86,97,108,0,0,18,105,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,1,0,1,
-1,1,0,0,9,0,102,0,0,0,1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,0,18,102,
-0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,1,0,1,1,1,0,0,1,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,
-0,18,98,0,20,0,0,1,90,95,0,0,9,0,1,1,1,0,0,5,0,105,0,0,0,1,4,105,118,101,99,52,95,116,111,95,118,
-101,99,52,0,18,95,95,114,101,116,86,97,108,0,0,18,105,0,0,0,0,1,90,95,0,0,9,0,1,1,1,0,0,1,0,98,0,0,
-0,1,4,105,118,101,99,52,95,116,111,95,118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,0,18,98,0,0,
-0,0,1,90,95,0,0,9,0,1,1,1,0,0,9,0,102,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,102,0,20,0,0,1,
-90,95,0,0,10,0,1,1,1,0,0,9,0,120,0,0,1,1,0,0,9,0,121,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,
-120,0,18,120,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,121,0,20,0,0,1,90,95,0,0,10,0,1,
-1,1,0,0,9,0,102,0,0,0,1,4,118,101,99,52,95,109,111,118,101,0,18,95,95,114,101,116,86,97,108,0,59,
-120,121,0,0,18,102,0,0,0,0,1,90,95,0,0,10,0,1,1,1,0,0,5,0,105,0,0,0,1,4,105,118,101,99,52,95,116,
-111,95,118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,105,0,0,0,0,1,90,95,0,0,
-10,0,1,1,1,0,0,1,0,98,0,0,0,1,4,105,118,101,99,52,95,116,111,95,118,101,99,52,0,18,95,95,114,101,
-116,86,97,108,0,59,120,121,0,0,18,98,0,0,0,0,1,90,95,0,0,10,0,1,1,1,0,0,2,0,98,0,0,0,1,4,105,118,
-101,99,52,95,116,111,95,118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,98,0,0,
-0,0,1,90,95,0,0,10,0,1,1,1,0,0,11,0,118,0,0,0,1,4,118,101,99,52,95,109,111,118,101,0,18,95,95,114,
-101,116,86,97,108,0,59,120,121,0,0,18,118,0,59,120,121,0,0,0,0,1,90,95,0,0,10,0,1,1,1,0,0,12,0,118,
-0,0,0,1,4,118,101,99,52,95,109,111,118,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,
-118,0,59,120,121,0,0,0,0,1,90,95,0,0,11,0,1,1,1,0,0,9,0,120,0,0,1,1,0,0,9,0,121,0,0,1,1,0,0,9,0,
-122,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,120,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,59,121,0,18,121,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,122,0,20,0,0,1,90,95,0,
-0,11,0,1,1,1,0,0,9,0,102,0,0,0,1,4,118,101,99,52,95,109,111,118,101,0,18,95,95,114,101,116,86,97,
-108,0,59,120,121,122,0,0,18,102,0,0,0,0,1,90,95,0,0,11,0,1,1,1,0,0,5,0,105,0,0,0,1,4,105,118,101,
-99,52,95,116,111,95,118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,105,0,0,
-0,0,1,90,95,0,0,11,0,1,1,1,0,0,1,0,98,0,0,0,1,4,105,118,101,99,52,95,116,111,95,118,101,99,52,0,18,
-95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,98,0,0,0,0,1,90,95,0,0,11,0,1,1,1,0,0,3,0,98,0,
-0,0,1,4,105,118,101,99,52,95,116,111,95,118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,59,120,
-121,122,0,0,18,98,0,0,0,0,1,90,95,0,0,11,0,1,1,1,0,0,12,0,118,0,0,0,1,4,118,101,99,52,95,109,111,
-118,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,118,0,0,0,0,1,90,95,0,0,12,0,1,1,
-1,0,0,9,0,120,0,0,1,1,0,0,9,0,121,0,0,1,1,0,0,9,0,122,0,0,1,1,0,0,9,0,119,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,59,120,0,18,120,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,121,0,20,
-0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,122,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,
-119,0,18,119,0,20,0,0,1,90,95,0,0,12,0,1,1,1,0,0,9,0,102,0,0,0,1,4,118,101,99,52,95,109,111,118,
-101,0,18,95,95,114,101,116,86,97,108,0,0,18,102,0,0,0,0,1,90,95,0,0,12,0,1,1,1,0,0,5,0,105,0,0,0,1,
-4,105,118,101,99,52,95,116,111,95,118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,0,18,105,0,0,0,
-0,1,90,95,0,0,12,0,1,1,1,0,0,1,0,98,0,0,0,1,4,105,118,101,99,52,95,116,111,95,118,101,99,52,0,18,
-95,95,114,101,116,86,97,108,0,0,18,98,0,0,0,0,1,90,95,0,0,12,0,1,1,1,0,0,4,0,98,0,0,0,1,4,105,118,
-101,99,52,95,116,111,95,118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,0,18,98,0,0,0,0,1,90,95,0,
-0,12,0,1,1,1,0,0,8,0,105,0,0,0,1,4,105,118,101,99,52,95,116,111,95,118,101,99,52,0,18,95,95,114,
-101,116,86,97,108,0,0,18,105,0,0,0,0,1,90,95,0,0,12,0,1,1,1,0,0,11,0,118,51,0,0,1,1,0,0,9,0,102,0,
-0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,18,118,51,0,20,0,9,18,95,95,114,101,116,
-86,97,108,0,59,119,0,18,102,0,20,0,0,1,90,95,0,0,12,0,1,1,1,0,0,10,0,118,50,0,0,1,1,0,0,9,0,102,49,
-0,0,1,1,0,0,9,0,102,50,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,121,0,18,118,50,0,20,0,9,
-18,95,95,114,101,116,86,97,108,0,59,122,0,18,102,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,
-119,0,18,102,50,0,20,0,0,1,90,95,0,0,6,0,1,1,1,0,0,5,0,105,0,0,1,1,0,0,5,0,106,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,59,120,0,18,105,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,106,
-0,20,0,0,1,90,95,0,0,6,0,1,1,1,0,0,5,0,105,0,0,0,1,4,118,101,99,52,95,109,111,118,101,0,18,95,95,
-114,101,116,86,97,108,0,59,120,121,0,0,18,105,0,0,0,0,1,90,95,0,0,6,0,1,1,1,0,0,9,0,102,0,0,0,1,4,
-118,101,99,52,95,116,111,95,105,118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,
-102,0,0,0,0,1,90,95,0,0,6,0,1,1,1,0,0,1,0,98,0,0,0,1,4,118,101,99,52,95,116,111,95,105,118,101,99,
-52,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,98,0,0,0,0,1,90,95,0,0,7,0,1,1,1,0,0,5,0,
-105,0,0,1,1,0,0,5,0,106,0,0,1,1,0,0,5,0,107,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,
-105,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,18,106,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,59,122,0,18,107,0,20,0,0,1,90,95,0,0,7,0,1,1,1,0,0,5,0,105,0,0,0,1,4,118,101,99,52,95,109,
-111,118,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,105,0,0,0,0,1,90,95,0,0,7,0,1,
-1,1,0,0,9,0,102,0,0,0,1,4,118,101,99,52,95,116,111,95,105,118,101,99,52,0,18,95,95,114,101,116,86,
-97,108,0,59,120,121,122,0,0,18,102,0,0,0,0,1,90,95,0,0,7,0,1,1,1,0,0,1,0,98,0,0,0,1,4,118,101,99,
-52,95,109,111,118,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,98,0,0,0,0,1,90,95,
-0,0,8,0,1,1,1,0,0,5,0,120,0,0,1,1,0,0,5,0,121,0,0,1,1,0,0,5,0,122,0,0,1,1,0,0,5,0,119,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,59,120,0,18,120,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,
-18,121,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,122,0,20,0,9,18,95,95,114,101,116,86,
-97,108,0,59,119,0,18,119,0,20,0,0,1,90,95,0,0,8,0,1,1,1,0,0,5,0,105,0,0,0,1,4,118,101,99,52,95,109,
-111,118,101,0,18,95,95,114,101,116,86,97,108,0,0,18,105,0,0,0,0,1,90,95,0,0,8,0,1,1,1,0,0,9,0,102,
-0,0,0,1,4,118,101,99,52,95,116,111,95,105,118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,0,18,
-102,0,0,0,0,1,90,95,0,0,8,0,1,1,1,0,0,1,0,98,0,0,0,1,4,118,101,99,52,95,116,111,95,105,118,101,99,
-52,0,18,95,95,114,101,116,86,97,108,0,0,18,98,0,0,0,0,1,90,95,0,0,2,0,1,1,1,0,0,1,0,98,49,0,0,1,1,
-0,0,1,0,98,50,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,98,49,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,59,121,0,18,98,50,0,20,0,0,1,90,95,0,0,2,0,1,1,1,0,0,5,0,105,49,0,0,1,1,0,0,5,
-0,105,50,0,0,0,1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,
-105,49,0,0,17,48,0,48,0,0,0,0,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,
-121,0,0,18,105,50,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,2,0,1,1,1,0,0,1,0,98,0,0,0,1,4,118,101,99,
-52,95,109,111,118,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,98,0,0,0,0,1,90,95,0,0,
-2,0,1,1,1,0,0,9,0,102,0,0,0,1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,
-120,121,0,0,18,102,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,2,0,1,1,1,0,0,5,0,105,0,0,0,1,4,118,101,99,
-52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,105,0,0,17,48,0,48,0,0,0,0,
-0,1,90,95,0,0,2,0,1,1,1,0,0,10,0,118,0,0,0,1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,
-86,97,108,0,59,120,121,0,0,18,118,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,2,0,1,1,1,0,0,6,0,118,0,0,0,
-1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,118,0,0,17,
-48,0,48,0,0,0,0,0,1,90,95,0,0,3,0,1,1,1,0,0,1,0,98,49,0,0,1,1,0,0,1,0,98,50,0,0,1,1,0,0,1,0,98,51,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,98,49,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,59,121,0,18,98,50,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,98,51,0,20,0,0,1,90,
-95,0,0,3,0,1,1,1,0,0,9,0,102,49,0,0,1,1,0,0,9,0,102,50,0,0,1,1,0,0,9,0,102,51,0,0,0,1,4,118,101,99,
-52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,102,49,0,0,17,48,0,48,0,0,0,0,4,
-118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,102,50,0,0,17,48,0,
-48,0,0,0,0,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,18,102,51,
-0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,3,0,1,1,1,0,0,1,0,98,0,0,0,1,4,118,101,99,52,95,109,111,118,
-101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,98,0,0,0,0,1,90,95,0,0,3,0,1,1,1,0,0,
-9,0,102,0,0,0,1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,
-0,18,102,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,3,0,1,1,1,0,0,5,0,105,0,0,0,1,4,118,101,99,52,95,115,
-110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,105,0,0,17,48,0,48,0,0,0,0,0,1,90,
-95,0,0,3,0,1,1,1,0,0,11,0,118,0,0,0,1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,
-108,0,59,120,121,122,0,0,18,118,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,3,0,1,1,1,0,0,7,0,118,0,0,0,1,
-4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,118,0,0,17,
-48,0,48,0,0,0,0,0,1,90,95,0,0,4,0,1,1,1,0,0,1,0,98,49,0,0,1,1,0,0,1,0,98,50,0,0,1,1,0,0,1,0,98,51,
-0,0,1,1,0,0,1,0,98,52,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,18,98,49,0,20,0,9,18,95,
-95,114,101,116,86,97,108,0,59,121,0,18,98,50,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,18,
-98,51,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,119,0,18,98,52,0,20,0,0,1,90,95,0,0,4,0,1,1,1,0,
-0,9,0,102,49,0,0,1,1,0,0,9,0,102,50,0,0,1,1,0,0,9,0,102,51,0,0,1,1,0,0,9,0,102,52,0,0,0,1,3,2,90,
-95,1,0,9,0,1,122,101,114,111,0,2,17,48,0,48,0,0,0,0,4,118,101,99,52,95,115,110,101,0,18,95,95,114,
-101,116,86,97,108,0,59,120,0,0,18,102,49,0,0,18,122,101,114,111,0,0,0,4,118,101,99,52,95,115,110,
-101,0,18,95,95,114,101,116,86,97,108,0,59,121,0,0,18,102,50,0,0,18,122,101,114,111,0,0,0,4,118,101,
-99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,122,0,0,18,102,51,0,0,18,122,101,114,
-111,0,0,0,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,119,0,0,18,102,52,0,
-0,18,122,101,114,111,0,0,0,0,1,90,95,0,0,4,0,1,1,1,0,0,1,0,98,0,0,0,1,4,118,101,99,52,95,109,111,
-118,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,119,0,0,18,98,0,0,0,0,1,90,95,0,0,4,0,1,
-1,1,0,0,9,0,102,0,0,0,1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,
-121,122,119,0,0,18,102,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,4,0,1,1,1,0,0,5,0,105,0,0,0,1,4,118,
-101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,119,0,0,18,105,0,0,17,
-48,0,48,0,0,0,0,0,1,90,95,0,0,4,0,1,1,1,0,0,12,0,118,0,0,0,1,4,118,101,99,52,95,115,110,101,0,18,
-95,95,114,101,116,86,97,108,0,59,120,121,122,119,0,0,18,118,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,4,
-0,1,1,1,0,0,8,0,118,0,0,0,1,4,118,101,99,52,95,115,110,101,0,18,95,95,114,101,116,86,97,108,0,59,
-120,121,122,119,0,0,18,118,0,0,17,48,0,48,0,0,0,0,0,1,90,95,0,0,13,0,1,1,1,0,0,9,0,109,48,48,0,0,1,
-1,0,0,9,0,109,49,48,0,0,1,1,0,0,9,0,109,48,49,0,0,1,1,0,0,9,0,109,49,49,0,0,0,1,9,18,95,95,114,101,
-116,86,97,108,0,16,8,48,0,57,59,120,0,18,109,48,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,
-48,0,57,59,121,0,18,109,49,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,120,0,18,
-109,48,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,121,0,18,109,49,49,0,20,0,0,1,
-90,95,0,0,13,0,1,1,1,0,0,9,0,102,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,120,0,
-18,102,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,121,0,17,48,0,48,0,0,20,0,9,18,95,
-95,114,101,116,86,97,108,0,16,10,49,0,57,59,120,0,17,48,0,48,0,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,49,0,57,59,121,0,18,102,0,20,0,0,1,90,95,0,0,13,0,1,1,1,0,0,5,0,105,0,0,0,1,8,58,109,
-97,116,50,0,0,58,102,108,111,97,116,0,0,18,105,0,0,0,0,0,0,0,1,90,95,0,0,13,0,1,1,1,0,0,1,0,98,0,0,
-0,1,8,58,109,97,116,50,0,0,58,102,108,111,97,116,0,0,18,98,0,0,0,0,0,0,0,1,90,95,0,0,13,0,1,1,1,0,
-0,10,0,99,48,0,0,1,1,0,0,10,0,99,49,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,99,
-48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,99,49,0,20,0,0,1,90,95,0,0,14,0,1,1,
-1,0,0,9,0,109,48,48,0,0,1,1,0,0,9,0,109,49,48,0,0,1,1,0,0,9,0,109,50,48,0,0,1,1,0,0,9,0,109,48,49,
-0,0,1,1,0,0,9,0,109,49,49,0,0,1,1,0,0,9,0,109,50,49,0,0,1,1,0,0,9,0,109,48,50,0,0,1,1,0,0,9,0,109,
-49,50,0,0,1,1,0,0,9,0,109,50,50,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,120,0,
-18,109,48,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,121,0,18,109,49,48,0,20,0,9,
-18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,122,0,18,109,50,48,0,20,0,9,18,95,95,114,101,116,
-86,97,108,0,16,10,49,0,57,59,120,0,18,109,48,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,
-0,57,59,121,0,18,109,49,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,122,0,18,109,
-50,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,59,120,0,18,109,48,50,0,20,0,9,18,95,
-95,114,101,116,86,97,108,0,16,10,50,0,57,59,121,0,18,109,49,50,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,50,0,57,59,122,0,18,109,50,50,0,20,0,0,1,90,95,0,0,14,0,1,1,1,0,0,9,0,102,0,0,0,1,3,2,
-90,95,0,0,10,0,1,118,0,2,58,118,101,99,50,0,0,18,102,0,0,17,48,0,48,0,0,0,0,0,0,9,18,95,95,114,101,
-116,86,97,108,0,16,8,48,0,57,18,118,0,59,120,121,121,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,
-10,49,0,57,18,118,0,59,121,120,121,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,118,
-0,59,121,121,120,0,20,0,0,1,90,95,0,0,14,0,1,1,1,0,0,5,0,105,0,0,0,1,8,58,109,97,116,51,0,0,58,102,
-108,111,97,116,0,0,18,105,0,0,0,0,0,0,0,1,90,95,0,0,14,0,1,1,1,0,0,1,0,98,0,0,0,1,8,58,109,97,116,
-51,0,0,58,102,108,111,97,116,0,0,18,98,0,0,0,0,0,0,0,1,90,95,0,0,14,0,1,1,1,0,0,11,0,99,48,0,0,1,1,
-0,0,11,0,99,49,0,0,1,1,0,0,11,0,99,50,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-99,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,99,49,0,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,50,0,57,18,99,50,0,20,0,0,1,90,95,0,0,15,0,1,1,1,0,0,9,0,109,48,48,0,0,1,1,0,
-0,9,0,109,49,48,0,0,1,1,0,0,9,0,109,50,48,0,0,1,1,0,0,9,0,109,51,48,0,0,1,1,0,0,9,0,109,48,49,0,0,
-1,1,0,0,9,0,109,49,49,0,0,1,1,0,0,9,0,109,50,49,0,0,1,1,0,0,9,0,109,51,49,0,0,1,1,0,0,9,0,109,48,
-50,0,0,1,1,0,0,9,0,109,49,50,0,0,1,1,0,0,9,0,109,50,50,0,0,1,1,0,0,9,0,109,51,50,0,0,1,1,0,0,9,0,
-109,48,51,0,0,1,1,0,0,9,0,109,49,51,0,0,1,1,0,0,9,0,109,50,51,0,0,1,1,0,0,9,0,109,51,51,0,0,0,1,9,
-18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,120,0,18,109,48,48,0,20,0,9,18,95,95,114,101,116,
-86,97,108,0,16,8,48,0,57,59,121,0,18,109,49,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,
-57,59,122,0,18,109,50,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,59,119,0,18,109,51,
-48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,59,120,0,18,109,48,49,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,49,0,57,59,121,0,18,109,49,49,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,49,0,57,59,122,0,18,109,50,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,
-59,119,0,18,109,51,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,59,120,0,18,109,48,
-50,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,59,121,0,18,109,49,50,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,50,0,57,59,122,0,18,109,50,50,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,50,0,57,59,119,0,18,109,51,50,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,
-59,120,0,18,109,48,51,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,59,121,0,18,109,49,
-51,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,59,122,0,18,109,50,51,0,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,51,0,57,59,119,0,18,109,51,51,0,20,0,0,1,90,95,0,0,15,0,1,1,1,0,0,9,
-0,102,0,0,0,1,3,2,90,95,0,0,10,0,1,118,0,2,58,118,101,99,50,0,0,18,102,0,0,17,48,0,48,0,0,0,0,0,0,
-9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,118,0,59,120,121,121,121,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,16,10,49,0,57,18,118,0,59,121,120,121,121,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,50,0,57,18,118,0,59,121,121,120,121,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,
-0,57,18,118,0,59,121,121,121,120,0,20,0,0,1,90,95,0,0,15,0,1,1,1,0,0,5,0,105,0,0,0,1,8,58,109,97,
-116,52,0,0,58,102,108,111,97,116,0,0,18,105,0,0,0,0,0,0,0,1,90,95,0,0,15,0,1,1,1,0,0,1,0,98,0,0,0,
-1,8,58,109,97,116,52,0,0,58,102,108,111,97,116,0,0,18,98,0,0,0,0,0,0,0,1,90,95,0,0,15,0,1,1,1,0,0,
-12,0,99,48,0,0,1,1,0,0,12,0,99,49,0,0,1,1,0,0,12,0,99,50,0,0,1,1,0,0,12,0,99,51,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,16,8,48,0,57,18,99,48,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,
-57,18,99,49,0,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,99,50,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,16,10,51,0,57,18,99,51,0,20,0,0,1,90,95,0,0,5,0,2,26,1,1,0,0,5,0,97,0,0,1,1,0,
-0,5,0,98,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,
-98,0,0,0,0,1,90,95,0,0,5,0,2,27,1,1,0,0,5,0,97,0,0,1,1,0,0,5,0,98,0,0,0,1,4,118,101,99,52,95,115,
-117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,
-5,0,2,21,1,1,0,0,5,0,97,0,0,1,1,0,0,5,0,98,0,0,0,1,4,118,101,99,52,95,109,117,108,116,105,112,108,
-121,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,5,0,2,22,1,1,0,0,5,0,
-97,0,0,1,1,0,0,5,0,98,0,0,0,1,3,2,90,95,0,0,9,0,1,98,73,110,118,0,0,1,1,120,0,0,0,4,102,108,111,97,
-116,95,114,99,112,0,18,98,73,110,118,0,0,18,98,0,0,0,4,118,101,99,52,95,109,117,108,116,105,112,
-108,121,0,18,120,0,0,18,97,0,0,18,98,73,110,118,0,0,0,4,118,101,99,52,95,116,111,95,105,118,101,99,
-52,0,18,95,95,114,101,116,86,97,108,0,0,18,120,0,0,0,0,1,90,95,0,0,6,0,2,26,1,1,0,0,6,0,97,0,0,1,1,
-0,0,6,0,98,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,
-98,0,0,0,0,1,90,95,0,0,6,0,2,27,1,1,0,0,6,0,97,0,0,1,1,0,0,6,0,98,0,0,0,1,4,118,101,99,52,95,115,
-117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,
-6,0,2,21,1,1,0,0,6,0,97,0,0,1,1,0,0,6,0,98,0,0,0,1,4,118,101,99,52,95,109,117,108,116,105,112,108,
-121,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,6,0,2,22,1,1,0,0,6,0,
-97,0,0,1,1,0,0,6,0,98,0,0,0,1,3,2,90,95,0,0,10,0,1,98,73,110,118,0,0,1,1,120,0,0,0,4,102,108,111,
-97,116,95,114,99,112,0,18,98,73,110,118,0,59,120,0,0,18,98,0,59,120,0,0,0,4,102,108,111,97,116,95,
-114,99,112,0,18,98,73,110,118,0,59,121,0,0,18,98,0,59,121,0,0,0,4,118,101,99,52,95,109,117,108,116,
-105,112,108,121,0,18,120,0,0,18,97,0,0,18,98,73,110,118,0,0,0,4,118,101,99,52,95,116,111,95,105,
-118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,0,18,120,0,0,0,0,1,90,95,0,0,7,0,2,26,1,1,0,0,7,0,
-97,0,0,1,1,0,0,7,0,98,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,101,116,86,97,108,0,0,
-18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,7,0,2,27,1,1,0,0,7,0,97,0,0,1,1,0,0,7,0,98,0,0,0,1,4,118,101,
-99,52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,
-1,90,95,0,0,7,0,2,21,1,1,0,0,7,0,97,0,0,1,1,0,0,7,0,98,0,0,0,1,4,118,101,99,52,95,109,117,108,116,
-105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,7,0,2,22,
-1,1,0,0,7,0,97,0,0,1,1,0,0,7,0,98,0,0,0,1,3,2,90,95,0,0,11,0,1,98,73,110,118,0,0,1,1,120,0,0,0,4,
-102,108,111,97,116,95,114,99,112,0,18,98,73,110,118,0,59,120,0,0,18,98,0,59,120,0,0,0,4,102,108,
-111,97,116,95,114,99,112,0,18,98,73,110,118,0,59,121,0,0,18,98,0,59,121,0,0,0,4,102,108,111,97,116,
-95,114,99,112,0,18,98,73,110,118,0,59,122,0,0,18,98,0,59,122,0,0,0,4,118,101,99,52,95,109,117,108,
-116,105,112,108,121,0,18,120,0,0,18,97,0,0,18,98,73,110,118,0,0,0,4,118,101,99,52,95,116,111,95,
-105,118,101,99,52,0,18,95,95,114,101,116,86,97,108,0,0,18,120,0,0,0,0,1,90,95,0,0,8,0,2,26,1,1,0,0,
-8,0,97,0,0,1,1,0,0,8,0,98,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,101,116,86,97,108,0,
-0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,8,0,2,27,1,1,0,0,8,0,97,0,0,1,1,0,0,8,0,98,0,0,0,1,4,118,101,
-99,52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,
-1,90,95,0,0,8,0,2,21,1,1,0,0,8,0,97,0,0,1,1,0,0,8,0,98,0,0,0,1,4,118,101,99,52,95,109,117,108,116,
-105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,8,0,2,22,
-1,1,0,0,8,0,97,0,0,1,1,0,0,8,0,98,0,0,0,1,3,2,90,95,0,0,12,0,1,98,73,110,118,0,0,1,1,120,0,0,0,4,
-102,108,111,97,116,95,114,99,112,0,18,98,73,110,118,0,59,120,0,0,18,98,0,59,120,0,0,0,4,102,108,
-111,97,116,95,114,99,112,0,18,98,73,110,118,0,59,121,0,0,18,98,0,59,121,0,0,0,4,102,108,111,97,116,
-95,114,99,112,0,18,98,73,110,118,0,59,122,0,0,18,98,0,59,122,0,0,0,4,102,108,111,97,116,95,114,99,
-112,0,18,98,73,110,118,0,59,119,0,0,18,98,0,59,119,0,0,0,4,118,101,99,52,95,109,117,108,116,105,
-112,108,121,0,18,120,0,0,18,97,0,0,18,98,73,110,118,0,0,0,4,118,101,99,52,95,116,111,95,105,118,
-101,99,52,0,18,95,95,114,101,116,86,97,108,0,0,18,120,0,0,0,0,1,90,95,0,0,9,0,2,26,1,1,0,0,9,0,97,
-0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,101,116,86,97,108,0,0,18,
-97,0,0,18,98,0,0,0,0,1,90,95,0,0,9,0,2,27,1,1,0,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,
-52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,
-90,95,0,0,9,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,109,117,108,116,
-105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,9,0,2,22,
-1,1,0,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,0,0,9,0,1,98,73,110,118,0,0,0,4,102,108,111,97,
-116,95,114,99,112,0,18,98,73,110,118,0,59,120,0,0,18,98,0,0,0,4,118,101,99,52,95,109,117,108,116,
-105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,73,110,118,0,0,0,0,1,90,95,0,
-0,10,0,2,26,1,1,0,0,10,0,118,0,0,1,1,0,0,10,0,117,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,
-114,101,116,86,97,108,0,59,120,121,0,0,18,118,0,0,18,117,0,0,0,0,1,90,95,0,0,10,0,2,27,1,1,0,0,10,
-0,118,0,0,1,1,0,0,10,0,117,0,0,0,1,4,118,101,99,52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,
-101,116,86,97,108,0,59,120,121,0,0,18,118,0,0,18,117,0,0,0,0,1,90,95,0,0,10,0,2,21,1,1,0,0,10,0,
-118,0,0,1,1,0,0,10,0,117,0,0,0,1,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,
-101,116,86,97,108,0,59,120,121,0,0,18,118,0,0,18,117,0,0,0,0,1,90,95,0,0,10,0,2,22,1,1,0,0,10,0,
-118,0,0,1,1,0,0,10,0,117,0,0,0,1,3,2,90,95,0,0,10,0,1,119,0,0,0,4,102,108,111,97,116,95,114,99,112,
-0,18,119,0,59,120,0,0,18,117,0,59,120,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,119,0,59,121,0,
-0,18,117,0,59,121,0,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,
-86,97,108,0,59,120,121,0,0,18,118,0,0,18,119,0,0,0,0,1,90,95,0,0,11,0,2,26,1,1,0,0,11,0,118,0,0,1,
-1,0,0,11,0,117,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,101,116,86,97,108,0,59,120,121,
-122,0,0,18,118,0,0,18,117,0,0,0,0,1,90,95,0,0,11,0,2,27,1,1,0,0,11,0,118,0,0,1,1,0,0,11,0,117,0,0,
-0,1,4,118,101,99,52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,59,120,121,
-122,0,0,18,118,0,0,18,117,0,0,0,0,1,90,95,0,0,11,0,2,21,1,1,0,0,11,0,118,0,0,1,1,0,0,11,0,117,0,0,
-0,1,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,120,
-121,122,0,0,18,118,0,0,18,117,0,0,0,0,1,90,95,0,0,11,0,2,22,1,1,0,0,11,0,118,0,0,1,1,0,0,11,0,117,
-0,0,0,1,3,2,90,95,0,0,11,0,1,119,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,119,0,59,120,0,0,18,
-117,0,59,120,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,119,0,59,121,0,0,18,117,0,59,121,0,0,0,
-4,102,108,111,97,116,95,114,99,112,0,18,119,0,59,122,0,0,18,117,0,59,122,0,0,0,4,118,101,99,52,95,
-109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,118,0,0,
-18,119,0,0,0,0,1,90,95,0,0,12,0,2,26,1,1,0,0,12,0,118,0,0,1,1,0,0,12,0,117,0,0,0,1,4,118,101,99,52,
-95,97,100,100,0,18,95,95,114,101,116,86,97,108,0,0,18,118,0,0,18,117,0,0,0,0,1,90,95,0,0,12,0,2,27,
-1,1,0,0,12,0,118,0,0,1,1,0,0,12,0,117,0,0,0,1,4,118,101,99,52,95,115,117,98,116,114,97,99,116,0,18,
-95,95,114,101,116,86,97,108,0,0,18,118,0,0,18,117,0,0,0,0,1,90,95,0,0,12,0,2,21,1,1,0,0,12,0,118,0,
-0,1,1,0,0,12,0,117,0,0,0,1,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,
-116,86,97,108,0,0,18,118,0,0,18,117,0,0,0,0,1,90,95,0,0,12,0,2,22,1,1,0,0,12,0,118,0,0,1,1,0,0,12,
-0,117,0,0,0,1,3,2,90,95,0,0,12,0,1,119,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,119,0,59,120,
-0,0,18,117,0,59,120,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,119,0,59,121,0,0,18,117,0,59,121,
-0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,119,0,59,122,0,0,18,117,0,59,122,0,0,0,4,102,108,111,
-97,116,95,114,99,112,0,18,119,0,59,119,0,0,18,117,0,59,119,0,0,0,4,118,101,99,52,95,109,117,108,
-116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,18,118,0,0,18,119,0,0,0,0,1,90,95,0,0,10,
-0,2,26,1,1,0,0,9,0,97,0,0,1,1,0,0,10,0,117,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,
-101,116,86,97,108,0,59,120,121,0,0,18,97,0,0,18,117,0,59,120,121,0,0,0,0,1,90,95,0,0,10,0,2,26,1,1,
-0,0,10,0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,101,116,86,97,
-108,0,59,120,121,0,0,18,118,0,59,120,121,0,0,18,98,0,0,0,0,1,90,95,0,0,10,0,2,27,1,1,0,0,9,0,97,0,
-0,1,1,0,0,10,0,117,0,0,0,1,4,118,101,99,52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,101,116,
-86,97,108,0,59,120,121,0,0,18,97,0,0,18,117,0,59,120,121,0,0,0,0,1,90,95,0,0,10,0,2,27,1,1,0,0,10,
-0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,
-101,116,86,97,108,0,59,120,121,0,0,18,118,0,59,120,121,0,0,18,98,0,0,0,0,1,90,95,0,0,10,0,2,21,1,1,
-0,0,9,0,97,0,0,1,1,0,0,10,0,117,0,0,0,1,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,
-95,114,101,116,86,97,108,0,59,120,121,0,0,18,97,0,0,18,117,0,59,120,121,0,0,0,0,1,90,95,0,0,10,0,2,
-21,1,1,0,0,10,0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,109,117,108,116,105,112,108,121,
-0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,118,0,59,120,121,0,0,18,98,0,0,0,0,1,90,95,0,
-0,10,0,2,22,1,1,0,0,9,0,97,0,0,1,1,0,0,10,0,117,0,0,0,1,3,2,90,95,0,0,10,0,1,105,110,118,85,0,0,0,
-4,102,108,111,97,116,95,114,99,112,0,18,105,110,118,85,0,59,120,0,0,18,117,0,59,120,0,0,0,4,102,
-108,111,97,116,95,114,99,112,0,18,105,110,118,85,0,59,121,0,0,18,117,0,59,121,0,0,0,4,118,101,99,
-52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,97,0,0,
-18,105,110,118,85,0,59,120,121,0,0,0,0,1,90,95,0,0,10,0,2,22,1,1,0,0,10,0,118,0,0,1,1,0,0,9,0,98,0,
-0,0,1,3,2,90,95,0,0,9,0,1,105,110,118,66,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,105,110,118,
-66,0,0,18,98,0,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,
-108,0,59,120,121,0,0,18,118,0,59,120,121,0,0,18,105,110,118,66,0,0,0,0,1,90,95,0,0,11,0,2,26,1,1,0,
-0,9,0,97,0,0,1,1,0,0,11,0,117,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,101,116,86,97,
-108,0,59,120,121,122,0,0,18,97,0,0,18,117,0,59,120,121,122,0,0,0,0,1,90,95,0,0,11,0,2,26,1,1,0,0,
-11,0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,101,116,86,97,108,
-0,59,120,121,122,0,0,18,118,0,59,120,121,122,0,0,18,98,0,0,0,0,1,90,95,0,0,11,0,2,27,1,1,0,0,9,0,
-97,0,0,1,1,0,0,11,0,117,0,0,0,1,4,118,101,99,52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,101,
-116,86,97,108,0,59,120,121,122,0,0,18,97,0,0,18,117,0,59,120,121,122,0,0,0,0,1,90,95,0,0,11,0,2,27,
-1,1,0,0,11,0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,115,117,98,116,114,97,99,116,0,18,
-95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,118,0,59,120,121,122,0,0,18,98,0,0,0,0,1,90,95,
-0,0,11,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,0,11,0,117,0,0,0,1,4,118,101,99,52,95,109,117,108,116,105,
-112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,97,0,0,18,117,0,59,120,121,
-122,0,0,0,0,1,90,95,0,0,11,0,2,21,1,1,0,0,11,0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,
-109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,118,0,59,
-120,121,122,0,0,18,98,0,0,0,0,1,90,95,0,0,11,0,2,22,1,1,0,0,9,0,97,0,0,1,1,0,0,11,0,117,0,0,0,1,3,
-2,90,95,0,0,11,0,1,105,110,118,85,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,105,110,118,85,0,
-59,120,0,0,18,117,0,59,120,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,105,110,118,85,0,59,121,0,
-0,18,117,0,59,121,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,105,110,118,85,0,59,122,0,0,18,117,
-0,59,122,0,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,
-0,59,120,121,122,0,0,18,97,0,0,18,105,110,118,85,0,59,120,121,122,0,0,0,0,1,90,95,0,0,11,0,2,22,1,
-1,0,0,11,0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,0,0,9,0,1,105,110,118,66,0,0,0,4,102,108,111,
-97,116,95,114,99,112,0,18,105,110,118,66,0,0,18,98,0,0,0,4,118,101,99,52,95,109,117,108,116,105,
-112,108,121,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,118,0,59,120,121,122,0,0,18,
-105,110,118,66,0,0,0,0,1,90,95,0,0,12,0,2,26,1,1,0,0,9,0,97,0,0,1,1,0,0,12,0,117,0,0,0,1,4,118,101,
-99,52,95,97,100,100,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,117,0,0,0,0,1,90,95,0,0,12,0,
-2,26,1,1,0,0,12,0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,97,100,100,0,18,95,95,114,101,
-116,86,97,108,0,0,18,118,0,0,18,98,0,0,0,0,1,90,95,0,0,12,0,2,27,1,1,0,0,9,0,97,0,0,1,1,0,0,12,0,
-117,0,0,0,1,4,118,101,99,52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,0,
-18,97,0,0,18,117,0,0,0,0,1,90,95,0,0,12,0,2,27,1,1,0,0,12,0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,
-101,99,52,95,115,117,98,116,114,97,99,116,0,18,95,95,114,101,116,86,97,108,0,0,18,118,0,0,18,98,0,
-0,0,0,1,90,95,0,0,12,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,0,12,0,117,0,0,0,1,4,118,101,99,52,95,109,117,
-108,116,105,112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,117,0,0,0,0,1,90,95,0,0,
-12,0,2,21,1,1,0,0,12,0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,109,117,108,116,105,112,
-108,121,0,18,95,95,114,101,116,86,97,108,0,0,18,118,0,0,18,98,0,0,0,0,1,90,95,0,0,12,0,2,22,1,1,0,
-0,9,0,97,0,0,1,1,0,0,12,0,117,0,0,0,1,3,2,90,95,0,0,12,0,1,105,110,118,85,0,0,0,4,102,108,111,97,
-116,95,114,99,112,0,18,105,110,118,85,0,59,120,0,0,18,117,0,59,120,0,0,0,4,102,108,111,97,116,95,
-114,99,112,0,18,105,110,118,85,0,59,121,0,0,18,117,0,59,121,0,0,0,4,102,108,111,97,116,95,114,99,
-112,0,18,105,110,118,85,0,59,122,0,0,18,117,0,59,122,0,0,0,4,102,108,111,97,116,95,114,99,112,0,18,
-105,110,118,85,0,59,119,0,0,18,117,0,59,119,0,0,0,4,118,101,99,52,95,109,117,108,116,105,112,108,
-121,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,105,110,118,85,0,0,0,0,1,90,95,0,0,12,0,2,22,
-1,1,0,0,12,0,118,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,0,0,9,0,1,105,110,118,66,0,0,0,4,102,108,111,
-97,116,95,114,99,112,0,18,105,110,118,66,0,0,18,98,0,0,0,4,118,101,99,52,95,109,117,108,116,105,
-112,108,121,0,18,95,95,114,101,116,86,97,108,0,0,18,118,0,0,18,105,110,118,66,0,0,0,0,1,90,95,0,0,
-6,0,2,26,1,1,0,0,5,0,97,0,0,1,1,0,0,6,0,117,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,105,118,
-101,99,50,0,0,18,97,0,0,0,18,117,0,46,20,0,0,1,90,95,0,0,6,0,2,26,1,1,0,0,6,0,118,0,0,1,1,0,0,5,0,
-98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,58,105,118,101,99,50,0,0,18,98,0,0,0,46,20,
-0,0,1,90,95,0,0,6,0,2,27,1,1,0,0,5,0,97,0,0,1,1,0,0,6,0,117,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,58,105,118,101,99,50,0,0,18,97,0,0,0,18,117,0,47,20,0,0,1,90,95,0,0,6,0,2,27,1,1,0,0,6,0,118,
-0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,58,105,118,101,99,50,0,0,18,
-98,0,0,0,47,20,0,0,1,90,95,0,0,6,0,2,21,1,1,0,0,5,0,97,0,0,1,1,0,0,6,0,117,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,58,105,118,101,99,50,0,0,18,97,0,0,0,18,117,0,48,20,0,0,1,90,95,0,0,6,0,2,21,1,
-1,0,0,6,0,118,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,58,105,118,
-101,99,50,0,0,18,98,0,0,0,48,20,0,0,1,90,95,0,0,6,0,2,22,1,1,0,0,5,0,97,0,0,1,1,0,0,6,0,117,0,0,0,
-1,9,18,95,95,114,101,116,86,97,108,0,58,105,118,101,99,50,0,0,18,97,0,0,0,18,117,0,49,20,0,0,1,90,
-95,0,0,6,0,2,22,1,1,0,0,6,0,118,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,
-118,0,58,105,118,101,99,50,0,0,18,98,0,0,0,49,20,0,0,1,90,95,0,0,7,0,2,26,1,1,0,0,5,0,97,0,0,1,1,0,
-0,7,0,117,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,105,118,101,99,51,0,0,18,97,0,0,0,18,117,0,
-46,20,0,0,1,90,95,0,0,7,0,2,26,1,1,0,0,7,0,118,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,95,95,114,101,116,
-86,97,108,0,18,118,0,58,105,118,101,99,51,0,0,18,98,0,0,0,46,20,0,0,1,90,95,0,0,7,0,2,27,1,1,0,0,5,
-0,97,0,0,1,1,0,0,7,0,117,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,105,118,101,99,51,0,0,18,97,
-0,0,0,18,117,0,47,20,0,0,1,90,95,0,0,7,0,2,27,1,1,0,0,7,0,118,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,95,
-95,114,101,116,86,97,108,0,18,118,0,58,105,118,101,99,51,0,0,18,98,0,0,0,47,20,0,0,1,90,95,0,0,7,0,
-2,21,1,1,0,0,5,0,97,0,0,1,1,0,0,7,0,117,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,105,118,101,
-99,51,0,0,18,97,0,0,0,18,117,0,48,20,0,0,1,90,95,0,0,7,0,2,21,1,1,0,0,7,0,118,0,0,1,1,0,0,5,0,98,0,
-0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,58,105,118,101,99,51,0,0,18,98,0,0,0,48,20,0,0,1,
-90,95,0,0,7,0,2,22,1,1,0,0,5,0,97,0,0,1,1,0,0,7,0,117,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
-58,105,118,101,99,51,0,0,18,97,0,0,0,18,117,0,49,20,0,0,1,90,95,0,0,7,0,2,22,1,1,0,0,7,0,118,0,0,1,
-1,0,0,5,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,58,105,118,101,99,51,0,0,18,98,0,
-0,0,49,20,0,0,1,90,95,0,0,8,0,2,26,1,1,0,0,5,0,97,0,0,1,1,0,0,8,0,117,0,0,0,1,9,18,95,95,114,101,
-116,86,97,108,0,58,105,118,101,99,52,0,0,18,97,0,0,0,18,117,0,46,20,0,0,1,90,95,0,0,8,0,2,26,1,1,0,
-0,8,0,118,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,58,105,118,101,99,
-52,0,0,18,98,0,0,0,46,20,0,0,1,90,95,0,0,8,0,2,27,1,1,0,0,5,0,97,0,0,1,1,0,0,8,0,117,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,58,105,118,101,99,52,0,0,18,97,0,0,0,18,117,0,47,20,0,0,1,90,95,0,0,
-8,0,2,27,1,1,0,0,8,0,118,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,58,
-105,118,101,99,52,0,0,18,98,0,0,0,47,20,0,0,1,90,95,0,0,8,0,2,21,1,1,0,0,5,0,97,0,0,1,1,0,0,8,0,
-117,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,105,118,101,99,52,0,0,18,97,0,0,0,18,117,0,48,20,
-0,0,1,90,95,0,0,8,0,2,21,1,1,0,0,8,0,118,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,18,118,0,58,105,118,101,99,52,0,0,18,98,0,0,0,48,20,0,0,1,90,95,0,0,8,0,2,22,1,1,0,0,5,0,97,
-0,0,1,1,0,0,8,0,117,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,58,105,118,101,99,52,0,0,18,97,0,0,
-0,18,117,0,49,20,0,0,1,90,95,0,0,8,0,2,22,1,1,0,0,8,0,118,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,18,118,0,58,105,118,101,99,52,0,0,18,98,0,0,0,49,20,0,0,1,90,95,0,0,5,0,2,
-27,1,1,0,0,5,0,97,0,0,0,1,4,118,101,99,52,95,110,101,103,97,116,101,0,18,95,95,114,101,116,86,97,
-108,0,59,120,0,0,18,97,0,0,0,0,1,90,95,0,0,6,0,2,27,1,1,0,0,6,0,118,0,0,0,1,4,118,101,99,52,95,110,
-101,103,97,116,101,0,18,95,95,114,101,116,86,97,108,0,0,18,118,0,0,0,0,1,90,95,0,0,7,0,2,27,1,1,0,
-0,7,0,118,0,0,0,1,4,118,101,99,52,95,110,101,103,97,116,101,0,18,95,95,114,101,116,86,97,108,0,0,
-18,118,0,0,0,0,1,90,95,0,0,8,0,2,27,1,1,0,0,8,0,118,0,0,0,1,4,118,101,99,52,95,110,101,103,97,116,
-101,0,18,95,95,114,101,116,86,97,108,0,0,18,118,0,0,0,0,1,90,95,0,0,9,0,2,27,1,1,0,0,9,0,97,0,0,0,
-1,4,118,101,99,52,95,110,101,103,97,116,101,0,18,95,95,114,101,116,86,97,108,0,59,120,0,0,18,97,0,
-0,0,0,1,90,95,0,0,10,0,2,27,1,1,0,0,10,0,118,0,0,0,1,4,118,101,99,52,95,110,101,103,97,116,101,0,
-18,95,95,114,101,116,86,97,108,0,59,120,121,0,0,18,118,0,59,120,121,0,0,0,0,1,90,95,0,0,11,0,2,27,
-1,1,0,0,11,0,118,0,0,0,1,4,118,101,99,52,95,110,101,103,97,116,101,0,18,95,95,114,101,116,86,97,
-108,0,59,120,121,122,0,0,18,118,0,59,120,121,122,0,0,0,0,1,90,95,0,0,12,0,2,27,1,1,0,0,12,0,118,0,
-0,0,1,4,118,101,99,52,95,110,101,103,97,116,101,0,18,95,95,114,101,116,86,97,108,0,0,18,118,0,0,0,
-0,1,90,95,0,0,13,0,2,27,1,1,0,0,13,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,
-18,109,0,16,8,48,0,57,54,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,
-57,54,20,0,0,1,90,95,0,0,14,0,2,27,1,1,0,0,14,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,
-8,48,0,57,18,109,0,16,8,48,0,57,54,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,
-16,10,49,0,57,54,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,54,
-20,0,0,1,90,95,0,0,15,0,2,27,1,1,0,0,15,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,
-57,18,109,0,16,8,48,0,57,54,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,
-49,0,57,54,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,54,20,0,9,
-18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,16,10,51,0,57,54,20,0,0,1,90,95,0,0,9,0,0,
-100,111,116,0,1,1,0,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,97,0,
-18,98,0,48,20,0,0,1,90,95,0,0,9,0,0,100,111,116,0,1,1,0,0,10,0,97,0,0,1,1,0,0,10,0,98,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,18,97,0,59,120,0,18,98,0,59,120,0,48,18,97,0,59,121,0,18,98,0,59,121,
-0,48,46,20,0,0,1,90,95,0,0,9,0,0,100,111,116,0,1,1,0,0,11,0,97,0,0,1,1,0,0,11,0,98,0,0,0,1,4,118,
-101,99,51,95,100,111,116,0,18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,
-9,0,0,100,111,116,0,1,1,0,0,12,0,97,0,0,1,1,0,0,12,0,98,0,0,0,1,4,118,101,99,52,95,100,111,116,0,
-18,95,95,114,101,116,86,97,108,0,0,18,97,0,0,18,98,0,0,0,0,1,90,95,0,0,5,0,2,1,1,0,2,0,5,0,97,0,0,
-1,1,0,0,5,0,98,0,0,0,1,9,18,97,0,18,97,0,18,98,0,46,20,0,8,18,97,0,0,0,1,90,95,0,0,5,0,2,2,1,0,2,0,
-5,0,97,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,97,0,18,97,0,18,98,0,47,20,0,8,18,97,0,0,0,1,90,95,0,0,5,0,
-2,3,1,0,2,0,5,0,97,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,97,0,18,97,0,18,98,0,48,20,0,8,18,97,0,0,0,1,90,
-95,0,0,5,0,2,4,1,0,2,0,5,0,97,0,0,1,1,0,0,5,0,98,0,0,0,1,9,18,97,0,18,97,0,18,98,0,49,20,0,8,18,97,
-0,0,0,1,90,95,0,0,6,0,2,1,1,0,2,0,6,0,118,0,0,1,1,0,0,6,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,
-46,20,0,8,18,118,0,0,0,1,90,95,0,0,6,0,2,2,1,0,2,0,6,0,118,0,0,1,1,0,0,6,0,117,0,0,0,1,9,18,118,0,
-18,118,0,18,117,0,47,20,0,8,18,118,0,0,0,1,90,95,0,0,6,0,2,3,1,0,2,0,6,0,118,0,0,1,1,0,0,6,0,117,0,
-0,0,1,9,18,118,0,18,118,0,18,117,0,48,20,0,8,18,118,0,0,0,1,90,95,0,0,6,0,2,4,1,0,2,0,6,0,118,0,0,
-1,1,0,0,6,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,49,20,0,8,18,118,0,0,0,1,90,95,0,0,7,0,2,1,1,
-0,2,0,7,0,118,0,0,1,1,0,0,7,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,46,20,0,8,18,118,0,0,0,1,90,
-95,0,0,7,0,2,2,1,0,2,0,7,0,118,0,0,1,1,0,0,7,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,47,20,0,8,
-18,118,0,0,0,1,90,95,0,0,7,0,2,3,1,0,2,0,7,0,118,0,0,1,1,0,0,7,0,117,0,0,0,1,9,18,118,0,18,118,0,
-18,117,0,48,20,0,8,18,118,0,0,0,1,90,95,0,0,7,0,2,4,1,0,2,0,7,0,118,0,0,1,1,0,0,7,0,117,0,0,0,1,9,
-18,118,0,18,118,0,18,117,0,49,20,0,8,18,118,0,0,0,1,90,95,0,0,8,0,2,1,1,0,2,0,8,0,118,0,0,1,1,0,0,
-8,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,46,20,0,8,18,118,0,0,0,1,90,95,0,0,8,0,2,2,1,0,2,0,8,
-0,118,0,0,1,1,0,0,8,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,47,20,0,8,18,118,0,0,0,1,90,95,0,0,
-8,0,2,3,1,0,2,0,8,0,118,0,0,1,1,0,0,8,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,48,20,0,8,18,118,
-0,0,0,1,90,95,0,0,8,0,2,4,1,0,2,0,8,0,118,0,0,1,1,0,0,8,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,
-49,20,0,8,18,118,0,0,0,1,90,95,0,0,9,0,2,1,1,0,2,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,97,0,18,
-97,0,18,98,0,46,20,0,8,18,97,0,0,0,1,90,95,0,0,9,0,2,2,1,0,2,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,9,
-18,97,0,18,97,0,18,98,0,47,20,0,8,18,97,0,0,0,1,90,95,0,0,9,0,2,3,1,0,2,0,9,0,97,0,0,1,1,0,0,9,0,
-98,0,0,0,1,9,18,97,0,18,97,0,18,98,0,48,20,0,8,18,97,0,0,0,1,90,95,0,0,9,0,2,4,1,0,2,0,9,0,97,0,0,
-1,1,0,0,9,0,98,0,0,0,1,9,18,97,0,18,97,0,18,98,0,49,20,0,8,18,97,0,0,0,1,90,95,0,0,10,0,2,1,1,0,2,
-0,10,0,118,0,0,1,1,0,0,10,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,46,20,0,8,18,118,0,0,0,1,90,
-95,0,0,10,0,2,2,1,0,2,0,10,0,118,0,0,1,1,0,0,10,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,47,20,0,
-8,18,118,0,0,0,1,90,95,0,0,10,0,2,3,1,0,2,0,10,0,118,0,0,1,1,0,0,10,0,117,0,0,0,1,9,18,118,0,18,
-118,0,18,117,0,48,20,0,8,18,118,0,0,0,1,90,95,0,0,10,0,2,4,1,0,2,0,10,0,118,0,0,1,1,0,0,10,0,117,0,
-0,0,1,9,18,118,0,18,118,0,18,117,0,49,20,0,8,18,118,0,0,0,1,90,95,0,0,11,0,2,1,1,0,2,0,11,0,118,0,
-0,1,1,0,0,11,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,46,20,0,8,18,118,0,0,0,1,90,95,0,0,11,0,2,
-2,1,0,2,0,11,0,118,0,0,1,1,0,0,11,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,47,20,0,8,18,118,0,0,
-0,1,90,95,0,0,11,0,2,3,1,0,2,0,11,0,118,0,0,1,1,0,0,11,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,
-48,20,0,8,18,118,0,0,0,1,90,95,0,0,11,0,2,4,1,0,2,0,11,0,118,0,0,1,1,0,0,11,0,117,0,0,0,1,9,18,118,
-0,18,118,0,18,117,0,49,20,0,8,18,118,0,0,0,1,90,95,0,0,12,0,2,1,1,0,2,0,12,0,118,0,0,1,1,0,0,12,0,
-117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,46,20,0,8,18,118,0,0,0,1,90,95,0,0,12,0,2,2,1,0,2,0,12,0,
-118,0,0,1,1,0,0,12,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,47,20,0,8,18,118,0,0,0,1,90,95,0,0,
-12,0,2,3,1,0,2,0,12,0,118,0,0,1,1,0,0,12,0,117,0,0,0,1,9,18,118,0,18,118,0,18,117,0,48,20,0,8,18,
-118,0,0,0,1,90,95,0,0,12,0,2,4,1,0,2,0,12,0,118,0,0,1,1,0,0,12,0,117,0,0,0,1,9,18,118,0,18,118,0,
-18,117,0,49,20,0,8,18,118,0,0,0,1,90,95,0,0,6,0,2,1,1,0,2,0,6,0,118,0,0,1,1,0,0,5,0,97,0,0,0,1,9,
-18,118,0,18,118,0,58,105,118,101,99,50,0,0,18,97,0,0,0,46,20,0,8,18,118,0,0,0,1,90,95,0,0,6,0,2,2,
-1,0,2,0,6,0,118,0,0,1,1,0,0,5,0,97,0,0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,50,0,0,18,97,0,0,
-0,47,20,0,8,18,118,0,0,0,1,90,95,0,0,6,0,2,3,1,0,2,0,6,0,118,0,0,1,1,0,0,5,0,97,0,0,0,1,9,18,118,0,
-18,118,0,58,105,118,101,99,50,0,0,18,97,0,0,0,48,20,0,8,18,118,0,0,0,1,90,95,0,0,6,0,2,4,1,0,2,0,6,
-0,118,0,0,1,1,0,0,5,0,97,0,0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,50,0,0,18,97,0,0,0,49,20,0,
-8,18,118,0,0,0,1,90,95,0,0,7,0,2,1,1,0,2,0,7,0,118,0,0,1,1,0,0,5,0,97,0,0,0,1,9,18,118,0,18,118,0,
-58,105,118,101,99,51,0,0,18,97,0,0,0,46,20,0,8,18,118,0,0,0,1,90,95,0,0,7,0,2,2,1,0,2,0,7,0,118,0,
-0,1,1,0,0,5,0,97,0,0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,51,0,0,18,97,0,0,0,47,20,0,8,18,118,
-0,0,0,1,90,95,0,0,7,0,2,3,1,0,2,0,7,0,118,0,0,1,1,0,0,5,0,97,0,0,0,1,9,18,118,0,18,118,0,58,105,
-118,101,99,51,0,0,18,97,0,0,0,48,20,0,8,18,118,0,0,0,1,90,95,0,0,8,0,2,4,1,0,2,0,7,0,118,0,0,1,1,0,
-0,5,0,97,0,0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,51,0,0,18,97,0,0,0,49,20,0,8,18,118,0,0,0,1,
-90,95,0,0,8,0,2,1,1,0,2,0,8,0,118,0,0,1,1,0,0,5,0,97,0,0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,
-52,0,0,18,97,0,0,0,46,20,0,8,18,118,0,0,0,1,90,95,0,0,8,0,2,2,1,0,2,0,8,0,118,0,0,1,1,0,0,5,0,97,0,
-0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,52,0,0,18,97,0,0,0,47,20,0,8,18,118,0,0,0,1,90,95,0,0,
-8,0,2,3,1,0,2,0,8,0,118,0,0,1,1,0,0,5,0,97,0,0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,52,0,0,18,
-97,0,0,0,48,20,0,8,18,118,0,0,0,1,90,95,0,0,8,0,2,4,1,0,2,0,8,0,118,0,0,1,1,0,0,5,0,97,0,0,0,1,9,
-18,118,0,18,118,0,58,105,118,101,99,52,0,0,18,97,0,0,0,49,20,0,8,18,118,0,0,0,1,90,95,0,0,10,0,2,1,
-1,0,2,0,10,0,118,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,50,0,0,18,97,0,0,0,
-46,20,0,8,18,118,0,0,0,1,90,95,0,0,10,0,2,2,1,0,2,0,10,0,118,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,118,0,
-18,118,0,58,118,101,99,50,0,0,18,97,0,0,0,47,20,0,8,18,118,0,0,0,1,90,95,0,0,10,0,2,3,1,0,2,0,10,0,
-118,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,50,0,0,18,97,0,0,0,48,20,0,8,18,
-118,0,0,0,1,90,95,0,0,10,0,2,4,1,0,2,0,10,0,118,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,118,0,18,118,0,58,
-118,101,99,50,0,0,18,97,0,0,0,49,20,0,8,18,118,0,0,0,1,90,95,0,0,11,0,2,1,1,0,2,0,11,0,118,0,0,1,1,
-0,0,9,0,97,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,51,0,0,18,97,0,0,0,46,20,0,8,18,118,0,0,0,1,
-90,95,0,0,11,0,2,2,1,0,2,0,11,0,118,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,
-51,0,0,18,97,0,0,0,47,20,0,8,18,118,0,0,0,1,90,95,0,0,11,0,2,3,1,0,2,0,11,0,118,0,0,1,1,0,0,9,0,97,
-0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,51,0,0,18,97,0,0,0,48,20,0,8,18,118,0,0,0,1,90,95,0,0,11,
-0,2,4,1,0,2,0,11,0,118,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,51,0,0,18,97,0,
-0,0,49,20,0,8,18,118,0,0,0,1,90,95,0,0,12,0,2,1,1,0,2,0,12,0,118,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,
-118,0,18,118,0,58,118,101,99,52,0,0,18,97,0,0,0,46,20,0,8,18,118,0,0,0,1,90,95,0,0,12,0,2,2,1,0,2,
-0,12,0,118,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,52,0,0,18,97,0,0,0,47,20,0,
-8,18,118,0,0,0,1,90,95,0,0,12,0,2,3,1,0,2,0,12,0,118,0,0,1,1,0,0,9,0,97,0,0,0,1,9,18,118,0,18,118,
-0,58,118,101,99,52,0,0,18,97,0,0,0,48,20,0,8,18,118,0,0,0,1,90,95,0,0,12,0,2,4,1,0,2,0,12,0,118,0,
-0,1,1,0,0,9,0,97,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,52,0,0,18,97,0,0,0,49,20,0,8,18,118,0,0,
-0,1,90,95,0,0,13,0,2,26,1,1,0,0,13,0,109,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,46,20,0,9,18,95,95,114,101,116,86,
-97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,46,20,0,0,1,90,95,0,0,13,0,2,
-27,1,1,0,0,13,0,109,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,
-18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,
-57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,47,20,0,0,1,90,95,0,0,13,0,2,21,1,1,0,0,13,0,109,
-0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,
-18,110,0,16,8,48,0,57,59,120,120,0,48,18,109,0,16,10,49,0,57,18,110,0,16,8,48,0,57,59,121,121,0,48,
-46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,8,48,0,57,18,110,0,16,10,49,0,
-57,59,120,120,0,48,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,59,121,121,0,48,46,20,0,0,1,90,95,
-0,0,13,0,2,22,1,1,0,0,13,0,109,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,
-8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,49,20,0,9,18,95,95,114,101,116,86,97,108,0,
-16,10,49,0,57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,49,20,0,0,1,90,95,0,0,14,0,2,26,1,1,0,
-0,14,0,109,0,0,1,1,0,0,14,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,
-16,8,48,0,57,18,110,0,16,8,48,0,57,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,
-0,16,10,49,0,57,18,110,0,16,10,49,0,57,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,
-109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,46,20,0,0,1,90,95,0,0,14,0,2,27,1,1,0,0,14,0,109,0,0,1,
-1,0,0,14,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,
-110,0,16,8,48,0,57,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,
-18,110,0,16,10,49,0,57,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,
-0,57,18,110,0,16,10,50,0,57,47,20,0,0,1,90,95,0,0,14,0,2,21,1,1,0,0,14,0,109,0,0,1,1,0,0,14,0,110,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,
-57,59,120,120,120,0,48,18,109,0,16,10,49,0,57,18,110,0,16,8,48,0,57,59,121,121,121,0,48,46,18,109,
-0,16,10,50,0,57,18,110,0,16,8,48,0,57,59,122,122,122,0,48,46,20,0,9,18,95,95,114,101,116,86,97,108,
-0,16,10,49,0,57,18,109,0,16,8,48,0,57,18,110,0,16,10,49,0,57,59,120,120,120,0,48,18,109,0,16,10,49,
-0,57,18,110,0,16,10,49,0,57,59,121,121,121,0,48,46,18,109,0,16,10,50,0,57,18,110,0,16,10,49,0,57,
-59,122,122,122,0,48,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,8,48,0,57,
-18,110,0,16,10,50,0,57,59,120,120,120,0,48,18,109,0,16,10,49,0,57,18,110,0,16,10,50,0,57,59,121,
-121,121,0,48,46,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,59,122,122,122,0,48,46,20,0,0,1,90,
-95,0,0,14,0,2,22,1,1,0,0,14,0,109,0,0,1,1,0,0,14,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
-16,8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,49,20,0,9,18,95,95,114,101,116,86,97,108,
-0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,49,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,49,20,0,0,1,90,95,0,0,15,0,2,26,
-1,1,0,0,15,0,109,0,0,1,1,0,0,15,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,
-18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,
-0,57,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,
-10,51,0,57,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,46,20,0,0,1,90,95,0,0,15,0,2,27,1,1,0,0,
-15,0,109,0,0,1,1,0,0,15,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,
-8,48,0,57,18,110,0,16,8,48,0,57,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,
-16,10,49,0,57,18,110,0,16,10,49,0,57,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,
-109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,
-57,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,47,20,0,0,1,90,95,0,0,15,0,2,21,1,1,0,0,15,0,109,
-0,0,1,1,0,0,15,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,
-18,110,0,16,8,48,0,57,59,120,120,120,120,0,48,18,109,0,16,10,49,0,57,18,110,0,16,8,48,0,57,59,121,
-121,121,121,0,48,46,18,109,0,16,10,50,0,57,18,110,0,16,8,48,0,57,59,122,122,122,122,0,48,46,18,109,
-0,16,10,51,0,57,18,110,0,16,8,48,0,57,59,119,119,119,119,0,48,46,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,49,0,57,18,109,0,16,8,48,0,57,18,110,0,16,10,49,0,57,59,120,120,120,120,0,48,18,109,0,
-16,10,49,0,57,18,110,0,16,10,49,0,57,59,121,121,121,121,0,48,46,18,109,0,16,10,50,0,57,18,110,0,16,
-10,49,0,57,59,122,122,122,122,0,48,46,18,109,0,16,10,51,0,57,18,110,0,16,10,49,0,57,59,119,119,119,
-119,0,48,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,8,48,0,57,18,110,0,
-16,10,50,0,57,59,120,120,120,120,0,48,18,109,0,16,10,49,0,57,18,110,0,16,10,50,0,57,59,121,121,121,
-121,0,48,46,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,59,122,122,122,122,0,48,46,18,109,0,16,
-10,51,0,57,18,110,0,16,10,50,0,57,59,119,119,119,119,0,48,46,20,0,9,18,95,95,114,101,116,86,97,108,
-0,16,10,51,0,57,18,109,0,16,8,48,0,57,18,110,0,16,10,51,0,57,59,120,120,120,120,0,48,18,109,0,16,
-10,49,0,57,18,110,0,16,10,51,0,57,59,121,121,121,121,0,48,46,18,109,0,16,10,50,0,57,18,110,0,16,10,
-51,0,57,59,122,122,122,122,0,48,46,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,59,119,119,119,
-119,0,48,46,20,0,0,1,90,95,0,0,15,0,2,22,1,1,0,0,15,0,109,0,0,1,1,0,0,15,0,110,0,0,0,1,9,18,95,95,
-114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,49,20,0,9,18,95,
-95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,49,20,0,9,
-18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,49,20,
-0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,16,10,51,0,57,18,110,0,16,10,51,0,57,
-49,20,0,0,1,90,95,0,0,13,0,2,26,1,1,0,0,9,0,97,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,95,95,114,101,116,
-86,97,108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,46,20,0,9,18,95,95,114,101,116,86,97,108,0,
-16,10,49,0,57,18,97,0,18,110,0,16,10,49,0,57,46,20,0,0,1,90,95,0,0,13,0,2,26,1,1,0,0,13,0,109,0,0,
-1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,
-0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,46,20,0,
-0,1,90,95,0,0,13,0,2,27,1,1,0,0,9,0,97,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,
-49,0,57,18,97,0,18,110,0,16,10,49,0,57,47,20,0,0,1,90,95,0,0,13,0,2,27,1,1,0,0,13,0,109,0,0,1,1,0,
-0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,47,
-20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,47,20,0,0,1,
-90,95,0,0,13,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
-16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,
-57,18,97,0,18,110,0,16,10,49,0,57,48,20,0,0,1,90,95,0,0,13,0,2,21,1,1,0,0,13,0,109,0,0,1,1,0,0,9,0,
-98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,48,20,0,9,
-18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,48,20,0,0,1,90,95,0,
-0,13,0,2,22,1,1,0,0,9,0,97,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,
-0,57,18,97,0,18,110,0,16,8,48,0,57,49,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,97,
-0,18,110,0,16,10,49,0,57,49,20,0,0,1,90,95,0,0,13,0,2,22,1,1,0,0,13,0,109,0,0,1,1,0,0,9,0,98,0,0,0,
-1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,49,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,49,20,0,0,1,90,95,0,0,14,0,2,
-26,1,1,0,0,9,0,97,0,0,1,1,0,0,14,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-97,0,18,110,0,16,8,48,0,57,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,
-0,16,10,49,0,57,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,97,0,18,110,0,16,10,50,
-0,57,46,20,0,0,1,90,95,0,0,14,0,2,26,1,1,0,0,14,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,46,20,0,9,18,95,95,114,101,116,86,
-97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,
-16,10,50,0,57,18,109,0,16,10,50,0,57,18,98,0,46,20,0,0,1,90,95,0,0,14,0,2,27,1,1,0,0,9,0,97,0,0,1,
-1,0,0,14,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,
-57,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,0,16,10,49,0,57,47,20,0,
-9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,97,0,18,110,0,16,10,50,0,57,47,20,0,0,1,90,95,
-0,0,14,0,2,27,1,1,0,0,14,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,
-48,0,57,18,109,0,16,8,48,0,57,18,98,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,
-109,0,16,10,49,0,57,18,98,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,
-10,50,0,57,18,98,0,47,20,0,0,1,90,95,0,0,14,0,2,21,1,1,0,0,9,0,97,0,0,1,1,0,0,14,0,110,0,0,0,1,9,
-18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,
-101,116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,
-97,108,0,16,10,50,0,57,18,97,0,18,110,0,16,10,50,0,57,48,20,0,0,1,90,95,0,0,14,0,2,21,1,1,0,0,14,0,
-109,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,
-57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,
-48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,98,0,48,20,0,0,
-1,90,95,0,0,14,0,2,22,1,1,0,0,9,0,97,0,0,1,1,0,0,14,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,
-0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,49,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,
-57,18,97,0,18,110,0,16,10,49,0,57,49,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,97,0,
-18,110,0,16,10,50,0,57,49,20,0,0,1,90,95,0,0,14,0,2,22,1,1,0,0,14,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,
-9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,49,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,49,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,98,0,49,20,0,0,1,90,95,0,0,15,0,2,26,1,1,0,
-0,9,0,97,0,0,1,1,0,0,15,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,97,0,18,
-110,0,16,8,48,0,57,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,0,16,10,
-49,0,57,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,97,0,18,110,0,16,10,50,0,57,46,
-20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,97,0,18,110,0,16,10,51,0,57,46,20,0,0,1,
-90,95,0,0,15,0,2,26,1,1,0,0,15,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
-16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,
-57,18,109,0,16,10,49,0,57,18,98,0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,
-0,16,10,50,0,57,18,98,0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,16,10,51,
-0,57,18,98,0,46,20,0,0,1,90,95,0,0,15,0,2,27,1,1,0,0,9,0,97,0,0,1,1,0,0,15,0,110,0,0,0,1,9,18,95,
-95,114,101,116,86,97,108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,47,20,0,9,18,95,95,114,101,
-116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,0,16,10,49,0,57,47,20,0,9,18,95,95,114,101,116,86,97,
-108,0,16,10,50,0,57,18,97,0,18,110,0,16,10,50,0,57,47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,
-10,51,0,57,18,97,0,18,110,0,16,10,51,0,57,47,20,0,0,1,90,95,0,0,15,0,2,27,1,1,0,0,15,0,109,0,0,1,1,
-0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,
-47,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,47,20,0,9,
-18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,98,0,47,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,16,10,51,0,57,18,98,0,47,20,0,0,1,90,95,0,0,15,0,2,
-21,1,1,0,0,9,0,97,0,0,1,1,0,0,15,0,110,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,
-97,0,18,110,0,16,8,48,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,
-0,16,10,49,0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,97,0,18,110,0,16,10,50,
-0,57,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,97,0,18,110,0,16,10,51,0,57,48,20,
-0,0,1,90,95,0,0,15,0,2,21,1,1,0,0,15,0,109,0,0,1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,
-49,0,57,18,109,0,16,10,49,0,57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,
-109,0,16,10,50,0,57,18,98,0,48,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,16,
-10,51,0,57,18,98,0,48,20,0,0,1,90,95,0,0,15,0,2,22,1,1,0,0,9,0,97,0,0,1,1,0,0,15,0,110,0,0,0,1,9,
-18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,97,0,18,110,0,16,8,48,0,57,49,20,0,9,18,95,95,114,
-101,116,86,97,108,0,16,10,49,0,57,18,97,0,18,110,0,16,10,49,0,57,49,20,0,9,18,95,95,114,101,116,86,
-97,108,0,16,10,50,0,57,18,97,0,18,110,0,16,10,50,0,57,49,20,0,9,18,95,95,114,101,116,86,97,108,0,
-16,10,51,0,57,18,97,0,18,110,0,16,10,51,0,57,49,20,0,0,1,90,95,0,0,15,0,2,22,1,1,0,0,15,0,109,0,0,
-1,1,0,0,9,0,98,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,98,
-0,49,20,0,9,18,95,95,114,101,116,86,97,108,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,98,0,49,20,0,
-9,18,95,95,114,101,116,86,97,108,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,98,0,49,20,0,9,18,95,95,
-114,101,116,86,97,108,0,16,10,51,0,57,18,109,0,16,10,51,0,57,18,98,0,49,20,0,0,1,90,95,0,0,10,0,2,
-21,1,1,0,0,13,0,109,0,0,1,1,0,0,10,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,16,8,
-48,0,57,18,118,0,59,120,120,0,48,18,109,0,16,10,49,0,57,18,118,0,59,121,121,0,48,46,20,0,0,1,90,95,
-0,0,10,0,2,21,1,1,0,0,10,0,118,0,0,1,1,0,0,13,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,
-120,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,8,48,0,57,0,0,20,0,9,18,95,95,114,101,116,86,97,
-108,0,59,121,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,49,0,57,0,0,20,0,0,1,90,95,0,0,11,0,2,
-21,1,1,0,0,14,0,109,0,0,1,1,0,0,11,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,16,8,
-48,0,57,18,118,0,59,120,120,120,0,48,18,109,0,16,10,49,0,57,18,118,0,59,121,121,121,0,48,46,18,109,
-0,16,10,50,0,57,18,118,0,59,122,122,122,0,48,46,20,0,0,1,90,95,0,0,11,0,2,21,1,1,0,0,11,0,118,0,0,
-1,1,0,0,14,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,59,120,0,58,100,111,116,0,0,18,118,0,0,
-18,109,0,16,8,48,0,57,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,121,0,58,100,111,116,0,0,18,
-118,0,0,18,109,0,16,10,49,0,57,0,0,20,0,9,18,95,95,114,101,116,86,97,108,0,59,122,0,58,100,111,116,
-0,0,18,118,0,0,18,109,0,16,10,50,0,57,0,0,20,0,0,1,90,95,0,0,12,0,2,21,1,1,0,0,15,0,109,0,0,1,1,0,
-0,12,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,16,8,48,0,57,18,118,0,59,120,120,
-120,120,0,48,18,109,0,16,10,49,0,57,18,118,0,59,121,121,121,121,0,48,46,18,109,0,16,10,50,0,57,18,
-118,0,59,122,122,122,122,0,48,46,18,109,0,16,10,51,0,57,18,118,0,59,119,119,119,119,0,48,46,20,0,0,
-1,90,95,0,0,12,0,2,21,1,1,0,0,12,0,118,0,0,1,1,0,0,15,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,59,120,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,8,48,0,57,0,0,20,0,9,18,95,95,114,101,116,
-86,97,108,0,59,121,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,49,0,57,0,0,20,0,9,18,95,95,114,
-101,116,86,97,108,0,59,122,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,50,0,57,0,0,20,0,9,18,95,
-95,114,101,116,86,97,108,0,59,119,0,58,100,111,116,0,0,18,118,0,0,18,109,0,16,10,51,0,57,0,0,20,0,
-0,1,90,95,0,0,13,0,2,1,1,0,2,0,13,0,109,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,
-109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,46,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,
-18,110,0,16,10,49,0,57,46,20,0,8,18,109,0,0,0,1,90,95,0,0,13,0,2,2,1,0,2,0,13,0,109,0,0,1,1,0,0,13,
-0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,47,20,0,9,18,109,
-0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,47,20,0,8,18,109,0,0,0,1,90,95,0,0,
-13,0,2,3,1,0,2,0,13,0,109,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,109,0,18,109,0,18,110,0,48,20,0,8,18,
-109,0,0,0,1,90,95,0,0,13,0,2,4,1,0,2,0,13,0,109,0,0,1,1,0,0,13,0,110,0,0,0,1,9,18,109,0,16,8,48,0,
-57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,49,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,
-0,57,18,110,0,16,10,49,0,57,49,20,0,8,18,109,0,0,0,1,90,95,0,0,14,0,2,1,1,0,2,0,14,0,109,0,0,1,1,0,
-0,14,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,46,20,0,9,
-18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,46,20,0,9,18,109,0,16,10,50,0,
-57,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,46,20,0,8,18,109,0,0,0,1,90,95,0,0,14,0,2,2,1,0,2,
-0,14,0,109,0,0,1,1,0,0,14,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,
-8,48,0,57,47,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,47,20,0,9,
-18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,47,20,0,8,18,109,0,0,0,1,90,
-95,0,0,14,0,2,3,1,0,2,0,14,0,109,0,0,1,1,0,0,14,0,110,0,0,0,1,9,18,109,0,18,109,0,18,110,0,48,20,0,
-8,18,109,0,0,0,1,90,95,0,0,14,0,2,4,1,0,2,0,14,0,109,0,0,1,1,0,0,14,0,110,0,0,0,1,9,18,109,0,16,8,
-48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,49,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,
-10,49,0,57,18,110,0,16,10,49,0,57,49,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,110,0,
-16,10,50,0,57,49,20,0,8,18,109,0,0,0,1,90,95,0,0,15,0,2,1,1,0,2,0,15,0,109,0,0,1,1,0,0,15,0,110,0,
-0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,46,20,0,9,18,109,0,16,10,
-49,0,57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,46,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,
-10,50,0,57,18,110,0,16,10,50,0,57,46,20,0,9,18,109,0,16,10,51,0,57,18,109,0,16,10,51,0,57,18,110,0,
-16,10,51,0,57,46,20,0,8,18,109,0,0,0,1,90,95,0,0,15,0,2,2,1,0,2,0,15,0,109,0,0,1,1,0,0,15,0,110,0,
-0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,47,20,0,9,18,109,0,16,10,
-49,0,57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,47,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,
-10,50,0,57,18,110,0,16,10,50,0,57,47,20,0,9,18,109,0,16,10,51,0,57,18,109,0,16,10,51,0,57,18,110,0,
-16,10,51,0,57,47,20,0,8,18,109,0,0,0,1,90,95,0,0,15,0,2,3,1,0,2,0,15,0,109,0,0,1,1,0,0,15,0,110,0,
-0,0,1,9,18,109,0,18,109,0,18,110,0,48,20,0,8,18,109,0,0,0,1,90,95,0,0,15,0,2,4,1,0,2,0,15,0,109,0,
-0,1,1,0,0,15,0,110,0,0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,110,0,16,8,48,0,57,49,
-20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,110,0,16,10,49,0,57,49,20,0,9,18,109,0,16,
-10,50,0,57,18,109,0,16,10,50,0,57,18,110,0,16,10,50,0,57,49,20,0,9,18,109,0,16,10,51,0,57,18,109,0,
-16,10,51,0,57,18,110,0,16,10,51,0,57,49,20,0,8,18,109,0,0,0,1,90,95,0,0,13,0,2,1,1,0,2,0,13,0,109,
-0,0,1,1,0,0,9,0,97,0,0,0,1,3,2,90,95,0,0,10,0,1,118,0,2,58,118,101,99,50,0,0,18,97,0,0,0,0,0,9,18,
-109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,118,0,46,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,
-49,0,57,18,118,0,46,20,0,8,18,109,0,0,0,1,90,95,0,0,13,0,2,2,1,0,2,0,13,0,109,0,0,1,1,0,0,9,0,97,0,
-0,0,1,3,2,90,95,0,0,10,0,1,118,0,2,58,118,101,99,50,0,0,18,97,0,0,0,0,0,9,18,109,0,16,8,48,0,57,18,
-109,0,16,8,48,0,57,18,118,0,47,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,118,0,47,20,
-0,8,18,109,0,0,0,1,90,95,0,0,13,0,2,3,1,0,2,0,13,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,3,2,90,95,0,0,10,
-0,1,118,0,2,58,118,101,99,50,0,0,18,97,0,0,0,0,0,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,
-118,0,48,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,118,0,48,20,0,8,18,109,0,0,0,1,90,
-95,0,0,13,0,2,4,1,0,2,0,13,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,3,2,90,95,0,0,10,0,1,118,0,2,58,118,
-101,99,50,0,0,17,49,0,48,0,0,18,97,0,49,0,0,0,0,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,
-118,0,48,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,118,0,48,20,0,8,18,109,0,0,0,1,90,
-95,0,0,14,0,2,1,1,0,2,0,14,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,3,2,90,95,0,0,11,0,1,118,0,2,58,118,
-101,99,51,0,0,18,97,0,0,0,0,0,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,118,0,46,20,0,9,18,
-109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,118,0,46,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,
-10,50,0,57,18,118,0,46,20,0,8,18,109,0,0,0,1,90,95,0,0,14,0,2,2,1,0,2,0,14,0,109,0,0,1,1,0,0,9,0,
-97,0,0,0,1,3,2,90,95,0,0,11,0,1,118,0,2,58,118,101,99,51,0,0,18,97,0,0,0,0,0,9,18,109,0,16,8,48,0,
-57,18,109,0,16,8,48,0,57,18,118,0,47,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,118,0,
-47,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,118,0,47,20,0,8,18,109,0,0,0,1,90,95,0,
-0,14,0,2,3,1,0,2,0,14,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,3,2,90,95,0,0,11,0,1,118,0,2,58,118,101,99,
-51,0,0,18,97,0,0,0,0,0,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,118,0,48,20,0,9,18,109,0,
-16,10,49,0,57,18,109,0,16,10,49,0,57,18,118,0,48,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,
-57,18,118,0,48,20,0,8,18,109,0,0,0,1,90,95,0,0,14,0,2,4,1,0,2,0,14,0,109,0,0,1,1,0,0,9,0,97,0,0,0,
-1,3,2,90,95,0,0,11,0,1,118,0,2,58,118,101,99,51,0,0,17,49,0,48,0,0,18,97,0,49,0,0,0,0,9,18,109,0,
-16,8,48,0,57,18,109,0,16,8,48,0,57,18,118,0,48,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,
-57,18,118,0,48,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,118,0,48,20,0,8,18,109,0,0,
-0,1,90,95,0,0,15,0,2,1,1,0,2,0,15,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,3,2,90,95,0,0,12,0,1,118,0,2,58,
-118,101,99,52,0,0,18,97,0,0,0,0,0,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,118,0,46,20,0,9,
-18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,118,0,46,20,0,9,18,109,0,16,10,50,0,57,18,109,0,
-16,10,50,0,57,18,118,0,46,20,0,9,18,109,0,16,10,51,0,57,18,109,0,16,10,51,0,57,18,118,0,46,20,0,8,
-18,109,0,0,0,1,90,95,0,0,15,0,2,2,1,0,2,0,15,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,3,2,90,95,0,0,12,0,1,
-118,0,2,58,118,101,99,52,0,0,18,97,0,0,0,0,0,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,18,118,
-0,47,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,118,0,47,20,0,9,18,109,0,16,10,50,0,
-57,18,109,0,16,10,50,0,57,18,118,0,47,20,0,9,18,109,0,16,10,51,0,57,18,109,0,16,10,51,0,57,18,118,
-0,47,20,0,8,18,109,0,0,0,1,90,95,0,0,15,0,2,3,1,0,2,0,15,0,109,0,0,1,1,0,0,9,0,97,0,0,0,1,3,2,90,
-95,0,0,12,0,1,118,0,2,58,118,101,99,52,0,0,18,97,0,0,0,0,0,9,18,109,0,16,8,48,0,57,18,109,0,16,8,
-48,0,57,18,118,0,48,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,18,118,0,48,20,0,9,18,109,
-0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,118,0,48,20,0,9,18,109,0,16,10,51,0,57,18,109,0,16,10,51,
-0,57,18,118,0,48,20,0,8,18,109,0,0,0,1,90,95,0,0,15,0,2,4,1,0,2,0,15,0,109,0,0,1,1,0,0,9,0,97,0,0,
-0,1,3,2,90,95,0,0,12,0,1,118,0,2,58,118,101,99,52,0,0,17,49,0,48,0,0,18,97,0,49,0,0,0,0,9,18,109,0,
-16,8,48,0,57,18,109,0,16,8,48,0,57,18,118,0,48,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,
-57,18,118,0,48,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,18,118,0,48,20,0,9,18,109,0,16,
-10,51,0,57,18,109,0,16,10,51,0,57,18,118,0,48,20,0,8,18,109,0,0,0,1,90,95,0,0,10,0,2,3,1,0,2,0,10,
-0,118,0,0,1,1,0,0,13,0,109,0,0,0,1,9,18,118,0,18,118,0,18,109,0,48,20,0,8,18,118,0,0,0,1,90,95,0,0,
-11,0,2,3,1,0,2,0,11,0,118,0,0,1,1,0,0,14,0,109,0,0,0,1,9,18,118,0,18,118,0,18,109,0,48,20,0,8,18,
-118,0,0,0,1,90,95,0,0,12,0,2,3,1,0,2,0,12,0,118,0,0,1,1,0,0,15,0,109,0,0,0,1,9,18,118,0,18,118,0,
-18,109,0,48,20,0,8,18,118,0,0,0,1,90,95,0,0,5,0,2,25,1,0,2,0,5,0,97,0,0,0,1,9,18,97,0,18,97,0,16,
-10,49,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,18,97,0,20,0,0,1,90,95,0,0,6,0,2,25,1,0,2,0,6,0,
-118,0,0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,50,0,0,16,10,49,0,0,0,47,20,0,9,18,95,95,114,101,
-116,86,97,108,0,18,118,0,20,0,0,1,90,95,0,0,7,0,2,25,1,0,2,0,7,0,118,0,0,0,1,9,18,118,0,18,118,0,
-58,105,118,101,99,51,0,0,16,10,49,0,0,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,0,
-1,90,95,0,0,8,0,2,25,1,0,2,0,8,0,118,0,0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,52,0,0,16,10,49,
-0,0,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,0,1,90,95,0,0,9,0,2,25,1,0,2,0,9,0,
-97,0,0,0,1,9,18,97,0,18,97,0,17,49,0,48,0,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,18,97,0,20,
-0,0,1,90,95,0,0,10,0,2,25,1,0,2,0,10,0,118,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,50,0,0,17,49,
-0,48,0,0,0,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,0,1,90,95,0,0,11,0,2,25,1,0,
-2,0,11,0,118,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,51,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,95,
-95,114,101,116,86,97,108,0,18,118,0,20,0,0,1,90,95,0,0,12,0,2,25,1,0,2,0,12,0,118,0,0,0,1,9,18,118,
-0,18,118,0,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,18,
-118,0,20,0,0,1,90,95,0,0,13,0,2,25,1,0,2,0,13,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,
-48,0,57,58,118,101,99,50,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,
-0,57,58,118,101,99,50,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,18,109,0,
-20,0,0,1,90,95,0,0,14,0,2,25,1,0,2,0,14,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,
-57,58,118,101,99,51,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,
-58,118,101,99,51,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,58,
-118,101,99,51,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,
-90,95,0,0,15,0,2,25,1,0,2,0,15,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,58,118,
-101,99,52,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,58,118,
-101,99,52,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,58,118,
-101,99,52,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,109,0,16,10,51,0,57,18,109,0,16,10,51,0,57,58,118,
-101,99,52,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,
-95,0,0,5,0,2,24,1,0,2,0,5,0,97,0,0,0,1,9,18,97,0,18,97,0,16,10,49,0,46,20,0,9,18,95,95,114,101,116,
-86,97,108,0,18,97,0,20,0,0,1,90,95,0,0,6,0,2,24,1,0,2,0,6,0,118,0,0,0,1,9,18,118,0,18,118,0,58,105,
-118,101,99,50,0,0,16,10,49,0,0,0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,0,1,90,
-95,0,0,7,0,2,24,1,0,2,0,7,0,118,0,0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,51,0,0,16,10,49,0,0,
-0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,0,1,90,95,0,0,8,0,2,24,1,0,2,0,8,0,118,
-0,0,0,1,9,18,118,0,18,118,0,58,105,118,101,99,52,0,0,16,10,49,0,0,0,46,20,0,9,18,95,95,114,101,116,
-86,97,108,0,18,118,0,20,0,0,1,90,95,0,0,9,0,2,24,1,0,2,0,9,0,97,0,0,0,1,9,18,97,0,18,97,0,17,49,0,
-48,0,0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,18,97,0,20,0,0,1,90,95,0,0,10,0,2,24,1,0,2,0,10,
-0,118,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,50,0,0,17,49,0,48,0,0,0,0,46,20,0,9,18,95,95,114,
-101,116,86,97,108,0,18,118,0,20,0,0,1,90,95,0,0,11,0,2,24,1,0,2,0,11,0,118,0,0,0,1,9,18,118,0,18,
-118,0,58,118,101,99,51,0,0,17,49,0,48,0,0,0,0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,18,118,0,
-20,0,0,1,90,95,0,0,12,0,2,24,1,0,2,0,12,0,118,0,0,0,1,9,18,118,0,18,118,0,58,118,101,99,52,0,0,17,
-49,0,48,0,0,0,0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,0,1,90,95,0,0,13,0,2,24,1,
-0,2,0,13,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,58,118,101,99,50,0,0,17,49,0,
-48,0,0,0,0,46,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,58,118,101,99,50,0,0,17,49,0,48,
-0,0,0,0,46,20,0,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,14,0,2,24,1,0,2,0,
-14,0,109,0,0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,58,118,101,99,51,0,0,17,49,0,48,0,0,
-0,0,46,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,58,118,101,99,51,0,0,17,49,0,48,0,0,0,
-0,46,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,58,118,101,99,51,0,0,17,49,0,48,0,0,0,0,
-46,20,0,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,15,0,2,24,1,0,2,0,15,0,109,
-0,0,0,1,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,46,
-20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,46,20,
-0,9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,46,20,0,
-9,18,109,0,16,10,51,0,57,18,109,0,16,10,51,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,46,20,0,9,
-18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,0,1,90,95,0,0,5,0,0,95,95,112,111,115,116,68,101,99,
-114,0,1,0,2,0,5,0,97,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,97,0,20,0,9,18,97,0,18,97,0,16,
-10,49,0,47,20,0,0,1,90,95,0,0,6,0,0,95,95,112,111,115,116,68,101,99,114,0,1,0,2,0,6,0,118,0,0,0,1,
-9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,9,18,118,0,18,118,0,58,105,118,101,99,50,0,0,16,
-10,49,0,0,0,47,20,0,0,1,90,95,0,0,7,0,0,95,95,112,111,115,116,68,101,99,114,0,1,0,2,0,7,0,118,0,0,
-0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,9,18,118,0,18,118,0,58,105,118,101,99,51,0,0,
-16,10,49,0,0,0,47,20,0,0,1,90,95,0,0,8,0,0,95,95,112,111,115,116,68,101,99,114,0,1,0,2,0,8,0,118,0,
-0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,9,18,118,0,18,118,0,58,105,118,101,99,52,0,
-0,16,10,49,0,0,0,47,20,0,0,1,90,95,0,0,9,0,0,95,95,112,111,115,116,68,101,99,114,0,1,0,2,0,9,0,97,
-0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,97,0,20,0,9,18,97,0,18,97,0,17,49,0,48,0,0,47,20,0,0,
-1,90,95,0,0,10,0,0,95,95,112,111,115,116,68,101,99,114,0,1,0,2,0,10,0,118,0,0,0,1,9,18,95,95,114,
-101,116,86,97,108,0,18,118,0,20,0,9,18,118,0,18,118,0,58,118,101,99,50,0,0,17,49,0,48,0,0,0,0,47,
-20,0,0,1,90,95,0,0,11,0,0,95,95,112,111,115,116,68,101,99,114,0,1,0,2,0,11,0,118,0,0,0,1,9,18,95,
-95,114,101,116,86,97,108,0,18,118,0,20,0,9,18,118,0,18,118,0,58,118,101,99,51,0,0,17,49,0,48,0,0,0,
-0,47,20,0,0,1,90,95,0,0,12,0,0,95,95,112,111,115,116,68,101,99,114,0,1,0,2,0,12,0,118,0,0,0,1,9,18,
-95,95,114,101,116,86,97,108,0,18,118,0,20,0,9,18,118,0,18,118,0,58,118,101,99,52,0,0,17,49,0,48,0,
-0,0,0,47,20,0,0,1,90,95,0,0,13,0,0,95,95,112,111,115,116,68,101,99,114,0,1,0,2,0,13,0,109,0,0,0,1,
-9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,58,
-118,101,99,50,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,58,
-118,101,99,50,0,0,17,49,0,48,0,0,0,0,47,20,0,0,1,90,95,0,0,14,0,0,95,95,112,111,115,116,68,101,99,
-114,0,1,0,2,0,14,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,0,20,0,9,18,109,0,16,8,48,
-0,57,18,109,0,16,8,48,0,57,58,118,101,99,51,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,109,0,16,10,49,0,
-57,18,109,0,16,10,49,0,57,58,118,101,99,51,0,0,17,49,0,48,0,0,0,0,47,20,0,9,18,109,0,16,10,50,0,57,
-18,109,0,16,10,50,0,57,58,118,101,99,51,0,0,17,49,0,48,0,0,0,0,47,20,0,0,1,90,95,0,0,15,0,0,95,95,
-112,111,115,116,68,101,99,114,0,1,0,2,0,15,0,109,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,109,
-0,20,0,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,47,20,
-0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,47,20,0,
-9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,47,20,0,9,
-18,109,0,16,10,51,0,57,18,109,0,16,10,51,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,47,20,0,0,1,
-90,95,0,0,9,0,0,95,95,112,111,115,116,73,110,99,114,0,1,0,2,0,9,0,97,0,0,0,1,9,18,95,95,114,101,
-116,86,97,108,0,18,97,0,20,0,9,18,97,0,18,97,0,16,10,49,0,46,20,0,0,1,90,95,0,0,10,0,0,95,95,112,
-111,115,116,73,110,99,114,0,1,0,2,0,10,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,
-20,0,9,18,118,0,18,118,0,58,118,101,99,50,0,0,17,49,0,48,0,0,0,0,46,20,0,0,1,90,95,0,0,11,0,0,95,
-95,112,111,115,116,73,110,99,114,0,1,0,2,0,11,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,
-118,0,20,0,9,18,118,0,18,118,0,58,118,101,99,51,0,0,17,49,0,48,0,0,0,0,46,20,0,0,1,90,95,0,0,12,0,
-0,95,95,112,111,115,116,73,110,99,114,0,1,0,2,0,12,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,
-0,18,118,0,20,0,9,18,118,0,18,118,0,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,46,20,0,0,1,90,95,0,0,
-5,0,0,95,95,112,111,115,116,73,110,99,114,0,1,0,2,0,5,0,97,0,0,0,1,9,18,95,95,114,101,116,86,97,
-108,0,18,97,0,20,0,9,18,97,0,18,97,0,16,10,49,0,46,20,0,0,1,90,95,0,0,6,0,0,95,95,112,111,115,116,
-73,110,99,114,0,1,0,2,0,6,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,9,18,118,
-0,18,118,0,58,105,118,101,99,50,0,0,16,10,49,0,0,0,46,20,0,0,1,90,95,0,0,7,0,0,95,95,112,111,115,
-116,73,110,99,114,0,1,0,2,0,7,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,9,18,
-118,0,18,118,0,58,105,118,101,99,51,0,0,16,10,49,0,0,0,46,20,0,0,1,90,95,0,0,8,0,0,95,95,112,111,
-115,116,73,110,99,114,0,1,0,2,0,8,0,118,0,0,0,1,9,18,95,95,114,101,116,86,97,108,0,18,118,0,20,0,9,
-18,118,0,18,118,0,58,105,118,101,99,51,0,0,16,10,49,0,0,0,46,20,0,0,1,90,95,0,0,13,0,0,95,95,112,
-111,115,116,73,110,99,114,0,1,0,2,0,13,0,109,0,0,0,1,3,2,90,95,0,0,13,0,1,110,0,2,18,109,0,0,0,9,
-18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,58,118,101,99,50,0,0,17,49,0,48,0,0,0,0,46,20,0,9,18,
-109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,58,118,101,99,50,0,0,17,49,0,48,0,0,0,0,46,20,0,8,18,
-110,0,0,0,1,90,95,0,0,14,0,0,95,95,112,111,115,116,73,110,99,114,0,1,0,2,0,14,0,109,0,0,0,1,3,2,90,
-95,0,0,14,0,1,110,0,2,18,109,0,0,0,9,18,109,0,16,8,48,0,57,18,109,0,16,8,48,0,57,58,118,101,99,51,
-0,0,17,49,0,48,0,0,0,0,46,20,0,9,18,109,0,16,10,49,0,57,18,109,0,16,10,49,0,57,58,118,101,99,51,0,
-0,17,49,0,48,0,0,0,0,46,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,10,50,0,57,58,118,101,99,51,0,0,
-17,49,0,48,0,0,0,0,46,20,0,8,18,110,0,0,0,1,90,95,0,0,15,0,0,95,95,112,111,115,116,73,110,99,114,0,
-1,0,2,0,15,0,109,0,0,0,1,3,2,90,95,0,0,15,0,1,110,0,2,18,109,0,0,0,9,18,109,0,16,8,48,0,57,18,109,
-0,16,8,48,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,46,20,0,9,18,109,0,16,10,49,0,57,18,109,0,
-16,10,49,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,46,20,0,9,18,109,0,16,10,50,0,57,18,109,0,16,
-10,50,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,46,20,0,9,18,109,0,16,10,51,0,57,18,109,0,16,10,
-51,0,57,58,118,101,99,52,0,0,17,49,0,48,0,0,0,0,46,20,0,8,18,110,0,0,0,1,90,95,0,0,1,0,2,15,1,1,0,
-0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,4,118,101,99,52,95,115,103,116,0,18,95,95,114,101,116,86,97,
-108,0,59,120,0,0,18,98,0,0,18,97,0,0,0,0,1,90,95,0,0,1,0,2,15,1,1,0,0,5,0,97,0,0,1,1,0,0,5,0,98,0,
-0,0,1,8,58,102,108,111,97,116,0,0,18,97,0,0,0,58,102,108,111,97,116,0,0,18,98,0,0,0,40,0,0,1,90,95,
-0,0,1,0,2,16,1,1,0,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,0,0,1,0,1,99,0,0,0,4,102,108,111,
-97,116,95,108,101,115,115,0,18,99,0,0,18,98,0,0,18,97,0,0,0,8,18,99,0,0,0,1,90,95,0,0,1,0,2,16,1,1,
-0,0,5,0,97,0,0,1,1,0,0,5,0,98,0,0,0,1,8,58,102,108,111,97,116,0,0,18,97,0,0,0,58,102,108,111,97,
-116,0,0,18,98,0,0,0,41,0,0,1,90,95,0,0,1,0,2,18,1,1,0,0,9,0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,
-95,0,0,1,0,1,103,0,0,1,1,101,0,0,0,4,102,108,111,97,116,95,108,101,115,115,0,18,103,0,0,18,98,0,0,
-18,97,0,0,0,4,102,108,111,97,116,95,101,113,117,97,108,0,18,101,0,0,18,97,0,0,18,98,0,0,0,8,18,103,
-0,18,101,0,32,0,0,1,90,95,0,0,1,0,2,18,1,1,0,0,5,0,97,0,0,1,1,0,0,5,0,98,0,0,0,1,8,58,102,108,111,
-97,116,0,0,18,97,0,0,0,58,102,108,111,97,116,0,0,18,98,0,0,0,43,0,0,1,90,95,0,0,1,0,2,17,1,1,0,0,9,
-0,97,0,0,1,1,0,0,9,0,98,0,0,0,1,3,2,90,95,0,0,1,0,1,103,0,0,1,1,101,0,0,0,4,102,108,111,97,116,95,
-108,101,115,115,0,18,103,0,0,18,97,0,0,18,98,0,0,0,4,102,108,111,97,116,95,101,113,117,97,108,0,18,
-101,0,0,18,97,0,0,18,98,0,0,0,8,18,103,0,18,101,0,32,0,0,1,90,95,0,0,1,0,2,17,1,1,0,0,5,0,97,0,0,1,
-1,0,0,5,0,98,0,0,0,1,8,58,102,108,111,97,116,0,0,18,97,0,0,0,58,102,108,111,97,116,0,0,18,98,0,0,0,
-42,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,9,0,102,0,0,0,1,4,102,108,111,
-97,116,95,112,114,105,110,116,0,18,102,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,
-1,1,0,0,5,0,105,0,0,0,1,4,105,110,116,95,112,114,105,110,116,0,18,105,0,0,0,0,1,90,95,0,0,0,0,0,
-112,114,105,110,116,77,69,83,65,0,1,1,0,0,1,0,98,0,0,0,1,4,98,111,111,108,95,112,114,105,110,116,0,
-18,98,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,10,0,118,0,0,0,1,9,58,
-112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,120,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,
-0,18,118,0,59,121,0,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,11,0,118,0,
-0,0,1,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,120,0,0,0,0,9,58,112,114,105,110,116,77,
-69,83,65,0,0,18,118,0,59,121,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,122,0,0,
-0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,12,0,118,0,0,0,1,9,58,112,114,
-105,110,116,77,69,83,65,0,0,18,118,0,59,120,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,
-118,0,59,121,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,122,0,0,0,0,9,58,112,114,
-105,110,116,77,69,83,65,0,0,18,118,0,59,119,0,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,
-83,65,0,1,1,0,0,6,0,118,0,0,0,1,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,120,0,0,0,0,9,
-58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,121,0,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,
-116,77,69,83,65,0,1,1,0,0,7,0,118,0,0,0,1,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,120,
-0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,121,0,0,0,0,9,58,112,114,105,110,116,
-77,69,83,65,0,0,18,118,0,59,122,0,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,
-0,0,8,0,118,0,0,0,1,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,120,0,0,0,0,9,58,112,114,
-105,110,116,77,69,83,65,0,0,18,118,0,59,121,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,
-118,0,59,122,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,119,0,0,0,0,0,1,90,95,0,
-0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,2,0,118,0,0,0,1,9,58,112,114,105,110,116,77,69,
-83,65,0,0,18,118,0,59,120,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,121,0,0,0,0,
-0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,3,0,118,0,0,0,1,9,58,112,114,105,110,
-116,77,69,83,65,0,0,18,118,0,59,120,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,
-121,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,122,0,0,0,0,0,1,90,95,0,0,0,0,0,
-112,114,105,110,116,77,69,83,65,0,1,1,0,0,4,0,118,0,0,0,1,9,58,112,114,105,110,116,77,69,83,65,0,0,
-18,118,0,59,120,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,118,0,59,121,0,0,0,0,9,58,112,
-114,105,110,116,77,69,83,65,0,0,18,118,0,59,122,0,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,
-18,118,0,59,119,0,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,13,0,109,0,0,
-0,1,9,58,112,114,105,110,116,77,69,83,65,0,0,18,109,0,16,8,48,0,57,0,0,0,9,58,112,114,105,110,116,
-77,69,83,65,0,0,18,109,0,16,10,49,0,57,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,
-1,1,0,0,14,0,109,0,0,0,1,9,58,112,114,105,110,116,77,69,83,65,0,0,18,109,0,16,8,48,0,57,0,0,0,9,58,
-112,114,105,110,116,77,69,83,65,0,0,18,109,0,16,10,49,0,57,0,0,0,9,58,112,114,105,110,116,77,69,83,
-65,0,0,18,109,0,16,10,50,0,57,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,
-15,0,109,0,0,0,1,9,58,112,114,105,110,116,77,69,83,65,0,0,18,109,0,16,8,48,0,57,0,0,0,9,58,112,114,
-105,110,116,77,69,83,65,0,0,18,109,0,16,10,49,0,57,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,
-18,109,0,16,10,50,0,57,0,0,0,9,58,112,114,105,110,116,77,69,83,65,0,0,18,109,0,16,10,51,0,57,0,0,0,
-0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,16,0,101,0,0,0,1,4,105,110,116,95,
-112,114,105,110,116,0,18,101,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,
-17,0,101,0,0,0,1,4,105,110,116,95,112,114,105,110,116,0,18,101,0,0,0,0,1,90,95,0,0,0,0,0,112,114,
-105,110,116,77,69,83,65,0,1,1,0,0,18,0,101,0,0,0,1,4,105,110,116,95,112,114,105,110,116,0,18,101,0,
-0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,0,19,0,101,0,0,0,1,4,105,110,116,
-95,112,114,105,110,116,0,18,101,0,0,0,0,1,90,95,0,0,0,0,0,112,114,105,110,116,77,69,83,65,0,1,1,0,
-0,20,0,101,0,0,0,1,4,105,110,116,95,112,114,105,110,116,0,18,101,0,0,0,0,1,90,95,0,0,0,0,0,112,114,
-105,110,116,77,69,83,65,0,1,1,0,0,21,0,101,0,0,0,1,4,105,110,116,95,112,114,105,110,116,0,18,101,0,
-0,0,0,0
diff --git a/src/mesa/shader/slang/library/slang_fragment_builtin_gc.h b/src/mesa/shader/slang/library/slang_fragment_builtin_gc.h
deleted file mode 100644 (file)
index c5a1cce..0000000
+++ /dev/null
@@ -1,110 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE FOLLOWING FILE: */
-/* slang_fragment_builtin.gc */
-
-5,2,2,90,95,6,0,12,0,1,103,108,95,70,114,97,103,67,111,111,114,100,0,0,0,2,2,90,95,6,0,1,0,1,103,
-108,95,70,114,111,110,116,70,97,99,105,110,103,0,0,0,2,2,90,95,5,0,12,0,1,103,108,95,70,114,97,103,
-67,111,108,111,114,0,0,0,2,2,90,95,5,0,12,0,1,103,108,95,70,114,97,103,68,97,116,97,0,3,18,103,108,
-95,77,97,120,68,114,97,119,66,117,102,102,101,114,115,0,0,0,2,2,90,95,5,0,9,0,1,103,108,95,70,114,
-97,103,68,101,112,116,104,0,0,0,2,2,90,95,3,0,12,0,1,103,108,95,67,111,108,111,114,0,0,0,2,2,90,95,
-3,0,12,0,1,103,108,95,83,101,99,111,110,100,97,114,121,67,111,108,111,114,0,0,0,2,2,90,95,3,0,12,0,
-1,103,108,95,84,101,120,67,111,111,114,100,0,3,18,103,108,95,77,97,120,84,101,120,116,117,114,101,
-67,111,111,114,100,115,0,0,0,2,2,90,95,3,0,9,0,1,103,108,95,70,111,103,70,114,97,103,67,111,111,
-114,100,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,49,68,0,1,1,0,0,16,0,115,97,109,112,
-108,101,114,0,0,1,1,0,0,9,0,99,111,111,114,100,0,0,1,1,0,0,9,0,98,105,97,115,0,0,0,1,3,2,90,95,0,0,
-12,0,1,99,111,111,114,100,52,0,0,0,9,18,99,111,111,114,100,52,0,59,120,0,18,99,111,111,114,100,0,
-20,0,9,18,99,111,111,114,100,52,0,59,119,0,18,98,105,97,115,0,20,0,4,118,101,99,52,95,116,101,120,
-95,49,100,95,98,105,97,115,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,
-18,99,111,111,114,100,52,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,49,68,80,114,111,
-106,0,1,1,0,0,16,0,115,97,109,112,108,101,114,0,0,1,1,0,0,10,0,99,111,111,114,100,0,0,1,1,0,0,9,0,
-98,105,97,115,0,0,0,1,3,2,90,95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,99,111,111,114,
-100,0,59,120,0,18,99,111,111,114,100,0,59,120,0,18,99,111,111,114,100,0,59,121,0,49,20,0,9,18,112,
-99,111,111,114,100,0,59,119,0,18,98,105,97,115,0,20,0,4,118,101,99,52,95,116,101,120,95,49,100,95,
-98,105,97,115,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,112,99,111,
-111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,49,68,80,114,111,106,0,1,1,0,0,
-16,0,115,97,109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,1,1,0,0,9,0,98,105,97,115,
-0,0,0,1,3,2,90,95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,99,111,111,114,100,0,59,120,0,
-18,99,111,111,114,100,0,59,120,0,18,99,111,111,114,100,0,59,122,0,49,20,0,9,18,112,99,111,111,114,
-100,0,59,119,0,18,98,105,97,115,0,20,0,4,118,101,99,52,95,116,101,120,95,49,100,95,98,105,97,115,0,
-18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,112,99,111,111,114,100,0,0,
-0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,50,68,0,1,1,0,0,17,0,115,97,109,112,108,101,114,
-0,0,1,1,0,0,10,0,99,111,111,114,100,0,0,1,1,0,0,9,0,98,105,97,115,0,0,0,1,3,2,90,95,0,0,12,0,1,99,
-111,111,114,100,52,0,0,0,9,18,99,111,111,114,100,52,0,59,120,121,0,18,99,111,111,114,100,0,59,120,
-121,0,20,0,9,18,99,111,111,114,100,52,0,59,119,0,18,98,105,97,115,0,20,0,4,118,101,99,52,95,116,
-101,120,95,50,100,95,98,105,97,115,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,
-114,0,0,18,99,111,111,114,100,52,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,50,68,80,
-114,111,106,0,1,1,0,0,17,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,0,0,1,1,
-0,0,9,0,98,105,97,115,0,0,0,1,3,2,90,95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,99,111,
-111,114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,122,0,
-49,20,0,9,18,112,99,111,111,114,100,0,59,119,0,18,98,105,97,115,0,20,0,4,118,101,99,52,95,116,101,
-120,95,50,100,95,98,105,97,115,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,
-0,0,18,112,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,50,68,80,114,
-111,106,0,1,1,0,0,17,0,115,97,109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,1,1,0,0,
-9,0,98,105,97,115,0,0,0,1,3,2,90,95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,99,111,111,
-114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,119,0,49,20,
-0,9,18,112,99,111,111,114,100,0,59,119,0,18,98,105,97,115,0,20,0,4,118,101,99,52,95,116,101,120,95,
-50,100,95,98,105,97,115,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,
-112,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,51,68,0,1,1,0,0,18,0,
-115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,0,0,1,1,0,0,9,0,98,105,97,115,0,0,0,
-1,3,2,90,95,0,0,12,0,1,99,111,111,114,100,52,0,0,0,9,18,99,111,111,114,100,52,0,59,120,121,122,0,
-18,99,111,111,114,100,0,59,120,121,122,0,20,0,9,18,99,111,111,114,100,52,0,59,119,0,18,98,105,97,
-115,0,20,0,4,118,101,99,52,95,116,101,120,95,51,100,95,98,105,97,115,0,18,95,95,114,101,116,86,97,
-108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,52,0,0,0,0,1,90,95,0,0,12,0,0,116,
-101,120,116,117,114,101,51,68,80,114,111,106,0,1,1,0,0,18,0,115,97,109,112,108,101,114,0,0,1,1,0,0,
-12,0,99,111,111,114,100,0,0,1,1,0,0,9,0,98,105,97,115,0,0,0,1,3,2,90,95,0,0,12,0,1,112,99,111,111,
-114,100,0,0,0,9,18,112,99,111,111,114,100,0,59,120,121,122,0,18,99,111,111,114,100,0,59,120,121,
-122,0,18,99,111,111,114,100,0,59,119,0,49,20,0,9,18,112,99,111,111,114,100,0,59,119,0,18,98,105,97,
-115,0,20,0,4,118,101,99,52,95,116,101,120,95,51,100,95,98,105,97,115,0,18,95,95,114,101,116,86,97,
-108,0,0,18,115,97,109,112,108,101,114,0,0,18,112,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,
-101,120,116,117,114,101,67,117,98,101,0,1,1,0,0,19,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,
-99,111,111,114,100,0,0,1,1,0,0,9,0,98,105,97,115,0,0,0,1,3,2,90,95,0,0,12,0,1,99,111,111,114,100,
-52,0,0,0,9,18,99,111,111,114,100,52,0,59,120,121,122,0,18,99,111,111,114,100,0,20,0,9,18,99,111,
-111,114,100,52,0,59,119,0,18,98,105,97,115,0,20,0,4,118,101,99,52,95,116,101,120,95,99,117,98,101,
-0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,52,0,
-0,0,0,1,90,95,0,0,12,0,0,115,104,97,100,111,119,49,68,0,1,1,0,0,20,0,115,97,109,112,108,101,114,0,
-0,1,1,0,0,11,0,99,111,111,114,100,0,0,1,1,0,0,9,0,98,105,97,115,0,0,0,1,3,2,90,95,0,0,12,0,1,99,
-111,111,114,100,52,0,0,0,9,18,99,111,111,114,100,52,0,59,120,121,122,0,18,99,111,111,114,100,0,20,
-0,9,18,99,111,111,114,100,52,0,59,119,0,18,98,105,97,115,0,20,0,4,118,101,99,52,95,116,101,120,95,
-49,100,95,98,105,97,115,95,115,104,97,100,111,119,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,
-109,112,108,101,114,0,0,18,99,111,111,114,100,52,0,0,0,0,1,90,95,0,0,12,0,0,115,104,97,100,111,119,
-49,68,80,114,111,106,0,1,1,0,0,20,0,115,97,109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,
-0,0,1,1,0,0,9,0,98,105,97,115,0,0,0,1,3,2,90,95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,
-99,111,111,114,100,0,59,120,0,18,99,111,111,114,100,0,59,120,0,18,99,111,111,114,100,0,59,119,0,49,
-20,0,9,18,112,99,111,111,114,100,0,59,122,0,18,99,111,111,114,100,0,59,122,0,20,0,9,18,112,99,111,
-111,114,100,0,59,119,0,18,98,105,97,115,0,20,0,4,118,101,99,52,95,116,101,120,95,49,100,95,98,105,
-97,115,95,115,104,97,100,111,119,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,
-114,0,0,18,112,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,115,104,97,100,111,119,50,68,0,1,1,0,
-0,21,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,0,0,1,1,0,0,9,0,98,105,97,
-115,0,0,0,1,3,2,90,95,0,0,12,0,1,99,111,111,114,100,52,0,0,0,9,18,99,111,111,114,100,52,0,59,120,
-121,122,0,18,99,111,111,114,100,0,20,0,9,18,99,111,111,114,100,52,0,59,119,0,18,98,105,97,115,0,20,
-0,4,118,101,99,52,95,116,101,120,95,50,100,95,98,105,97,115,95,115,104,97,100,111,119,0,18,95,95,
-114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,52,0,0,0,0,1,90,
-95,0,0,12,0,0,115,104,97,100,111,119,50,68,80,114,111,106,0,1,1,0,0,21,0,115,97,109,112,108,101,
-114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,1,1,0,0,9,0,98,105,97,115,0,0,0,1,3,2,90,95,0,0,12,0,1,
-112,99,111,111,114,100,0,0,0,9,18,112,99,111,111,114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,
-120,121,0,18,99,111,111,114,100,0,59,119,0,49,20,0,9,18,112,99,111,111,114,100,0,59,122,0,18,99,
-111,111,114,100,0,59,122,0,20,0,9,18,112,99,111,111,114,100,0,59,119,0,18,98,105,97,115,0,20,0,4,
-118,101,99,52,95,116,101,120,95,50,100,95,98,105,97,115,95,115,104,97,100,111,119,0,18,95,95,114,
-101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,112,99,111,111,114,100,0,0,0,0,1,90,95,
-0,0,9,0,0,100,70,100,120,0,1,1,0,0,9,0,112,0,0,0,1,4,118,101,99,52,95,100,100,120,0,18,95,95,114,
-101,116,86,97,108,0,59,120,0,0,18,112,0,59,120,120,120,120,0,0,0,0,1,90,95,0,0,10,0,0,100,70,100,
-120,0,1,1,0,0,10,0,112,0,0,0,1,4,118,101,99,52,95,100,100,120,0,18,95,95,114,101,116,86,97,108,0,
-59,120,121,0,0,18,112,0,59,120,121,121,121,0,0,0,0,1,90,95,0,0,11,0,0,100,70,100,120,0,1,1,0,0,11,
-0,112,0,0,0,1,4,118,101,99,52,95,100,100,120,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,
-18,112,0,59,120,121,122,122,0,0,0,0,1,90,95,0,0,12,0,0,100,70,100,120,0,1,1,0,0,12,0,112,0,0,0,1,4,
-118,101,99,52,95,100,100,120,0,18,95,95,114,101,116,86,97,108,0,0,18,112,0,0,0,0,1,90,95,0,0,9,0,0,
-100,70,100,121,0,1,1,0,0,9,0,112,0,0,0,1,4,118,101,99,52,95,100,100,121,0,18,95,95,114,101,116,86,
-97,108,0,59,120,0,0,18,112,0,59,120,120,120,120,0,0,0,0,1,90,95,0,0,10,0,0,100,70,100,121,0,1,1,0,
-0,10,0,112,0,0,0,1,4,118,101,99,52,95,100,100,121,0,18,95,95,114,101,116,86,97,108,0,59,120,121,0,
-0,18,112,0,59,120,121,121,121,0,0,0,0,1,90,95,0,0,11,0,0,100,70,100,121,0,1,1,0,0,11,0,112,0,0,0,1,
-4,118,101,99,52,95,100,100,121,0,18,95,95,114,101,116,86,97,108,0,59,120,121,122,0,0,18,112,0,59,
-120,121,122,122,0,0,0,0,1,90,95,0,0,12,0,0,100,70,100,121,0,1,1,0,0,12,0,112,0,0,0,1,4,118,101,99,
-52,95,100,100,121,0,18,95,95,114,101,116,86,97,108,0,0,18,112,0,0,0,0,1,90,95,0,0,9,0,0,102,119,
-105,100,116,104,0,1,1,0,0,9,0,112,0,0,0,1,8,58,97,98,115,0,0,58,100,70,100,120,0,0,18,112,0,0,0,0,
-0,58,97,98,115,0,0,58,100,70,100,121,0,0,18,112,0,0,0,0,0,46,0,0,1,90,95,0,0,10,0,0,102,119,105,
-100,116,104,0,1,1,0,0,10,0,112,0,0,0,1,8,58,97,98,115,0,0,58,100,70,100,120,0,0,18,112,0,0,0,0,0,
-58,97,98,115,0,0,58,100,70,100,121,0,0,18,112,0,0,0,0,0,46,0,0,1,90,95,0,0,11,0,0,102,119,105,100,
-116,104,0,1,1,0,0,11,0,112,0,0,0,1,8,58,97,98,115,0,0,58,100,70,100,120,0,0,18,112,0,0,0,0,0,58,97,
-98,115,0,0,58,100,70,100,121,0,0,18,112,0,0,0,0,0,46,0,0,1,90,95,0,0,12,0,0,102,119,105,100,116,
-104,0,1,1,0,0,12,0,112,0,0,0,1,8,58,97,98,115,0,0,58,100,70,100,120,0,0,18,112,0,0,0,0,0,58,97,98,
-115,0,0,58,100,70,100,121,0,0,18,112,0,0,0,0,0,46,0,0,0
diff --git a/src/mesa/shader/slang/library/slang_pp_directives.syn b/src/mesa/shader/slang/library/slang_pp_directives.syn
deleted file mode 100644 (file)
index b51d168..0000000
+++ /dev/null
@@ -1,405 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  6.6
- *
- * Copyright (C) 2006  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/**
- * \file slang_pp_directives.syn
- * slang preprocessor directives parser
- * \author Michal Krol
- */
-
-.syntax source;
-
-/*
- * This syntax script preprocesses a GLSL shader.
- * It is assumed, that the #version directive has been parsed. Separate pass for parsing
- * version gives better control on behavior depending on the version number given.
- *
- * The output is a source string with comments and directives removed. White spaces and comments
- * are replaced with on or more spaces. All new-lines are preserved and converted to Linux format.
- * Directives are escaped with a null character. The end of the source string is marked by
- * two consecutive null characters. The consumer is responsible for executing the escaped
- * directives, removing dead portions of code and expanding macros.
- */
-
-.emtcode ESCAPE_TOKEN 0
-
-/*
- * The TOKEN_* symbols follow the ESCAPE_TOKEN.
- *
- * NOTE:
- * There is no TOKEN_IFDEF and neither is TOKEN_IFNDEF. They are handled with TOKEN_IF and
- * operator defined.
- * The "#ifdef SYMBOL" is replaced with "#if defined SYMBOL"
- * The "#ifndef SYMBOL" is replaced with "#if !defined SYMBOL"
- */
-.emtcode TOKEN_END       0
-.emtcode TOKEN_DEFINE    1
-.emtcode TOKEN_UNDEF     2
-.emtcode TOKEN_IF        3
-.emtcode TOKEN_ELSE      4
-.emtcode TOKEN_ELIF      5
-.emtcode TOKEN_ENDIF     6
-.emtcode TOKEN_ERROR     7
-.emtcode TOKEN_PRAGMA    8
-.emtcode TOKEN_EXTENSION 9
-.emtcode TOKEN_LINE      10
-
-/*
- * The PARAM_* symbols follow the TOKEN_DEFINE.
- */
-.emtcode PARAM_END       0
-.emtcode PARAM_PARAMETER 1
-
-/*
- * The BEHAVIOR_* symbols follow the TOKEN_EXTENSION.
- */
-.emtcode BEHAVIOR_REQUIRE 1
-.emtcode BEHAVIOR_ENABLE  2
-.emtcode BEHAVIOR_WARN    3
-.emtcode BEHAVIOR_DISABLE 4
-
-/*
- * The PRAGMA_* symbols follow TOKEN_PRAGMA
- */
-.emtcode PRAGMA_NO_PARAM 0
-.emtcode PRAGMA_PARAM    1
-
-source
-   optional_directive .and .loop source_element .and '\0' .emit ESCAPE_TOKEN .emit TOKEN_END;
-
-source_element
-   c_style_comment_block .or cpp_style_comment_block .or new_line_directive .or source_token;
-
-c_style_comment_block
-   '/' .and '*' .and c_style_comment_rest .and .true .emit ' ';
-
-c_style_comment_rest
-   .loop c_style_comment_body .and c_style_comment_end;
-
-c_style_comment_body
-   c_style_comment_char_nostar .or c_style_comment_char_star_noslashstar;
-
-c_style_comment_char_nostar
-   new_line .or '\x2B'-'\xFF' .or '\x01'-'\x29';
-
-c_style_comment_char_star_noslashstar
-   '*' .and c_style_comment_char_star_noslashstar_1;
-c_style_comment_char_star_noslashstar_1
-   c_style_comment_char_noslashstar .or c_style_comment_char_star_noslashstar;
-
-c_style_comment_char_noslashstar
-   new_line .or '\x30'-'\xFF' .or '\x01'-'\x29' .or '\x2B'-'\x2E';
-
-c_style_comment_end
-   '*' .and .loop c_style_comment_char_star .and '/';
-
-c_style_comment_char_star
-   '*';
-
-cpp_style_comment_block
-   '/' .and '/' .and cpp_style_comment_block_1;
-cpp_style_comment_block_1
-   cpp_style_comment_block_2 .or cpp_style_comment_block_3;
-cpp_style_comment_block_2
-   .loop cpp_style_comment_char .and new_line_directive;
-cpp_style_comment_block_3
-   .loop cpp_style_comment_char;
-
-cpp_style_comment_char
-   '\x0E'-'\xFF' .or '\x01'-'\x09' .or '\x0B'-'\x0C';
-
-new_line_directive
-   new_line .and optional_directive;
-
-new_line
-   generic_new_line .emit '\n';
-
-generic_new_line
-   carriage_return_line_feed .or line_feed_carriage_return .or '\n' .or '\r';
-
-carriage_return_line_feed
-   '\r' .and '\n';
-
-line_feed_carriage_return
-   '\n' .and '\r';
-
-optional_directive
-   directive .emit ESCAPE_TOKEN .or .true;
-
-directive
-   dir_define .emit TOKEN_DEFINE .or
-   dir_undef .emit TOKEN_UNDEF .or
-   dir_if .emit TOKEN_IF .or
-   dir_ifdef .emit TOKEN_IF .emit 'd' .emit 'e' .emit 'f' .emit 'i' .emit 'n' .emit 'e' .emit 'd'
-             .emit ' ' .or
-   dir_ifndef .emit TOKEN_IF .emit '!' .emit 'd' .emit 'e' .emit 'f' .emit 'i' .emit 'n' .emit 'e'
-              .emit 'd' .emit ' ' .or
-   dir_else .emit TOKEN_ELSE .or
-   dir_elif .emit TOKEN_ELIF .or
-   dir_endif .emit TOKEN_ENDIF .or
-   dir_ext .emit TOKEN_EXTENSION .or
-   dir_pragma .emit TOKEN_PRAGMA .or
-   dir_line .emit TOKEN_LINE;
-
-dir_define
-   optional_space .and '#' .and optional_space .and "define" .and symbol .and opt_parameters .and
-   definition;
-
-dir_undef
-   optional_space .and '#' .and optional_space .and "undef" .and symbol;
-
-dir_if
-   optional_space .and '#' .and optional_space .and "if" .and expression;
-
-dir_ifdef
-   optional_space .and '#' .and optional_space .and "ifdef" .and symbol;
-
-dir_ifndef
-   optional_space .and '#' .and optional_space .and "ifndef" .and symbol;
-
-dir_else
-   optional_space .and '#' .and optional_space .and "else";
-
-dir_elif
-   optional_space .and '#' .and optional_space .and "elif" .and expression;
-
-dir_endif
-   optional_space .and '#' .and optional_space .and "endif";
-
-dir_ext
-   optional_space .and '#' .and optional_space .and "extension" .and space .and extension_name .and
-   optional_space .and ':' .and optional_space .and extension_behavior;
-
-dir_line
-   optional_space .and '#' .and optional_space .and "line" .and expression;
-
-dir_pragma
-   optional_space .and '#' .and optional_space .and "pragma" .and symbol .and opt_pragma_param;
-
-
-opt_pragma_param
-   pragma_param .or .true .emit PRAGMA_NO_PARAM;
-
-pragma_param
-   optional_space .and '(' .emit PRAGMA_PARAM .and optional_space .and symbol_no_space .and optional_space .and ')';
-
-symbol_no_space
-   symbol_character .emit * .and .loop symbol_character2 .emit * .and .true .emit '\0';
-
-symbol
-   space .and symbol_character .emit * .and .loop symbol_character2 .emit * .and .true .emit '\0';
-
-opt_parameters
-   parameters .or .true .emit PARAM_END;
-
-parameters
-   '(' .and parameters_1 .and optional_space .and ')' .emit PARAM_END;
-parameters_1
-   parameters_2 .or .true;
-parameters_2
-   parameter .emit PARAM_PARAMETER .and .loop parameters_3;
-parameters_3
-   optional_space .and ',' .and parameter .emit PARAM_PARAMETER;
-
-parameter
-   optional_space .and symbol_character .emit * .and .loop symbol_character2 .emit * .and
-   .true .emit '\0';
-
-definition
-   .loop definition_character .emit * .and .true .emit '\0';
-
-definition_character
-   '\x0E'-'\xFF' .or '\x01'-'\x09' .or '\x0B'-'\x0C';
-
-expression
-   expression_element .and .loop expression_element .and .true .emit '\0';
-
-expression_element
-   expression_character .emit *;
-
-expression_character
-   '\x0E'-'\xFF' .or '\x01'-'\x09' .or '\x0B'-'\x0C';
-
-extension_name
-   symbol_character .emit * .and .loop symbol_character2 .emit * .and .true .emit '\0';
-
-extension_behavior
-   "require" .emit BEHAVIOR_REQUIRE .or
-   "enable" .emit BEHAVIOR_ENABLE .or
-   "warn" .emit BEHAVIOR_WARN .or
-   "disable" .emit BEHAVIOR_DISABLE;
-
-optional_space
-   .loop single_space;
-
-space
-   single_space .and .loop single_space;
-
-single_space
-   ' ' .or '\t';
-
-source_token
-   space .emit ' ' .or complex_token .or source_token_1;
-source_token_1
-   simple_token .emit ' ' .and .true .emit ' ';
-
-/*
- * All possible tokens.
- */
-
-complex_token
-   identifier .or number;
-
-simple_token
-   increment .or decrement .or lequal .or gequal .or equal .or nequal .or and .or xor .or or .or
-   addto .or subtractfrom .or multiplyto .or divideto .or other;
-
-identifier
-   identifier_char1 .emit * .and .loop identifier_char2 .emit *;
-identifier_char1
-   'a'-'z' .or 'A'-'Z' .or '_';
-identifier_char2
-   'a'-'z' .or 'A'-'Z' .or '0'-'9' .or '_';
-
-number
-   float .or integer;
-
-digit_oct
-   '0'-'7';
-
-digit_dec
-   '0'-'9';
-
-digit_hex
-   '0'-'9' .or 'A'-'F' .or 'a'-'f';
-
-float
-   float_1 .or float_2;
-float_1
-   float_fractional_constant .and float_optional_exponent_part;
-float_2
-   float_digit_sequence .and float_exponent_part;
-
-float_fractional_constant
-   float_fractional_constant_1 .or float_fractional_constant_2 .or float_fractional_constant_3;
-float_fractional_constant_1
-   float_digit_sequence .and '.' .emit '.' .and float_digit_sequence;
-float_fractional_constant_2
-   float_digit_sequence .and '.' .emit '.';
-float_fractional_constant_3
-   '.' .emit '.' .and float_digit_sequence;
-
-float_optional_exponent_part
-   float_exponent_part .or .true;
-
-float_digit_sequence
-   digit_dec .emit * .and .loop digit_dec .emit *;
-
-float_exponent_part
-   float_exponent_part_1 .or float_exponent_part_2;
-float_exponent_part_1
-   'e' .emit 'e' .and float_optional_sign .and float_digit_sequence;
-float_exponent_part_2
-   'E' .emit 'E' .and float_optional_sign .and float_digit_sequence;
-
-float_optional_sign
-   '+' .emit '+' .or '-' .emit '-' .or .true;
-
-integer
-   integer_hex .or integer_oct .or integer_dec;
-
-integer_hex
-   '0' .emit '0' .and integer_hex_1 .emit * .and digit_hex .emit * .and
-   .loop digit_hex .emit *;
-integer_hex_1
-   'x' .or 'X';
-
-integer_oct
-   '0' .emit '0' .and .loop digit_oct .emit *;
-
-integer_dec
-   digit_dec .emit * .and .loop digit_dec .emit *;
-
-increment
-   '+' .emit * .and '+' .emit *;
-
-decrement
-   '-' .emit * .and '-' .emit *;
-
-lequal
-   '<' .emit * .and '=' .emit *;
-
-gequal
-   '>' .emit * .and '=' .emit *;
-
-equal
-   '=' .emit * .and '=' .emit *;
-
-nequal
-   '!' .emit * .and '=' .emit *;
-
-and
-   '&' .emit * .and '&' .emit *;
-
-xor
-   '^' .emit * .and '^' .emit *;
-
-or
-   '|' .emit * .and '|' .emit *;
-
-addto
-   '+' .emit * .and '=' .emit *;
-
-subtractfrom
-   '-' .emit * .and '=' .emit *;
-
-multiplyto
-   '*' .emit * .and '=' .emit *;
-
-divideto
-   '/' .emit * .and '=' .emit *;
-
-/*
- * All characters except '\0' and '#'.
- */
-other
-   '\x24'-'\xFF' .emit * .or '\x01'-'\x22' .emit *;
-
-symbol_character
-   'A'-'Z' .or 'a'-'z' .or '_';
-
-symbol_character2
-   'A'-'Z' .or 'a'-'z' .or '0'-'9' .or '_';
-
-.string string_lexer;
-
-string_lexer
-   lex_first_identifier_character .and .loop lex_next_identifier_character;
-
-lex_first_identifier_character
-   'a'-'z' .or 'A'-'Z' .or '_';
-
-lex_next_identifier_character
-   'a'-'z' .or 'A'-'Z' .or '0'-'9' .or '_';
-
diff --git a/src/mesa/shader/slang/library/slang_pp_directives_syn.h b/src/mesa/shader/slang/library/slang_pp_directives_syn.h
deleted file mode 100644 (file)
index 430f8d8..0000000
+++ /dev/null
@@ -1,250 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE .syn FILE */
-
-".syntax source;\n"
-".emtcode ESCAPE_TOKEN 0\n"
-".emtcode TOKEN_END 0\n"
-".emtcode TOKEN_DEFINE 1\n"
-".emtcode TOKEN_UNDEF 2\n"
-".emtcode TOKEN_IF 3\n"
-".emtcode TOKEN_ELSE 4\n"
-".emtcode TOKEN_ELIF 5\n"
-".emtcode TOKEN_ENDIF 6\n"
-".emtcode TOKEN_ERROR 7\n"
-".emtcode TOKEN_PRAGMA 8\n"
-".emtcode TOKEN_EXTENSION 9\n"
-".emtcode TOKEN_LINE 10\n"
-".emtcode PARAM_END 0\n"
-".emtcode PARAM_PARAMETER 1\n"
-".emtcode BEHAVIOR_REQUIRE 1\n"
-".emtcode BEHAVIOR_ENABLE 2\n"
-".emtcode BEHAVIOR_WARN 3\n"
-".emtcode BEHAVIOR_DISABLE 4\n"
-".emtcode PRAGMA_NO_PARAM 0\n"
-".emtcode PRAGMA_PARAM 1\n"
-"source\n"
-" optional_directive .and .loop source_element .and '\\0' .emit ESCAPE_TOKEN .emit TOKEN_END;\n"
-"source_element\n"
-" c_style_comment_block .or cpp_style_comment_block .or new_line_directive .or source_token;\n"
-"c_style_comment_block\n"
-" '/' .and '*' .and c_style_comment_rest .and .true .emit ' ';\n"
-"c_style_comment_rest\n"
-" .loop c_style_comment_body .and c_style_comment_end;\n"
-"c_style_comment_body\n"
-" c_style_comment_char_nostar .or c_style_comment_char_star_noslashstar;\n"
-"c_style_comment_char_nostar\n"
-" new_line .or '\\x2B'-'\\xFF' .or '\\x01'-'\\x29';\n"
-"c_style_comment_char_star_noslashstar\n"
-" '*' .and c_style_comment_char_star_noslashstar_1;\n"
-"c_style_comment_char_star_noslashstar_1\n"
-" c_style_comment_char_noslashstar .or c_style_comment_char_star_noslashstar;\n"
-"c_style_comment_char_noslashstar\n"
-" new_line .or '\\x30'-'\\xFF' .or '\\x01'-'\\x29' .or '\\x2B'-'\\x2E';\n"
-"c_style_comment_end\n"
-" '*' .and .loop c_style_comment_char_star .and '/';\n"
-"c_style_comment_char_star\n"
-" '*';\n"
-"cpp_style_comment_block\n"
-" '/' .and '/' .and cpp_style_comment_block_1;\n"
-"cpp_style_comment_block_1\n"
-" cpp_style_comment_block_2 .or cpp_style_comment_block_3;\n"
-"cpp_style_comment_block_2\n"
-" .loop cpp_style_comment_char .and new_line_directive;\n"
-"cpp_style_comment_block_3\n"
-" .loop cpp_style_comment_char;\n"
-"cpp_style_comment_char\n"
-" '\\x0E'-'\\xFF' .or '\\x01'-'\\x09' .or '\\x0B'-'\\x0C';\n"
-"new_line_directive\n"
-" new_line .and optional_directive;\n"
-"new_line\n"
-" generic_new_line .emit '\\n';\n"
-"generic_new_line\n"
-" carriage_return_line_feed .or line_feed_carriage_return .or '\\n' .or '\\r';\n"
-"carriage_return_line_feed\n"
-" '\\r' .and '\\n';\n"
-"line_feed_carriage_return\n"
-" '\\n' .and '\\r';\n"
-"optional_directive\n"
-" directive .emit ESCAPE_TOKEN .or .true;\n"
-"directive\n"
-" dir_define .emit TOKEN_DEFINE .or\n"
-" dir_undef .emit TOKEN_UNDEF .or\n"
-" dir_if .emit TOKEN_IF .or\n"
-" dir_ifdef .emit TOKEN_IF .emit 'd' .emit 'e' .emit 'f' .emit 'i' .emit 'n' .emit 'e' .emit 'd'\n"
-" .emit ' ' .or\n"
-" dir_ifndef .emit TOKEN_IF .emit '!' .emit 'd' .emit 'e' .emit 'f' .emit 'i' .emit 'n' .emit 'e'\n"
-" .emit 'd' .emit ' ' .or\n"
-" dir_else .emit TOKEN_ELSE .or\n"
-" dir_elif .emit TOKEN_ELIF .or\n"
-" dir_endif .emit TOKEN_ENDIF .or\n"
-" dir_ext .emit TOKEN_EXTENSION .or\n"
-" dir_pragma .emit TOKEN_PRAGMA .or\n"
-" dir_line .emit TOKEN_LINE;\n"
-"dir_define\n"
-" optional_space .and '#' .and optional_space .and \"define\" .and symbol .and opt_parameters .and\n"
-" definition;\n"
-"dir_undef\n"
-" optional_space .and '#' .and optional_space .and \"undef\" .and symbol;\n"
-"dir_if\n"
-" optional_space .and '#' .and optional_space .and \"if\" .and expression;\n"
-"dir_ifdef\n"
-" optional_space .and '#' .and optional_space .and \"ifdef\" .and symbol;\n"
-"dir_ifndef\n"
-" optional_space .and '#' .and optional_space .and \"ifndef\" .and symbol;\n"
-"dir_else\n"
-" optional_space .and '#' .and optional_space .and \"else\";\n"
-"dir_elif\n"
-" optional_space .and '#' .and optional_space .and \"elif\" .and expression;\n"
-"dir_endif\n"
-" optional_space .and '#' .and optional_space .and \"endif\";\n"
-"dir_ext\n"
-" optional_space .and '#' .and optional_space .and \"extension\" .and space .and extension_name .and\n"
-" optional_space .and ':' .and optional_space .and extension_behavior;\n"
-"dir_line\n"
-" optional_space .and '#' .and optional_space .and \"line\" .and expression;\n"
-"dir_pragma\n"
-" optional_space .and '#' .and optional_space .and \"pragma\" .and symbol .and opt_pragma_param;\n"
-"opt_pragma_param\n"
-" pragma_param .or .true .emit PRAGMA_NO_PARAM;\n"
-"pragma_param\n"
-" optional_space .and '(' .emit PRAGMA_PARAM .and optional_space .and symbol_no_space .and optional_space .and ')';\n"
-"symbol_no_space\n"
-" symbol_character .emit * .and .loop symbol_character2 .emit * .and .true .emit '\\0';\n"
-"symbol\n"
-" space .and symbol_character .emit * .and .loop symbol_character2 .emit * .and .true .emit '\\0';\n"
-"opt_parameters\n"
-" parameters .or .true .emit PARAM_END;\n"
-"parameters\n"
-" '(' .and parameters_1 .and optional_space .and ')' .emit PARAM_END;\n"
-"parameters_1\n"
-" parameters_2 .or .true;\n"
-"parameters_2\n"
-" parameter .emit PARAM_PARAMETER .and .loop parameters_3;\n"
-"parameters_3\n"
-" optional_space .and ',' .and parameter .emit PARAM_PARAMETER;\n"
-"parameter\n"
-" optional_space .and symbol_character .emit * .and .loop symbol_character2 .emit * .and\n"
-" .true .emit '\\0';\n"
-"definition\n"
-" .loop definition_character .emit * .and .true .emit '\\0';\n"
-"definition_character\n"
-" '\\x0E'-'\\xFF' .or '\\x01'-'\\x09' .or '\\x0B'-'\\x0C';\n"
-"expression\n"
-" expression_element .and .loop expression_element .and .true .emit '\\0';\n"
-"expression_element\n"
-" expression_character .emit *;\n"
-"expression_character\n"
-" '\\x0E'-'\\xFF' .or '\\x01'-'\\x09' .or '\\x0B'-'\\x0C';\n"
-"extension_name\n"
-" symbol_character .emit * .and .loop symbol_character2 .emit * .and .true .emit '\\0';\n"
-"extension_behavior\n"
-" \"require\" .emit BEHAVIOR_REQUIRE .or\n"
-" \"enable\" .emit BEHAVIOR_ENABLE .or\n"
-" \"warn\" .emit BEHAVIOR_WARN .or\n"
-" \"disable\" .emit BEHAVIOR_DISABLE;\n"
-"optional_space\n"
-" .loop single_space;\n"
-"space\n"
-" single_space .and .loop single_space;\n"
-"single_space\n"
-" ' ' .or '\\t';\n"
-"source_token\n"
-" space .emit ' ' .or complex_token .or source_token_1;\n"
-"source_token_1\n"
-" simple_token .emit ' ' .and .true .emit ' ';\n"
-"complex_token\n"
-" identifier .or number;\n"
-"simple_token\n"
-" increment .or decrement .or lequal .or gequal .or equal .or nequal .or and .or xor .or or .or\n"
-" addto .or subtractfrom .or multiplyto .or divideto .or other;\n"
-"identifier\n"
-" identifier_char1 .emit * .and .loop identifier_char2 .emit *;\n"
-"identifier_char1\n"
-" 'a'-'z' .or 'A'-'Z' .or '_';\n"
-"identifier_char2\n"
-" 'a'-'z' .or 'A'-'Z' .or '0'-'9' .or '_';\n"
-"number\n"
-" float .or integer;\n"
-"digit_oct\n"
-" '0'-'7';\n"
-"digit_dec\n"
-" '0'-'9';\n"
-"digit_hex\n"
-" '0'-'9' .or 'A'-'F' .or 'a'-'f';\n"
-"float\n"
-" float_1 .or float_2;\n"
-"float_1\n"
-" float_fractional_constant .and float_optional_exponent_part;\n"
-"float_2\n"
-" float_digit_sequence .and float_exponent_part;\n"
-"float_fractional_constant\n"
-" float_fractional_constant_1 .or float_fractional_constant_2 .or float_fractional_constant_3;\n"
-"float_fractional_constant_1\n"
-" float_digit_sequence .and '.' .emit '.' .and float_digit_sequence;\n"
-"float_fractional_constant_2\n"
-" float_digit_sequence .and '.' .emit '.';\n"
-"float_fractional_constant_3\n"
-" '.' .emit '.' .and float_digit_sequence;\n"
-"float_optional_exponent_part\n"
-" float_exponent_part .or .true;\n"
-"float_digit_sequence\n"
-" digit_dec .emit * .and .loop digit_dec .emit *;\n"
-"float_exponent_part\n"
-" float_exponent_part_1 .or float_exponent_part_2;\n"
-"float_exponent_part_1\n"
-" 'e' .emit 'e' .and float_optional_sign .and float_digit_sequence;\n"
-"float_exponent_part_2\n"
-" 'E' .emit 'E' .and float_optional_sign .and float_digit_sequence;\n"
-"float_optional_sign\n"
-" '+' .emit '+' .or '-' .emit '-' .or .true;\n"
-"integer\n"
-" integer_hex .or integer_oct .or integer_dec;\n"
-"integer_hex\n"
-" '0' .emit '0' .and integer_hex_1 .emit * .and digit_hex .emit * .and\n"
-" .loop digit_hex .emit *;\n"
-"integer_hex_1\n"
-" 'x' .or 'X';\n"
-"integer_oct\n"
-" '0' .emit '0' .and .loop digit_oct .emit *;\n"
-"integer_dec\n"
-" digit_dec .emit * .and .loop digit_dec .emit *;\n"
-"increment\n"
-" '+' .emit * .and '+' .emit *;\n"
-"decrement\n"
-" '-' .emit * .and '-' .emit *;\n"
-"lequal\n"
-" '<' .emit * .and '=' .emit *;\n"
-"gequal\n"
-" '>' .emit * .and '=' .emit *;\n"
-"equal\n"
-" '=' .emit * .and '=' .emit *;\n"
-"nequal\n"
-" '!' .emit * .and '=' .emit *;\n"
-"and\n"
-" '&' .emit * .and '&' .emit *;\n"
-"xor\n"
-" '^' .emit * .and '^' .emit *;\n"
-"or\n"
-" '|' .emit * .and '|' .emit *;\n"
-"addto\n"
-" '+' .emit * .and '=' .emit *;\n"
-"subtractfrom\n"
-" '-' .emit * .and '=' .emit *;\n"
-"multiplyto\n"
-" '*' .emit * .and '=' .emit *;\n"
-"divideto\n"
-" '/' .emit * .and '=' .emit *;\n"
-"other\n"
-" '\\x24'-'\\xFF' .emit * .or '\\x01'-'\\x22' .emit *;\n"
-"symbol_character\n"
-" 'A'-'Z' .or 'a'-'z' .or '_';\n"
-"symbol_character2\n"
-" 'A'-'Z' .or 'a'-'z' .or '0'-'9' .or '_';\n"
-".string string_lexer;\n"
-"string_lexer\n"
-" lex_first_identifier_character .and .loop lex_next_identifier_character;\n"
-"lex_first_identifier_character\n"
-" 'a'-'z' .or 'A'-'Z' .or '_';\n"
-"lex_next_identifier_character\n"
-" 'a'-'z' .or 'A'-'Z' .or '0'-'9' .or '_';\n"
-""
diff --git a/src/mesa/shader/slang/library/slang_pp_expression.syn b/src/mesa/shader/slang/library/slang_pp_expression.syn
deleted file mode 100644 (file)
index bfdb220..0000000
+++ /dev/null
@@ -1,265 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  6.6
- *
- * Copyright (C) 2006  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/**
- * \file slang_pp_expression.syn
- * slang preprocessor expression parser
- * \author Michal Krol
- */
-
-/*
- * Parses one or two (optional) expressions on literal integer constants. Those expressions come
- * from #if #elif and #line directives. The preprocessor already parsed those directives and
- * expanded the expression (expressions). All occurences of the operator "defined" are already
- * replaced with either "0" or "1" literals.
- */
-
-.syntax expression;
-
-/*
- * Those separate individual expressions.
- * For #if/#elif case it is: EXP_EXPRESSION ... EXP_END
- * For #line case it may be: EXP_EXPRESSION ... EXP_EXPRESSION ... EXP_END
- */
-.emtcode EXP_END        0
-.emtcode EXP_EXPRESSION 1
-
-.emtcode OP_END          0
-.emtcode OP_PUSHINT      1
-.emtcode OP_LOGICALOR    2
-.emtcode OP_LOGICALAND   3
-.emtcode OP_OR           4
-.emtcode OP_XOR          5
-.emtcode OP_AND          6
-.emtcode OP_EQUAL        7
-.emtcode OP_NOTEQUAL     8
-.emtcode OP_LESSEQUAL    9
-.emtcode OP_GREATEREQUAL 10
-.emtcode OP_LESS         11
-.emtcode OP_GREATER      12
-.emtcode OP_LEFTSHIFT    13
-.emtcode OP_RIGHTSHIFT   14
-.emtcode OP_ADD          15
-.emtcode OP_SUBTRACT     16
-.emtcode OP_MULTIPLY     17
-.emtcode OP_DIVIDE       18
-.emtcode OP_MODULUS      19
-.emtcode OP_PLUS         20
-.emtcode OP_MINUS        21
-.emtcode OP_NEGATE       22
-.emtcode OP_COMPLEMENT   23
-
-expression
-   first_expression .and optional_second_expression .and optional_space .and '\0' .emit EXP_END;
-
-first_expression
-   optional_space .and logical_or_expression .emit EXP_EXPRESSION .and .true .emit OP_END;
-
-optional_second_expression
-   second_expression .or .true;
-
-second_expression
-   space .and logical_or_expression .emit EXP_EXPRESSION .and .true .emit OP_END;
-
-logical_or_expression
-   logical_and_expression .and .loop logical_or_expression_1;
-logical_or_expression_1
-   barbar .and logical_and_expression .and .true .emit OP_LOGICALOR;
-
-logical_and_expression
-   or_expression .and .loop logical_and_expression_1;
-logical_and_expression_1
-   ampersandampersand .and or_expression .and .true .emit OP_LOGICALAND;
-
-or_expression
-   xor_expression .and .loop or_expression_1;
-or_expression_1
-   bar .and xor_expression .and .true .emit OP_OR;
-
-xor_expression
-    and_expression .and .loop xor_expression_1;
-xor_expression_1
-    caret .and and_expression .and .true .emit OP_XOR;
-
-and_expression
-    equality_expression .and .loop and_expression_1;
-and_expression_1
-    ampersand .and equality_expression .and .true .emit OP_AND;
-
-equality_expression
-    relational_expression .and .loop equality_expression_1;
-equality_expression_1
-    equality_expression_2 .or equality_expression_3;
-equality_expression_2
-    equalsequals .and relational_expression .and .true .emit OP_EQUAL;
-equality_expression_3
-    bangequals .and relational_expression .and .true .emit OP_NOTEQUAL;
-
-relational_expression
-    shift_expression .and .loop relational_expression_1;
-relational_expression_1
-    relational_expression_2 .or relational_expression_3 .or relational_expression_4 .or
-    relational_expression_5;
-relational_expression_2
-    lessequals .and shift_expression .and .true .emit OP_LESSEQUAL;
-relational_expression_3
-    greaterequals .and shift_expression .and .true .emit OP_GREATEREQUAL;
-relational_expression_4
-    less .and shift_expression .and .true .emit OP_LESS;
-relational_expression_5
-    greater .and shift_expression .and .true .emit OP_GREATER;
-
-shift_expression
-    additive_expression .and .loop shift_expression_1;
-shift_expression_1
-    shift_expression_2 .or shift_expression_3;
-shift_expression_2
-    lessless .and additive_expression .and .true .emit OP_LEFTSHIFT;
-shift_expression_3
-    greatergreater .and additive_expression .and .true .emit OP_RIGHTSHIFT;
-
-additive_expression
-    multiplicative_expression .and .loop additive_expression_1;
-additive_expression_1
-    additive_expression_2 .or additive_expression_3;
-additive_expression_2
-    plus .and multiplicative_expression .and .true .emit OP_ADD;
-additive_expression_3
-    dash .and multiplicative_expression .and .true .emit OP_SUBTRACT;
-
-multiplicative_expression
-    unary_expression .and .loop multiplicative_expression_1;
-multiplicative_expression_1
-    multiplicative_expression_2 .or multiplicative_expression_3 .or multiplicative_expression_4;
-multiplicative_expression_2
-    star .and unary_expression .and .true .emit OP_MULTIPLY;
-multiplicative_expression_3
-    slash .and unary_expression .and .true .emit OP_DIVIDE;
-multiplicative_expression_4
-    percent .and unary_expression .and .true .emit OP_MODULUS;
-
-unary_expression
-    primary_expression .or unary_expression_1 .or unary_expression_2 .or unary_expression_3 .or
-    unary_expression_4;
-unary_expression_1
-    plus .and unary_expression .and .true .emit OP_PLUS;
-unary_expression_2
-    dash .and unary_expression .and .true .emit OP_MINUS;
-unary_expression_3
-    bang .and unary_expression .and .true .emit OP_NEGATE;
-unary_expression_4
-    tilda .and unary_expression .and .true .emit OP_COMPLEMENT;
-
-primary_expression
-   intconstant .or primary_expression_1;
-primary_expression_1
-   lparen .and logical_or_expression .and rparen;
-
-intconstant
-   integer .emit OP_PUSHINT;
-
-integer
-   integer_dec;
-
-integer_dec
-   digit_dec .emit 10 .emit * .and .loop digit_dec .emit * .and .true .emit '\0';
-
-digit_dec
-   '0'-'9';
-
-optional_space
-   .loop single_space;
-
-space
-   single_space .and .loop single_space;
-
-single_space
-   ' ' .or '\t';
-
-ampersand
-   optional_space .and '&' .and optional_space;
-
-ampersandampersand
-   optional_space .and '&' .and '&' .and optional_space;
-
-bang
-   optional_space .and '!' .and optional_space;
-
-bangequals
-   optional_space .and '!' .and '=' .and optional_space;
-
-bar
-   optional_space .and '|' .and optional_space;
-
-barbar
-   optional_space .and '|' .and '|' .and optional_space;
-
-caret
-   optional_space .and '^' .and optional_space;
-
-dash
-   optional_space .and '-' .and optional_space;
-
-equalsequals
-   optional_space .and '=' .and '=' .and optional_space;
-
-greater
-   optional_space .and '>' .and optional_space;
-
-greaterequals
-   optional_space .and '>' .and '=' .and optional_space;
-
-greatergreater
-   optional_space .and '>' .and '>' .and optional_space;
-
-less
-   optional_space .and '<' .and optional_space;
-
-lessequals
-   optional_space .and '<' .and '=' .and optional_space;
-
-lessless
-   optional_space .and '<' .and '<' .and optional_space;
-
-lparen
-   optional_space .and '(' .and optional_space;
-
-percent
-   optional_space .and '%' .and optional_space;
-
-plus
-   optional_space .and '+' .and optional_space;
-
-rparen
-   optional_space .and ')' .and optional_space;
-
-slash
-   optional_space .and '/' .and optional_space;
-
-star
-   optional_space .and '*' .and optional_space;
-
-tilda
-   optional_space .and '~' .and optional_space;
-
diff --git a/src/mesa/shader/slang/library/slang_pp_expression_syn.h b/src/mesa/shader/slang/library/slang_pp_expression_syn.h
deleted file mode 100644 (file)
index f3e9ef6..0000000
+++ /dev/null
@@ -1,179 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE .syn FILE */
-
-".syntax expression;\n"
-".emtcode EXP_END 0\n"
-".emtcode EXP_EXPRESSION 1\n"
-".emtcode OP_END 0\n"
-".emtcode OP_PUSHINT 1\n"
-".emtcode OP_LOGICALOR 2\n"
-".emtcode OP_LOGICALAND 3\n"
-".emtcode OP_OR 4\n"
-".emtcode OP_XOR 5\n"
-".emtcode OP_AND 6\n"
-".emtcode OP_EQUAL 7\n"
-".emtcode OP_NOTEQUAL 8\n"
-".emtcode OP_LESSEQUAL 9\n"
-".emtcode OP_GREATEREQUAL 10\n"
-".emtcode OP_LESS 11\n"
-".emtcode OP_GREATER 12\n"
-".emtcode OP_LEFTSHIFT 13\n"
-".emtcode OP_RIGHTSHIFT 14\n"
-".emtcode OP_ADD 15\n"
-".emtcode OP_SUBTRACT 16\n"
-".emtcode OP_MULTIPLY 17\n"
-".emtcode OP_DIVIDE 18\n"
-".emtcode OP_MODULUS 19\n"
-".emtcode OP_PLUS 20\n"
-".emtcode OP_MINUS 21\n"
-".emtcode OP_NEGATE 22\n"
-".emtcode OP_COMPLEMENT 23\n"
-"expression\n"
-" first_expression .and optional_second_expression .and optional_space .and '\\0' .emit EXP_END;\n"
-"first_expression\n"
-" optional_space .and logical_or_expression .emit EXP_EXPRESSION .and .true .emit OP_END;\n"
-"optional_second_expression\n"
-" second_expression .or .true;\n"
-"second_expression\n"
-" space .and logical_or_expression .emit EXP_EXPRESSION .and .true .emit OP_END;\n"
-"logical_or_expression\n"
-" logical_and_expression .and .loop logical_or_expression_1;\n"
-"logical_or_expression_1\n"
-" barbar .and logical_and_expression .and .true .emit OP_LOGICALOR;\n"
-"logical_and_expression\n"
-" or_expression .and .loop logical_and_expression_1;\n"
-"logical_and_expression_1\n"
-" ampersandampersand .and or_expression .and .true .emit OP_LOGICALAND;\n"
-"or_expression\n"
-" xor_expression .and .loop or_expression_1;\n"
-"or_expression_1\n"
-" bar .and xor_expression .and .true .emit OP_OR;\n"
-"xor_expression\n"
-" and_expression .and .loop xor_expression_1;\n"
-"xor_expression_1\n"
-" caret .and and_expression .and .true .emit OP_XOR;\n"
-"and_expression\n"
-" equality_expression .and .loop and_expression_1;\n"
-"and_expression_1\n"
-" ampersand .and equality_expression .and .true .emit OP_AND;\n"
-"equality_expression\n"
-" relational_expression .and .loop equality_expression_1;\n"
-"equality_expression_1\n"
-" equality_expression_2 .or equality_expression_3;\n"
-"equality_expression_2\n"
-" equalsequals .and relational_expression .and .true .emit OP_EQUAL;\n"
-"equality_expression_3\n"
-" bangequals .and relational_expression .and .true .emit OP_NOTEQUAL;\n"
-"relational_expression\n"
-" shift_expression .and .loop relational_expression_1;\n"
-"relational_expression_1\n"
-" relational_expression_2 .or relational_expression_3 .or relational_expression_4 .or\n"
-" relational_expression_5;\n"
-"relational_expression_2\n"
-" lessequals .and shift_expression .and .true .emit OP_LESSEQUAL;\n"
-"relational_expression_3\n"
-" greaterequals .and shift_expression .and .true .emit OP_GREATEREQUAL;\n"
-"relational_expression_4\n"
-" less .and shift_expression .and .true .emit OP_LESS;\n"
-"relational_expression_5\n"
-" greater .and shift_expression .and .true .emit OP_GREATER;\n"
-"shift_expression\n"
-" additive_expression .and .loop shift_expression_1;\n"
-"shift_expression_1\n"
-" shift_expression_2 .or shift_expression_3;\n"
-"shift_expression_2\n"
-" lessless .and additive_expression .and .true .emit OP_LEFTSHIFT;\n"
-"shift_expression_3\n"
-" greatergreater .and additive_expression .and .true .emit OP_RIGHTSHIFT;\n"
-"additive_expression\n"
-" multiplicative_expression .and .loop additive_expression_1;\n"
-"additive_expression_1\n"
-" additive_expression_2 .or additive_expression_3;\n"
-"additive_expression_2\n"
-" plus .and multiplicative_expression .and .true .emit OP_ADD;\n"
-"additive_expression_3\n"
-" dash .and multiplicative_expression .and .true .emit OP_SUBTRACT;\n"
-"multiplicative_expression\n"
-" unary_expression .and .loop multiplicative_expression_1;\n"
-"multiplicative_expression_1\n"
-" multiplicative_expression_2 .or multiplicative_expression_3 .or multiplicative_expression_4;\n"
-"multiplicative_expression_2\n"
-" star .and unary_expression .and .true .emit OP_MULTIPLY;\n"
-"multiplicative_expression_3\n"
-" slash .and unary_expression .and .true .emit OP_DIVIDE;\n"
-"multiplicative_expression_4\n"
-" percent .and unary_expression .and .true .emit OP_MODULUS;\n"
-"unary_expression\n"
-" primary_expression .or unary_expression_1 .or unary_expression_2 .or unary_expression_3 .or\n"
-" unary_expression_4;\n"
-"unary_expression_1\n"
-" plus .and unary_expression .and .true .emit OP_PLUS;\n"
-"unary_expression_2\n"
-" dash .and unary_expression .and .true .emit OP_MINUS;\n"
-"unary_expression_3\n"
-" bang .and unary_expression .and .true .emit OP_NEGATE;\n"
-"unary_expression_4\n"
-" tilda .and unary_expression .and .true .emit OP_COMPLEMENT;\n"
-"primary_expression\n"
-" intconstant .or primary_expression_1;\n"
-"primary_expression_1\n"
-" lparen .and logical_or_expression .and rparen;\n"
-"intconstant\n"
-" integer .emit OP_PUSHINT;\n"
-"integer\n"
-" integer_dec;\n"
-"integer_dec\n"
-" digit_dec .emit 10 .emit * .and .loop digit_dec .emit * .and .true .emit '\\0';\n"
-"digit_dec\n"
-" '0'-'9';\n"
-"optional_space\n"
-" .loop single_space;\n"
-"space\n"
-" single_space .and .loop single_space;\n"
-"single_space\n"
-" ' ' .or '\\t';\n"
-"ampersand\n"
-" optional_space .and '&' .and optional_space;\n"
-"ampersandampersand\n"
-" optional_space .and '&' .and '&' .and optional_space;\n"
-"bang\n"
-" optional_space .and '!' .and optional_space;\n"
-"bangequals\n"
-" optional_space .and '!' .and '=' .and optional_space;\n"
-"bar\n"
-" optional_space .and '|' .and optional_space;\n"
-"barbar\n"
-" optional_space .and '|' .and '|' .and optional_space;\n"
-"caret\n"
-" optional_space .and '^' .and optional_space;\n"
-"dash\n"
-" optional_space .and '-' .and optional_space;\n"
-"equalsequals\n"
-" optional_space .and '=' .and '=' .and optional_space;\n"
-"greater\n"
-" optional_space .and '>' .and optional_space;\n"
-"greaterequals\n"
-" optional_space .and '>' .and '=' .and optional_space;\n"
-"greatergreater\n"
-" optional_space .and '>' .and '>' .and optional_space;\n"
-"less\n"
-" optional_space .and '<' .and optional_space;\n"
-"lessequals\n"
-" optional_space .and '<' .and '=' .and optional_space;\n"
-"lessless\n"
-" optional_space .and '<' .and '<' .and optional_space;\n"
-"lparen\n"
-" optional_space .and '(' .and optional_space;\n"
-"percent\n"
-" optional_space .and '%' .and optional_space;\n"
-"plus\n"
-" optional_space .and '+' .and optional_space;\n"
-"rparen\n"
-" optional_space .and ')' .and optional_space;\n"
-"slash\n"
-" optional_space .and '/' .and optional_space;\n"
-"star\n"
-" optional_space .and '*' .and optional_space;\n"
-"tilda\n"
-" optional_space .and '~' .and optional_space;\n"
-""
diff --git a/src/mesa/shader/slang/library/slang_pp_version.syn b/src/mesa/shader/slang/library/slang_pp_version.syn
deleted file mode 100644 (file)
index 3fe1a57..0000000
+++ /dev/null
@@ -1,122 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  6.6
- *
- * Copyright (C) 2005-2006  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/**
- * \file slang_pp_version.syn
- * slang #version directive syntax
- * \author Michal Krol
- */
-
-.syntax version_directive;
-
-version_directive
-       version_directive_1;
-version_directive_1
-       prior_optional_spaces .and optional_version_directive .and .true .emit $;
-
-optional_version_directive
-       version_directive_body .or .true .emit 10 .emit 1;
-
-version_directive_body
-       '#' .and optional_space .and "version" .and space .and version_number .and optional_space .and
-       new_line;
-
-version_number
-       version_number_100 .or version_number_110 .or version_number_120;
-
-version_number_100
-       leading_zeroes .and "100" .emit 0 .emit 1;
-
-version_number_110
-       leading_zeroes .and "110" .emit 10 .emit 1;
-
-version_number_120
-       leading_zeroes .and "120" .emit 20 .emit 1;
-
-leading_zeroes
-       .loop zero;
-
-zero
-       '0';
-
-space
-    single_space .and .loop single_space;
-
-optional_space
-    .loop single_space;
-
-single_space
-    ' ' .or '\t';
-
-prior_optional_spaces
-       .loop prior_space;
-
-prior_space
-       c_style_comment_block .or cpp_style_comment_block .or space .or new_line;
-
-c_style_comment_block
-    '/' .and '*' .and c_style_comment_rest;
-
-c_style_comment_rest
-    .loop c_style_comment_char_no_star .and c_style_comment_rest_1;
-c_style_comment_rest_1
-    c_style_comment_end .or c_style_comment_rest_2;
-c_style_comment_rest_2
-    '*' .and c_style_comment_rest;
-
-c_style_comment_char_no_star
-    '\x2B'-'\xFF' .or '\x01'-'\x29';
-
-c_style_comment_end
-    '*' .and '/';
-
-cpp_style_comment_block
-    '/' .and '/' .and cpp_style_comment_block_1;
-cpp_style_comment_block_1
-    cpp_style_comment_block_2 .or cpp_style_comment_block_3;
-cpp_style_comment_block_2
-    .loop cpp_style_comment_char .and new_line;
-cpp_style_comment_block_3
-    .loop cpp_style_comment_char;
-
-cpp_style_comment_char
-    '\x0E'-'\xFF' .or '\x01'-'\x09' .or '\x0B'-'\x0C';
-
-new_line
-    cr_lf .or lf_cr .or '\n' .or '\r';
-
-cr_lf
-    '\r' .and '\n';
-
-lf_cr
-       '\n' .and '\r';
-
-.string __string_filter;
-
-__string_filter
-    .loop __identifier_char;
-
-__identifier_char
-    'a'-'z' .or 'A'-'Z' .or '_' .or '0'-'9';
-
diff --git a/src/mesa/shader/slang/library/slang_pp_version_syn.h b/src/mesa/shader/slang/library/slang_pp_version_syn.h
deleted file mode 100644 (file)
index 54aee3f..0000000
+++ /dev/null
@@ -1,69 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE .syn FILE */
-
-".syntax version_directive;\n"
-"version_directive\n"
-"      version_directive_1;\n"
-"version_directive_1\n"
-"      prior_optional_spaces .and optional_version_directive .and .true .emit $;\n"
-"optional_version_directive\n"
-"      version_directive_body .or .true .emit 10 .emit 1;\n"
-"version_directive_body\n"
-"      '#' .and optional_space .and \"version\" .and space .and version_number .and optional_space .and\n"
-"      new_line;\n"
-"version_number\n"
-"      version_number_100 .or version_number_110 .or version_number_120;\n"
-"version_number_100\n"
-"      leading_zeroes .and \"100\" .emit 0 .emit 1;\n"
-"version_number_110\n"
-"      leading_zeroes .and \"110\" .emit 10 .emit 1;\n"
-"version_number_120\n"
-"      leading_zeroes .and \"120\" .emit 20 .emit 1;\n"
-"leading_zeroes\n"
-"      .loop zero;\n"
-"zero\n"
-"      '0';\n"
-"space\n"
-" single_space .and .loop single_space;\n"
-"optional_space\n"
-" .loop single_space;\n"
-"single_space\n"
-" ' ' .or '\\t';\n"
-"prior_optional_spaces\n"
-"      .loop prior_space;\n"
-"prior_space\n"
-"      c_style_comment_block .or cpp_style_comment_block .or space .or new_line;\n"
-"c_style_comment_block\n"
-" '/' .and '*' .and c_style_comment_rest;\n"
-"c_style_comment_rest\n"
-" .loop c_style_comment_char_no_star .and c_style_comment_rest_1;\n"
-"c_style_comment_rest_1\n"
-" c_style_comment_end .or c_style_comment_rest_2;\n"
-"c_style_comment_rest_2\n"
-" '*' .and c_style_comment_rest;\n"
-"c_style_comment_char_no_star\n"
-" '\\x2B'-'\\xFF' .or '\\x01'-'\\x29';\n"
-"c_style_comment_end\n"
-" '*' .and '/';\n"
-"cpp_style_comment_block\n"
-" '/' .and '/' .and cpp_style_comment_block_1;\n"
-"cpp_style_comment_block_1\n"
-" cpp_style_comment_block_2 .or cpp_style_comment_block_3;\n"
-"cpp_style_comment_block_2\n"
-" .loop cpp_style_comment_char .and new_line;\n"
-"cpp_style_comment_block_3\n"
-" .loop cpp_style_comment_char;\n"
-"cpp_style_comment_char\n"
-" '\\x0E'-'\\xFF' .or '\\x01'-'\\x09' .or '\\x0B'-'\\x0C';\n"
-"new_line\n"
-" cr_lf .or lf_cr .or '\\n' .or '\\r';\n"
-"cr_lf\n"
-" '\\r' .and '\\n';\n"
-"lf_cr\n"
-"      '\\n' .and '\\r';\n"
-".string __string_filter;\n"
-"__string_filter\n"
-" .loop __identifier_char;\n"
-"__identifier_char\n"
-" 'a'-'z' .or 'A'-'Z' .or '_' .or '0'-'9';\n"
-""
diff --git a/src/mesa/shader/slang/library/slang_shader.syn b/src/mesa/shader/slang/library/slang_shader.syn
deleted file mode 100644 (file)
index cc5c70a..0000000
+++ /dev/null
@@ -1,1716 +0,0 @@
-/*
- * Mesa 3-D graphics library
- *
- * Copyright (C) 2004-2006  Brian Paul   All Rights Reserved.
- * Copyright (C) 2008 VMware, Inc.  All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/*
- * \file slang_shader.syn
- * slang vertex/fragment shader syntax
- * \author Michal Krol
- */
-
-/*
- * usage:
- *   syn2c slang_shader.syn > slang_shader_syn.h
- *
- * when modifying or extending this file, several things must be taken into
- * consideration:
- *
- * - when adding new operators that were marked as reserved in the
- *   initial specification, one must only uncomment particular lines of
- *   code that refer to operators being added;
- *
- * - when adding new shader targets, one must reserve a new value for
- *   shader_type register and use it in .if constructs for symbols that
- *   are exclusive for that shader;
- *
- * - some symbols mimic output of other symbols - the best example is
- *   the "for" construct: expression "for (foo(); ; bar())" is seen as
- *   "for (foo(); true; bar())" by the output processor - hence, special
- *   care must be taken when rearranging output of essential symbols;
- *
- * - order of single-quoted tokens does matter in alternatives - so do not
- *   parse "<" operator before "<<" and "<<" before "<<=";
- *
- * - all double-quoted tokens are internally preprocessed to eliminate
- *   problems with parsing strings that are prefixes of other strings,
- *   like "sampler1D" and "sampler1DShadow";
- */
-
-.syntax translation_unit;
-
-/* revision number - increment after each change affecting emitted output */
-.emtcode REVISION                                   5
-
-/* external declaration (or precision or invariant stmt) */
-.emtcode EXTERNAL_NULL                              0
-.emtcode EXTERNAL_FUNCTION_DEFINITION               1
-.emtcode EXTERNAL_DECLARATION                       2
-.emtcode DEFAULT_PRECISION                          3
-.emtcode INVARIANT_STMT                             4
-
-/* precision */
-.emtcode PRECISION_DEFAULT                          0
-.emtcode PRECISION_LOW                              1
-.emtcode PRECISION_MEDIUM                           2
-.emtcode PRECISION_HIGH                             3
-
-/* declaration */
-.emtcode DECLARATION_FUNCTION_PROTOTYPE             1
-.emtcode DECLARATION_INIT_DECLARATOR_LIST           2
-
-/* function type */
-.emtcode FUNCTION_ORDINARY                          0
-.emtcode FUNCTION_CONSTRUCTOR                       1
-.emtcode FUNCTION_OPERATOR                          2
-
-/* function call type */
-.emtcode FUNCTION_CALL_NONARRAY                     0
-.emtcode FUNCTION_CALL_ARRAY                        1
-
-/* operator type */
-.emtcode OPERATOR_ADDASSIGN                         1
-.emtcode OPERATOR_SUBASSIGN                         2
-.emtcode OPERATOR_MULASSIGN                         3
-.emtcode OPERATOR_DIVASSIGN                         4
-/*.emtcode OPERATOR_MODASSIGN                         5*/
-/*.emtcode OPERATOR_LSHASSIGN                         6*/
-/*.emtcode OPERATOR_RSHASSIGN                         7*/
-/*.emtcode OPERATOR_ORASSIGN                          8*/
-/*.emtcode OPERATOR_XORASSIGN                         9*/
-/*.emtcode OPERATOR_ANDASSIGN                         10*/
-.emtcode OPERATOR_LOGICALXOR                        11
-/*.emtcode OPERATOR_BITOR                             12*/
-/*.emtcode OPERATOR_BITXOR                            13*/
-/*.emtcode OPERATOR_BITAND                            14*/
-.emtcode OPERATOR_LESS                              15
-.emtcode OPERATOR_GREATER                           16
-.emtcode OPERATOR_LESSEQUAL                         17
-.emtcode OPERATOR_GREATEREQUAL                      18
-/*.emtcode OPERATOR_LSHIFT                            19*/
-/*.emtcode OPERATOR_RSHIFT                            20*/
-.emtcode OPERATOR_MULTIPLY                          21
-.emtcode OPERATOR_DIVIDE                            22
-/*.emtcode OPERATOR_MODULUS                           23*/
-.emtcode OPERATOR_INCREMENT                         24
-.emtcode OPERATOR_DECREMENT                         25
-.emtcode OPERATOR_PLUS                              26
-.emtcode OPERATOR_MINUS                             27
-/*.emtcode OPERATOR_COMPLEMENT                        28*/
-.emtcode OPERATOR_NOT                               29
-
-/* init declarator list */
-.emtcode DECLARATOR_NONE                            0
-.emtcode DECLARATOR_NEXT                            1
-
-/* variable declaration */
-.emtcode VARIABLE_NONE                              0
-.emtcode VARIABLE_IDENTIFIER                        1
-.emtcode VARIABLE_INITIALIZER                       2
-.emtcode VARIABLE_ARRAY_EXPLICIT                    3
-.emtcode VARIABLE_ARRAY_UNKNOWN                     4
-
-/* type qualifier */
-.emtcode TYPE_QUALIFIER_NONE                        0
-.emtcode TYPE_QUALIFIER_CONST                       1
-.emtcode TYPE_QUALIFIER_ATTRIBUTE                   2
-.emtcode TYPE_QUALIFIER_VARYING                     3
-.emtcode TYPE_QUALIFIER_UNIFORM                     4
-.emtcode TYPE_QUALIFIER_FIXEDOUTPUT                 5
-.emtcode TYPE_QUALIFIER_FIXEDINPUT                  6
-
-/* invariant qualifier */
-.emtcode TYPE_VARIANT                               90
-.emtcode TYPE_INVARIANT                             91
-
-/* centroid qualifier */
-.emtcode TYPE_CENTER                                95
-.emtcode TYPE_CENTROID                              96
-
-/* type specifier */
-.emtcode TYPE_SPECIFIER_VOID                        0
-.emtcode TYPE_SPECIFIER_BOOL                        1
-.emtcode TYPE_SPECIFIER_BVEC2                       2
-.emtcode TYPE_SPECIFIER_BVEC3                       3
-.emtcode TYPE_SPECIFIER_BVEC4                       4
-.emtcode TYPE_SPECIFIER_INT                         5
-.emtcode TYPE_SPECIFIER_IVEC2                       6
-.emtcode TYPE_SPECIFIER_IVEC3                       7
-.emtcode TYPE_SPECIFIER_IVEC4                       8
-.emtcode TYPE_SPECIFIER_FLOAT                       9
-.emtcode TYPE_SPECIFIER_VEC2                        10
-.emtcode TYPE_SPECIFIER_VEC3                        11
-.emtcode TYPE_SPECIFIER_VEC4                        12
-.emtcode TYPE_SPECIFIER_MAT2                        13
-.emtcode TYPE_SPECIFIER_MAT3                        14
-.emtcode TYPE_SPECIFIER_MAT4                        15
-.emtcode TYPE_SPECIFIER_SAMPLER1D                   16
-.emtcode TYPE_SPECIFIER_SAMPLER2D                   17
-.emtcode TYPE_SPECIFIER_SAMPLER3D                   18
-.emtcode TYPE_SPECIFIER_SAMPLERCUBE                 19
-.emtcode TYPE_SPECIFIER_SAMPLER1DSHADOW             20
-.emtcode TYPE_SPECIFIER_SAMPLER2DSHADOW             21
-.emtcode TYPE_SPECIFIER_SAMPLER2DRECT               22
-.emtcode TYPE_SPECIFIER_SAMPLER2DRECTSHADOW         23
-.emtcode TYPE_SPECIFIER_STRUCT                      24
-.emtcode TYPE_SPECIFIER_TYPENAME                    25
-
-/* OpenGL 2.1 */
-.emtcode TYPE_SPECIFIER_MAT23                       26
-.emtcode TYPE_SPECIFIER_MAT32                       27
-.emtcode TYPE_SPECIFIER_MAT24                       28
-.emtcode TYPE_SPECIFIER_MAT42                       29
-.emtcode TYPE_SPECIFIER_MAT34                       30
-.emtcode TYPE_SPECIFIER_MAT43                       31
-
-/* type specifier array */
-.emtcode TYPE_SPECIFIER_NONARRAY                    0
-.emtcode TYPE_SPECIFIER_ARRAY                       1
-
-/* structure field */
-.emtcode FIELD_NONE                                 0
-.emtcode FIELD_NEXT                                 1
-.emtcode FIELD_ARRAY                                2
-
-/* operation */
-.emtcode OP_END                                     0
-.emtcode OP_BLOCK_BEGIN_NO_NEW_SCOPE                1
-.emtcode OP_BLOCK_BEGIN_NEW_SCOPE                   2
-.emtcode OP_DECLARE                                 3
-.emtcode OP_ASM                                     4
-.emtcode OP_BREAK                                   5
-.emtcode OP_CONTINUE                                6
-.emtcode OP_DISCARD                                 7
-.emtcode OP_RETURN                                  8
-.emtcode OP_EXPRESSION                              9
-.emtcode OP_IF                                      10
-.emtcode OP_WHILE                                   11
-.emtcode OP_DO                                      12
-.emtcode OP_FOR                                     13
-.emtcode OP_PUSH_VOID                               14
-.emtcode OP_PUSH_BOOL                               15
-.emtcode OP_PUSH_INT                                16
-.emtcode OP_PUSH_FLOAT                              17
-.emtcode OP_PUSH_IDENTIFIER                         18
-.emtcode OP_SEQUENCE                                19
-.emtcode OP_ASSIGN                                  20
-.emtcode OP_ADDASSIGN                               21
-.emtcode OP_SUBASSIGN                               22
-.emtcode OP_MULASSIGN                               23
-.emtcode OP_DIVASSIGN                               24
-/*.emtcode OP_MODASSIGN                               25*/
-/*.emtcode OP_LSHASSIGN                               26*/
-/*.emtcode OP_RSHASSIGN                               27*/
-/*.emtcode OP_ORASSIGN                                28*/
-/*.emtcode OP_XORASSIGN                               29*/
-/*.emtcode OP_ANDASSIGN                               30*/
-.emtcode OP_SELECT                                  31
-.emtcode OP_LOGICALOR                               32
-.emtcode OP_LOGICALXOR                              33
-.emtcode OP_LOGICALAND                              34
-/*.emtcode OP_BITOR                                   35*/
-/*.emtcode OP_BITXOR                                  36*/
-/*.emtcode OP_BITAND                                  37*/
-.emtcode OP_EQUAL                                   38
-.emtcode OP_NOTEQUAL                                39
-.emtcode OP_LESS                                    40
-.emtcode OP_GREATER                                 41
-.emtcode OP_LESSEQUAL                               42
-.emtcode OP_GREATEREQUAL                            43
-/*.emtcode OP_LSHIFT                                  44*/
-/*.emtcode OP_RSHIFT                                  45*/
-.emtcode OP_ADD                                     46
-.emtcode OP_SUBTRACT                                47
-.emtcode OP_MULTIPLY                                48
-.emtcode OP_DIVIDE                                  49
-/*.emtcode OP_MODULUS                                 50*/
-.emtcode OP_PREINCREMENT                            51
-.emtcode OP_PREDECREMENT                            52
-.emtcode OP_PLUS                                    53
-.emtcode OP_MINUS                                   54
-/*.emtcode OP_COMPLEMENT                              55*/
-.emtcode OP_NOT                                     56
-.emtcode OP_SUBSCRIPT                               57
-.emtcode OP_CALL                                    58
-.emtcode OP_FIELD                                   59
-.emtcode OP_POSTINCREMENT                           60
-.emtcode OP_POSTDECREMENT                           61
-.emtcode OP_PRECISION                               62
-.emtcode OP_METHOD                                  63
-
-/* parameter qualifier */
-.emtcode PARAM_QUALIFIER_IN                         0
-.emtcode PARAM_QUALIFIER_OUT                        1
-.emtcode PARAM_QUALIFIER_INOUT                      2
-
-/* function parameter */
-.emtcode PARAMETER_NONE                             0
-.emtcode PARAMETER_NEXT                             1
-
-/* function parameter array presence */
-.emtcode PARAMETER_ARRAY_NOT_PRESENT                0
-.emtcode PARAMETER_ARRAY_PRESENT                    1
-
-/* INVALID_EXTERNAL_DECLARATION seems to be reported when there's */
-/* any syntax errors... */
-.errtext INVALID_EXTERNAL_DECLARATION               "2001: Syntax error."
-.errtext INVALID_OPERATOR_OVERRIDE                  "2002: Invalid operator override."
-.errtext LBRACE_EXPECTED                            "2003: '{' expected but '$err_token$' found."
-.errtext LPAREN_EXPECTED                            "2004: '(' expected but '$err_token$' found."
-.errtext RPAREN_EXPECTED                            "2005: ')' expected but '$err_token$' found."
-.errtext INVALID_PRECISION                          "2006: Invalid precision specifier '$err_token$'."
-.errtext INVALID_PRECISION_TYPE                     "2007: Invalid precision type '$err_token$'."
-
-
-/*
- * tells whether the shader that is being parsed is a built-in shader or not
- *  0 - normal behaviour
- *  1 - accepts constructor and operator definitions and __asm statements
- * the implementation will set it to 1 when compiling internal built-in shaders
- */
-.regbyte parsing_builtin                            0
-
-/*
- * holds the type of the shader being parsed; possible values are
- * listed below.
- *   FRAGMENT_SHADER            1
- *   VERTEX_SHADER              2
- * shader type is set by the caller before parsing
- */
-.regbyte shader_type                                0
-
-/*
- * <variable_identifier> ::= <identifier>
- */
-variable_identifier
-    identifier .emit OP_PUSH_IDENTIFIER;
-
-/*
- *  <primary_expression> ::= <variable_identifier>
- *                         | <intconstant>
- *                         | <floatconstant>
- *                         | <boolconstant>
- *                         | "(" <expression> ")"
- */
-primary_expression
-    floatconstant .or boolconstant .or intconstant .or variable_identifier .or primary_expression_1;
-primary_expression_1
-    lparen .and expression .and rparen;
-
-/*
- * <postfix_expression> ::= <primary_expression>
- *                        | <postfix_expression> "[" <integer_expression> "]"
- *                        | <function_call>
- *                        | <postfix_expression> "." <field_selection>
- *                        | <postfix_expression> "++"
- *                        | <postfix_expression> "--"
- */
-postfix_expression
-    postfix_expression_1 .and .loop postfix_expression_2;
-postfix_expression_1
-    function_call .or primary_expression;
-postfix_expression_2
-    postfix_expression_3 .or postfix_expression_4 .or
-    plusplus .emit OP_POSTINCREMENT .or
-    minusminus .emit OP_POSTDECREMENT;
-postfix_expression_3
-    lbracket .and integer_expression .and rbracket .emit OP_SUBSCRIPT;
-postfix_expression_4
-    dot .and field_selection .emit OP_FIELD;
-
-/*
- * <integer_expression> ::= <expression>
- */
-integer_expression
-    expression;
-
-/*
- * <function_call> ::= <function_call_generic>
- */
-function_call
-    function_call_or_method;
-
-/*
- * <function_call_or_method> ::= <regular_function_call>
- *                             | <postfix_expression> "." <function_call_generic>
- */
-function_call_or_method
-    regular_function_call .or method_call;
-
-/*
- * <method_call> ::= <identifier> "." <function_call_generic>
- */
-method_call
-    identifier .emit OP_METHOD .and dot .and function_call_generic .and .true .emit OP_END;
-
-/*
- * <regular_function_call> ::= <function_call_generic>
- */
-regular_function_call
-    function_call_generic .emit OP_CALL .and .true .emit OP_END;
-
-/*
- * <function_call_generic> ::= <function_call_header_with_parameters> ")"
- *                           | <function_call_header_no_parameters> ")"
- */
-function_call_generic
-    function_call_generic_1 .or function_call_generic_2;
-function_call_generic_1
-    function_call_header_with_parameters .and rparen .error RPAREN_EXPECTED;
-function_call_generic_2
-    function_call_header_no_parameters .and rparen .error RPAREN_EXPECTED;
-
-/*
- * <function_call_header_no_parameters>::= <function_call_header> "void"
- *                                        | <function_call_header>
- */
-function_call_header_no_parameters
-    function_call_header .and function_call_header_no_parameters_1;
-function_call_header_no_parameters_1
-    "void" .or .true;
-
-/*
- * <function_call_header_with_parameters> ::= <function_call_header> <assignment_expression>
- *                                          | <function_call_header_with_parameters> "," <assignment_expression>
- */
-function_call_header_with_parameters
-    function_call_header .and assignment_expression .and .true .emit OP_END .and
-    .loop function_call_header_with_parameters_1;
-function_call_header_with_parameters_1
-    comma .and assignment_expression .and .true .emit OP_END;
-
-/*
- * <function_call_header> ::= <function_identifier> "("
- */
-function_call_header
-    function_identifier .and lparen;
-
-/*
- * <function_identifier> ::= <constructor_identifier>
- *                         | <identifier>
- *                         | <type_specifier>
- *
- * note: <constructor_identifier> and <type_specifier> have been deleted
- */
-function_identifier
-    identifier .and function_identifier_opt_array;
-function_identifier_opt_array
-    function_identifier_array .emit FUNCTION_CALL_ARRAY .or
-    .true .emit FUNCTION_CALL_NONARRAY;
-function_identifier_array
-    lbracket .and constant_expression .and rbracket;
-
-/*
- * <unary_expression> ::= <postfix_expression>
- *                      | "++" <unary_expression>
- *                      | "--" <unary_expression>
- *                      | <unary_operator> <unary_expression>
- *
- * <unary_operator>   ::= "+"
- *                      | "-"
- *                      | "!"
- *                      | "~" // reserved
- */
-unary_expression
-    postfix_expression .or unary_expression_1 .or unary_expression_2 .or unary_expression_3 .or
-    unary_expression_4 .or unary_expression_5/* .or unary_expression_6*/;
-unary_expression_1
-    plusplus .and unary_expression .and .true .emit OP_PREINCREMENT;
-unary_expression_2
-    minusminus .and unary_expression .and .true .emit OP_PREDECREMENT;
-unary_expression_3
-    plus .and unary_expression .and .true .emit OP_PLUS;
-unary_expression_4
-    minus .and unary_expression .and .true .emit OP_MINUS;
-unary_expression_5
-    bang .and unary_expression .and .true .emit OP_NOT;
-/*unary_expression_6
-    tilde .and unary_expression .and .true .emit OP_COMPLEMENT;*/
-
-/*
- * <multiplicative_expression> ::= <unary_expression>
- *                               | <multiplicative_expression> "*" <unary_expression>
- *                               | <multiplicative_expression> "/" <unary_expression>
- *                               | <multiplicative_expression> "%" <unary_expression> // reserved
- */
-multiplicative_expression
-    unary_expression .and .loop multiplicative_expression_1;
-multiplicative_expression_1
-    multiplicative_expression_2 .or multiplicative_expression_3/* .or multiplicative_expression_4*/;
-multiplicative_expression_2
-    star .and unary_expression .and .true .emit OP_MULTIPLY;
-multiplicative_expression_3
-    slash .and unary_expression .and .true .emit OP_DIVIDE;
-/*multiplicative_expression_4
-    percent .and unary_expression .and .true .emit OP_MODULUS;*/
-
-/*
- * <additive_expression> ::= <multiplicative_expression>
- *                         | <additive_expression> "+" <multiplicative_expression>
- *                         | <additive_expression> "-" <multiplicative_expression>
- */
-additive_expression
-    multiplicative_expression .and .loop additive_expression_1;
-additive_expression_1
-    additive_expression_2 .or additive_expression_3;
-additive_expression_2
-    plus .and multiplicative_expression .and .true .emit OP_ADD;
-additive_expression_3
-    minus .and multiplicative_expression .and .true .emit OP_SUBTRACT;
-
-/*
- * <shift_expression> ::= <additive_expression>
- *                      | <shift_expression> "<<" <additive_expression> // reserved
- *                      | <shift_expression> ">>" <additive_expression> // reserved
- */
-shift_expression
-    additive_expression/* .and .loop shift_expression_1*/;
-/*shift_expression_1
-    shift_expression_2 .or shift_expression_3;*/
-/*shift_expression_2
-    lessless .and additive_expression .and .true .emit OP_LSHIFT;*/
-/*shift_expression_3
-    greatergreater .and additive_expression .and .true .emit OP_RSHIFT;*/
-
-/*
- * <relational_expression> ::= <shift_expression>
- *                           | <relational_expression> "<" <shift_expression>
- *                           | <relational_expression> ">" <shift_expression>
- *                           | <relational_expression> "<=" <shift_expression>
- *                           | <relational_expression> ">=" <shift_expression>
- */
-relational_expression
-    shift_expression .and .loop relational_expression_1;
-relational_expression_1
-    relational_expression_2 .or relational_expression_3 .or relational_expression_4 .or
-    relational_expression_5;
-relational_expression_2
-    lessequals .and shift_expression .and .true .emit OP_LESSEQUAL;
-relational_expression_3
-    greaterequals .and shift_expression .and .true .emit OP_GREATEREQUAL;
-relational_expression_4
-    less .and shift_expression .and .true .emit OP_LESS;
-relational_expression_5
-    greater .and shift_expression .and .true .emit OP_GREATER;
-
-/*
- * <equality_expression> ::= <relational_expression>
- *                         | <equality_expression> "==" <relational_expression>
- *                         | <equality_expression> "!=" <relational_expression>
- */
-equality_expression
-    relational_expression .and .loop equality_expression_1;
-equality_expression_1
-    equality_expression_2 .or equality_expression_3;
-equality_expression_2
-    equalsequals .and relational_expression .and .true .emit OP_EQUAL;
-equality_expression_3
-    bangequals .and relational_expression .and .true .emit OP_NOTEQUAL;
-
-/*
- * <and_expression> ::= <equality_expression>
- *                    | <and_expression> "&" <equality_expression> // reserved
- */
-and_expression
-    equality_expression/* .and .loop and_expression_1*/;
-/*and_expression_1
-    ampersand .and equality_expression .and .true .emit OP_BITAND;*/
-
-/*
- * <exclusive_or_expression> ::= <and_expression>
- *                             | <exclusive_or_expression> "^" <and_expression> // reserved
- */
-exclusive_or_expression
-    and_expression/* .and .loop exclusive_or_expression_1*/;
-/*exclusive_or_expression_1
-    caret .and and_expression .and .true .emit OP_BITXOR;*/
-
-/*
- * <inclusive_or_expression> ::= <exclusive_or_expression>
- *                             | <inclusive_or_expression> "|" <exclusive_or_expression> // reserved
- */
-inclusive_or_expression
-    exclusive_or_expression/* .and .loop inclusive_or_expression_1*/;
-/*inclusive_or_expression_1
-    bar .and exclusive_or_expression .and .true .emit OP_BITOR;*/
-
-/*
- * <logical_and_expression> ::= <inclusive_or_expression>
- *                            | <logical_and_expression> "&&" <inclusive_or_expression>
- */
-logical_and_expression
-    inclusive_or_expression .and .loop logical_and_expression_1;
-logical_and_expression_1
-    ampersandampersand .and inclusive_or_expression .and .true .emit OP_LOGICALAND;
-
-/*
- * <logical_xor_expression> ::= <logical_and_expression>
- *                            | <logical_xor_expression> "^^" <logical_and_expression>
- */
-logical_xor_expression
-    logical_and_expression .and .loop logical_xor_expression_1;
-logical_xor_expression_1
-    caretcaret .and logical_and_expression .and .true .emit OP_LOGICALXOR;
-
-/*
- * <logical_or_expression> ::= <logical_xor_expression>
- *                           | <logical_or_expression> "||" <logical_xor_expression>
- */
-logical_or_expression
-    logical_xor_expression .and .loop logical_or_expression_1;
-logical_or_expression_1
-    barbar .and logical_xor_expression .and .true .emit OP_LOGICALOR;
-
-/*
- * <conditional_expression> ::= <logical_or_expression>
- *                            | <logical_or_expression> "?" <expression> ":" <conditional_expression>
- */
-conditional_expression
-    logical_or_expression .and .loop conditional_expression_1;
-conditional_expression_1
-    question .and expression .and colon .and conditional_expression .and .true .emit OP_SELECT;
-
-/*
- * <assignment_expression> ::= <conditional_expression>
- *                           | <unary_expression> <assignment_operator> <assignment_expression>
- *
- * <assignment_operator> ::= "="
- *                         | "*="
- *                         | "/="
- *                         | "+="
- *                         | "-="
- *                         | "%=" // reserved
- *                         | "<<=" // reserved
- *                         | ">>=" // reserved
- *                         | "&=" // reserved
- *                         | "^=" // reserved
- *                         | "|=" // reserved
- */
-assignment_expression
-    assignment_expression_1 .or assignment_expression_2 .or assignment_expression_3 .or
-    assignment_expression_4 .or assignment_expression_5/* .or assignment_expression_6 .or
-    assignment_expression_7 .or assignment_expression_8 .or assignment_expression_9 .or
-    assignment_expression_10 .or assignment_expression_11*/ .or conditional_expression;
-assignment_expression_1
-    unary_expression .and equals .and assignment_expression .and .true .emit OP_ASSIGN;
-assignment_expression_2
-    unary_expression .and starequals .and assignment_expression .and .true .emit OP_MULASSIGN;
-assignment_expression_3
-    unary_expression .and slashequals .and assignment_expression .and .true .emit OP_DIVASSIGN;
-assignment_expression_4
-    unary_expression .and plusequals .and assignment_expression .and .true .emit OP_ADDASSIGN;
-assignment_expression_5
-    unary_expression .and minusequals .and assignment_expression .and .true .emit OP_SUBASSIGN;
-/*assignment_expression_6
-    unary_expression .and percentequals .and assignment_expression .and .true .emit OP_MODASSIGN;*/
-/*assignment_expression_7
-    unary_expression .and lesslessequals .and assignment_expression .and .true .emit OP_LSHASSIGN;*/
-/*assignment_expression_8
-    unary_expression .and greatergreaterequals .and assignment_expression .and
-    .true .emit OP_RSHASSIGN;*/
-/*assignment_expression_9
-    unary_expression .and ampersandequals .and assignment_expression .and .true .emit OP_ANDASSIGN;*/
-/*assignment_expression_10
-    unary_expression .and caretequals .and assignment_expression .and .true .emit OP_XORASSIGN;*/
-/*assignment_expression_11
-    unary_expression .and barequals .and assignment_expression .and .true .emit OP_ORASSIGN;*/
-
-/*
- * <expression> ::= <assignment_expression>
- *                | <expression> "," <assignment_expression>
- */
-expression
-    assignment_expression .and .loop expression_1;
-expression_1
-    comma .and assignment_expression .and .true .emit OP_SEQUENCE;
-
-/*
- * <constant_expression> ::= <conditional_expression>
- */
-constant_expression
-    conditional_expression .and .true .emit OP_END;
-
-/*
- * <declaration> ::= <function_prototype> ";"
- *                 | <init_declarator_list> ";"
- */
-declaration
-    declaration_1 .or declaration_2;
-declaration_1
-    function_prototype .emit DECLARATION_FUNCTION_PROTOTYPE .and semicolon;
-declaration_2
-    init_declarator_list .emit DECLARATION_INIT_DECLARATOR_LIST .and semicolon;
-
-/*
- * <function_prototype> ::= <function_header> "void" ")"
- *                        | <function_declarator> ")"
- */
-function_prototype
-    function_prototype_1 .or function_prototype_2;
-function_prototype_1
-    function_header .and "void" .and rparen .error RPAREN_EXPECTED .emit PARAMETER_NONE;
-function_prototype_2
-    function_declarator .and rparen .error RPAREN_EXPECTED .emit PARAMETER_NONE;
-
-/*
- * <function_declarator> ::= <function_header>
- *                         | <function_header_with_parameters>
- */
-function_declarator
-    function_header_with_parameters .or function_header;
-
-/*
- * <function_header_with_parameters> ::= <function_header> <parameter_declaration>
- *                                     | <function_header_with_parameters> ","
- *                                       <parameter_declaration>
- */
-function_header_with_parameters
-    function_header .and parameter_declaration .and .loop function_header_with_parameters_1;
-function_header_with_parameters_1
-    comma .and parameter_declaration;
-
-/*
- * <function_header> ::= <fully_specified_type> <identifier> "("
- */
-function_header
-    function_header_nospace .or function_header_space;
-function_header_space
-    fully_specified_type_space .and space .and function_decl_identifier .and lparen;
-function_header_nospace
-    fully_specified_type_nospace .and function_decl_identifier .and lparen;
-
-/*
- * <function_decl_identifier> ::= "__constructor"
- *                              | <__operator>
- *                              | <identifier>
- *
- * note: this is an extension to the standard language specification.
- * normally slang disallows operator and constructor prototypes and definitions
- */
-function_decl_identifier
-    .if (parsing_builtin != 0) __operator .emit FUNCTION_OPERATOR .or
-    .if (parsing_builtin != 0) "__constructor" .emit FUNCTION_CONSTRUCTOR .or
-    identifier .emit FUNCTION_ORDINARY;
-
-/*
- * <__operator> ::= "__operator" <overriden_op>
- *
- * note: this is an extension to the standard language specification.
- * normally slang disallows operator prototypes and definitions
- */
-__operator
-    "__operator" .and overriden_operator .error INVALID_OPERATOR_OVERRIDE;
-
-/*
- * <overriden_op> ::= "="
- *                  | "+="
- *                  | "-="
- *                  | "*="
- *                  | "/="
- *                  | "%=" // reserved
- *                  | "<<=" // reserved
- *                  | ">>=" // reserved
- *                  | "&=" // reserved
- *                  | "^=" // reserved
- *                  | "|=" // reserved
- *                  | "^^"
- *                  | "|" // reserved
- *                  | "^" // reserved
- *                  | "&" // reserved
- *                  | "=="
- *                  | "!="
- *                  | "<"
- *                  | ">"
- *                  | "<="
- *                  | ">="
- *                  | "<<" // reserved
- *                  | ">>" // reserved
- *                  | "*"
- *                  | "/"
- *                  | "%" // reserved
- *                  | "++"
- *                  | "--"
- *                  | "+"
- *                  | "-"
- *                  | "~" // reserved
- *                  | "!"
- *
- * note: this is an extension to the standard language specification.
- * normally slang disallows operator prototypes and definitions
- */
-overriden_operator
-    plusplus .emit OPERATOR_INCREMENT .or
-    plusequals .emit OPERATOR_ADDASSIGN .or
-    plus .emit OPERATOR_PLUS .or
-    minusminus .emit OPERATOR_DECREMENT .or
-    minusequals .emit OPERATOR_SUBASSIGN .or
-    minus .emit OPERATOR_MINUS .or
-    bang .emit OPERATOR_NOT .or
-    starequals .emit OPERATOR_MULASSIGN .or
-    star .emit OPERATOR_MULTIPLY .or
-    slashequals .emit OPERATOR_DIVASSIGN .or
-    slash .emit OPERATOR_DIVIDE .or
-    lessequals .emit OPERATOR_LESSEQUAL .or
-    /*lesslessequals .emit OPERATOR_LSHASSIGN .or*/
-    /*lessless .emit OPERATOR_LSHIFT .or*/
-    less .emit OPERATOR_LESS .or
-    greaterequals .emit OPERATOR_GREATEREQUAL .or
-    /*greatergreaterequals .emit OPERATOR_RSHASSIGN .or*/
-    /*greatergreater .emit OPERATOR_RSHIFT .or*/
-    greater .emit OPERATOR_GREATER .or
-    /*percentequals .emit OPERATOR_MODASSIGN .or*/
-    /*percent .emit OPERATOR_MODULUS .or*/
-    /*ampersandequals .emit OPERATOR_ANDASSIGN */
-    /*ampersand .emit OPERATOR_BITAND .or*/
-    /*barequals .emit OPERATOR_ORASSIGN .or*/
-    /*bar .emit OPERATOR_BITOR .or*/
-    /*tilde .emit OPERATOR_COMPLEMENT .or*/
-    /*caretequals .emit OPERATOR_XORASSIGN .or*/
-    caretcaret .emit OPERATOR_LOGICALXOR /*.or
-    caret .emit OPERATOR_BITXOR*/;
-
-/*
- * <parameter_declarator> ::= <type_specifier> <identifier>
- *                          | <type_specifier> <identifier> "[" <constant_expression> "]"
- */
-parameter_declarator
-    parameter_declarator_nospace .or parameter_declarator_space;
-parameter_declarator_nospace
-    type_specifier_nospace .and identifier .and parameter_declarator_1;
-parameter_declarator_space
-    type_specifier_space .and space .and identifier .and parameter_declarator_1;
-parameter_declarator_1
-    parameter_declarator_2 .emit PARAMETER_ARRAY_PRESENT .or
-    .true .emit PARAMETER_ARRAY_NOT_PRESENT;
-parameter_declarator_2
-    lbracket .and constant_expression .and rbracket;
-
-/*
- * <parameter_declaration> ::= <type_qualifier> <parameter_qualifier>
- *                             <precision> <parameter_declarator>
- *                           | <type_qualifier> <parameter_qualifier>
- *                             <precision> <parameter_type_specifier>
- *                           | <type_qualifier> <parameter_qualifier>
- *                             <parameter_declarator>
- *                           | <type_qualifier> <parameter_qualifier>
- *                             <parameter_type_specifier>
- *                           | <parameter_qualifier> <precision>
- *                             <parameter_declarator>
- *                           | <parameter_qualifier> <precision>
- *                             <parameter_type_specifier>
- *                           | <parameter_qualifier> <parameter_declarator>
- *                           | <parameter_qualifier> <parameter_type_specifier>
- */
-parameter_declaration
-    parameter_declaration_1 .emit PARAMETER_NEXT;
-parameter_declaration_1
-    parameter_declaration_2 .or parameter_declaration_3;
-parameter_declaration_2
-    type_qualifier .and space .and parameter_qualifier .and parameter_declaration_4;
-parameter_declaration_3
-    parameter_qualifier .emit TYPE_QUALIFIER_NONE .and parameter_declaration_4;
-parameter_declaration_4
-    parameter_declaration_optprec .and parameter_declaration_rest;
-parameter_declaration_optprec
-    parameter_declaration_prec .or .true .emit PRECISION_DEFAULT;
-parameter_declaration_prec
-    precision .and space;
-parameter_declaration_rest
-    parameter_declarator .or parameter_type_specifier;
-
-/*
- * <parameter_qualifier> ::= "in"
- *                         | "out"
- *                         | "inout"
- *                         | ""
- */
-parameter_qualifier
-    parameter_qualifier_1 .or .true .emit PARAM_QUALIFIER_IN;
-parameter_qualifier_1
-    parameter_qualifier_2 .and space;
-parameter_qualifier_2
-    "in" .emit PARAM_QUALIFIER_IN .or
-    "out" .emit PARAM_QUALIFIER_OUT .or
-    "inout" .emit PARAM_QUALIFIER_INOUT;
-
-/*
- * <parameter_type_specifier> ::= <type_specifier>
- *                              | <type_specifier> "[" <constant_expression> "]"
- */
-parameter_type_specifier
-    parameter_type_specifier_1 .and .true .emit '\0' .and parameter_type_specifier_2;
-parameter_type_specifier_1
-    type_specifier_nospace .or type_specifier_space;
-parameter_type_specifier_2
-    parameter_type_specifier_3 .emit PARAMETER_ARRAY_PRESENT .or
-    .true .emit PARAMETER_ARRAY_NOT_PRESENT;
-parameter_type_specifier_3
-    lbracket .and constant_expression .and rbracket;
-
-/*
- * <init_declarator_list> ::= <single_declaration>
- *                          | <init_declarator_list> "," <identifier>
- *                          | <init_declarator_list> "," <identifier> "[" "]"
- *                          | <init_declarator_list> "," <identifier> "[" <constant_expression> "]"
- *                          | <init_declarator_list> "," <identifier> "=" <initializer>
- */
-init_declarator_list
-    single_declaration .and .loop init_declarator_list_1 .emit DECLARATOR_NEXT .and
-    .true .emit DECLARATOR_NONE;
-init_declarator_list_1
-    comma .and identifier .emit VARIABLE_IDENTIFIER .and init_declarator_list_2;
-init_declarator_list_2
-    init_declarator_list_3 .or init_declarator_list_4 .or .true .emit VARIABLE_NONE;
-init_declarator_list_3
-    equals .and initializer .emit VARIABLE_INITIALIZER;
-init_declarator_list_4
-    lbracket .and init_declarator_list_5 .and rbracket;
-init_declarator_list_5
-    constant_expression .emit VARIABLE_ARRAY_EXPLICIT .or .true .emit VARIABLE_ARRAY_UNKNOWN;
-
-/*
- * <single_declaration> ::= <fully_specified_type>
- *                        | <fully_specified_type> <identifier>
- *                        | <fully_specified_type> <identifier> "[" "]"
- *                        | <fully_specified_type> <identifier> "[" <constant_expression> "]"
- *                        | <fully_specified_type> <identifier> "=" <initializer>
- */
-single_declaration
-    single_declaration_nospace .or single_declaration_space;
-single_declaration_space
-    fully_specified_type_space .and single_declaration_space_1;
-single_declaration_nospace
-    fully_specified_type_nospace .and single_declaration_nospace_1;
-single_declaration_space_1
-    single_declaration_space_2 .emit VARIABLE_IDENTIFIER .or .true .emit VARIABLE_NONE;
-single_declaration_nospace_1
-    single_declaration_nospace_2 .emit VARIABLE_IDENTIFIER .or .true .emit VARIABLE_NONE;
-single_declaration_space_2
-    space .and identifier .and single_declaration_3;
-single_declaration_nospace_2
-    identifier .and single_declaration_3;
-single_declaration_3
-    single_declaration_4 .or single_declaration_5 .or .true .emit VARIABLE_NONE;
-single_declaration_4
-    equals .and initializer .emit VARIABLE_INITIALIZER;
-single_declaration_5
-    lbracket .and single_declaration_6 .and rbracket;
-single_declaration_6
-    constant_expression .emit VARIABLE_ARRAY_EXPLICIT .or .true .emit VARIABLE_ARRAY_UNKNOWN;
-
-/*
- * <fully_specified_type> ::= <opt_invariant> <opt_centroid> <opt_qualifer> <opt_precision> <type_specifier>
- *
- * Example: "invariant varying highp vec3"
- */
-fully_specified_type_space
-    fully_specified_type_optinvariant .and fully_specified_type_optcentroid .and fully_specified_type_optqual .and fully_specified_type_optprec .and type_specifier_space;
-fully_specified_type_nospace
-    fully_specified_type_optinvariant .and fully_specified_type_optcentroid .and fully_specified_type_optqual .and fully_specified_type_optprec .and type_specifier_nospace;
-fully_specified_type_optinvariant
-    fully_specified_type_invariant .or .true .emit TYPE_VARIANT;
-fully_specified_type_invariant
-    invariant_qualifier .and space;
-fully_specified_type_optcentroid
-    fully_specified_type_centroid .or .true .emit TYPE_CENTER;
-fully_specified_type_centroid
-    centroid_qualifier .and space;
-fully_specified_type_optqual
-    fully_specified_type_qual .or .true .emit TYPE_QUALIFIER_NONE;
-fully_specified_type_qual
-    type_qualifier .and space;
-fully_specified_type_optprec
-    fully_specified_type_prec .or .true .emit PRECISION_DEFAULT;
-fully_specified_type_prec
-    precision .and space;
-
-/*
- * <invariant_qualifier> ::= "invariant"
- */
-invariant_qualifier
-    "invariant" .emit TYPE_INVARIANT;
-
-centroid_qualifier
-    "centroid" .emit TYPE_CENTROID;
-
-
-/*
- * <type_qualifier> ::= "const"
- *                    | "attribute" // Vertex only.
- *                    | "varying"
- *                    | "uniform"
- *                    | "__fixed_output"
- *                    | "__fixed_input"
- *
- * note: this is an extension to the standard language specification,
- * normally slang disallows __fixed_output and __fixed_input type qualifiers
- */
-type_qualifier
-    "const" .emit TYPE_QUALIFIER_CONST .or
-    .if (shader_type == 2) "attribute" .emit TYPE_QUALIFIER_ATTRIBUTE .or
-    "varying" .emit TYPE_QUALIFIER_VARYING .or
-    "uniform" .emit TYPE_QUALIFIER_UNIFORM .or
-    .if (parsing_builtin != 0) "__fixed_output" .emit TYPE_QUALIFIER_FIXEDOUTPUT .or
-    .if (parsing_builtin != 0) "__fixed_input" .emit TYPE_QUALIFIER_FIXEDINPUT;
-
-/*
- * <type_specifier_nonarray> ::= "void"
- *                             | "float"
- *                             | "int"
- *                             | "bool"
- *                             | "vec2"
- *                             | "vec3"
- *                             | "vec4"
- *                             | "bvec2"
- *                             | "bvec3"
- *                             | "bvec4"
- *                             | "ivec2"
- *                             | "ivec3"
- *                             | "ivec4"
- *                             | "mat2"
- *                             | "mat3"
- *                             | "mat4"
- *                             | "mat2x3"
- *                             | "mat3x2"
- *                             | "mat2x4"
- *                             | "mat4x2"
- *                             | "mat3x4"
- *                             | "mat4x3"
- *                             | "sampler1D"
- *                             | "sampler2D"
- *                             | "sampler3D"
- *                             | "samplerCube"
- *                             | "sampler1DShadow"
- *                             | "sampler2DShadow"
- *                             | "sampler2DRect"
- *                             | "sampler2DRectShadow"
- *                             | <struct_specifier>
- *                             | <type_name>
- */
-type_specifier_nonarray_space
-    "void" .emit TYPE_SPECIFIER_VOID .or
-    "float" .emit TYPE_SPECIFIER_FLOAT .or
-    "int" .emit TYPE_SPECIFIER_INT .or
-    "bool" .emit TYPE_SPECIFIER_BOOL .or
-    "vec2" .emit TYPE_SPECIFIER_VEC2 .or
-    "vec3" .emit TYPE_SPECIFIER_VEC3 .or
-    "vec4" .emit TYPE_SPECIFIER_VEC4 .or
-    "bvec2" .emit TYPE_SPECIFIER_BVEC2 .or
-    "bvec3" .emit TYPE_SPECIFIER_BVEC3 .or
-    "bvec4" .emit TYPE_SPECIFIER_BVEC4 .or
-    "ivec2" .emit TYPE_SPECIFIER_IVEC2 .or
-    "ivec3" .emit TYPE_SPECIFIER_IVEC3 .or
-    "ivec4" .emit TYPE_SPECIFIER_IVEC4 .or
-    "mat2" .emit TYPE_SPECIFIER_MAT2 .or
-    "mat3" .emit TYPE_SPECIFIER_MAT3 .or
-    "mat4" .emit TYPE_SPECIFIER_MAT4 .or
-    "mat2x3" .emit TYPE_SPECIFIER_MAT23 .or
-    "mat3x2" .emit TYPE_SPECIFIER_MAT32 .or
-    "mat2x4" .emit TYPE_SPECIFIER_MAT24 .or
-    "mat4x2" .emit TYPE_SPECIFIER_MAT42 .or
-    "mat3x4" .emit TYPE_SPECIFIER_MAT34 .or
-    "mat4x3" .emit TYPE_SPECIFIER_MAT43 .or
-    "sampler1D" .emit TYPE_SPECIFIER_SAMPLER1D .or
-    "sampler2D" .emit TYPE_SPECIFIER_SAMPLER2D .or
-    "sampler3D" .emit TYPE_SPECIFIER_SAMPLER3D .or
-    "samplerCube" .emit TYPE_SPECIFIER_SAMPLERCUBE .or
-    "sampler1DShadow" .emit TYPE_SPECIFIER_SAMPLER1DSHADOW .or
-    "sampler2DShadow" .emit TYPE_SPECIFIER_SAMPLER2DSHADOW .or
-    "sampler2DRect" .emit TYPE_SPECIFIER_SAMPLER2DRECT .or
-    "sampler2DRectShadow" .emit TYPE_SPECIFIER_SAMPLER2DRECTSHADOW .or
-    type_name .emit TYPE_SPECIFIER_TYPENAME;
-type_specifier_nonarray_nospace
-    struct_specifier .emit TYPE_SPECIFIER_STRUCT;
-type_specifier_nonarray
-    type_specifier_nonarray_nospace .or type_specifier_nonarray_space;
-
-/*
- * <type_specifier> ::= <type_specifier_nonarray>
- *                    | <type_specifier_nonarray> "[" <constant_expression> "]"
- */
-type_specifier_space
-    type_specifier_nonarray_space .and .true .emit TYPE_SPECIFIER_NONARRAY;
-type_specifier_nospace
-    type_specifier_nospace_array .or type_specifier_nospace_1;
-type_specifier_nospace_1
-    type_specifier_nonarray_nospace .and .true .emit TYPE_SPECIFIER_NONARRAY;
-type_specifier_nospace_array
-    type_specifier_nonarray .and lbracket .emit TYPE_SPECIFIER_ARRAY .and constant_expression .and rbracket;
-
-/*
- * <struct_specifier> ::= "struct" <identifier> "{" <struct_declaration_list> "}"
- *                      | "struct" "{" <struct_declaration_list> "}"
- */
-struct_specifier
-    "struct" .and struct_specifier_1 .and optional_space .and lbrace .error LBRACE_EXPECTED .and
-    struct_declaration_list .and rbrace .emit FIELD_NONE;
-struct_specifier_1
-    struct_specifier_2 .or .true .emit '\0';
-struct_specifier_2
-    space .and identifier;
-
-/*
- * <struct_declaration_list> ::= <struct_declaration>
- *                             | <struct_declaration_list> <struct_declaration>
- */
-struct_declaration_list
-    struct_declaration .and .loop struct_declaration .emit FIELD_NEXT;
-
-/*
- * <struct_declaration> ::= <type_specifier> <struct_declarator_list> ";"
- */
-struct_declaration
-    struct_declaration_nospace .or struct_declaration_space;
-struct_declaration_space
-    type_specifier_space .and space .and struct_declarator_list .and semicolon .emit FIELD_NONE;
-struct_declaration_nospace
-    type_specifier_nospace .and struct_declarator_list .and semicolon .emit FIELD_NONE;
-
-/*
- * <struct_declarator_list> ::= <struct_declarator>
- *                            | <struct_declarator_list> "," <struct_declarator>
- */
-struct_declarator_list
-    struct_declarator .and .loop struct_declarator_list_1 .emit FIELD_NEXT;
-struct_declarator_list_1
-    comma .and struct_declarator;
-
-/*
- * <struct_declarator> ::= <identifier>
- *                       | <identifier> "[" <constant_expression> "]"
- */
-struct_declarator
-    identifier .and struct_declarator_1;
-struct_declarator_1
-    struct_declarator_2 .emit FIELD_ARRAY .or .true .emit FIELD_NONE;
-struct_declarator_2
-    lbracket .and constant_expression .and rbracket;
-
-/*
- * <initializer> ::= <assignment_expression>
- */
-initializer
-    assignment_expression .and .true .emit OP_END;
-
-/*
- * <declaration_statement> ::= <declaration>
- */
-declaration_statement
-    declaration;
-
-/*
- * <statement> ::= <compound_statement>
- *               | <simple_statement>
- */
-statement
-    compound_statement .or simple_statement;
-statement_space
-    compound_statement .or statement_space_1;
-statement_space_1
-    space .and simple_statement;
-
-/*
- * <simple_statement> ::= <__asm_statement>
- *                      | <selection_statement>
- *                      | <iteration_statement>
- *                      | <jump_statement>
- *                      | <expression_statement>
- *                      | <declaration_statement>
- *
- * note: this is an extension to the standard language specification.
- * normally slang disallows use of __asm statements
- */
-simple_statement
-    .if (parsing_builtin != 0) __asm_statement .emit OP_ASM .or
-    selection_statement .or
-    iteration_statement .or
-    precision_stmt .emit OP_PRECISION .or
-    jump_statement .or
-    expression_statement .emit OP_EXPRESSION .or
-    declaration_statement .emit OP_DECLARE;
-
-/*
- * <compound_statement> ::= "{" "}"
- *                        | "{" <statement_list> "}"
- */
-compound_statement
-    compound_statement_1 .emit OP_BLOCK_BEGIN_NEW_SCOPE .and .true .emit OP_END;
-compound_statement_1
-    compound_statement_2 .or compound_statement_3;
-compound_statement_2
-    lbrace .and rbrace;
-compound_statement_3
-    lbrace .and statement_list .and rbrace;
-
-/*
- * <compound_statement_no_new_scope> ::= "{" "}"
- *                                     | "{" <statement_list> "}"
- */
-compound_statement_no_new_scope
-    compound_statement_no_new_scope_1 .emit OP_BLOCK_BEGIN_NO_NEW_SCOPE .and .true .emit OP_END;
-compound_statement_no_new_scope_1
-    compound_statement_no_new_scope_2 .or compound_statement_no_new_scope_3;
-compound_statement_no_new_scope_2
-    lbrace .and rbrace;
-compound_statement_no_new_scope_3
-    lbrace .and statement_list .and rbrace;
-
-
-/*
- * <statement_list> ::= <statement>
- *                    | <statement_list> <statement>
- */
-statement_list
-    statement .and .loop statement;
-
-/*
- * <expression_statement> ::= ";"
- *                          | <expression> ";"
- */
-expression_statement
-    expression_statement_1 .or expression_statement_2;
-expression_statement_1
-    semicolon .emit OP_PUSH_VOID .emit OP_END;
-expression_statement_2
-    expression .and semicolon .emit OP_END;
-
-/*
- * <selection_statement> ::= "if" "(" <expression> ")" <selection_rest_statement>
- */
-selection_statement
-    "if" .emit OP_IF .and lparen .error LPAREN_EXPECTED .and expression .and
-    rparen .error RPAREN_EXPECTED .emit OP_END .and selection_rest_statement;
-
-/*
- * <selection_rest_statement> ::= <statement> "else" <statement>
- *                              | <statement>
- */
-selection_rest_statement
-    statement .and selection_rest_statement_1;
-selection_rest_statement_1
-    selection_rest_statement_2 .or .true .emit OP_EXPRESSION .emit OP_PUSH_VOID .emit OP_END;
-selection_rest_statement_2
-    "else" .and optional_space .and statement;
-
-/*
- * <condition> ::= <expression>
- *               | <fully_specified_type> <identifier> "=" <initializer>
- *
- * note: if <condition_1> is executed, the emit format must
- * match <declaration> emit format
- */
-condition
-    condition_1 .emit OP_DECLARE .emit DECLARATION_INIT_DECLARATOR_LIST .or
-    condition_3 .emit OP_EXPRESSION;
-condition_1
-    condition_1_nospace .or condition_1_space;
-condition_1_nospace
-    fully_specified_type_nospace .and condition_2;
-condition_1_space
-    fully_specified_type_space .and space .and condition_2;
-condition_2
-    identifier .emit VARIABLE_IDENTIFIER .and equals .emit VARIABLE_INITIALIZER .and
-    initializer .and .true .emit DECLARATOR_NONE;
-condition_3
-    expression .and .true .emit OP_END;
-
-/*
- * <iteration_statement> ::= "while" "(" <condition> ")" <statement>
- *                         | "do" <statement> "while" "(" <expression> ")" ";"
- *                         | "for" "(" <for_init_statement> <for_rest_statement> ")" <statement>
- */
-iteration_statement
-    iteration_statement_1 .or iteration_statement_2 .or iteration_statement_3;
-iteration_statement_1
-    "while" .emit OP_WHILE .and lparen .error LPAREN_EXPECTED .and condition .and
-    rparen .error RPAREN_EXPECTED .and statement;
-iteration_statement_2
-    "do" .emit OP_DO .and statement_space .and "while" .and lparen .error LPAREN_EXPECTED .and
-    expression .and rparen .error RPAREN_EXPECTED .emit OP_END .and semicolon;
-iteration_statement_3
-    "for" .emit OP_FOR .and lparen .error LPAREN_EXPECTED .and for_init_statement .and
-    for_rest_statement .and rparen .error RPAREN_EXPECTED .and statement;
-
-/*
- * <for_init_statement> ::= <expression_statement>
- *                        | <declaration_statement>
- */
-for_init_statement
-    expression_statement .emit OP_EXPRESSION .or declaration_statement .emit OP_DECLARE;
-
-/*
- * <conditionopt> ::= <condition>
- *                  | ""
- *
- * note: <conditionopt> is used only by "for" statement.
- * if <condition> is ommitted, parser simulates default behaviour,
- * that is simulates "true" expression
- */
-conditionopt
-    condition .or
-    .true .emit OP_EXPRESSION .emit OP_PUSH_BOOL .emit 2 .emit '1' .emit '\0' .emit OP_END;
-
-/*
- * <for_rest_statement> ::= <conditionopt> ";"
- *                        | <conditionopt> ";" <expression>
- */
-for_rest_statement
-    conditionopt .and semicolon .and for_rest_statement_1;
-for_rest_statement_1
-    for_rest_statement_2 .or .true .emit OP_PUSH_VOID .emit OP_END;
-for_rest_statement_2
-    expression .and .true .emit OP_END;
-
-/*
- * <jump_statement> ::= "continue" ";"
- *                    | "break" ";"
- *                    | "return" ";"
- *                    | "return" <expression> ";"
- *                    | "discard" ";" // Fragment shader only.
- */
-jump_statement
-    jump_statement_1 .or jump_statement_2 .or jump_statement_3 .or jump_statement_4 .or
-    .if (shader_type == 1) jump_statement_5;
-jump_statement_1
-    "continue" .and semicolon .emit OP_CONTINUE;
-jump_statement_2
-    "break" .and semicolon .emit OP_BREAK;
-jump_statement_3
-    "return" .emit OP_RETURN .and optional_space .and expression .and semicolon .emit OP_END;
-jump_statement_4
-    "return" .emit OP_RETURN .and semicolon .emit OP_PUSH_VOID .emit OP_END;
-jump_statement_5
-    "discard" .and semicolon .emit OP_DISCARD;
-
-/*
- * <__asm_statement> ::= "__asm" <identifier> <asm_arguments> ";"
- *
- * note: this is an extension to the standard language specification.
- * normally slang disallows __asm statements
- */
-__asm_statement
-    "__asm" .and space .and identifier .and space .and asm_arguments .and semicolon .emit OP_END;
-
-/*
- * <asm_arguments> ::= <asm_argument>
- *                   | <asm_arguments> "," <asm_argument>
- *
- * note: this is an extension to the standard language specification.
- * normally slang disallows __asm statements
- */
-asm_arguments
-    asm_argument .and .true .emit OP_END .and .loop asm_arguments_1;
-asm_arguments_1
-    comma .and asm_argument .and .true .emit OP_END;
-
-/*
- * <asm_argument> ::= <variable_identifier>
- *                  | <floatconstant>
- *
- * note: this is an extension to the standard language specification.
- * normally slang disallows __asm statements
- */
-asm_argument
-    var_with_field .or
-    variable_identifier .or
-    floatconstant;
-
-var_with_field
-    variable_identifier .and dot .and field_selection .emit OP_FIELD;
-
-
-/*
- * <translation_unit> ::= <external_declaration>
- *                      | <translation_unit> <external_declaration>
- */
-translation_unit
-    optional_space .emit REVISION .and external_declaration .error INVALID_EXTERNAL_DECLARATION .and
-    .loop external_declaration .and optional_space .and
-    '\0' .error INVALID_EXTERNAL_DECLARATION .emit EXTERNAL_NULL;
-
-
-/*
- * <external_declaration> ::= <function_definition>
- *                          | <declaration>
- */
-external_declaration
-    precision_stmt .emit DEFAULT_PRECISION .or
-    function_definition .emit EXTERNAL_FUNCTION_DEFINITION .or
-    invariant_stmt .emit INVARIANT_STMT .or
-    declaration .emit EXTERNAL_DECLARATION;
-
-
-/*
- * <precision_stmt> ::= "precision" <precision> <prectype>
- */
-precision_stmt
-    "precision" .and space .and precision .error INVALID_PRECISION .and space .and prectype .error INVALID_PRECISION_TYPE .and semicolon;
-
-/*
- * <precision> ::= "lowp"
- *               | "mediump"
- *               | "highp"
- */
-precision
-    "lowp" .emit PRECISION_LOW .or
-    "mediump" .emit PRECISION_MEDIUM .or
-    "highp" .emit PRECISION_HIGH;
-
-/*
- * <prectype> ::= "int"
- *              | "float"
- *              | "a sampler type"
- */
-prectype
-    "int" .emit TYPE_SPECIFIER_INT .or
-    "float" .emit TYPE_SPECIFIER_FLOAT .or
-    "sampler1D" .emit TYPE_SPECIFIER_SAMPLER1D .or
-    "sampler2D" .emit TYPE_SPECIFIER_SAMPLER2D .or
-    "sampler3D" .emit TYPE_SPECIFIER_SAMPLER3D .or
-    "samplerCube" .emit TYPE_SPECIFIER_SAMPLERCUBE .or
-    "sampler1DShadow" .emit TYPE_SPECIFIER_SAMPLER1DSHADOW .or
-    "sampler2DShadow" .emit TYPE_SPECIFIER_SAMPLER2DSHADOW .or
-    "sampler2DRect" .emit TYPE_SPECIFIER_SAMPLER2DRECT .or
-    "sampler2DRectShadow" .emit TYPE_SPECIFIER_SAMPLER2DRECTSHADOW;
-
-
-/*
- * <invariant_stmt> ::= "invariant" identifier;
- */
-invariant_stmt
-    "invariant" .and space .and identifier .and semicolon;
-
-
-/*
- * <function_definition> :: <function_prototype> <compound_statement_no_new_scope>
- */
-function_definition
-    function_prototype .and compound_statement_no_new_scope;
-
-
-
-/*
- * helper rules, not part of the official language syntax
- */
-
-digit_oct
-    '0'-'7';
-
-digit_dec
-    '0'-'9';
-
-digit_hex
-    '0'-'9' .or 'A'-'F' .or 'a'-'f';
-
-id_character_first
-    'a'-'z' .or 'A'-'Z' .or '_';
-
-id_character_next
-    id_character_first .or digit_dec;
-
-identifier
-    id_character_first .emit * .and .loop id_character_next .emit * .and .true .emit '\0';
-
-float
-    float_1 .or float_2 .or float_3;
-float_1
-    float_fractional_constant .and float_optional_exponent_part .and optional_f_suffix;
-float_2
-    float_digit_sequence .and .true .emit '\0' .and float_exponent_part .and optional_f_suffix;
-float_3
-    float_digit_sequence .and .true .emit '\0' .and 'f' .emit '\0';
-
-float_fractional_constant
-    float_fractional_constant_1 .or float_fractional_constant_2 .or float_fractional_constant_3;
-float_fractional_constant_1
-    float_digit_sequence .and '.' .and float_digit_sequence;
-float_fractional_constant_2
-    float_digit_sequence .and '.' .and .true .emit '\0';
-float_fractional_constant_3
-    '.' .emit '\0' .and float_digit_sequence;
-
-float_optional_exponent_part
-    float_exponent_part .or .true .emit '\0';
-
-float_digit_sequence
-    digit_dec .emit * .and .loop digit_dec .emit * .and .true .emit '\0';
-
-float_exponent_part
-    float_exponent_part_1 .or float_exponent_part_2;
-float_exponent_part_1
-    'e' .and float_optional_sign .and float_digit_sequence;
-float_exponent_part_2
-    'E' .and float_optional_sign .and float_digit_sequence;
-
-float_optional_sign
-    float_sign .or .true;
-
-float_sign
-    '+' .or '-' .emit '-';
-
-optional_f_suffix
-    'f' .or .true;
-
-
-integer
-    integer_hex .or integer_oct .or integer_dec;
-
-integer_hex
-    '0' .and integer_hex_1 .emit 0x10 .and digit_hex .emit * .and .loop digit_hex .emit * .and
-    .true .emit '\0';
-integer_hex_1
-    'x' .or 'X';
-
-integer_oct
-    '0' .emit 8 .emit * .and .loop digit_oct .emit * .and .true .emit '\0';
-
-integer_dec
-    digit_dec .emit 10 .emit * .and .loop digit_dec .emit * .and .true .emit '\0';
-
-boolean
-    "true" .emit 2 .emit '1' .emit '\0' .or
-    "false" .emit 2 .emit '0' .emit '\0';
-
-type_name
-    identifier;
-
-field_selection
-    identifier;
-
-floatconstant
-    float .emit OP_PUSH_FLOAT;
-
-intconstant
-    integer .emit OP_PUSH_INT;
-
-boolconstant
-    boolean .emit OP_PUSH_BOOL;
-
-optional_space
-    .loop single_space;
-
-space
-    single_space .and .loop single_space;
-
-single_space
-    white_char .or c_style_comment_block .or cpp_style_comment_block;
-
-white_char
-    ' ' .or '\t' .or new_line .or '\v' .or '\f';
-
-new_line
-    cr_lf .or lf_cr .or '\n' .or '\r';
-
-cr_lf
-    '\r' .and '\n';
-
-lf_cr
-    '\n' .and '\r';
-
-c_style_comment_block
-    '/' .and '*' .and c_style_comment_rest;
-
-c_style_comment_rest
-    .loop c_style_comment_char_no_star .and c_style_comment_rest_1;
-c_style_comment_rest_1
-    c_style_comment_end .or c_style_comment_rest_2;
-c_style_comment_rest_2
-    '*' .and c_style_comment_rest;
-
-c_style_comment_char_no_star
-    '\x2B'-'\xFF' .or '\x01'-'\x29';
-
-c_style_comment_end
-    '*' .and '/';
-
-cpp_style_comment_block
-    '/' .and '/' .and cpp_style_comment_block_1;
-cpp_style_comment_block_1
-    cpp_style_comment_block_2 .or cpp_style_comment_block_3;
-cpp_style_comment_block_2
-    .loop cpp_style_comment_char .and new_line;
-cpp_style_comment_block_3
-    .loop cpp_style_comment_char;
-
-cpp_style_comment_char
-    '\x0E'-'\xFF' .or '\x01'-'\x09' .or '\x0B'-'\x0C';
-
-/* lexical rules */
-
-/*ampersand
-    optional_space .and '&' .and optional_space;*/
-
-ampersandampersand
-    optional_space .and '&' .and '&' .and optional_space;
-
-/*ampersandequals
-    optional_space .and '&' .and '=' .and optional_space;*/
-
-/*bar
-    optional_space .and '|' .and optional_space;*/
-
-barbar
-    optional_space .and '|' .and '|' .and optional_space;
-
-/*barequals
-    optional_space .and '|' .and '=' .and optional_space;*/
-
-bang
-    optional_space .and '!' .and optional_space;
-
-bangequals
-    optional_space .and '!' .and '=' .and optional_space;
-
-/*caret
-    optional_space .and '^' .and optional_space;*/
-
-caretcaret
-    optional_space .and '^' .and '^' .and optional_space;
-
-/*caretequals
-    optional_space .and '^' .and '=' .and optional_space;*/
-
-colon
-    optional_space .and ':' .and optional_space;
-
-comma
-    optional_space .and ',' .and optional_space;
-
-dot
-    optional_space .and '.' .and optional_space;
-
-equals
-    optional_space .and '=' .and optional_space;
-
-equalsequals
-    optional_space .and '=' .and '=' .and optional_space;
-
-greater
-    optional_space .and '>' .and optional_space;
-
-greaterequals
-    optional_space .and '>' .and '=' .and optional_space;
-
-/*greatergreater
-    optional_space .and '>' .and '>' .and optional_space;*/
-
-/*greatergreaterequals
-    optional_space .and '>' .and '>' .and '=' .and optional_space;*/
-
-lbrace
-    optional_space .and '{' .and optional_space;
-
-lbracket
-    optional_space .and '[' .and optional_space;
-
-less
-    optional_space .and '<' .and optional_space;
-
-lessequals
-    optional_space .and '<' .and '=' .and optional_space;
-
-/*lessless
-    optional_space .and '<' .and '<' .and optional_space;*/
-
-/*lesslessequals
-    optional_space .and '<' .and '<' .and '=' .and optional_space;*/
-
-lparen
-    optional_space .and '(' .and optional_space;
-
-minus
-    optional_space .and '-' .and optional_space;
-
-minusequals
-    optional_space .and '-' .and '=' .and optional_space;
-
-minusminus
-    optional_space .and '-' .and '-' .and optional_space;
-
-/*percent
-    optional_space .and '%' .and optional_space;*/
-
-/*percentequals
-    optional_space .and '%' .and '=' .and optional_space;*/
-
-plus
-    optional_space .and '+' .and optional_space;
-
-plusequals
-    optional_space .and '+' .and '=' .and optional_space;
-
-plusplus
-    optional_space .and '+' .and '+' .and optional_space;
-
-question
-    optional_space .and '?' .and optional_space;
-
-rbrace
-    optional_space .and '}' .and optional_space;
-
-rbracket
-    optional_space .and ']' .and optional_space;
-
-rparen
-    optional_space .and ')' .and optional_space;
-
-semicolon
-    optional_space .and ';' .and optional_space;
-
-slash
-    optional_space .and '/' .and optional_space;
-
-slashequals
-    optional_space .and '/' .and '=' .and optional_space;
-
-star
-    optional_space .and '*' .and optional_space;
-
-starequals
-    optional_space .and '*' .and '=' .and optional_space;
-
-/*tilde
-    optional_space .and '~' .and optional_space;*/
-
-/* string rules - these are used internally by the parser when parsing quoted strings */
-
-.string string_lexer;
-
-string_lexer
-    lex_first_identifier_character .and .loop lex_next_identifier_character;
-
-lex_first_identifier_character
-    'a'-'z' .or 'A'-'Z' .or '_';
-
-lex_next_identifier_character
-    'a'-'z' .or 'A'-'Z' .or '0'-'9' .or '_';
-
-/* error rules - these are used by error messages */
-
-err_token
-    '~' .or '`' .or '!' .or '@' .or '#' .or '$' .or '%' .or '^' .or '&' .or '*' .or '(' .or ')' .or
-    '-' .or '+' .or '=' .or '|' .or '\\' .or '[' .or ']' .or '{' .or '}' .or ':' .or ';' .or '"' .or
-    '\'' .or '<' .or ',' .or '>' .or '.' .or '/' .or '?' .or err_identifier;
-
-err_identifier
-    id_character_first .and .loop id_character_next;
-
diff --git a/src/mesa/shader/slang/library/slang_shader_syn.h b/src/mesa/shader/slang/library/slang_shader_syn.h
deleted file mode 100644 (file)
index 6a38297..0000000
+++ /dev/null
@@ -1,866 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE .syn FILE */
-
-".syntax translation_unit;\n"
-".emtcode REVISION 5\n"
-".emtcode EXTERNAL_NULL 0\n"
-".emtcode EXTERNAL_FUNCTION_DEFINITION 1\n"
-".emtcode EXTERNAL_DECLARATION 2\n"
-".emtcode DEFAULT_PRECISION 3\n"
-".emtcode INVARIANT_STMT 4\n"
-".emtcode PRECISION_DEFAULT 0\n"
-".emtcode PRECISION_LOW 1\n"
-".emtcode PRECISION_MEDIUM 2\n"
-".emtcode PRECISION_HIGH 3\n"
-".emtcode DECLARATION_FUNCTION_PROTOTYPE 1\n"
-".emtcode DECLARATION_INIT_DECLARATOR_LIST 2\n"
-".emtcode FUNCTION_ORDINARY 0\n"
-".emtcode FUNCTION_CONSTRUCTOR 1\n"
-".emtcode FUNCTION_OPERATOR 2\n"
-".emtcode FUNCTION_CALL_NONARRAY 0\n"
-".emtcode FUNCTION_CALL_ARRAY 1\n"
-".emtcode OPERATOR_ADDASSIGN 1\n"
-".emtcode OPERATOR_SUBASSIGN 2\n"
-".emtcode OPERATOR_MULASSIGN 3\n"
-".emtcode OPERATOR_DIVASSIGN 4\n"
-".emtcode OPERATOR_LOGICALXOR 11\n"
-".emtcode OPERATOR_LESS 15\n"
-".emtcode OPERATOR_GREATER 16\n"
-".emtcode OPERATOR_LESSEQUAL 17\n"
-".emtcode OPERATOR_GREATEREQUAL 18\n"
-".emtcode OPERATOR_MULTIPLY 21\n"
-".emtcode OPERATOR_DIVIDE 22\n"
-".emtcode OPERATOR_INCREMENT 24\n"
-".emtcode OPERATOR_DECREMENT 25\n"
-".emtcode OPERATOR_PLUS 26\n"
-".emtcode OPERATOR_MINUS 27\n"
-".emtcode OPERATOR_NOT 29\n"
-".emtcode DECLARATOR_NONE 0\n"
-".emtcode DECLARATOR_NEXT 1\n"
-".emtcode VARIABLE_NONE 0\n"
-".emtcode VARIABLE_IDENTIFIER 1\n"
-".emtcode VARIABLE_INITIALIZER 2\n"
-".emtcode VARIABLE_ARRAY_EXPLICIT 3\n"
-".emtcode VARIABLE_ARRAY_UNKNOWN 4\n"
-".emtcode TYPE_QUALIFIER_NONE 0\n"
-".emtcode TYPE_QUALIFIER_CONST 1\n"
-".emtcode TYPE_QUALIFIER_ATTRIBUTE 2\n"
-".emtcode TYPE_QUALIFIER_VARYING 3\n"
-".emtcode TYPE_QUALIFIER_UNIFORM 4\n"
-".emtcode TYPE_QUALIFIER_FIXEDOUTPUT 5\n"
-".emtcode TYPE_QUALIFIER_FIXEDINPUT 6\n"
-".emtcode TYPE_VARIANT 90\n"
-".emtcode TYPE_INVARIANT 91\n"
-".emtcode TYPE_CENTER 95\n"
-".emtcode TYPE_CENTROID 96\n"
-".emtcode TYPE_SPECIFIER_VOID 0\n"
-".emtcode TYPE_SPECIFIER_BOOL 1\n"
-".emtcode TYPE_SPECIFIER_BVEC2 2\n"
-".emtcode TYPE_SPECIFIER_BVEC3 3\n"
-".emtcode TYPE_SPECIFIER_BVEC4 4\n"
-".emtcode TYPE_SPECIFIER_INT 5\n"
-".emtcode TYPE_SPECIFIER_IVEC2 6\n"
-".emtcode TYPE_SPECIFIER_IVEC3 7\n"
-".emtcode TYPE_SPECIFIER_IVEC4 8\n"
-".emtcode TYPE_SPECIFIER_FLOAT 9\n"
-".emtcode TYPE_SPECIFIER_VEC2 10\n"
-".emtcode TYPE_SPECIFIER_VEC3 11\n"
-".emtcode TYPE_SPECIFIER_VEC4 12\n"
-".emtcode TYPE_SPECIFIER_MAT2 13\n"
-".emtcode TYPE_SPECIFIER_MAT3 14\n"
-".emtcode TYPE_SPECIFIER_MAT4 15\n"
-".emtcode TYPE_SPECIFIER_SAMPLER1D 16\n"
-".emtcode TYPE_SPECIFIER_SAMPLER2D 17\n"
-".emtcode TYPE_SPECIFIER_SAMPLER3D 18\n"
-".emtcode TYPE_SPECIFIER_SAMPLERCUBE 19\n"
-".emtcode TYPE_SPECIFIER_SAMPLER1DSHADOW 20\n"
-".emtcode TYPE_SPECIFIER_SAMPLER2DSHADOW 21\n"
-".emtcode TYPE_SPECIFIER_SAMPLER2DRECT 22\n"
-".emtcode TYPE_SPECIFIER_SAMPLER2DRECTSHADOW 23\n"
-".emtcode TYPE_SPECIFIER_STRUCT 24\n"
-".emtcode TYPE_SPECIFIER_TYPENAME 25\n"
-".emtcode TYPE_SPECIFIER_MAT23 26\n"
-".emtcode TYPE_SPECIFIER_MAT32 27\n"
-".emtcode TYPE_SPECIFIER_MAT24 28\n"
-".emtcode TYPE_SPECIFIER_MAT42 29\n"
-".emtcode TYPE_SPECIFIER_MAT34 30\n"
-".emtcode TYPE_SPECIFIER_MAT43 31\n"
-".emtcode TYPE_SPECIFIER_NONARRAY 0\n"
-".emtcode TYPE_SPECIFIER_ARRAY 1\n"
-".emtcode FIELD_NONE 0\n"
-".emtcode FIELD_NEXT 1\n"
-".emtcode FIELD_ARRAY 2\n"
-".emtcode OP_END 0\n"
-".emtcode OP_BLOCK_BEGIN_NO_NEW_SCOPE 1\n"
-".emtcode OP_BLOCK_BEGIN_NEW_SCOPE 2\n"
-".emtcode OP_DECLARE 3\n"
-".emtcode OP_ASM 4\n"
-".emtcode OP_BREAK 5\n"
-".emtcode OP_CONTINUE 6\n"
-".emtcode OP_DISCARD 7\n"
-".emtcode OP_RETURN 8\n"
-".emtcode OP_EXPRESSION 9\n"
-".emtcode OP_IF 10\n"
-".emtcode OP_WHILE 11\n"
-".emtcode OP_DO 12\n"
-".emtcode OP_FOR 13\n"
-".emtcode OP_PUSH_VOID 14\n"
-".emtcode OP_PUSH_BOOL 15\n"
-".emtcode OP_PUSH_INT 16\n"
-".emtcode OP_PUSH_FLOAT 17\n"
-".emtcode OP_PUSH_IDENTIFIER 18\n"
-".emtcode OP_SEQUENCE 19\n"
-".emtcode OP_ASSIGN 20\n"
-".emtcode OP_ADDASSIGN 21\n"
-".emtcode OP_SUBASSIGN 22\n"
-".emtcode OP_MULASSIGN 23\n"
-".emtcode OP_DIVASSIGN 24\n"
-".emtcode OP_SELECT 31\n"
-".emtcode OP_LOGICALOR 32\n"
-".emtcode OP_LOGICALXOR 33\n"
-".emtcode OP_LOGICALAND 34\n"
-".emtcode OP_EQUAL 38\n"
-".emtcode OP_NOTEQUAL 39\n"
-".emtcode OP_LESS 40\n"
-".emtcode OP_GREATER 41\n"
-".emtcode OP_LESSEQUAL 42\n"
-".emtcode OP_GREATEREQUAL 43\n"
-".emtcode OP_ADD 46\n"
-".emtcode OP_SUBTRACT 47\n"
-".emtcode OP_MULTIPLY 48\n"
-".emtcode OP_DIVIDE 49\n"
-".emtcode OP_PREINCREMENT 51\n"
-".emtcode OP_PREDECREMENT 52\n"
-".emtcode OP_PLUS 53\n"
-".emtcode OP_MINUS 54\n"
-".emtcode OP_NOT 56\n"
-".emtcode OP_SUBSCRIPT 57\n"
-".emtcode OP_CALL 58\n"
-".emtcode OP_FIELD 59\n"
-".emtcode OP_POSTINCREMENT 60\n"
-".emtcode OP_POSTDECREMENT 61\n"
-".emtcode OP_PRECISION 62\n"
-".emtcode OP_METHOD 63\n"
-".emtcode PARAM_QUALIFIER_IN 0\n"
-".emtcode PARAM_QUALIFIER_OUT 1\n"
-".emtcode PARAM_QUALIFIER_INOUT 2\n"
-".emtcode PARAMETER_NONE 0\n"
-".emtcode PARAMETER_NEXT 1\n"
-".emtcode PARAMETER_ARRAY_NOT_PRESENT 0\n"
-".emtcode PARAMETER_ARRAY_PRESENT 1\n"
-".errtext INVALID_EXTERNAL_DECLARATION \"2001: Syntax error.\"\n"
-".errtext INVALID_OPERATOR_OVERRIDE \"2002: Invalid operator override.\"\n"
-".errtext LBRACE_EXPECTED \"2003: '{' expected but '$err_token$' found.\"\n"
-".errtext LPAREN_EXPECTED \"2004: '(' expected but '$err_token$' found.\"\n"
-".errtext RPAREN_EXPECTED \"2005: ')' expected but '$err_token$' found.\"\n"
-".errtext INVALID_PRECISION \"2006: Invalid precision specifier '$err_token$'.\"\n"
-".errtext INVALID_PRECISION_TYPE \"2007: Invalid precision type '$err_token$'.\"\n"
-".regbyte parsing_builtin 0\n"
-".regbyte shader_type 0\n"
-"variable_identifier\n"
-" identifier .emit OP_PUSH_IDENTIFIER;\n"
-"primary_expression\n"
-" floatconstant .or boolconstant .or intconstant .or variable_identifier .or primary_expression_1;\n"
-"primary_expression_1\n"
-" lparen .and expression .and rparen;\n"
-"postfix_expression\n"
-" postfix_expression_1 .and .loop postfix_expression_2;\n"
-"postfix_expression_1\n"
-" function_call .or primary_expression;\n"
-"postfix_expression_2\n"
-" postfix_expression_3 .or postfix_expression_4 .or\n"
-" plusplus .emit OP_POSTINCREMENT .or\n"
-" minusminus .emit OP_POSTDECREMENT;\n"
-"postfix_expression_3\n"
-" lbracket .and integer_expression .and rbracket .emit OP_SUBSCRIPT;\n"
-"postfix_expression_4\n"
-" dot .and field_selection .emit OP_FIELD;\n"
-"integer_expression\n"
-" expression;\n"
-"function_call\n"
-" function_call_or_method;\n"
-"function_call_or_method\n"
-" regular_function_call .or method_call;\n"
-"method_call\n"
-" identifier .emit OP_METHOD .and dot .and function_call_generic .and .true .emit OP_END;\n"
-"regular_function_call\n"
-" function_call_generic .emit OP_CALL .and .true .emit OP_END;\n"
-"function_call_generic\n"
-" function_call_generic_1 .or function_call_generic_2;\n"
-"function_call_generic_1\n"
-" function_call_header_with_parameters .and rparen .error RPAREN_EXPECTED;\n"
-"function_call_generic_2\n"
-" function_call_header_no_parameters .and rparen .error RPAREN_EXPECTED;\n"
-"function_call_header_no_parameters\n"
-" function_call_header .and function_call_header_no_parameters_1;\n"
-"function_call_header_no_parameters_1\n"
-" \"void\" .or .true;\n"
-"function_call_header_with_parameters\n"
-" function_call_header .and assignment_expression .and .true .emit OP_END .and\n"
-" .loop function_call_header_with_parameters_1;\n"
-"function_call_header_with_parameters_1\n"
-" comma .and assignment_expression .and .true .emit OP_END;\n"
-"function_call_header\n"
-" function_identifier .and lparen;\n"
-"function_identifier\n"
-" identifier .and function_identifier_opt_array;\n"
-"function_identifier_opt_array\n"
-" function_identifier_array .emit FUNCTION_CALL_ARRAY .or\n"
-" .true .emit FUNCTION_CALL_NONARRAY;\n"
-"function_identifier_array\n"
-" lbracket .and constant_expression .and rbracket;\n"
-"unary_expression\n"
-" postfix_expression .or unary_expression_1 .or unary_expression_2 .or unary_expression_3 .or\n"
-" unary_expression_4 .or unary_expression_5;\n"
-"unary_expression_1\n"
-" plusplus .and unary_expression .and .true .emit OP_PREINCREMENT;\n"
-"unary_expression_2\n"
-" minusminus .and unary_expression .and .true .emit OP_PREDECREMENT;\n"
-"unary_expression_3\n"
-" plus .and unary_expression .and .true .emit OP_PLUS;\n"
-"unary_expression_4\n"
-" minus .and unary_expression .and .true .emit OP_MINUS;\n"
-"unary_expression_5\n"
-" bang .and unary_expression .and .true .emit OP_NOT;\n"
-"multiplicative_expression\n"
-" unary_expression .and .loop multiplicative_expression_1;\n"
-"multiplicative_expression_1\n"
-" multiplicative_expression_2 .or multiplicative_expression_3;\n"
-"multiplicative_expression_2\n"
-" star .and unary_expression .and .true .emit OP_MULTIPLY;\n"
-"multiplicative_expression_3\n"
-" slash .and unary_expression .and .true .emit OP_DIVIDE;\n"
-"additive_expression\n"
-" multiplicative_expression .and .loop additive_expression_1;\n"
-"additive_expression_1\n"
-" additive_expression_2 .or additive_expression_3;\n"
-"additive_expression_2\n"
-" plus .and multiplicative_expression .and .true .emit OP_ADD;\n"
-"additive_expression_3\n"
-" minus .and multiplicative_expression .and .true .emit OP_SUBTRACT;\n"
-"shift_expression\n"
-" additive_expression;\n"
-"relational_expression\n"
-" shift_expression .and .loop relational_expression_1;\n"
-"relational_expression_1\n"
-" relational_expression_2 .or relational_expression_3 .or relational_expression_4 .or\n"
-" relational_expression_5;\n"
-"relational_expression_2\n"
-" lessequals .and shift_expression .and .true .emit OP_LESSEQUAL;\n"
-"relational_expression_3\n"
-" greaterequals .and shift_expression .and .true .emit OP_GREATEREQUAL;\n"
-"relational_expression_4\n"
-" less .and shift_expression .and .true .emit OP_LESS;\n"
-"relational_expression_5\n"
-" greater .and shift_expression .and .true .emit OP_GREATER;\n"
-"equality_expression\n"
-" relational_expression .and .loop equality_expression_1;\n"
-"equality_expression_1\n"
-" equality_expression_2 .or equality_expression_3;\n"
-"equality_expression_2\n"
-" equalsequals .and relational_expression .and .true .emit OP_EQUAL;\n"
-"equality_expression_3\n"
-" bangequals .and relational_expression .and .true .emit OP_NOTEQUAL;\n"
-"and_expression\n"
-" equality_expression;\n"
-"exclusive_or_expression\n"
-" and_expression;\n"
-"inclusive_or_expression\n"
-" exclusive_or_expression;\n"
-"logical_and_expression\n"
-" inclusive_or_expression .and .loop logical_and_expression_1;\n"
-"logical_and_expression_1\n"
-" ampersandampersand .and inclusive_or_expression .and .true .emit OP_LOGICALAND;\n"
-"logical_xor_expression\n"
-" logical_and_expression .and .loop logical_xor_expression_1;\n"
-"logical_xor_expression_1\n"
-" caretcaret .and logical_and_expression .and .true .emit OP_LOGICALXOR;\n"
-"logical_or_expression\n"
-" logical_xor_expression .and .loop logical_or_expression_1;\n"
-"logical_or_expression_1\n"
-" barbar .and logical_xor_expression .and .true .emit OP_LOGICALOR;\n"
-"conditional_expression\n"
-" logical_or_expression .and .loop conditional_expression_1;\n"
-"conditional_expression_1\n"
-" question .and expression .and colon .and conditional_expression .and .true .emit OP_SELECT;\n"
-"assignment_expression\n"
-" assignment_expression_1 .or assignment_expression_2 .or assignment_expression_3 .or\n"
-" assignment_expression_4 .or assignment_expression_5 .or conditional_expression;\n"
-"assignment_expression_1\n"
-" unary_expression .and equals .and assignment_expression .and .true .emit OP_ASSIGN;\n"
-"assignment_expression_2\n"
-" unary_expression .and starequals .and assignment_expression .and .true .emit OP_MULASSIGN;\n"
-"assignment_expression_3\n"
-" unary_expression .and slashequals .and assignment_expression .and .true .emit OP_DIVASSIGN;\n"
-"assignment_expression_4\n"
-" unary_expression .and plusequals .and assignment_expression .and .true .emit OP_ADDASSIGN;\n"
-"assignment_expression_5\n"
-" unary_expression .and minusequals .and assignment_expression .and .true .emit OP_SUBASSIGN;\n"
-"expression\n"
-" assignment_expression .and .loop expression_1;\n"
-"expression_1\n"
-" comma .and assignment_expression .and .true .emit OP_SEQUENCE;\n"
-"constant_expression\n"
-" conditional_expression .and .true .emit OP_END;\n"
-"declaration\n"
-" declaration_1 .or declaration_2;\n"
-"declaration_1\n"
-" function_prototype .emit DECLARATION_FUNCTION_PROTOTYPE .and semicolon;\n"
-"declaration_2\n"
-" init_declarator_list .emit DECLARATION_INIT_DECLARATOR_LIST .and semicolon;\n"
-"function_prototype\n"
-" function_prototype_1 .or function_prototype_2;\n"
-"function_prototype_1\n"
-" function_header .and \"void\" .and rparen .error RPAREN_EXPECTED .emit PARAMETER_NONE;\n"
-"function_prototype_2\n"
-" function_declarator .and rparen .error RPAREN_EXPECTED .emit PARAMETER_NONE;\n"
-"function_declarator\n"
-" function_header_with_parameters .or function_header;\n"
-"function_header_with_parameters\n"
-" function_header .and parameter_declaration .and .loop function_header_with_parameters_1;\n"
-"function_header_with_parameters_1\n"
-" comma .and parameter_declaration;\n"
-"function_header\n"
-" function_header_nospace .or function_header_space;\n"
-"function_header_space\n"
-" fully_specified_type_space .and space .and function_decl_identifier .and lparen;\n"
-"function_header_nospace\n"
-" fully_specified_type_nospace .and function_decl_identifier .and lparen;\n"
-"function_decl_identifier\n"
-" .if (parsing_builtin != 0) __operator .emit FUNCTION_OPERATOR .or\n"
-" .if (parsing_builtin != 0) \"__constructor\" .emit FUNCTION_CONSTRUCTOR .or\n"
-" identifier .emit FUNCTION_ORDINARY;\n"
-"__operator\n"
-" \"__operator\" .and overriden_operator .error INVALID_OPERATOR_OVERRIDE;\n"
-"overriden_operator\n"
-" plusplus .emit OPERATOR_INCREMENT .or\n"
-" plusequals .emit OPERATOR_ADDASSIGN .or\n"
-" plus .emit OPERATOR_PLUS .or\n"
-" minusminus .emit OPERATOR_DECREMENT .or\n"
-" minusequals .emit OPERATOR_SUBASSIGN .or\n"
-" minus .emit OPERATOR_MINUS .or\n"
-" bang .emit OPERATOR_NOT .or\n"
-" starequals .emit OPERATOR_MULASSIGN .or\n"
-" star .emit OPERATOR_MULTIPLY .or\n"
-" slashequals .emit OPERATOR_DIVASSIGN .or\n"
-" slash .emit OPERATOR_DIVIDE .or\n"
-" lessequals .emit OPERATOR_LESSEQUAL .or\n"
-" \n"
-" \n"
-" less .emit OPERATOR_LESS .or\n"
-" greaterequals .emit OPERATOR_GREATEREQUAL .or\n"
-" \n"
-" \n"
-" greater .emit OPERATOR_GREATER .or\n"
-" \n"
-" \n"
-" \n"
-" \n"
-" \n"
-" \n"
-" \n"
-" \n"
-" caretcaret .emit OPERATOR_LOGICALXOR ;\n"
-"parameter_declarator\n"
-" parameter_declarator_nospace .or parameter_declarator_space;\n"
-"parameter_declarator_nospace\n"
-" type_specifier_nospace .and identifier .and parameter_declarator_1;\n"
-"parameter_declarator_space\n"
-" type_specifier_space .and space .and identifier .and parameter_declarator_1;\n"
-"parameter_declarator_1\n"
-" parameter_declarator_2 .emit PARAMETER_ARRAY_PRESENT .or\n"
-" .true .emit PARAMETER_ARRAY_NOT_PRESENT;\n"
-"parameter_declarator_2\n"
-" lbracket .and constant_expression .and rbracket;\n"
-"parameter_declaration\n"
-" parameter_declaration_1 .emit PARAMETER_NEXT;\n"
-"parameter_declaration_1\n"
-" parameter_declaration_2 .or parameter_declaration_3;\n"
-"parameter_declaration_2\n"
-" type_qualifier .and space .and parameter_qualifier .and parameter_declaration_4;\n"
-"parameter_declaration_3\n"
-" parameter_qualifier .emit TYPE_QUALIFIER_NONE .and parameter_declaration_4;\n"
-"parameter_declaration_4\n"
-" parameter_declaration_optprec .and parameter_declaration_rest;\n"
-"parameter_declaration_optprec\n"
-" parameter_declaration_prec .or .true .emit PRECISION_DEFAULT;\n"
-"parameter_declaration_prec\n"
-" precision .and space;\n"
-"parameter_declaration_rest\n"
-" parameter_declarator .or parameter_type_specifier;\n"
-"parameter_qualifier\n"
-" parameter_qualifier_1 .or .true .emit PARAM_QUALIFIER_IN;\n"
-"parameter_qualifier_1\n"
-" parameter_qualifier_2 .and space;\n"
-"parameter_qualifier_2\n"
-" \"in\" .emit PARAM_QUALIFIER_IN .or\n"
-" \"out\" .emit PARAM_QUALIFIER_OUT .or\n"
-" \"inout\" .emit PARAM_QUALIFIER_INOUT;\n"
-"parameter_type_specifier\n"
-" parameter_type_specifier_1 .and .true .emit '\\0' .and parameter_type_specifier_2;\n"
-"parameter_type_specifier_1\n"
-" type_specifier_nospace .or type_specifier_space;\n"
-"parameter_type_specifier_2\n"
-" parameter_type_specifier_3 .emit PARAMETER_ARRAY_PRESENT .or\n"
-" .true .emit PARAMETER_ARRAY_NOT_PRESENT;\n"
-"parameter_type_specifier_3\n"
-" lbracket .and constant_expression .and rbracket;\n"
-"init_declarator_list\n"
-" single_declaration .and .loop init_declarator_list_1 .emit DECLARATOR_NEXT .and\n"
-" .true .emit DECLARATOR_NONE;\n"
-"init_declarator_list_1\n"
-" comma .and identifier .emit VARIABLE_IDENTIFIER .and init_declarator_list_2;\n"
-"init_declarator_list_2\n"
-" init_declarator_list_3 .or init_declarator_list_4 .or .true .emit VARIABLE_NONE;\n"
-"init_declarator_list_3\n"
-" equals .and initializer .emit VARIABLE_INITIALIZER;\n"
-"init_declarator_list_4\n"
-" lbracket .and init_declarator_list_5 .and rbracket;\n"
-"init_declarator_list_5\n"
-" constant_expression .emit VARIABLE_ARRAY_EXPLICIT .or .true .emit VARIABLE_ARRAY_UNKNOWN;\n"
-"single_declaration\n"
-" single_declaration_nospace .or single_declaration_space;\n"
-"single_declaration_space\n"
-" fully_specified_type_space .and single_declaration_space_1;\n"
-"single_declaration_nospace\n"
-" fully_specified_type_nospace .and single_declaration_nospace_1;\n"
-"single_declaration_space_1\n"
-" single_declaration_space_2 .emit VARIABLE_IDENTIFIER .or .true .emit VARIABLE_NONE;\n"
-"single_declaration_nospace_1\n"
-" single_declaration_nospace_2 .emit VARIABLE_IDENTIFIER .or .true .emit VARIABLE_NONE;\n"
-"single_declaration_space_2\n"
-" space .and identifier .and single_declaration_3;\n"
-"single_declaration_nospace_2\n"
-" identifier .and single_declaration_3;\n"
-"single_declaration_3\n"
-" single_declaration_4 .or single_declaration_5 .or .true .emit VARIABLE_NONE;\n"
-"single_declaration_4\n"
-" equals .and initializer .emit VARIABLE_INITIALIZER;\n"
-"single_declaration_5\n"
-" lbracket .and single_declaration_6 .and rbracket;\n"
-"single_declaration_6\n"
-" constant_expression .emit VARIABLE_ARRAY_EXPLICIT .or .true .emit VARIABLE_ARRAY_UNKNOWN;\n"
-"fully_specified_type_space\n"
-" fully_specified_type_optinvariant .and fully_specified_type_optcentroid .and fully_specified_type_optqual .and fully_specified_type_optprec .and type_specifier_space;\n"
-"fully_specified_type_nospace\n"
-" fully_specified_type_optinvariant .and fully_specified_type_optcentroid .and fully_specified_type_optqual .and fully_specified_type_optprec .and type_specifier_nospace;\n"
-"fully_specified_type_optinvariant\n"
-" fully_specified_type_invariant .or .true .emit TYPE_VARIANT;\n"
-"fully_specified_type_invariant\n"
-" invariant_qualifier .and space;\n"
-"fully_specified_type_optcentroid\n"
-" fully_specified_type_centroid .or .true .emit TYPE_CENTER;\n"
-"fully_specified_type_centroid\n"
-" centroid_qualifier .and space;\n"
-"fully_specified_type_optqual\n"
-" fully_specified_type_qual .or .true .emit TYPE_QUALIFIER_NONE;\n"
-"fully_specified_type_qual\n"
-" type_qualifier .and space;\n"
-"fully_specified_type_optprec\n"
-" fully_specified_type_prec .or .true .emit PRECISION_DEFAULT;\n"
-"fully_specified_type_prec\n"
-" precision .and space;\n"
-"invariant_qualifier\n"
-" \"invariant\" .emit TYPE_INVARIANT;\n"
-"centroid_qualifier\n"
-" \"centroid\" .emit TYPE_CENTROID;\n"
-"type_qualifier\n"
-" \"const\" .emit TYPE_QUALIFIER_CONST .or\n"
-" .if (shader_type == 2) \"attribute\" .emit TYPE_QUALIFIER_ATTRIBUTE .or\n"
-" \"varying\" .emit TYPE_QUALIFIER_VARYING .or\n"
-" \"uniform\" .emit TYPE_QUALIFIER_UNIFORM .or\n"
-" .if (parsing_builtin != 0) \"__fixed_output\" .emit TYPE_QUALIFIER_FIXEDOUTPUT .or\n"
-" .if (parsing_builtin != 0) \"__fixed_input\" .emit TYPE_QUALIFIER_FIXEDINPUT;\n"
-"type_specifier_nonarray_space\n"
-" \"void\" .emit TYPE_SPECIFIER_VOID .or\n"
-" \"float\" .emit TYPE_SPECIFIER_FLOAT .or\n"
-" \"int\" .emit TYPE_SPECIFIER_INT .or\n"
-" \"bool\" .emit TYPE_SPECIFIER_BOOL .or\n"
-" \"vec2\" .emit TYPE_SPECIFIER_VEC2 .or\n"
-" \"vec3\" .emit TYPE_SPECIFIER_VEC3 .or\n"
-" \"vec4\" .emit TYPE_SPECIFIER_VEC4 .or\n"
-" \"bvec2\" .emit TYPE_SPECIFIER_BVEC2 .or\n"
-" \"bvec3\" .emit TYPE_SPECIFIER_BVEC3 .or\n"
-" \"bvec4\" .emit TYPE_SPECIFIER_BVEC4 .or\n"
-" \"ivec2\" .emit TYPE_SPECIFIER_IVEC2 .or\n"
-" \"ivec3\" .emit TYPE_SPECIFIER_IVEC3 .or\n"
-" \"ivec4\" .emit TYPE_SPECIFIER_IVEC4 .or\n"
-" \"mat2\" .emit TYPE_SPECIFIER_MAT2 .or\n"
-" \"mat3\" .emit TYPE_SPECIFIER_MAT3 .or\n"
-" \"mat4\" .emit TYPE_SPECIFIER_MAT4 .or\n"
-" \"mat2x3\" .emit TYPE_SPECIFIER_MAT23 .or\n"
-" \"mat3x2\" .emit TYPE_SPECIFIER_MAT32 .or\n"
-" \"mat2x4\" .emit TYPE_SPECIFIER_MAT24 .or\n"
-" \"mat4x2\" .emit TYPE_SPECIFIER_MAT42 .or\n"
-" \"mat3x4\" .emit TYPE_SPECIFIER_MAT34 .or\n"
-" \"mat4x3\" .emit TYPE_SPECIFIER_MAT43 .or\n"
-" \"sampler1D\" .emit TYPE_SPECIFIER_SAMPLER1D .or\n"
-" \"sampler2D\" .emit TYPE_SPECIFIER_SAMPLER2D .or\n"
-" \"sampler3D\" .emit TYPE_SPECIFIER_SAMPLER3D .or\n"
-" \"samplerCube\" .emit TYPE_SPECIFIER_SAMPLERCUBE .or\n"
-" \"sampler1DShadow\" .emit TYPE_SPECIFIER_SAMPLER1DSHADOW .or\n"
-" \"sampler2DShadow\" .emit TYPE_SPECIFIER_SAMPLER2DSHADOW .or\n"
-" \"sampler2DRect\" .emit TYPE_SPECIFIER_SAMPLER2DRECT .or\n"
-" \"sampler2DRectShadow\" .emit TYPE_SPECIFIER_SAMPLER2DRECTSHADOW .or\n"
-" type_name .emit TYPE_SPECIFIER_TYPENAME;\n"
-"type_specifier_nonarray_nospace\n"
-" struct_specifier .emit TYPE_SPECIFIER_STRUCT;\n"
-"type_specifier_nonarray\n"
-" type_specifier_nonarray_nospace .or type_specifier_nonarray_space;\n"
-"type_specifier_space\n"
-" type_specifier_nonarray_space .and .true .emit TYPE_SPECIFIER_NONARRAY;\n"
-"type_specifier_nospace\n"
-" type_specifier_nospace_array .or type_specifier_nospace_1;\n"
-"type_specifier_nospace_1\n"
-" type_specifier_nonarray_nospace .and .true .emit TYPE_SPECIFIER_NONARRAY;\n"
-"type_specifier_nospace_array\n"
-" type_specifier_nonarray .and lbracket .emit TYPE_SPECIFIER_ARRAY .and constant_expression .and rbracket;\n"
-"struct_specifier\n"
-" \"struct\" .and struct_specifier_1 .and optional_space .and lbrace .error LBRACE_EXPECTED .and\n"
-" struct_declaration_list .and rbrace .emit FIELD_NONE;\n"
-"struct_specifier_1\n"
-" struct_specifier_2 .or .true .emit '\\0';\n"
-"struct_specifier_2\n"
-" space .and identifier;\n"
-"struct_declaration_list\n"
-" struct_declaration .and .loop struct_declaration .emit FIELD_NEXT;\n"
-"struct_declaration\n"
-" struct_declaration_nospace .or struct_declaration_space;\n"
-"struct_declaration_space\n"
-" type_specifier_space .and space .and struct_declarator_list .and semicolon .emit FIELD_NONE;\n"
-"struct_declaration_nospace\n"
-" type_specifier_nospace .and struct_declarator_list .and semicolon .emit FIELD_NONE;\n"
-"struct_declarator_list\n"
-" struct_declarator .and .loop struct_declarator_list_1 .emit FIELD_NEXT;\n"
-"struct_declarator_list_1\n"
-" comma .and struct_declarator;\n"
-"struct_declarator\n"
-" identifier .and struct_declarator_1;\n"
-"struct_declarator_1\n"
-" struct_declarator_2 .emit FIELD_ARRAY .or .true .emit FIELD_NONE;\n"
-"struct_declarator_2\n"
-" lbracket .and constant_expression .and rbracket;\n"
-"initializer\n"
-" assignment_expression .and .true .emit OP_END;\n"
-"declaration_statement\n"
-" declaration;\n"
-"statement\n"
-" compound_statement .or simple_statement;\n"
-"statement_space\n"
-" compound_statement .or statement_space_1;\n"
-"statement_space_1\n"
-" space .and simple_statement;\n"
-"simple_statement\n"
-" .if (parsing_builtin != 0) __asm_statement .emit OP_ASM .or\n"
-" selection_statement .or\n"
-" iteration_statement .or\n"
-" precision_stmt .emit OP_PRECISION .or\n"
-" jump_statement .or\n"
-" expression_statement .emit OP_EXPRESSION .or\n"
-" declaration_statement .emit OP_DECLARE;\n"
-"compound_statement\n"
-" compound_statement_1 .emit OP_BLOCK_BEGIN_NEW_SCOPE .and .true .emit OP_END;\n"
-"compound_statement_1\n"
-" compound_statement_2 .or compound_statement_3;\n"
-"compound_statement_2\n"
-" lbrace .and rbrace;\n"
-"compound_statement_3\n"
-" lbrace .and statement_list .and rbrace;\n"
-"compound_statement_no_new_scope\n"
-" compound_statement_no_new_scope_1 .emit OP_BLOCK_BEGIN_NO_NEW_SCOPE .and .true .emit OP_END;\n"
-"compound_statement_no_new_scope_1\n"
-" compound_statement_no_new_scope_2 .or compound_statement_no_new_scope_3;\n"
-"compound_statement_no_new_scope_2\n"
-" lbrace .and rbrace;\n"
-"compound_statement_no_new_scope_3\n"
-" lbrace .and statement_list .and rbrace;\n"
-"statement_list\n"
-" statement .and .loop statement;\n"
-"expression_statement\n"
-" expression_statement_1 .or expression_statement_2;\n"
-"expression_statement_1\n"
-" semicolon .emit OP_PUSH_VOID .emit OP_END;\n"
-"expression_statement_2\n"
-" expression .and semicolon .emit OP_END;\n"
-"selection_statement\n"
-" \"if\" .emit OP_IF .and lparen .error LPAREN_EXPECTED .and expression .and\n"
-" rparen .error RPAREN_EXPECTED .emit OP_END .and selection_rest_statement;\n"
-"selection_rest_statement\n"
-" statement .and selection_rest_statement_1;\n"
-"selection_rest_statement_1\n"
-" selection_rest_statement_2 .or .true .emit OP_EXPRESSION .emit OP_PUSH_VOID .emit OP_END;\n"
-"selection_rest_statement_2\n"
-" \"else\" .and optional_space .and statement;\n"
-"condition\n"
-" condition_1 .emit OP_DECLARE .emit DECLARATION_INIT_DECLARATOR_LIST .or\n"
-" condition_3 .emit OP_EXPRESSION;\n"
-"condition_1\n"
-" condition_1_nospace .or condition_1_space;\n"
-"condition_1_nospace\n"
-" fully_specified_type_nospace .and condition_2;\n"
-"condition_1_space\n"
-" fully_specified_type_space .and space .and condition_2;\n"
-"condition_2\n"
-" identifier .emit VARIABLE_IDENTIFIER .and equals .emit VARIABLE_INITIALIZER .and\n"
-" initializer .and .true .emit DECLARATOR_NONE;\n"
-"condition_3\n"
-" expression .and .true .emit OP_END;\n"
-"iteration_statement\n"
-" iteration_statement_1 .or iteration_statement_2 .or iteration_statement_3;\n"
-"iteration_statement_1\n"
-" \"while\" .emit OP_WHILE .and lparen .error LPAREN_EXPECTED .and condition .and\n"
-" rparen .error RPAREN_EXPECTED .and statement;\n"
-"iteration_statement_2\n"
-" \"do\" .emit OP_DO .and statement_space .and \"while\" .and lparen .error LPAREN_EXPECTED .and\n"
-" expression .and rparen .error RPAREN_EXPECTED .emit OP_END .and semicolon;\n"
-"iteration_statement_3\n"
-" \"for\" .emit OP_FOR .and lparen .error LPAREN_EXPECTED .and for_init_statement .and\n"
-" for_rest_statement .and rparen .error RPAREN_EXPECTED .and statement;\n"
-"for_init_statement\n"
-" expression_statement .emit OP_EXPRESSION .or declaration_statement .emit OP_DECLARE;\n"
-"conditionopt\n"
-" condition .or\n"
-" .true .emit OP_EXPRESSION .emit OP_PUSH_BOOL .emit 2 .emit '1' .emit '\\0' .emit OP_END;\n"
-"for_rest_statement\n"
-" conditionopt .and semicolon .and for_rest_statement_1;\n"
-"for_rest_statement_1\n"
-" for_rest_statement_2 .or .true .emit OP_PUSH_VOID .emit OP_END;\n"
-"for_rest_statement_2\n"
-" expression .and .true .emit OP_END;\n"
-"jump_statement\n"
-" jump_statement_1 .or jump_statement_2 .or jump_statement_3 .or jump_statement_4 .or\n"
-" .if (shader_type == 1) jump_statement_5;\n"
-"jump_statement_1\n"
-" \"continue\" .and semicolon .emit OP_CONTINUE;\n"
-"jump_statement_2\n"
-" \"break\" .and semicolon .emit OP_BREAK;\n"
-"jump_statement_3\n"
-" \"return\" .emit OP_RETURN .and optional_space .and expression .and semicolon .emit OP_END;\n"
-"jump_statement_4\n"
-" \"return\" .emit OP_RETURN .and semicolon .emit OP_PUSH_VOID .emit OP_END;\n"
-"jump_statement_5\n"
-" \"discard\" .and semicolon .emit OP_DISCARD;\n"
-"__asm_statement\n"
-" \"__asm\" .and space .and identifier .and space .and asm_arguments .and semicolon .emit OP_END;\n"
-"asm_arguments\n"
-" asm_argument .and .true .emit OP_END .and .loop asm_arguments_1;\n"
-"asm_arguments_1\n"
-" comma .and asm_argument .and .true .emit OP_END;\n"
-"asm_argument\n"
-" var_with_field .or\n"
-" variable_identifier .or\n"
-" floatconstant;\n"
-"var_with_field\n"
-" variable_identifier .and dot .and field_selection .emit OP_FIELD;\n"
-"translation_unit\n"
-" optional_space .emit REVISION .and external_declaration .error INVALID_EXTERNAL_DECLARATION .and\n"
-" .loop external_declaration .and optional_space .and\n"
-" '\\0' .error INVALID_EXTERNAL_DECLARATION .emit EXTERNAL_NULL;\n"
-"external_declaration\n"
-" precision_stmt .emit DEFAULT_PRECISION .or\n"
-" function_definition .emit EXTERNAL_FUNCTION_DEFINITION .or\n"
-" invariant_stmt .emit INVARIANT_STMT .or\n"
-" declaration .emit EXTERNAL_DECLARATION;\n"
-"precision_stmt\n"
-" \"precision\" .and space .and precision .error INVALID_PRECISION .and space .and prectype .error INVALID_PRECISION_TYPE .and semicolon;\n"
-"precision\n"
-" \"lowp\" .emit PRECISION_LOW .or\n"
-" \"mediump\" .emit PRECISION_MEDIUM .or\n"
-" \"highp\" .emit PRECISION_HIGH;\n"
-"prectype\n"
-" \"int\" .emit TYPE_SPECIFIER_INT .or\n"
-" \"float\" .emit TYPE_SPECIFIER_FLOAT .or\n"
-" \"sampler1D\" .emit TYPE_SPECIFIER_SAMPLER1D .or\n"
-" \"sampler2D\" .emit TYPE_SPECIFIER_SAMPLER2D .or\n"
-" \"sampler3D\" .emit TYPE_SPECIFIER_SAMPLER3D .or\n"
-" \"samplerCube\" .emit TYPE_SPECIFIER_SAMPLERCUBE .or\n"
-" \"sampler1DShadow\" .emit TYPE_SPECIFIER_SAMPLER1DSHADOW .or\n"
-" \"sampler2DShadow\" .emit TYPE_SPECIFIER_SAMPLER2DSHADOW .or\n"
-" \"sampler2DRect\" .emit TYPE_SPECIFIER_SAMPLER2DRECT .or\n"
-" \"sampler2DRectShadow\" .emit TYPE_SPECIFIER_SAMPLER2DRECTSHADOW;\n"
-"invariant_stmt\n"
-" \"invariant\" .and space .and identifier .and semicolon;\n"
-"function_definition\n"
-" function_prototype .and compound_statement_no_new_scope;\n"
-"digit_oct\n"
-" '0'-'7';\n"
-"digit_dec\n"
-" '0'-'9';\n"
-"digit_hex\n"
-" '0'-'9' .or 'A'-'F' .or 'a'-'f';\n"
-"id_character_first\n"
-" 'a'-'z' .or 'A'-'Z' .or '_';\n"
-"id_character_next\n"
-" id_character_first .or digit_dec;\n"
-"identifier\n"
-" id_character_first .emit * .and .loop id_character_next .emit * .and .true .emit '\\0';\n"
-"float\n"
-" float_1 .or float_2 .or float_3;\n"
-"float_1\n"
-" float_fractional_constant .and float_optional_exponent_part .and optional_f_suffix;\n"
-"float_2\n"
-" float_digit_sequence .and .true .emit '\\0' .and float_exponent_part .and optional_f_suffix;\n"
-"float_3\n"
-" float_digit_sequence .and .true .emit '\\0' .and 'f' .emit '\\0';\n"
-"float_fractional_constant\n"
-" float_fractional_constant_1 .or float_fractional_constant_2 .or float_fractional_constant_3;\n"
-"float_fractional_constant_1\n"
-" float_digit_sequence .and '.' .and float_digit_sequence;\n"
-"float_fractional_constant_2\n"
-" float_digit_sequence .and '.' .and .true .emit '\\0';\n"
-"float_fractional_constant_3\n"
-" '.' .emit '\\0' .and float_digit_sequence;\n"
-"float_optional_exponent_part\n"
-" float_exponent_part .or .true .emit '\\0';\n"
-"float_digit_sequence\n"
-" digit_dec .emit * .and .loop digit_dec .emit * .and .true .emit '\\0';\n"
-"float_exponent_part\n"
-" float_exponent_part_1 .or float_exponent_part_2;\n"
-"float_exponent_part_1\n"
-" 'e' .and float_optional_sign .and float_digit_sequence;\n"
-"float_exponent_part_2\n"
-" 'E' .and float_optional_sign .and float_digit_sequence;\n"
-"float_optional_sign\n"
-" float_sign .or .true;\n"
-"float_sign\n"
-" '+' .or '-' .emit '-';\n"
-"optional_f_suffix\n"
-" 'f' .or .true;\n"
-"integer\n"
-" integer_hex .or integer_oct .or integer_dec;\n"
-"integer_hex\n"
-" '0' .and integer_hex_1 .emit 0x10 .and digit_hex .emit * .and .loop digit_hex .emit * .and\n"
-" .true .emit '\\0';\n"
-"integer_hex_1\n"
-" 'x' .or 'X';\n"
-"integer_oct\n"
-" '0' .emit 8 .emit * .and .loop digit_oct .emit * .and .true .emit '\\0';\n"
-"integer_dec\n"
-" digit_dec .emit 10 .emit * .and .loop digit_dec .emit * .and .true .emit '\\0';\n"
-"boolean\n"
-" \"true\" .emit 2 .emit '1' .emit '\\0' .or\n"
-" \"false\" .emit 2 .emit '0' .emit '\\0';\n"
-"type_name\n"
-" identifier;\n"
-"field_selection\n"
-" identifier;\n"
-"floatconstant\n"
-" float .emit OP_PUSH_FLOAT;\n"
-"intconstant\n"
-" integer .emit OP_PUSH_INT;\n"
-"boolconstant\n"
-" boolean .emit OP_PUSH_BOOL;\n"
-"optional_space\n"
-" .loop single_space;\n"
-"space\n"
-" single_space .and .loop single_space;\n"
-"single_space\n"
-" white_char .or c_style_comment_block .or cpp_style_comment_block;\n"
-"white_char\n"
-" ' ' .or '\\t' .or new_line .or '\\v' .or '\\f';\n"
-"new_line\n"
-" cr_lf .or lf_cr .or '\\n' .or '\\r';\n"
-"cr_lf\n"
-" '\\r' .and '\\n';\n"
-"lf_cr\n"
-" '\\n' .and '\\r';\n"
-"c_style_comment_block\n"
-" '/' .and '*' .and c_style_comment_rest;\n"
-"c_style_comment_rest\n"
-" .loop c_style_comment_char_no_star .and c_style_comment_rest_1;\n"
-"c_style_comment_rest_1\n"
-" c_style_comment_end .or c_style_comment_rest_2;\n"
-"c_style_comment_rest_2\n"
-" '*' .and c_style_comment_rest;\n"
-"c_style_comment_char_no_star\n"
-" '\\x2B'-'\\xFF' .or '\\x01'-'\\x29';\n"
-"c_style_comment_end\n"
-" '*' .and '/';\n"
-"cpp_style_comment_block\n"
-" '/' .and '/' .and cpp_style_comment_block_1;\n"
-"cpp_style_comment_block_1\n"
-" cpp_style_comment_block_2 .or cpp_style_comment_block_3;\n"
-"cpp_style_comment_block_2\n"
-" .loop cpp_style_comment_char .and new_line;\n"
-"cpp_style_comment_block_3\n"
-" .loop cpp_style_comment_char;\n"
-"cpp_style_comment_char\n"
-" '\\x0E'-'\\xFF' .or '\\x01'-'\\x09' .or '\\x0B'-'\\x0C';\n"
-"ampersandampersand\n"
-" optional_space .and '&' .and '&' .and optional_space;\n"
-"barbar\n"
-" optional_space .and '|' .and '|' .and optional_space;\n"
-"bang\n"
-" optional_space .and '!' .and optional_space;\n"
-"bangequals\n"
-" optional_space .and '!' .and '=' .and optional_space;\n"
-"caretcaret\n"
-" optional_space .and '^' .and '^' .and optional_space;\n"
-"colon\n"
-" optional_space .and ':' .and optional_space;\n"
-"comma\n"
-" optional_space .and ',' .and optional_space;\n"
-"dot\n"
-" optional_space .and '.' .and optional_space;\n"
-"equals\n"
-" optional_space .and '=' .and optional_space;\n"
-"equalsequals\n"
-" optional_space .and '=' .and '=' .and optional_space;\n"
-"greater\n"
-" optional_space .and '>' .and optional_space;\n"
-"greaterequals\n"
-" optional_space .and '>' .and '=' .and optional_space;\n"
-"lbrace\n"
-" optional_space .and '{' .and optional_space;\n"
-"lbracket\n"
-" optional_space .and '[' .and optional_space;\n"
-"less\n"
-" optional_space .and '<' .and optional_space;\n"
-"lessequals\n"
-" optional_space .and '<' .and '=' .and optional_space;\n"
-"lparen\n"
-" optional_space .and '(' .and optional_space;\n"
-"minus\n"
-" optional_space .and '-' .and optional_space;\n"
-"minusequals\n"
-" optional_space .and '-' .and '=' .and optional_space;\n"
-"minusminus\n"
-" optional_space .and '-' .and '-' .and optional_space;\n"
-"plus\n"
-" optional_space .and '+' .and optional_space;\n"
-"plusequals\n"
-" optional_space .and '+' .and '=' .and optional_space;\n"
-"plusplus\n"
-" optional_space .and '+' .and '+' .and optional_space;\n"
-"question\n"
-" optional_space .and '?' .and optional_space;\n"
-"rbrace\n"
-" optional_space .and '}' .and optional_space;\n"
-"rbracket\n"
-" optional_space .and ']' .and optional_space;\n"
-"rparen\n"
-" optional_space .and ')' .and optional_space;\n"
-"semicolon\n"
-" optional_space .and ';' .and optional_space;\n"
-"slash\n"
-" optional_space .and '/' .and optional_space;\n"
-"slashequals\n"
-" optional_space .and '/' .and '=' .and optional_space;\n"
-"star\n"
-" optional_space .and '*' .and optional_space;\n"
-"starequals\n"
-" optional_space .and '*' .and '=' .and optional_space;\n"
-".string string_lexer;\n"
-"string_lexer\n"
-" lex_first_identifier_character .and .loop lex_next_identifier_character;\n"
-"lex_first_identifier_character\n"
-" 'a'-'z' .or 'A'-'Z' .or '_';\n"
-"lex_next_identifier_character\n"
-" 'a'-'z' .or 'A'-'Z' .or '0'-'9' .or '_';\n"
-"err_token\n"
-" '~' .or '`' .or '!' .or '@' .or '#' .or '$' .or '%' .or '^' .or '&' .or '*' .or '(' .or ')' .or\n"
-" '-' .or '+' .or '=' .or '|' .or '\\\\' .or '[' .or ']' .or '{' .or '}' .or ':' .or ';' .or '\"' .or\n"
-" '\\'' .or '<' .or ',' .or '>' .or '.' .or '/' .or '?' .or err_identifier;\n"
-"err_identifier\n"
-" id_character_first .and .loop id_character_next;\n"
-""
diff --git a/src/mesa/shader/slang/library/slang_version.syn b/src/mesa/shader/slang/library/slang_version.syn
deleted file mode 100644 (file)
index aaf8bef..0000000
+++ /dev/null
@@ -1,118 +0,0 @@
-/*
- * Mesa 3-D graphics library
- * Version:  6.3
- *
- * Copyright (C) 2005  Brian Paul   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/**
- * \file slang_version.syn
- * slang #version directive syntax
- * \author Michal Krol
- */
-
-.syntax version_directive;
-
-version_directive
-       version_directive_1 .and .loop version_directive_2;
-version_directive_1
-       prior_optional_spaces .and optional_version_directive .and .true .emit $;
-version_directive_2
-       prior_optional_spaces .and version_directive_body .and .true .emit $;
-
-optional_version_directive
-       version_directive_body .or .true .emit 10 .emit 1;
-
-version_directive_body
-       '#' .and optional_space .and "version" .and space .and version_number .and optional_space .and
-       new_line;
-
-version_number
-       version_number_110;
-
-version_number_110
-       leading_zeroes .and "110" .emit 10 .emit 1;
-
-leading_zeroes
-       .loop zero;
-
-zero
-       '0';
-
-space
-    single_space .and .loop single_space;
-
-optional_space
-    .loop single_space;
-
-single_space
-    ' ' .or '\t';
-
-prior_optional_spaces
-       .loop prior_space;
-
-prior_space
-       c_style_comment_block .or cpp_style_comment_block .or space .or new_line;
-
-c_style_comment_block
-    '/' .and '*' .and c_style_comment_rest;
-
-c_style_comment_rest
-    .loop c_style_comment_char_no_star .and c_style_comment_rest_1;
-c_style_comment_rest_1
-    c_style_comment_end .or c_style_comment_rest_2;
-c_style_comment_rest_2
-    '*' .and c_style_comment_rest;
-
-c_style_comment_char_no_star
-    '\x2B'-'\xFF' .or '\x01'-'\x29';
-
-c_style_comment_end
-    '*' .and '/';
-
-cpp_style_comment_block
-    '/' .and '/' .and cpp_style_comment_block_1;
-cpp_style_comment_block_1
-    cpp_style_comment_block_2 .or cpp_style_comment_block_3;
-cpp_style_comment_block_2
-    .loop cpp_style_comment_char .and new_line;
-cpp_style_comment_block_3
-    .loop cpp_style_comment_char;
-
-cpp_style_comment_char
-    '\x0E'-'\xFF' .or '\x01'-'\x09' .or '\x0B'-'\x0C';
-
-new_line
-    cr_lf .or lf_cr .or '\n' .or '\r';
-
-cr_lf
-    '\r' .and '\n';
-
-lf_cr
-       '\n' .and '\r';
-
-.string __string_filter;
-
-__string_filter
-    .loop __identifier_char;
-
-__identifier_char
-    'a'-'z' .or 'A'-'Z' .or '_' .or '0'-'9';
-
diff --git a/src/mesa/shader/slang/library/slang_vertex_builtin_gc.h b/src/mesa/shader/slang/library/slang_vertex_builtin_gc.h
deleted file mode 100644 (file)
index e5a252b..0000000
+++ /dev/null
@@ -1,109 +0,0 @@
-
-/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE FOLLOWING FILE: */
-/* slang_vertex_builtin.gc */
-
-5,2,2,90,95,5,0,12,0,1,103,108,95,80,111,115,105,116,105,111,110,0,0,0,2,2,90,95,5,0,9,0,1,103,108,
-95,80,111,105,110,116,83,105,122,101,0,0,0,2,2,90,95,5,0,12,0,1,103,108,95,67,108,105,112,86,101,
-114,116,101,120,0,0,0,2,2,90,95,2,0,12,0,1,103,108,95,67,111,108,111,114,0,0,0,2,2,90,95,2,0,12,0,
-1,103,108,95,83,101,99,111,110,100,97,114,121,67,111,108,111,114,0,0,0,2,2,90,95,2,0,11,0,1,103,
-108,95,78,111,114,109,97,108,0,0,0,2,2,90,95,2,0,12,0,1,103,108,95,86,101,114,116,101,120,0,0,0,2,
-2,90,95,2,0,12,0,1,103,108,95,77,117,108,116,105,84,101,120,67,111,111,114,100,48,0,0,0,2,2,90,95,
-2,0,12,0,1,103,108,95,77,117,108,116,105,84,101,120,67,111,111,114,100,49,0,0,0,2,2,90,95,2,0,12,0,
-1,103,108,95,77,117,108,116,105,84,101,120,67,111,111,114,100,50,0,0,0,2,2,90,95,2,0,12,0,1,103,
-108,95,77,117,108,116,105,84,101,120,67,111,111,114,100,51,0,0,0,2,2,90,95,2,0,12,0,1,103,108,95,
-77,117,108,116,105,84,101,120,67,111,111,114,100,52,0,0,0,2,2,90,95,2,0,12,0,1,103,108,95,77,117,
-108,116,105,84,101,120,67,111,111,114,100,53,0,0,0,2,2,90,95,2,0,12,0,1,103,108,95,77,117,108,116,
-105,84,101,120,67,111,111,114,100,54,0,0,0,2,2,90,95,2,0,12,0,1,103,108,95,77,117,108,116,105,84,
-101,120,67,111,111,114,100,55,0,0,0,2,2,90,95,2,0,9,0,1,103,108,95,70,111,103,67,111,111,114,100,0,
-0,0,2,2,90,95,3,0,12,0,1,103,108,95,70,114,111,110,116,67,111,108,111,114,0,0,0,2,2,90,95,3,0,12,0,
-1,103,108,95,66,97,99,107,67,111,108,111,114,0,0,0,2,2,90,95,3,0,12,0,1,103,108,95,70,114,111,110,
-116,83,101,99,111,110,100,97,114,121,67,111,108,111,114,0,0,0,2,2,90,95,3,0,12,0,1,103,108,95,66,
-97,99,107,83,101,99,111,110,100,97,114,121,67,111,108,111,114,0,0,0,2,2,90,95,3,0,12,0,1,103,108,
-95,84,101,120,67,111,111,114,100,0,3,18,103,108,95,77,97,120,84,101,120,116,117,114,101,67,111,111,
-114,100,115,0,0,0,2,2,90,95,3,0,9,0,1,103,108,95,70,111,103,70,114,97,103,67,111,111,114,100,0,0,0,
-1,90,95,0,0,12,0,0,102,116,114,97,110,115,102,111,114,109,0,0,1,9,18,95,95,114,101,116,86,97,108,0,
-18,103,108,95,77,111,100,101,108,86,105,101,119,80,114,111,106,101,99,116,105,111,110,77,97,116,
-114,105,120,0,16,8,48,0,57,18,103,108,95,86,101,114,116,101,120,0,59,120,120,120,120,0,48,18,103,
-108,95,77,111,100,101,108,86,105,101,119,80,114,111,106,101,99,116,105,111,110,77,97,116,114,105,
-120,0,16,10,49,0,57,18,103,108,95,86,101,114,116,101,120,0,59,121,121,121,121,0,48,46,18,103,108,
-95,77,111,100,101,108,86,105,101,119,80,114,111,106,101,99,116,105,111,110,77,97,116,114,105,120,0,
-16,10,50,0,57,18,103,108,95,86,101,114,116,101,120,0,59,122,122,122,122,0,48,46,18,103,108,95,77,
-111,100,101,108,86,105,101,119,80,114,111,106,101,99,116,105,111,110,77,97,116,114,105,120,0,16,10,
-51,0,57,18,103,108,95,86,101,114,116,101,120,0,59,119,119,119,119,0,48,46,20,0,0,1,90,95,0,0,12,0,
-0,116,101,120,116,117,114,101,49,68,76,111,100,0,1,1,0,0,16,0,115,97,109,112,108,101,114,0,0,1,1,0,
-0,9,0,99,111,111,114,100,0,0,1,1,0,0,9,0,108,111,100,0,0,0,1,3,2,90,95,0,0,12,0,1,99,111,111,114,
-100,52,0,0,0,9,18,99,111,111,114,100,52,0,59,120,0,18,99,111,111,114,100,0,20,0,9,18,99,111,111,
-114,100,52,0,59,119,0,18,108,111,100,0,20,0,4,118,101,99,52,95,116,101,120,95,49,100,95,98,105,97,
-115,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,
-52,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,49,68,80,114,111,106,76,111,100,0,1,1,0,
-0,16,0,115,97,109,112,108,101,114,0,0,1,1,0,0,10,0,99,111,111,114,100,0,0,1,1,0,0,9,0,108,111,100,
-0,0,0,1,3,2,90,95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,99,111,111,114,100,0,59,120,0,
-18,99,111,111,114,100,0,59,120,0,18,99,111,111,114,100,0,59,121,0,49,20,0,9,18,112,99,111,111,114,
-100,0,59,119,0,18,108,111,100,0,20,0,4,118,101,99,52,95,116,101,120,95,49,100,95,98,105,97,115,0,
-18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,112,99,111,111,114,100,0,0,
-0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,49,68,80,114,111,106,76,111,100,0,1,1,0,0,16,0,
-115,97,109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,1,1,0,0,9,0,108,111,100,0,0,0,1,
-3,2,90,95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,99,111,111,114,100,0,59,120,0,18,99,111,
-111,114,100,0,59,120,0,18,99,111,111,114,100,0,59,122,0,49,20,0,9,18,112,99,111,111,114,100,0,59,
-119,0,18,108,111,100,0,20,0,4,118,101,99,52,95,116,101,120,95,49,100,95,98,105,97,115,0,18,95,95,
-114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,112,99,111,111,114,100,0,0,0,0,1,90,
-95,0,0,12,0,0,116,101,120,116,117,114,101,50,68,76,111,100,0,1,1,0,0,17,0,115,97,109,112,108,101,
-114,0,0,1,1,0,0,10,0,99,111,111,114,100,0,0,1,1,0,0,9,0,108,111,100,0,0,0,1,3,2,90,95,0,0,12,0,1,
-99,111,111,114,100,52,0,0,0,9,18,99,111,111,114,100,52,0,59,120,121,0,18,99,111,111,114,100,0,59,
-120,121,0,20,0,9,18,99,111,111,114,100,52,0,59,119,0,18,108,111,100,0,20,0,4,118,101,99,52,95,116,
-101,120,95,50,100,95,98,105,97,115,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,
-114,0,0,18,99,111,111,114,100,52,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,50,68,80,
-114,111,106,76,111,100,0,1,1,0,0,17,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,
-100,0,0,1,1,0,0,9,0,108,111,100,0,0,0,1,3,2,90,95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,
-99,111,111,114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,
-122,0,49,20,0,9,18,112,99,111,111,114,100,0,59,119,0,18,108,111,100,0,20,0,4,118,101,99,52,95,116,
-101,120,95,50,100,95,98,105,97,115,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,
-114,0,0,18,112,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,50,68,80,
-114,111,106,76,111,100,0,1,1,0,0,17,0,115,97,109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,
-100,0,0,1,1,0,0,9,0,108,111,100,0,0,0,1,3,2,90,95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,
-99,111,111,114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,
-122,0,49,20,0,9,18,112,99,111,111,114,100,0,59,119,0,18,108,111,100,0,20,0,4,118,101,99,52,95,116,
-101,120,95,50,100,95,98,105,97,115,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,
-114,0,0,18,112,99,111,111,114,100,0,0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,51,68,76,
-111,100,0,1,1,0,0,18,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,0,0,1,1,0,0,
-9,0,108,111,100,0,0,0,1,3,2,90,95,0,0,12,0,1,99,111,111,114,100,52,0,0,0,9,18,99,111,111,114,100,
-52,0,59,120,121,122,0,18,99,111,111,114,100,0,59,120,121,122,0,20,0,9,18,99,111,111,114,100,52,0,
-59,119,0,18,108,111,100,0,20,0,4,118,101,99,52,95,116,101,120,95,51,100,95,98,105,97,115,0,18,95,
-95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,52,0,0,0,0,1,
-90,95,0,0,12,0,0,116,101,120,116,117,114,101,51,68,80,114,111,106,76,111,100,0,1,1,0,0,18,0,115,97,
-109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,1,1,0,0,9,0,108,111,100,0,0,0,1,3,2,90,
-95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,99,111,111,114,100,0,59,120,121,122,0,18,99,
-111,111,114,100,0,59,120,121,122,0,18,99,111,111,114,100,0,59,119,0,49,20,0,9,18,112,99,111,111,
-114,100,0,59,119,0,18,108,111,100,0,20,0,4,118,101,99,52,95,116,101,120,95,51,100,95,98,105,97,115,
-0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,112,99,111,111,114,100,0,
-0,0,0,1,90,95,0,0,12,0,0,116,101,120,116,117,114,101,67,117,98,101,76,111,100,0,1,1,0,0,19,0,115,
-97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,0,0,1,1,0,0,9,0,108,111,100,0,0,0,1,3,2,
-90,95,0,0,12,0,1,99,111,111,114,100,52,0,0,0,9,18,99,111,111,114,100,52,0,59,120,121,122,0,18,99,
-111,111,114,100,0,20,0,9,18,99,111,111,114,100,52,0,59,119,0,18,108,111,100,0,20,0,4,118,101,99,52,
-95,116,101,120,95,99,117,98,101,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,
-0,0,18,99,111,111,114,100,52,0,0,0,0,1,90,95,0,0,12,0,0,115,104,97,100,111,119,49,68,76,111,100,0,
-1,1,0,0,20,0,115,97,109,112,108,101,114,0,0,1,1,0,0,11,0,99,111,111,114,100,0,0,1,1,0,0,9,0,108,
-111,100,0,0,0,1,3,2,90,95,0,0,12,0,1,99,111,111,114,100,52,0,0,0,9,18,99,111,111,114,100,52,0,59,
-120,121,122,0,18,99,111,111,114,100,0,20,0,9,18,99,111,111,114,100,52,0,59,119,0,18,108,111,100,0,
-20,0,4,118,101,99,52,95,116,101,120,95,49,100,95,98,105,97,115,95,115,104,97,100,111,119,0,18,95,
-95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,99,111,111,114,100,52,0,0,0,0,1,
-90,95,0,0,12,0,0,115,104,97,100,111,119,49,68,80,114,111,106,76,111,100,0,1,1,0,0,20,0,115,97,109,
-112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,100,0,0,1,1,0,0,9,0,108,111,100,0,0,0,1,3,2,90,95,
-0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,99,111,111,114,100,0,59,120,0,18,99,111,111,114,
-100,0,59,120,0,18,99,111,111,114,100,0,59,119,0,49,20,0,9,18,112,99,111,111,114,100,0,59,122,0,18,
-99,111,111,114,100,0,59,122,0,20,0,9,18,112,99,111,111,114,100,0,59,119,0,18,108,111,100,0,20,0,4,
-118,101,99,52,95,116,101,120,95,49,100,95,98,105,97,115,95,115,104,97,100,111,119,0,18,95,95,114,
-101,116,86,97,108,0,0,18,115,97,109,112,108,101,114,0,0,18,112,99,111,111,114,100,0,0,0,0,1,90,95,
-0,0,12,0,0,115,104,97,100,111,119,50,68,76,111,100,0,1,1,0,0,21,0,115,97,109,112,108,101,114,0,0,1,
-1,0,0,11,0,99,111,111,114,100,0,0,1,1,0,0,9,0,108,111,100,0,0,0,1,3,2,90,95,0,0,12,0,1,99,111,111,
-114,100,52,0,0,0,9,18,99,111,111,114,100,52,0,59,120,121,122,0,18,99,111,111,114,100,0,20,0,9,18,
-99,111,111,114,100,52,0,59,119,0,18,108,111,100,0,20,0,4,118,101,99,52,95,116,101,120,95,50,100,95,
-98,105,97,115,95,115,104,97,100,111,119,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,108,
-101,114,0,0,18,99,111,111,114,100,52,0,0,0,0,1,90,95,0,0,12,0,0,115,104,97,100,111,119,50,68,80,
-114,111,106,76,111,100,0,1,1,0,0,21,0,115,97,109,112,108,101,114,0,0,1,1,0,0,12,0,99,111,111,114,
-100,0,0,1,1,0,0,9,0,108,111,100,0,0,0,1,3,2,90,95,0,0,12,0,1,112,99,111,111,114,100,0,0,0,9,18,112,
-99,111,111,114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,120,121,0,18,99,111,111,114,100,0,59,
-119,0,49,20,0,9,18,112,99,111,111,114,100,0,59,122,0,18,99,111,111,114,100,0,59,122,0,20,0,9,18,
-112,99,111,111,114,100,0,59,119,0,18,108,111,100,0,20,0,4,118,101,99,52,95,116,101,120,95,50,100,
-95,98,105,97,115,95,115,104,97,100,111,119,0,18,95,95,114,101,116,86,97,108,0,0,18,115,97,109,112,
-108,101,114,0,0,18,112,99,111,111,114,100,0,0,0,0,0
diff --git a/src/mesa/shader/slang/library/syn_to_c.c b/src/mesa/shader/slang/library/syn_to_c.c
deleted file mode 100644 (file)
index f997edf..0000000
+++ /dev/null
@@ -1,72 +0,0 @@
-#include <stdio.h>
-
-static int was_space = 0;
-static int first_char = 1;
-
-static void put_char (int c)
-{
-   if (c == '\n') {
-      if (!first_char) {
-         fputs ("\\n\"\n\"", stdout);
-         first_char = 1;
-      }
-   }
-   else {
-      first_char = 0;
-      if (c == '\\')
-         fputs ("\\\\", stdout);
-      else if (c == '\"')
-         fputs ("\\\"", stdout);
-      else if (!was_space || !(c == ' ' || c == '\t'))
-         fputc (c, stdout);
-      was_space = (c == ' ' || c == '\t');
-   }
-}
-
-int main (int argc, char *argv[])
-{
-   int c;
-   FILE *f;
-
-   if (argc == 1)
-      return 1;
-   f = fopen (argv[1], "r");
-   if (f == NULL)
-      return 1;
-
-   fputs ("\n", stdout);
-   fputs ("/* DO NOT EDIT - THIS FILE IS AUTOMATICALLY GENERATED FROM THE .syn FILE */\n", stdout);
-   fputs ("\n", stdout);
-   fputs ("\"", stdout);
-   c = getc (f);
-   while (c != EOF) {
-      if (c == '/') {
-         int c2 = getc (f);
-         if (c2 == '*') {
-            was_space = 0;
-            c = getc (f);
-            for (;;) {
-               if (c == '*') {
-                  c2 = getc (f);
-                  if (c2 == '/')
-                     break;
-               }
-               c = getc (f);
-            }
-         }
-         else {
-            put_char (c);
-            put_char (c2);
-         }
-      }
-      else {
-         put_char (c);
-      }
-      c = getc (f);
-   }
-   fputs ("\"\n", stdout);
-
-   fclose (f);
-   return 0;
-}
-
index 499e16e2de9eb82f5fdd3a4ecfa218baf293a859..f60ad9ee385514c805461e324e676f99d942d4f3 100644 (file)
 #include "shader/prog_optimize.h"
 #include "shader/prog_print.h"
 #include "shader/prog_parameter.h"
-#include "shader/grammar/grammar_mesa.h"
+#include "../../glsl/pp/sl_pp_public.h"
+#include "../../glsl/cl/sl_cl_parse.h"
 #include "slang_codegen.h"
 #include "slang_compile.h"
-#include "slang_preprocess.h"
 #include "slang_storage.h"
 #include "slang_emit.h"
 #include "slang_log.h"
@@ -128,7 +128,7 @@ _slang_code_object_dtr(slang_code_object * self)
 
 typedef struct slang_parse_ctx_
 {
-   const byte *I;
+   const unsigned char *I;
    slang_info_log *L;
    int parsing_builtin;
    GLboolean global_scope;   /**< Is object being declared a global? */
@@ -183,23 +183,103 @@ parse_identifier(slang_parse_ctx * C)
    return slang_atom_pool_atom(C->atoms, id);
 }
 
+static int
+is_hex_digit(char c)
+{
+   return (c >= '0' && c <= '9') || (c >= 'a' && c <= 'f') || (c >= 'A' && c <= 'F');
+}
+
+static int
+parse_general_number(slang_parse_ctx *ctx, float *number)
+{
+   char *flt = NULL;
+
+   if (*ctx->I == '0') {
+      int value = 0;
+      const unsigned char *pi;
+
+      if (ctx->I[1] == 'x' || ctx->I[1] == 'X') {
+         ctx->I += 2;
+         if (!is_hex_digit(*ctx->I)) {
+            return 0;
+         }
+         do {
+            int digit;
+
+            if (*ctx->I >= '0' && *ctx->I <= '9') {
+               digit = (int)(*ctx->I - '0');
+            } else if (*ctx->I >= 'a' && *ctx->I <= 'f') {
+               digit = (int)(*ctx->I - 'a') + 10;
+            } else {
+               digit = (int)(*ctx->I - 'A') + 10;
+            }
+            value = value * 0x10 + digit;
+            ctx->I++;
+         } while (is_hex_digit(*ctx->I));
+         if (*ctx->I != '\0') {
+            return 0;
+         }
+         ctx->I++;
+         *number = (float)value;
+         return 1;
+      }
+
+      pi = ctx->I;
+      pi++;
+      while (*pi >= '0' && *pi <= '7') {
+         int digit;
+
+         digit = (int)(*pi - '0');
+         value = value * 010 + digit;
+         pi++;
+      }
+      if (*pi == '\0') {
+         pi++;
+         ctx->I = pi;
+         *number = (float)value;
+         return 1;
+      }
+   }
+
+   parse_identifier_str(ctx, &flt);
+   flt = strdup(flt);
+   if (!flt) {
+      return 0;
+   }
+   if (flt[strlen(flt) - 1] == 'f' || flt[strlen(flt) - 1] == 'F') {
+      flt[strlen(flt) - 1] = '\0';
+   }
+   *number = (float)_mesa_strtod(flt, (char **)NULL);
+   free(flt);
+
+   return 1;
+}
+
 static int
 parse_number(slang_parse_ctx * C, int *number)
 {
    const int radix = (int) (*C->I++);
-   *number = 0;
-   while (*C->I != '\0') {
-      int digit;
-      if (*C->I >= '0' && *C->I <= '9')
-         digit = (int) (*C->I - '0');
-      else if (*C->I >= 'A' && *C->I <= 'Z')
-         digit = (int) (*C->I - 'A') + 10;
-      else
-         digit = (int) (*C->I - 'a') + 10;
-      *number = *number * radix + digit;
+
+   if (radix == 1) {
+      float f = 0.0f;
+
+      parse_general_number(C, &f);
+      *number = (int)f;
+   } else {
+      *number = 0;
+      while (*C->I != '\0') {
+         int digit;
+         if (*C->I >= '0' && *C->I <= '9')
+            digit = (int) (*C->I - '0');
+         else if (*C->I >= 'A' && *C->I <= 'Z')
+            digit = (int) (*C->I - 'A') + 10;
+         else
+            digit = (int) (*C->I - 'a') + 10;
+         *number = *number * radix + digit;
+         C->I++;
+      }
       C->I++;
    }
-   C->I++;
    if (*number > 65535)
       slang_info_log_warning(C->L, "%d: literal integer overflow.", *number);
    return 1;
@@ -208,32 +288,37 @@ parse_number(slang_parse_ctx * C, int *number)
 static int
 parse_float(slang_parse_ctx * C, float *number)
 {
-   char *integral = NULL;
-   char *fractional = NULL;
-   char *exponent = NULL;
-   char *whole = NULL;
-
-   parse_identifier_str(C, &integral);
-   parse_identifier_str(C, &fractional);
-   parse_identifier_str(C, &exponent);
-
-   whole = (char *) _slang_alloc((_mesa_strlen(integral) +
-                                  _mesa_strlen(fractional) +
-                                  _mesa_strlen(exponent) + 3) * sizeof(char));
-   if (whole == NULL) {
-      slang_info_log_memory(C->L);
-      RETURN0;
-   }
+   if (*C->I == 1) {
+      C->I++;
+      parse_general_number(C, number);
+   } else {
+      char *integral = NULL;
+      char *fractional = NULL;
+      char *exponent = NULL;
+      char *whole = NULL;
+
+      parse_identifier_str(C, &integral);
+      parse_identifier_str(C, &fractional);
+      parse_identifier_str(C, &exponent);
+
+      whole = (char *) _slang_alloc((_mesa_strlen(integral) +
+                                     _mesa_strlen(fractional) +
+                                     _mesa_strlen(exponent) + 3) * sizeof(char));
+      if (whole == NULL) {
+         slang_info_log_memory(C->L);
+         RETURN0;
+      }
 
-   slang_string_copy(whole, integral);
-   slang_string_concat(whole, ".");
-   slang_string_concat(whole, fractional);
-   slang_string_concat(whole, "E");
-   slang_string_concat(whole, exponent);
+      slang_string_copy(whole, integral);
+      slang_string_concat(whole, ".");
+      slang_string_concat(whole, fractional);
+      slang_string_concat(whole, "E");
+      slang_string_concat(whole, exponent);
 
-   *number = (float) (_mesa_strtod(whole, (char **) NULL));
+      *number = (float) (_mesa_strtod(whole, (char **) NULL));
 
-   _slang_free(whole);
+      _slang_free(whole);
+   }
 
    return 1;
 }
@@ -2455,7 +2540,7 @@ parse_code_unit(slang_parse_ctx * C, slang_code_unit * unit,
 }
 
 static GLboolean
-compile_binary(const byte * prod, slang_code_unit * unit,
+compile_binary(const unsigned char * prod, slang_code_unit * unit,
                GLuint version,
                slang_unit_type type, slang_info_log * infolog,
                slang_code_unit * builtin, slang_code_unit * downlink,
@@ -2488,17 +2573,56 @@ compile_binary(const byte * prod, slang_code_unit * unit,
 }
 
 static GLboolean
-compile_with_grammar(grammar id, const char *source, slang_code_unit * unit,
-                     slang_unit_type type, slang_info_log * infolog,
-                     slang_code_unit * builtin,
+compile_with_grammar(const char *source,
+                     slang_code_unit *unit,
+                     slang_unit_type type,
+                     slang_info_log *infolog,
+                     slang_code_unit *builtin,
                      struct gl_shader *shader,
-                     const struct gl_extensions *extensions,
-                     struct gl_sl_pragmas *pragmas)
+                     struct gl_sl_pragmas *pragmas,
+                     unsigned int shader_type,
+                     unsigned int parsing_builtin)
 {
-   byte *prod;
-   GLuint size, start, version;
-   slang_string preprocessed;
-   GLuint maxVersion;
+   struct sl_pp_purify_options options;
+   struct sl_pp_context *context;
+   unsigned char *prod;
+   GLuint size;
+   unsigned int version;
+   unsigned int maxVersion;
+   int result;
+   char errmsg[200] = "";
+
+   assert(shader_type == 1 || shader_type == 2);
+
+   memset(&options, 0, sizeof(options));
+
+   context = sl_pp_context_create(source, &options);
+   if (!context) {
+      slang_info_log_error(infolog, "out of memory");
+      return GL_FALSE;
+   }
+
+   if (sl_pp_version(context, &version)) {
+      slang_info_log_error(infolog, "%s", sl_pp_context_error_message(context));
+      sl_pp_context_destroy(context);
+      return GL_FALSE;
+   }
+
+   if (sl_pp_context_add_extension(context, "ARB_draw_buffers", "GL_ARB_draw_buffers") ||
+       sl_pp_context_add_extension(context, "ARB_texture_rectangle", "GL_ARB_texture_rectangle")) {
+      slang_info_log_error(infolog, "%s", sl_pp_context_error_message(context));
+      sl_pp_context_destroy(context);
+      return GL_FALSE;
+   }
+
+#if FEATURE_es2_glsl
+   if (sl_pp_context_add_predefined(context, "GL_ES", "1") ||
+       sl_pp_context_add_predefined(context, "GL_FRAGMENT_PRECISION_HIGH", "1")) {
+      slang_info_log_error(infolog, "%s", sl_pp_context_error_message(context));
+      sl_pp_context_destroy(context);
+      return GL_FALSE;
+   }
+#endif
 
 #if FEATURE_ARB_shading_language_120
    maxVersion = 120;
@@ -2508,36 +2632,30 @@ compile_with_grammar(grammar id, const char *source, slang_code_unit * unit,
    maxVersion = 110;
 #endif
 
-   /* First retrieve the version number. */
-   if (!_slang_preprocess_version(source, &version, &start, infolog))
-      return GL_FALSE;
-
-   if (version > maxVersion) {
+   if (version > maxVersion ||
+       (version != 100 && version != 110 && version != 120)) {
       slang_info_log_error(infolog,
                            "language version %.2f is not supported.",
                            version * 0.01);
-      return GL_FALSE;
-   }
-
-   /* Now preprocess the source string. */
-   slang_string_init(&preprocessed);
-   if (!_slang_preprocess_directives(&preprocessed, &source[start],
-                                     infolog, extensions, pragmas)) {
-      slang_string_free(&preprocessed);
-      slang_info_log_error(infolog, "failed to preprocess the source.");
+      sl_pp_context_destroy(context);
       return GL_FALSE;
    }
 
    /* Finally check the syntax and generate its binary representation. */
-   if (!grammar_fast_check(id,
-                           (const byte *) (slang_string_cstr(&preprocessed)),
-                           &prod, &size, 65536)) {
-      char buf[1024];
-      GLint pos;
-
-      slang_string_free(&preprocessed);
-      grammar_get_last_error((byte *) (buf), sizeof(buf), &pos);
-      slang_info_log_error(infolog, buf);
+   result = sl_cl_compile(context,
+                          shader_type,
+                          parsing_builtin,
+                          &prod,
+                          &size,
+                          errmsg,
+                          sizeof(errmsg));
+
+   sl_pp_context_destroy(context);
+
+   if (result) {
+      /*GLint pos;*/
+
+      slang_info_log_error(infolog, errmsg);
       /* syntax error (possibly in library code) */
 #if 0
       {
@@ -2551,77 +2669,64 @@ compile_with_grammar(grammar id, const char *source, slang_code_unit * unit,
 #endif
       return GL_FALSE;
    }
-   slang_string_free(&preprocessed);
 
    /* Syntax is okay - translate it to internal representation. */
    if (!compile_binary(prod, unit, version, type, infolog, builtin,
                        &builtin[SLANG_BUILTIN_TOTAL - 1],
                        shader)) {
-      grammar_alloc_free(prod);
+      free(prod);
       return GL_FALSE;
    }
-   grammar_alloc_free(prod);
+   free(prod);
    return GL_TRUE;
 }
 
-LONGSTRING static const char *slang_shader_syn =
-#include "library/slang_shader_syn.h"
-   ;
-
-static const byte slang_core_gc[] = {
+static const unsigned char slang_core_gc[] = {
 #include "library/slang_core_gc.h"
 };
 
-static const byte slang_120_core_gc[] = {
+static const unsigned char slang_120_core_gc[] = {
 #include "library/slang_120_core_gc.h"
 };
 
-static const byte slang_120_fragment_gc[] = {
+static const unsigned char slang_120_fragment_gc[] = {
 #include "library/slang_builtin_120_fragment_gc.h"
 };
 
-static const byte slang_common_builtin_gc[] = {
+static const unsigned char slang_common_builtin_gc[] = {
 #include "library/slang_common_builtin_gc.h"
 };
 
-static const byte slang_fragment_builtin_gc[] = {
+static const unsigned char slang_fragment_builtin_gc[] = {
 #include "library/slang_fragment_builtin_gc.h"
 };
 
-static const byte slang_vertex_builtin_gc[] = {
+static const unsigned char slang_vertex_builtin_gc[] = {
 #include "library/slang_vertex_builtin_gc.h"
 };
 
 static GLboolean
-compile_object(grammar * id, const char *source, slang_code_object * object,
-               slang_unit_type type, slang_info_log * infolog,
+compile_object(const char *source,
+               slang_code_object *object,
+               slang_unit_type type,
+               slang_info_log *infolog,
                struct gl_shader *shader,
-               const struct gl_extensions *extensions,
                struct gl_sl_pragmas *pragmas)
 {
    slang_code_unit *builtins = NULL;
    GLuint base_version = 110;
-
-   /* load GLSL grammar */
-   *id = grammar_load_from_text((const byte *) (slang_shader_syn));
-   if (*id == 0) {
-      byte buf[1024];
-      int pos;
-
-      grammar_get_last_error(buf, 1024, &pos);
-      slang_info_log_error(infolog, (const char *) (buf));
-      return GL_FALSE;
-   }
+   unsigned int shader_type;
+   unsigned int parsing_builtin;
 
    /* set shader type - the syntax is slightly different for different shaders */
-   if (type == SLANG_UNIT_FRAGMENT_SHADER
-       || type == SLANG_UNIT_FRAGMENT_BUILTIN)
-      grammar_set_reg8(*id, (const byte *) "shader_type", 1);
-   else
-      grammar_set_reg8(*id, (const byte *) "shader_type", 2);
+   if (type == SLANG_UNIT_FRAGMENT_SHADER || type == SLANG_UNIT_FRAGMENT_BUILTIN) {
+      shader_type = 1;
+   } else {
+      shader_type = 2;
+   }
 
    /* enable language extensions */
-   grammar_set_reg8(*id, (const byte *) "parsing_builtin", 1);
+   parsing_builtin = 1;
 
    /* if parsing user-specified shader, load built-in library */
    if (type == SLANG_UNIT_FRAGMENT_SHADER || type == SLANG_UNIT_VERTEX_SHADER) {
@@ -2686,51 +2791,24 @@ compile_object(grammar * id, const char *source, slang_code_object * object,
       }
 
       /* disable language extensions */
-#if NEW_SLANG /* allow-built-ins */
-      grammar_set_reg8(*id, (const byte *) "parsing_builtin", 1);
-#else
-      grammar_set_reg8(*id, (const byte *) "parsing_builtin", 0);
-#endif
+      parsing_builtin = 0;
+
       builtins = object->builtin;
    }
 
    /* compile the actual shader - pass-in built-in library for external shader */
-   return compile_with_grammar(*id, source, &object->unit, type, infolog,
-                               builtins, shader, extensions, pragmas);
+   return compile_with_grammar(source,
+                               &object->unit,
+                               type,
+                               infolog,
+                               builtins,
+                               shader,
+                               pragmas,
+                               shader_type,
+                               parsing_builtin);
 }
 
 
-static GLboolean
-compile_shader(GLcontext *ctx, slang_code_object * object,
-               slang_unit_type type, slang_info_log * infolog,
-               struct gl_shader *shader)
-{
-   GLboolean success;
-   grammar id = 0;
-
-#if 0 /* for debug */
-   _mesa_printf("********* COMPILE SHADER ***********\n");
-   _mesa_printf("%s\n", shader->Source);
-   _mesa_printf("************************************\n");
-#endif
-
-   assert(shader->Program);
-
-   _slang_code_object_dtr(object);
-   _slang_code_object_ctr(object);
-
-   success = compile_object(&id, shader->Source, object, type, infolog, shader,
-                            &ctx->Extensions, &shader->Pragmas);
-   if (id != 0)
-      grammar_destroy(id);
-   if (!success)
-      return GL_FALSE;
-
-   return GL_TRUE;
-}
-
-
-
 GLboolean
 _slang_compile(GLcontext *ctx, struct gl_shader *shader)
 {
@@ -2771,7 +2849,12 @@ _slang_compile(GLcontext *ctx, struct gl_shader *shader)
    slang_info_log_construct(&info_log);
    _slang_code_object_ctr(&obj);
 
-   success = compile_shader(ctx, &obj, type, &info_log, shader);
+   success = compile_object(shader->Source,
+                            &obj,
+                            type,
+                            &info_log,
+                            shader,
+                            &shader->Pragmas);
 
    /* free shader's prev info log */
    if (shader->InfoLog) {
index c0e4b27aa5f6bd60095c95ea66fa3edd6db944a2..ce3f6ab7ea274d0b2b0117e7aeebdddfa24b7c12 100644 (file)
@@ -81,8 +81,8 @@ new_subroutine(slang_emit_info *emitInfo, GLuint *id)
 
    emitInfo->Subroutines = (struct gl_program **)
       _mesa_realloc(emitInfo->Subroutines,
-                    n * sizeof(struct gl_program),
-                    (n + 1) * sizeof(struct gl_program));
+                    n * sizeof(struct gl_program *),
+                    (n + 1) * sizeof(struct gl_program *));
    emitInfo->Subroutines[n] = ctx->Driver.NewProgram(ctx, emitInfo->prog->Target, 0);
    emitInfo->Subroutines[n]->Parameters = emitInfo->prog->Parameters;
    emitInfo->NumSubroutines++;
@@ -551,6 +551,9 @@ emit_instruction(slang_emit_info *emitInfo,
                                        &srcRelAddr,
                                        NULL,
                                        NULL);
+               if (!inst) {
+                  return NULL;
+               }
 
                src[i] = &newSrc[i];
             }
@@ -948,6 +951,9 @@ emit_compare(slang_emit_info *emitInfo, slang_ir_node *n)
                               n->Children[0]->Store,
                               n->Children[1]->Store,
                               NULL);
+      if (!inst) {
+         return NULL;
+      }
       inst_comment(inst, "Compare values");
 
       /* Compute val = DOT(temp, temp)  (reduction) */
@@ -957,6 +963,9 @@ emit_compare(slang_emit_info *emitInfo, slang_ir_node *n)
                               &tempStore,
                               &tempStore,
                               NULL);
+      if (!inst) {
+         return NULL;
+      }
       inst->SrcReg[0].Swizzle = inst->SrcReg[1].Swizzle = swizzle; /*override*/
       inst_comment(inst, "Reduce vec to bool");
 
@@ -972,6 +981,9 @@ emit_compare(slang_emit_info *emitInfo, slang_ir_node *n)
                                  n->Store,
                                  &zero,
                                  NULL);
+         if (!inst) {
+            return NULL;
+         }
          inst_comment(inst, "Invert true/false");
       }
    }
@@ -1001,6 +1013,9 @@ emit_compare(slang_emit_info *emitInfo, slang_ir_node *n)
                                     &srcStore0,
                                     &srcStore1,
                                     NULL);
+            if (!inst) {
+               return NULL;
+            }
             inst_comment(inst, "Begin struct/array comparison");
          }
          else {
@@ -1010,12 +1025,18 @@ emit_compare(slang_emit_info *emitInfo, slang_ir_node *n)
                                     &srcStore0,
                                     &srcStore1,
                                     NULL);
+            if (!inst) {
+               return NULL;
+            }
             /* ADD accTemp, accTemp, sneTemp; # like logical-OR */
             inst = emit_instruction(emitInfo, OPCODE_ADD,
                                     &accTemp, /* dest */
                                     &accTemp,
                                     &sneTemp,
                                     NULL);
+            if (!inst) {
+               return NULL;
+            }
          }
       }
 
@@ -1025,6 +1046,9 @@ emit_compare(slang_emit_info *emitInfo, slang_ir_node *n)
                               &accTemp,
                               &accTemp,
                               NULL);
+      if (!inst) {
+         return NULL;
+      }
       inst_comment(inst, "End struct/array comparison");
 
       if (n->Opcode == IR_EQUAL) {
@@ -1036,6 +1060,9 @@ emit_compare(slang_emit_info *emitInfo, slang_ir_node *n)
                                  n->Store,
                                  &zero,
                                  NULL);
+         if (!inst) {
+            return NULL;
+         }
          inst_comment(inst, "Invert true/false");
       }
 
@@ -1119,6 +1146,9 @@ emit_clamp(slang_emit_info *emitInfo, slang_ir_node *n)
                            n->Children[0]->Store,
                            n->Children[1]->Store,
                            NULL);
+   if (!inst) {
+      return NULL;
+   }
 
    /* n->dest = min(tmp, ch[2]) */
    inst = emit_instruction(emitInfo, OPCODE_MIN,
@@ -1153,7 +1183,9 @@ emit_negation(slang_emit_info *emitInfo, slang_ir_node *n)
                            n->Children[0]->Store,
                            NULL,
                            NULL);
-   inst->SrcReg[0].Negate = NEGATE_XYZW;
+   if (inst) {
+      inst->SrcReg[0].Negate = NEGATE_XYZW;
+   }
    return inst;
 }
 
@@ -1356,6 +1388,9 @@ emit_tex(slang_emit_info *emitInfo, slang_ir_node *n)
                            n->Children[1]->Store,
                            NULL,
                            NULL);
+   if (!inst) {
+      return NULL;
+   }
 
    inst->TexShadow = shadow;
 
@@ -1458,6 +1493,9 @@ emit_copy(slang_emit_info *emitInfo, slang_ir_node *n)
                                     &srcStore,
                                     NULL,
                                     NULL);
+            if (!inst) {
+               return NULL;
+            }
             inst_comment(inst, "IR_COPY block");
             srcStore.Index++;
             dstStore.Index++;
@@ -1473,6 +1511,9 @@ emit_copy(slang_emit_info *emitInfo, slang_ir_node *n)
                                  n->Children[1]->Store,
                                  NULL,
                                  NULL);
+         if (!inst) {
+            return NULL;
+         }
          dstAnnot = storage_annotation(n->Children[0], emitInfo->prog);
          srcAnnot = storage_annotation(n->Children[1], emitInfo->prog);
          inst->Comment = instruction_annotation(inst->Opcode, dstAnnot,
@@ -1534,6 +1575,9 @@ emit_cond(slang_emit_info *emitInfo, slang_ir_node *n)
                                  n->Children[0]->Store,
                                  NULL,
                                  NULL);
+         if (!inst) {
+            return NULL;
+         }
          inst->CondUpdate = GL_TRUE;
          inst_comment(inst, "COND expr");
          _slang_free_temp(emitInfo->vt, n->Store);
@@ -1596,6 +1640,9 @@ emit_not(slang_emit_info *emitInfo, slang_ir_node *n)
                            n->Children[0]->Store,
                            &zero,
                            NULL);
+   if (!inst) {
+      return NULL;
+   }
    inst_comment(inst, "NOT");
 
    free_node_storage(emitInfo->vt, n->Children[0]);
@@ -1646,12 +1693,17 @@ emit_if(slang_emit_info *emitInfo, slang_ir_node *n)
          ifInst->DstReg.CondSwizzle = writemask_to_swizzle(condWritemask);
       }
       else {
+         struct prog_instruction *inst;
+
          /* IF src[0] THEN ... */
-         emit_instruction(emitInfo, OPCODE_IF,
-                          NULL, /* dst */
-                          n->Children[0]->Store, /* op0 */
-                          NULL,
-                          NULL);
+         inst = emit_instruction(emitInfo, OPCODE_IF,
+                                 NULL, /* dst */
+                                 n->Children[0]->Store, /* op0 */
+                                 NULL,
+                                 NULL);
+         if (!inst) {
+            return NULL;
+         }
       }
    }
    else {
@@ -1676,6 +1728,7 @@ emit_if(slang_emit_info *emitInfo, slang_ir_node *n)
          if (!inst) {
             return NULL;
          }
+         prog->Instructions[ifInstLoc].BranchTarget = prog->NumInstructions - 1;
       }
       else {
          /* jump to endif instruction */
@@ -1685,8 +1738,8 @@ emit_if(slang_emit_info *emitInfo, slang_ir_node *n)
          }
          inst_comment(inst, "else");
          inst->DstReg.CondMask = COND_TR;  /* always branch */
+         prog->Instructions[ifInstLoc].BranchTarget = prog->NumInstructions;
       }
-      prog->Instructions[ifInstLoc].BranchTarget = prog->NumInstructions;
       emit(emitInfo, n->Children[2]);
    }
    else {
@@ -1701,8 +1754,14 @@ emit_if(slang_emit_info *emitInfo, slang_ir_node *n)
       }
    }
 
-   if (n->Children[2]) {
-      prog->Instructions[elseInstLoc].BranchTarget = prog->NumInstructions;
+   if (elseInstLoc) {
+      /* point ELSE instruction BranchTarget at ENDIF */
+      if (emitInfo->EmitHighLevelInstructions) {
+         prog->Instructions[elseInstLoc].BranchTarget = prog->NumInstructions - 1;
+      }
+      else {
+         prog->Instructions[elseInstLoc].BranchTarget = prog->NumInstructions;
+      }
    }
    return NULL;
 }
@@ -1762,7 +1821,7 @@ emit_loop(slang_emit_info *emitInfo, slang_ir_node *n)
 
    /* Done emitting loop code.  Now walk over the loop's linked list of
     * BREAK and CONT nodes, filling in their BranchTarget fields (which
-    * will point to the ENDLOOP+1 or BGNLOOP instructions, respectively).
+    * will point to the corresponding ENDLOOP instruction.
     */
    for (ir = n->List; ir; ir = ir->List) {
       struct prog_instruction *inst = prog->Instructions + ir->InstLocation;
@@ -1771,8 +1830,13 @@ emit_loop(slang_emit_info *emitInfo, slang_ir_node *n)
           ir->Opcode == IR_BREAK_IF_TRUE) {
          assert(inst->Opcode == OPCODE_BRK ||
                 inst->Opcode == OPCODE_BRA);
-         /* go to instruction after end of loop */
-         inst->BranchTarget = endInstLoc + 1;
+         /* go to instruction at end of loop */
+         if (emitInfo->EmitHighLevelInstructions) {
+            inst->BranchTarget = endInstLoc;
+         }
+         else {
+            inst->BranchTarget = endInstLoc + 1;
+         }
       }
       else {
          assert(ir->Opcode == IR_CONT ||
@@ -1875,6 +1939,9 @@ emit_cont_break_if_true(slang_emit_info *emitInfo, slang_ir_node *n)
                                  n->Children[0]->Store,
                                  NULL,
                                  NULL);
+         if (!inst) {
+            return NULL;
+         }
          n->InstLocation = emitInfo->prog->NumInstructions;
 
          inst = new_instruction(emitInfo, opcode);
@@ -1887,7 +1954,7 @@ emit_cont_break_if_true(slang_emit_info *emitInfo, slang_ir_node *n)
          }
 
          emitInfo->prog->Instructions[ifInstLoc].BranchTarget
-            = emitInfo->prog->NumInstructions;
+            = emitInfo->prog->NumInstructions - 1;
          return inst;
       }
    }
@@ -2045,6 +2112,9 @@ emit_array_element(slang_emit_info *emitInfo, slang_ir_node *n)
                                  indexStore, /* the index */
                                  &elemSizeStore,
                                  NULL);
+         if (!inst) {
+            return NULL;
+         }
 
          indexStore = indexTemp;
       }
@@ -2071,6 +2141,9 @@ emit_array_element(slang_emit_info *emitInfo, slang_ir_node *n)
                                  indexStore,     /* the index */
                                  &indirectArray, /* indirect array base */
                                  NULL);
+         if (!inst) {
+            return NULL;
+         }
 
          indexStore = indexTemp;
       }
index 0a2bc497803ea655f3618672b0b20cdae4798abb..ed27821a951bd0320acfb2aefcc63662512fcf4a 100644 (file)
@@ -590,11 +590,16 @@ concat_shaders(struct gl_shader_program *shProg, GLenum shaderType)
 {
    struct gl_shader *newShader;
    const struct gl_shader *firstShader = NULL;
-   GLuint shaderLengths[100];
+   GLuint *shaderLengths;
    GLchar *source;
    GLuint totalLen = 0, len = 0;
    GLuint i;
 
+   shaderLengths = (GLuint *)_mesa_malloc(shProg->NumShaders * sizeof(GLuint));
+   if (!shaderLengths) {
+      return NULL;
+   }
+
    /* compute total size of new shader source code */
    for (i = 0; i < shProg->NumShaders; i++) {
       const struct gl_shader *shader = shProg->Shaders[i];
@@ -606,12 +611,16 @@ concat_shaders(struct gl_shader_program *shProg, GLenum shaderType)
       }
    }
 
-   if (totalLen == 0)
+   if (totalLen == 0) {
+      _mesa_free(shaderLengths);
       return NULL;
+   }
 
    source = (GLchar *) _mesa_malloc(totalLen + 1);
-   if (!source)
+   if (!source) {
+      _mesa_free(shaderLengths);
       return NULL;
+   }
 
    /* concatenate shaders */
    for (i = 0; i < shProg->NumShaders; i++) {
@@ -626,9 +635,16 @@ concat_shaders(struct gl_shader_program *shProg, GLenum shaderType)
    _mesa_printf("---NEW CONCATENATED SHADER---:\n%s\n------------\n", source);
    */
 
+   _mesa_free(shaderLengths);
+
    remove_extra_version_directives(source);
 
    newShader = CALLOC_STRUCT(gl_shader);
+   if (!newShader) {
+      _mesa_free(source);
+      return NULL;
+   }
+
    newShader->Type = shaderType;
    newShader->Source = source;
    newShader->Pragmas = firstShader->Pragmas;
diff --git a/src/mesa/shader/slang/slang_preprocess.c b/src/mesa/shader/slang/slang_preprocess.c
deleted file mode 100644 (file)
index e9a24cc..0000000
+++ /dev/null
@@ -1,1406 +0,0 @@
-/*
- * Mesa 3-D graphics library
- *
- * Copyright (C) 2005-2008  Brian Paul   All Rights Reserved.
- * Copyright (C) 2009 VMware, Inc.   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-/**
- * \file slang_preprocess.c
- * slang preprocessor
- * \author Michal Krol
- */
-
-#include "main/imports.h"
-#include "shader/grammar/grammar_mesa.h"
-#include "slang_preprocess.h"
-
-LONGSTRING static const char *slang_pp_directives_syn =
-#include "library/slang_pp_directives_syn.h"
-;
-
-LONGSTRING static const char *slang_pp_expression_syn =
-#include "library/slang_pp_expression_syn.h"
-;
-
-LONGSTRING static const char *slang_pp_version_syn =
-#include "library/slang_pp_version_syn.h"
-;
-
-static GLvoid
-grammar_error_to_log (slang_info_log *log)
-{
-   char buf[1024];
-   GLint pos;
-
-   grammar_get_last_error ((byte *) (buf), sizeof (buf), &pos);
-   if (buf[0] == 0) {
-      _mesa_snprintf(buf, sizeof(buf), "Preprocessor error");
-   }
-   slang_info_log_error (log, buf);
-}
-
-GLboolean
-_slang_preprocess_version (const char *text, GLuint *version, GLuint *eaten, slang_info_log *log)
-{
-   grammar id;
-   byte *prod, *I;
-   unsigned int size;
-
-   id = grammar_load_from_text ((const byte *) (slang_pp_version_syn));
-   if (id == 0) {
-      grammar_error_to_log (log);
-      return GL_FALSE;
-   }
-
-   if (!grammar_fast_check (id, (const byte *) (text), &prod, &size, 8)) {
-      grammar_error_to_log (log);
-      grammar_destroy (id);
-      return GL_FALSE;
-   }
-
-   /* there can be multiple #version directives - grab the last one */
-   I = &prod[size - 6];
-   *version = (GLuint) (I[0]) + (GLuint) (I[1]) * 100;
-   *eaten = (GLuint) (I[2]) + ((GLuint) (I[3]) << 8) + ((GLuint) (I[4]) << 16) + ((GLuint) (I[5]) << 24);
-
-   grammar_destroy (id);
-   grammar_alloc_free (prod);
-   return GL_TRUE;
-}
-
-/*
- * The preprocessor does the following work.
- * 1. Remove comments. Each comment block is replaced with a single space and if the
- *    block contains new-lines, they are preserved. This ensures that line numbers
- *    stay the same and if a comment block delimits two tokens, the are delitmited
- *    by the space after comment removal.
- * 2. Remove preprocessor directives from the source string, checking their syntax and
- *    executing them if appropriate. Again, new-lines are preserved.
- * 3. Expand macros.
- * 4. Tokenize the source string by ensuring there is at least one space between every
- *    two adjacent tokens.
- */
-
-#define PP_ANNOTATE 0
-
-static GLvoid
-pp_annotate (slang_string *output, const char *fmt, ...)
-{
-#if PP_ANNOTATE
-   va_list va;
-   char buffer[1024];
-
-   va_start (va, fmt);
-   _mesa_vsprintf (buffer, fmt, va);
-   va_end (va);
-   slang_string_pushs (output, buffer, _mesa_strlen (buffer));
-#else
-   (GLvoid) (output);
-   (GLvoid) (fmt);
-#endif
-}
-
- /*
- * The expression is executed on a fixed-sized stack. The PUSH macro makes a runtime
- * check if the stack is not overflown by too complex expressions. In that situation the
- * GLSL preprocessor should report internal compiler error.
- * The BINARYDIV makes a runtime check if the divider is not 0. If it is, it reports
- * compilation error.
- */
-
-#define EXECUTION_STACK_SIZE 1024
-
-#define PUSH(x)\
-   do {\
-      if (sp == 0) {\
-         slang_info_log_error (elog, "internal compiler error: preprocessor execution stack overflow.");\
-         return GL_FALSE;\
-      }\
-      stack[--sp] = x;\
-   } while (GL_FALSE)
-
-#define POP(x)\
-   do {\
-      assert (sp < EXECUTION_STACK_SIZE);\
-      x = stack[sp++];\
-   } while (GL_FALSE)
-
-#define BINARY(op)\
-   do {\
-      GLint a, b;\
-      POP(b);\
-      POP(a);\
-      PUSH(a op b);\
-   } while (GL_FALSE)
-
-#define BINARYDIV(op)\
-   do {\
-      GLint a, b;\
-      POP(b);\
-      POP(a);\
-      if (b == 0) {\
-         slang_info_log_error (elog, "division by zero in preprocessor expression.");\
-         return GL_FALSE;\
-      }\
-      PUSH(a op b);\
-   } while (GL_FALSE)
-
-#define UNARY(op)\
-   do {\
-      GLint a;\
-      POP(a);\
-      PUSH(op a);\
-   } while (GL_FALSE)
-
-#define OP_END          0
-#define OP_PUSHINT      1
-#define OP_LOGICALOR    2
-#define OP_LOGICALAND   3
-#define OP_OR           4
-#define OP_XOR          5
-#define OP_AND          6
-#define OP_EQUAL        7
-#define OP_NOTEQUAL     8
-#define OP_LESSEQUAL    9
-#define OP_GREATEREQUAL 10
-#define OP_LESS         11
-#define OP_GREATER      12
-#define OP_LEFTSHIFT    13
-#define OP_RIGHTSHIFT   14
-#define OP_ADD          15
-#define OP_SUBTRACT     16
-#define OP_MULTIPLY     17
-#define OP_DIVIDE       18
-#define OP_MODULUS      19
-#define OP_PLUS         20
-#define OP_MINUS        21
-#define OP_NEGATE       22
-#define OP_COMPLEMENT   23
-
-static GLboolean
-execute_expression (slang_string *output, const byte *code, GLuint *pi, GLint *result,
-                    slang_info_log *elog)
-{
-   GLuint i = *pi;
-   GLint stack[EXECUTION_STACK_SIZE];
-   GLuint sp = EXECUTION_STACK_SIZE;
-
-   while (code[i] != OP_END) {
-      switch (code[i++]) {
-         case OP_PUSHINT:
-            i++;
-            PUSH(_mesa_atoi ((const char *) (&code[i])));
-            i += _mesa_strlen ((const char *) (&code[i])) + 1;
-            break;
-         case OP_LOGICALOR:
-            BINARY(||);
-            break;
-         case OP_LOGICALAND:
-            BINARY(&&);
-            break;
-         case OP_OR:
-            BINARY(|);
-            break;
-         case OP_XOR:
-            BINARY(^);
-            break;
-         case OP_AND:
-            BINARY(&);
-            break;
-         case OP_EQUAL:
-            BINARY(==);
-            break;
-         case OP_NOTEQUAL:
-            BINARY(!=);
-            break;
-         case OP_LESSEQUAL:
-            BINARY(<=);
-            break;
-         case OP_GREATEREQUAL:
-            BINARY(>=);
-            break;
-         case OP_LESS:
-            BINARY(<);
-            break;
-         case OP_GREATER:
-            BINARY(>);
-            break;
-         case OP_LEFTSHIFT:
-            BINARY(<<);
-            break;
-         case OP_RIGHTSHIFT:
-            BINARY(>>);
-            break;
-         case OP_ADD:
-            BINARY(+);
-            break;
-         case OP_SUBTRACT:
-            BINARY(-);
-            break;
-         case OP_MULTIPLY:
-            BINARY(*);
-            break;
-         case OP_DIVIDE:
-            BINARYDIV(/);
-            break;
-         case OP_MODULUS:
-            BINARYDIV(%);
-            break;
-         case OP_PLUS:
-            UNARY(+);
-            break;
-         case OP_MINUS:
-            UNARY(-);
-            break;
-         case OP_NEGATE:
-            UNARY(!);
-            break;
-         case OP_COMPLEMENT:
-            UNARY(~);
-            break;
-         default:
-            assert (0);
-      }
-   }
-
-   /* Write-back the index skipping the OP_END. */
-   *pi = i + 1;
-
-   /* There should be exactly one value left on the stack. This is our result. */
-   POP(*result);
-   pp_annotate (output, "%d ", *result);
-   assert (sp == EXECUTION_STACK_SIZE);
-   return GL_TRUE;
-}
-
-/*
- * Function execute_expressions() executes up to 2 expressions. The second expression is there
- * for the #line directive which takes 1 or 2 expressions that indicate line and file numbers.
- * If it fails, it returns 0. If it succeeds, it returns the number of executed expressions.
- */
-
-#define EXP_END        0
-#define EXP_EXPRESSION 1
-
-static GLuint
-execute_expressions (slang_string *output, grammar eid, const byte *expr, GLint results[2],
-                     slang_info_log *elog)
-{
-   GLint success;
-   byte *code;
-   GLuint size, count = 0;
-
-   success = grammar_fast_check (eid, expr, &code, &size, 64);
-   if (success) {
-      GLuint i = 0;
-
-      while (code[i++] == EXP_EXPRESSION) {
-         assert (count < 2);
-
-         if (!execute_expression (output, code, &i, &results[count], elog)) {
-            count = 0;
-            break;
-         }
-         count++;
-      }
-      grammar_alloc_free (code);
-   }
-   else {
-      slang_info_log_error (elog, "syntax error in preprocessor expression.");\
-   }
-   return count;
-}
-
-/*
- * The pp_symbol structure is used to hold macro definitions and macro formal parameters. The
- * pp_symbols strcture is a collection of pp_symbol. It is used both for storing macro formal
- * parameters and all global macro definitions. Making this unification wastes some memory,
- * becuse macro formal parameters don't need further lists of symbols. We lose 8 bytes per
- * formal parameter here, but making this we can use the same code to substitute macro parameters
- * as well as macros in the source string.
- */
-
-typedef struct
-{
-   struct pp_symbol_ *symbols;
-   GLuint count;
-} pp_symbols;
-
-static GLvoid
-pp_symbols_init (pp_symbols *self)
-{
-   self->symbols = NULL;
-   self->count = 0;
-}
-
-static GLvoid
-pp_symbols_free (pp_symbols *);
-
-typedef struct pp_symbol_
-{
-   slang_string name;
-   slang_string replacement;
-   pp_symbols parameters;
-} pp_symbol;
-
-static GLvoid
-pp_symbol_init (pp_symbol *self)
-{
-   slang_string_init (&self->name);
-   slang_string_init (&self->replacement);
-   pp_symbols_init (&self->parameters);
-}
-
-static GLvoid
-pp_symbol_free (pp_symbol *self)
-{
-   slang_string_free (&self->name);
-   slang_string_free (&self->replacement);
-   pp_symbols_free (&self->parameters);
-}
-
-static GLvoid
-pp_symbol_reset (pp_symbol *self)
-{
-   /* Leave symbol name intact. */
-   slang_string_reset (&self->replacement);
-   pp_symbols_free (&self->parameters);
-   pp_symbols_init (&self->parameters);
-}
-
-static GLvoid
-pp_symbols_free (pp_symbols *self)
-{
-   GLuint i;
-
-   for (i = 0; i < self->count; i++)
-      pp_symbol_free (&self->symbols[i]);
-   _mesa_free (self->symbols);
-}
-
-static pp_symbol *
-pp_symbols_push (pp_symbols *self)
-{
-   self->symbols = (pp_symbol *) (_mesa_realloc (self->symbols, self->count * sizeof (pp_symbol),
-                                                 (self->count + 1) * sizeof (pp_symbol)));
-   if (self->symbols == NULL)
-      return NULL;
-   pp_symbol_init (&self->symbols[self->count]);
-   return &self->symbols[self->count++];
-}
-
-static GLboolean
-pp_symbols_erase (pp_symbols *self, pp_symbol *symbol)
-{
-   assert (symbol >= self->symbols && symbol < self->symbols + self->count);
-
-   self->count--;
-   pp_symbol_free (symbol);
-   if (symbol < self->symbols + self->count)
-      _mesa_memcpy (symbol, symbol + 1, sizeof (pp_symbol) * (self->symbols + self->count - symbol));
-   self->symbols = (pp_symbol *) (_mesa_realloc (self->symbols, (self->count + 1) * sizeof (pp_symbol),
-                                                 self->count * sizeof (pp_symbol)));
-   return self->symbols != NULL;
-}
-
-static pp_symbol *
-pp_symbols_find (pp_symbols *self, const char *name)
-{
-   GLuint i;
-
-   for (i = 0; i < self->count; i++)
-      if (_mesa_strcmp (name, slang_string_cstr (&self->symbols[i].name)) == 0)
-         return &self->symbols[i];
-   return NULL;
-}
-
-/*
- * The condition context of a single #if/#else/#endif level. Those can be nested, so there
- * is a stack of condition contexts.
- * There is a special global context on the bottom of the stack. It is there to simplify
- * context handling.
- */
-
-typedef struct
-{
-   GLboolean current;         /* The condition value of this level. */
-   GLboolean effective;       /* The effective product of current condition, outer level conditions
-                               * and position within #if-#else-#endif sections. */
-   GLboolean else_allowed;    /* TRUE if in #if-#else section, FALSE if in #else-#endif section
-                               * and for global context. */
-   GLboolean endif_required;  /* FALSE for global context only. */
-} pp_cond_ctx;
-
-/* Should be enuff. */
-#define CONDITION_STACK_SIZE 64
-
-typedef struct
-{
-   pp_cond_ctx stack[CONDITION_STACK_SIZE];
-   pp_cond_ctx *top;
-} pp_cond_stack;
-
-static GLboolean
-pp_cond_stack_push (pp_cond_stack *self, slang_info_log *elog)
-{
-   if (self->top == self->stack) {
-      slang_info_log_error (elog, "internal compiler error: preprocessor condition stack overflow.");
-      return GL_FALSE;
-   }
-   self->top--;
-   return GL_TRUE;
-}
-
-static GLvoid
-pp_cond_stack_reevaluate (pp_cond_stack *self)
-{
-   /* There must be at least 2 conditions on the stack - one global and one being evaluated. */
-   assert (self->top <= &self->stack[CONDITION_STACK_SIZE - 2]);
-
-   self->top->effective = self->top->current && self->top[1].effective;
-}
-
-
-/**
- * Extension enables through #extension directive.
- * NOTE: Currently, only enable/disable state is stored.
- */
-typedef struct
-{
-   GLboolean ARB_draw_buffers;
-   GLboolean ARB_texture_rectangle;
-} pp_ext;
-
-
-/**
- * Disable all extensions. Called at startup and on #extension all: disable.
- */
-static GLvoid
-pp_ext_disable_all(pp_ext *self)
-{
-   _mesa_memset(self, 0, sizeof(self));
-}
-
-
-/**
- * Called during preprocessor initialization to set the initial enable/disable
- * state of extensions.
- */
-static GLvoid
-pp_ext_init(pp_ext *self, const struct gl_extensions *extensions)
-{
-   pp_ext_disable_all (self);
-   self->ARB_draw_buffers = GL_TRUE;
-   if (extensions->NV_texture_rectangle)
-      self->ARB_texture_rectangle = GL_TRUE;
-}
-
-/**
- * Called in response to #extension directives to enable/disable
- * the named extension.
- */
-static GLboolean
-pp_ext_set(pp_ext *self, const char *name, GLboolean enable)
-{
-   if (_mesa_strcmp (name, "GL_ARB_draw_buffers") == 0)
-      self->ARB_draw_buffers = enable;
-   else if (_mesa_strcmp (name, "GL_ARB_texture_rectangle") == 0)
-      self->ARB_texture_rectangle = enable;
-   else
-      return GL_FALSE;
-   return GL_TRUE;
-}
-
-
-/**
- * Called in response to #pragma.  For example, "#pragma debug(on)" would
- * call this function as pp_pragma("debug", "on").
- * \return GL_TRUE if pragma is valid, GL_FALSE if invalid
- */
-static GLboolean
-pp_pragma(struct gl_sl_pragmas *pragmas, const char *pragma, const char *param)
-{
-#if 0
-   printf("#pragma %s %s\n", pragma, param);
-#endif
-   if (_mesa_strcmp(pragma, "optimize") == 0) {
-      if (!param)
-         return GL_FALSE; /* missing required param */
-      if (_mesa_strcmp(param, "on") == 0) {
-         if (!pragmas->IgnoreOptimize)
-            pragmas->Optimize = GL_TRUE;
-      }
-      else if (_mesa_strcmp(param, "off") == 0) {
-         if (!pragmas->IgnoreOptimize)
-            pragmas->Optimize = GL_FALSE;
-      }
-      else {
-         return GL_FALSE; /* invalid param */
-      }
-   }
-   else if (_mesa_strcmp(pragma, "debug") == 0) {
-      if (!param)
-         return GL_FALSE; /* missing required param */
-      if (_mesa_strcmp(param, "on") == 0) {
-         if (!pragmas->IgnoreDebug)
-            pragmas->Debug = GL_TRUE;
-      }
-      else if (_mesa_strcmp(param, "off") == 0) {
-         if (!pragmas->IgnoreDebug)
-            pragmas->Debug = GL_FALSE;
-      }
-      else {
-         return GL_FALSE; /* invalid param */
-      }
-   }
-   /* all other pragmas are silently ignored */
-   return GL_TRUE;
-}
-
-
-/**
- * The state of preprocessor: current line, file and version number, list
- * of all defined macros and the #if/#endif context.
- */
-typedef struct
-{
-   GLint line;
-   GLint file;
-   GLint version;
-   pp_symbols symbols;
-   pp_ext ext;
-   slang_info_log *elog;
-   pp_cond_stack cond;
-} pp_state;
-
-static GLvoid
-pp_state_init (pp_state *self, slang_info_log *elog,
-               const struct gl_extensions *extensions)
-{
-   self->line = 0;
-   self->file = 1;
-#if FEATURE_es2_glsl
-   self->version = 100;
-#else
-   self->version = 110;
-#endif
-   pp_symbols_init (&self->symbols);
-   pp_ext_init (&self->ext, extensions);
-   self->elog = elog;
-
-   /* Initialize condition stack and create the global context. */
-   self->cond.top = &self->cond.stack[CONDITION_STACK_SIZE - 1];
-   self->cond.top->current = GL_TRUE;
-   self->cond.top->effective = GL_TRUE;
-   self->cond.top->else_allowed = GL_FALSE;
-   self->cond.top->endif_required = GL_FALSE;
-}
-
-static GLvoid
-pp_state_free (pp_state *self)
-{
-   pp_symbols_free (&self->symbols);
-}
-
-#define IS_FIRST_ID_CHAR(x) (((x) >= 'a' && (x) <= 'z') || ((x) >= 'A' && (x) <= 'Z') || (x) == '_')
-#define IS_NEXT_ID_CHAR(x) (IS_FIRST_ID_CHAR(x) || ((x) >= '0' && (x) <= '9'))
-#define IS_WHITE(x) ((x) == ' ' || (x) == '\n')
-#define IS_NULL(x) ((x) == '\0')
-
-#define SKIP_WHITE(x) do { while (IS_WHITE(*(x))) (x)++; } while (GL_FALSE)
-
-typedef struct
-{
-   slang_string *output;
-   const char *input;
-   pp_state *state;
-} expand_state;
-
-static GLboolean
-expand_defined (expand_state *e, slang_string *buffer)
-{
-   GLboolean in_paren = GL_FALSE;
-   const char *id;
-
-   /* Parse the optional opening parenthesis. */
-   SKIP_WHITE(e->input);
-   if (*e->input == '(') {
-      e->input++;
-      in_paren = GL_TRUE;
-      SKIP_WHITE(e->input);
-   }
-
-   /* Parse operand. */
-   if (!IS_FIRST_ID_CHAR(*e->input)) {
-      slang_info_log_error (e->state->elog,
-                            "preprocess error: identifier expected after operator 'defined'.");
-      return GL_FALSE;
-   }
-   slang_string_reset (buffer);
-   slang_string_pushc (buffer, *e->input++);
-   while (IS_NEXT_ID_CHAR(*e->input))
-      slang_string_pushc (buffer, *e->input++);
-   id = slang_string_cstr (buffer);
-
-   /* Check if the operand is defined. Output 1 if it is defined, output 0 if not. */
-   if (pp_symbols_find (&e->state->symbols, id) == NULL)
-      slang_string_pushs (e->output, " 0 ", 3);
-   else
-      slang_string_pushs (e->output, " 1 ", 3);
-
-   /* Parse the closing parentehesis if the opening one was there. */
-   if (in_paren) {
-      SKIP_WHITE(e->input);
-      if (*e->input != ')') {
-         slang_info_log_error (e->state->elog, "preprocess error: ')' expected.");
-         return GL_FALSE;
-      }
-      e->input++;
-      SKIP_WHITE(e->input);
-   }
-   return GL_TRUE;
-}
-
-static GLboolean
-expand (expand_state *, pp_symbols *);
-
-static GLboolean
-expand_symbol (expand_state *e, pp_symbol *symbol)
-{
-   expand_state es;
-
-   /* If the macro has some parameters, we need to parse them. */
-   if (symbol->parameters.count != 0) {
-      GLuint i;
-
-      /* Parse the opening parenthesis. */
-      SKIP_WHITE(e->input);
-      if (*e->input != '(') {
-         slang_info_log_error (e->state->elog, "preprocess error: '(' expected.");
-         return GL_FALSE;
-      }
-      e->input++;
-      SKIP_WHITE(e->input);
-
-      /* Parse macro actual parameters. This can be anything, separated by a colon.
-       */
-      for (i = 0; i < symbol->parameters.count; i++) {
-         GLuint nested_paren_count = 0; /* track number of nested parentheses */
-
-         if (*e->input == ')') {
-            slang_info_log_error (e->state->elog, "preprocess error: unexpected ')'.");
-            return GL_FALSE;
-         }
-
-         /* Eat all characters up to the comma or closing parentheses. */
-         pp_symbol_reset (&symbol->parameters.symbols[i]);
-         while (!IS_NULL(*e->input)) {
-            /* Exit loop only when all nested parens have been eaten. */
-            if (nested_paren_count == 0 && (*e->input == ',' || *e->input == ')'))
-               break;
-
-            /* Actually count nested parens here. */
-            if (*e->input == '(')
-               nested_paren_count++;
-            else if (*e->input == ')')
-               nested_paren_count--;
-
-            slang_string_pushc (&symbol->parameters.symbols[i].replacement, *e->input++);
-         }
-
-         /* If it was not the last paremeter, skip the comma. Otherwise, skip the
-          * closing parentheses. */
-         if (i + 1 == symbol->parameters.count) {
-            /* This is the last paremeter - skip the closing parentheses. */
-            if (*e->input != ')') {
-               slang_info_log_error (e->state->elog, "preprocess error: ')' expected.");
-               return GL_FALSE;
-            }
-            e->input++;
-            SKIP_WHITE(e->input);
-         }
-         else {
-            /* Skip the separating comma. */
-            if (*e->input != ',') {
-               slang_info_log_error (e->state->elog, "preprocess error: ',' expected.");
-               return GL_FALSE;
-            }
-            e->input++;
-            SKIP_WHITE(e->input);
-         }
-      }
-   }
-
-   /* Expand the macro. Use its parameters as a priority symbol list to expand
-    * macro parameters correctly. */
-   es.output = e->output;
-   es.input = slang_string_cstr (&symbol->replacement);
-   es.state = e->state;
-   slang_string_pushc (e->output, ' ');
-   if (!expand (&es, &symbol->parameters))
-      return GL_FALSE;
-   slang_string_pushc (e->output, ' ');
-   return GL_TRUE;
-}
-
-/*
- * Function expand() expands source text from <input> to <output>. The expansion is made using
- * the list passed in <symbols> parameter. It allows us to expand macro formal parameters with
- * actual parameters. The global list of symbols from pp state is used when doing a recursive
- * call of expand().
- */
-
-static GLboolean
-expand (expand_state *e, pp_symbols *symbols)
-{
-   while (!IS_NULL(*e->input)) {
-      if (IS_FIRST_ID_CHAR(*e->input)) {
-         slang_string buffer;
-         const char *id;
-
-         /* Parse the identifier. */
-         slang_string_init (&buffer);
-         slang_string_pushc (&buffer, *e->input++);
-         while (IS_NEXT_ID_CHAR(*e->input))
-            slang_string_pushc (&buffer, *e->input++);
-         id = slang_string_cstr (&buffer);
-
-         /* Now check if the identifier is special in some way. The "defined" identifier is
-          * actually an operator that we must handle here and expand it either to " 0 " or " 1 ".
-          * The other identifiers start with "__" and we expand it to appropriate values
-          * taken from the preprocessor state. */
-         if (_mesa_strcmp (id, "defined") == 0) {
-            if (!expand_defined (e, &buffer))
-               return GL_FALSE;
-         }
-         else if (_mesa_strcmp (id, "__LINE__") == 0) {
-            slang_string_pushc (e->output, ' ');
-            slang_string_pushi (e->output, e->state->line);
-            slang_string_pushc (e->output, ' ');
-         }
-         else if (_mesa_strcmp (id, "__FILE__") == 0) {
-            slang_string_pushc (e->output, ' ');
-            slang_string_pushi (e->output, e->state->file);
-            slang_string_pushc (e->output, ' ');
-         }
-         else if (_mesa_strcmp (id, "__VERSION__") == 0) {
-            slang_string_pushc (e->output, ' ');
-            slang_string_pushi (e->output, e->state->version);
-            slang_string_pushc (e->output, ' ');
-         }
-#if FEATURE_es2_glsl
-         else if (_mesa_strcmp (id, "GL_ES") == 0 ||
-                  _mesa_strcmp (id, "GL_FRAGMENT_PRECISION_HIGH") == 0) {
-            slang_string_pushc (e->output, ' ');
-            slang_string_pushi (e->output, '1');
-            slang_string_pushc (e->output, ' ');
-         }
-#endif
-         else {
-            pp_symbol *symbol;
-
-            /* The list of symbols from <symbols> take precedence over the list from <state>.
-             * Note that in some cases this is the same list so avoid double look-up. */
-            symbol = pp_symbols_find (symbols, id);
-            if (symbol == NULL && symbols != &e->state->symbols)
-               symbol = pp_symbols_find (&e->state->symbols, id);
-
-            /* If the symbol was found, recursively expand its definition. */
-            if (symbol != NULL) {
-               if (!expand_symbol (e, symbol)) {
-                  slang_string_free (&buffer);
-                  return GL_FALSE;
-               }
-            }
-            else {
-               slang_string_push (e->output, &buffer);
-            }
-         }
-         slang_string_free (&buffer);
-      }
-      else if (IS_WHITE(*e->input)) {
-         slang_string_pushc (e->output, *e->input++);
-      }
-      else {
-         while (!IS_WHITE(*e->input) && !IS_NULL(*e->input) && !IS_FIRST_ID_CHAR(*e->input))
-            slang_string_pushc (e->output, *e->input++);
-      }
-   }
-   return GL_TRUE;
-}
-
-static GLboolean
-parse_if (slang_string *output, const byte *prod, GLuint *pi, GLint *result, pp_state *state,
-          grammar eid)
-{
-   const char *text;
-   GLuint len;
-
-   text = (const char *) (&prod[*pi]);
-   len = _mesa_strlen (text);
-
-   if (state->cond.top->effective) {
-      slang_string expr;
-      GLuint count;
-      GLint results[2];
-      expand_state es;
-
-      /* Expand the expression. */
-      slang_string_init (&expr);
-      es.output = &expr;
-      es.input = text;
-      es.state = state;
-      if (!expand (&es, &state->symbols))
-         return GL_FALSE;
-
-      /* Execute the expression. */
-      count = execute_expressions (output, eid, (const byte *) (slang_string_cstr (&expr)),
-                                   results, state->elog);
-      slang_string_free (&expr);
-      if (count != 1)
-         return GL_FALSE;
-      *result = results[0];
-   }
-   else {
-      /* The directive is dead. */
-      *result = 0;
-   }
-
-   *pi += len + 1;
-   return GL_TRUE;
-}
-
-#define ESCAPE_TOKEN 0
-
-#define TOKEN_END       0
-#define TOKEN_DEFINE    1
-#define TOKEN_UNDEF     2
-#define TOKEN_IF        3
-#define TOKEN_ELSE      4
-#define TOKEN_ELIF      5
-#define TOKEN_ENDIF     6
-#define TOKEN_ERROR     7
-#define TOKEN_PRAGMA    8
-#define TOKEN_EXTENSION 9
-#define TOKEN_LINE      10
-
-#define PARAM_END       0
-#define PARAM_PARAMETER 1
-
-#define BEHAVIOR_REQUIRE 1
-#define BEHAVIOR_ENABLE  2
-#define BEHAVIOR_WARN    3
-#define BEHAVIOR_DISABLE 4
-
-#define PRAGMA_NO_PARAM  0
-#define PRAGMA_PARAM     1
-
-
-static GLboolean
-preprocess_source (slang_string *output, const char *source,
-                   grammar pid, grammar eid,
-                   slang_info_log *elog,
-                   const struct gl_extensions *extensions,
-                   struct gl_sl_pragmas *pragmas)
-{
-   static const char *predefined[] = {
-      "__FILE__",
-      "__LINE__",
-      "__VERSION__",
-#if FEATURE_es2_glsl
-      "GL_ES",
-      "GL_FRAGMENT_PRECISION_HIGH",
-#endif
-      NULL
-   };
-   byte *prod;
-   GLuint size, i;
-   pp_state state;
-
-   if (!grammar_fast_check (pid, (const byte *) (source), &prod, &size, 65536)) {
-      grammar_error_to_log (elog);
-      return GL_FALSE;
-   }
-
-   pp_state_init (&state, elog, extensions);
-
-   /* add the predefined symbols to the symbol table */
-   for (i = 0; predefined[i]; i++) {
-      pp_symbol *symbol = NULL;
-      symbol = pp_symbols_push(&state.symbols);
-      assert(symbol);
-      slang_string_pushs(&symbol->name,
-                         predefined[i], _mesa_strlen(predefined[i]));
-   }
-
-   i = 0;
-   while (i < size) {
-      if (prod[i] != ESCAPE_TOKEN) {
-         if (state.cond.top->effective) {
-            slang_string input;
-            expand_state es;
-
-            /* Eat only one line of source code to expand it.
-             * FIXME: This approach has one drawback. If a macro with parameters spans across
-             *        multiple lines, the preprocessor will raise an error. */
-            slang_string_init (&input);
-            while (prod[i] != '\0' && prod[i] != '\n')
-               slang_string_pushc (&input, prod[i++]);
-            if (prod[i] != '\0')
-               slang_string_pushc (&input, prod[i++]);
-
-            /* Increment line number. */
-            state.line++;
-
-            es.output = output;
-            es.input = slang_string_cstr (&input);
-            es.state = &state;
-            if (!expand (&es, &state.symbols))
-               goto error;
-
-            slang_string_free (&input);
-         }
-         else {
-            /* Condition stack is disabled - keep track on line numbers and output only newlines. */
-            if (prod[i] == '\n') {
-               state.line++;
-               /*pp_annotate (output, "%c", prod[i]);*/
-            }
-            else {
-               /*pp_annotate (output, "%c", prod[i]);*/
-            }
-            i++;
-         }
-      }
-      else {
-         const char *id;
-         GLuint idlen;
-         GLubyte token;
-
-         i++;
-         token = prod[i++];
-         switch (token) {
-
-         case TOKEN_END:
-            /* End of source string.
-               * Check if all #ifs have been terminated by matching #endifs.
-               * On condition stack there should be only the global condition context. */
-            if (state.cond.top->endif_required) {
-               slang_info_log_error (elog, "end of source without matching #endif.");
-               return GL_FALSE;
-            }
-            break;
-
-         case TOKEN_DEFINE:
-            {
-               pp_symbol *symbol = NULL;
-
-               /* Parse macro name. */
-               id = (const char *) (&prod[i]);
-               idlen = _mesa_strlen (id);
-               if (state.cond.top->effective) {
-                  pp_annotate (output, "// #define %s(", id);
-
-                  /* If the symbol is already defined, override it. */
-                  symbol = pp_symbols_find (&state.symbols, id);
-                  if (symbol == NULL) {
-                     symbol = pp_symbols_push (&state.symbols);
-                     if (symbol == NULL)
-                        goto error;
-                     slang_string_pushs (&symbol->name, id, idlen);
-                  }
-                  else {
-                     pp_symbol_reset (symbol);
-                  }
-               }
-               i += idlen + 1;
-
-               /* Parse optional macro parameters. */
-               while (prod[i++] != PARAM_END) {
-                  pp_symbol *param;
-
-                  id = (const char *) (&prod[i]);
-                  idlen = _mesa_strlen (id);
-                  if (state.cond.top->effective) {
-                     pp_annotate (output, "%s, ", id);
-                     param = pp_symbols_push (&symbol->parameters);
-                     if (param == NULL)
-                        goto error;
-                     slang_string_pushs (&param->name, id, idlen);
-                  }
-                  i += idlen + 1;
-               }
-
-               /* Parse macro replacement. */
-               id = (const char *) (&prod[i]);
-               idlen = _mesa_strlen (id);
-               if (state.cond.top->effective) {
-                  slang_string replacement;
-                  expand_state es;
-
-                  pp_annotate (output, ") %s", id);
-
-                  slang_string_init(&replacement);
-                  slang_string_pushs(&replacement, id, idlen);
-
-                  /* Expand macro replacement. */
-                  es.output = &symbol->replacement;
-                  es.input = slang_string_cstr(&replacement);
-                  es.state = &state;
-                  if (!expand(&es, &state.symbols)) {
-                     slang_string_free(&replacement);
-                     goto error;
-                  }
-                  slang_string_free(&replacement);
-               }
-               i += idlen + 1;
-            }
-            break;
-
-         case TOKEN_UNDEF:
-            id = (const char *) (&prod[i]);
-            i += _mesa_strlen (id) + 1;
-            if (state.cond.top->effective) {
-               pp_symbol *symbol;
-
-               pp_annotate (output, "// #undef %s", id);
-               /* Try to find symbol with given name and remove it. */
-               symbol = pp_symbols_find (&state.symbols, id);
-               if (symbol != NULL)
-                  if (!pp_symbols_erase (&state.symbols, symbol))
-                     goto error;
-            }
-            break;
-
-         case TOKEN_IF:
-            {
-               GLint result;
-
-               /* Parse #if expression end execute it. */
-               pp_annotate (output, "// #if ");
-               if (!parse_if (output, prod, &i, &result, &state, eid))
-                  goto error;
-
-               /* Push new condition on the stack. */
-               if (!pp_cond_stack_push (&state.cond, state.elog))
-                  goto error;
-               state.cond.top->current = result ? GL_TRUE : GL_FALSE;
-               state.cond.top->else_allowed = GL_TRUE;
-               state.cond.top->endif_required = GL_TRUE;
-               pp_cond_stack_reevaluate (&state.cond);
-            }
-            break;
-
-         case TOKEN_ELSE:
-            /* Check if #else is alloved here. */
-            if (!state.cond.top->else_allowed) {
-               slang_info_log_error (elog, "#else without matching #if.");
-               goto error;
-            }
-
-            /* Negate current condition and reevaluate it. */
-            state.cond.top->current = !state.cond.top->current;
-            state.cond.top->else_allowed = GL_FALSE;
-            pp_cond_stack_reevaluate (&state.cond);
-            if (state.cond.top->effective)
-               pp_annotate (output, "// #else");
-            break;
-
-         case TOKEN_ELIF:
-            /* Check if #elif is alloved here. */
-            if (!state.cond.top->else_allowed) {
-               slang_info_log_error (elog, "#elif without matching #if.");
-               goto error;
-            }
-
-            /* Negate current condition and reevaluate it. */
-            state.cond.top->current = !state.cond.top->current;
-            pp_cond_stack_reevaluate (&state.cond);
-
-            if (state.cond.top->effective)
-               pp_annotate (output, "// #elif ");
-
-            {
-               GLint result;
-
-               /* Parse #elif expression end execute it. */
-               if (!parse_if (output, prod, &i, &result, &state, eid))
-                  goto error;
-
-               /* Update current condition and reevaluate it. */
-               state.cond.top->current = result ? GL_TRUE : GL_FALSE;
-               pp_cond_stack_reevaluate (&state.cond);
-            }
-            break;
-
-         case TOKEN_ENDIF:
-            /* Check if #endif is alloved here. */
-            if (!state.cond.top->endif_required) {
-               slang_info_log_error (elog, "#endif without matching #if.");
-               goto error;
-            }
-
-            /* Pop the condition off the stack. */
-            state.cond.top++;
-            if (state.cond.top->effective)
-               pp_annotate (output, "// #endif");
-            break;
-
-         case TOKEN_EXTENSION:
-            /* Parse the extension name. */
-            id = (const char *) (&prod[i]);
-            i += _mesa_strlen (id) + 1;
-            if (state.cond.top->effective)
-               pp_annotate (output, "// #extension %s: ", id);
-
-            /* Parse and apply extension behavior. */
-            if (state.cond.top->effective) {
-               switch (prod[i++]) {
-
-               case BEHAVIOR_REQUIRE:
-                  pp_annotate (output, "require");
-                  if (!pp_ext_set (&state.ext, id, GL_TRUE)) {
-                     if (_mesa_strcmp (id, "all") == 0) {
-                        slang_info_log_error (elog, "require: bad behavior for #extension all.");
-                        goto error;
-                     }
-                     else {
-                        slang_info_log_error (elog, "%s: required extension is not supported.", id);
-                        goto error;
-                     }
-                  }
-                  break;
-
-               case BEHAVIOR_ENABLE:
-                  pp_annotate (output, "enable");
-                  if (!pp_ext_set (&state.ext, id, GL_TRUE)) {
-                     if (_mesa_strcmp (id, "all") == 0) {
-                        slang_info_log_error (elog, "enable: bad behavior for #extension all.");
-                        goto error;
-                     }
-                     else {
-                        slang_info_log_warning (elog, "%s: enabled extension is not supported.", id);
-                     }
-                  }
-                  break;
-
-               case BEHAVIOR_WARN:
-                  pp_annotate (output, "warn");
-                  if (!pp_ext_set (&state.ext, id, GL_TRUE)) {
-                     if (_mesa_strcmp (id, "all") != 0) {
-                        slang_info_log_warning (elog, "%s: enabled extension is not supported.", id);
-                     }
-                  }
-                  break;
-
-               case BEHAVIOR_DISABLE:
-                  pp_annotate (output, "disable");
-                  if (!pp_ext_set (&state.ext, id, GL_FALSE)) {
-                     if (_mesa_strcmp (id, "all") == 0) {
-                        pp_ext_disable_all (&state.ext);
-                     }
-                     else {
-                        slang_info_log_warning (elog, "%s: disabled extension is not supported.", id);
-                     }
-                  }
-                  break;
-
-               default:
-                  assert (0);
-               }
-            }
-            break;
-
-         case TOKEN_PRAGMA:
-            {
-               GLint have_param;
-               const char *pragma, *param;
-
-               pragma = (const char *) (&prod[i]);
-               i += _mesa_strlen(pragma) + 1;
-               have_param = (prod[i++] == PRAGMA_PARAM);
-               if (have_param) {
-                  param = (const char *) (&prod[i]);
-                  i += _mesa_strlen(param) + 1;
-               }
-               else {
-                  param = NULL;
-               }
-               pp_pragma(pragmas, pragma, param);
-            }
-            break;
-
-         case TOKEN_LINE:
-            id = (const char *) (&prod[i]);
-            i += _mesa_strlen (id) + 1;
-
-            if (state.cond.top->effective) {
-               slang_string buffer;
-               GLuint count;
-               GLint results[2];
-               expand_state es;
-
-               slang_string_init (&buffer);
-               state.line++;
-               es.output = &buffer;
-               es.input = id;
-               es.state = &state;
-               if (!expand (&es, &state.symbols))
-                  goto error;
-
-               pp_annotate (output, "// #line ");
-               count = execute_expressions (output, eid,
-                                             (const byte *) (slang_string_cstr (&buffer)),
-                                             results, state.elog);
-               slang_string_free (&buffer);
-               if (count == 0)
-                  goto error;
-
-               state.line = results[0] - 1;
-               if (count == 2)
-                  state.file = results[1];
-            }
-            break;
-         }
-      }
-   }
-
-   /* Check for missing #endifs. */
-   if (state.cond.top->endif_required) {
-      slang_info_log_error (elog, "#endif expected but end of source found.");
-      goto error;
-   }
-
-   grammar_alloc_free(prod);
-   pp_state_free (&state);
-   return GL_TRUE;
-
-error:
-   grammar_alloc_free(prod);
-   pp_state_free (&state);
-   return GL_FALSE;
-}
-
-
-/**
- * Remove the continuation characters from the input string.
- * This is the very first step in preprocessing and is effective
- * even inside comment blocks.
- * If there is a whitespace between a backslash and a newline,
- * this is not considered as a line continuation.
- * \return GL_TRUE for success, GL_FALSE otherwise.
- */
-static GLboolean
-_slang_preprocess_backslashes(slang_string *output,
-                              const char *input)
-{
-   while (*input) {
-      if (input[0] == '\\') {
-         /* If a newline follows, eat the backslash and the newline. */
-         if (input[1] == '\r') {
-            if (input[2] == '\n') {
-               input += 3;
-            } else {
-               input += 2;
-            }
-         } else if (input[1] == '\n') {
-            if (input[2] == '\r') {
-               input += 3;
-            } else {
-               input += 2;
-            }
-         } else {
-            /* Leave the backslash alone. */
-            slang_string_pushc(output, *input++);
-         }
-      } else {
-         slang_string_pushc(output, *input++);
-      }
-   }
-   return GL_TRUE;
-}
-
-
-/**
- * Run preprocessor on source code.
- * \param extensions  indicates which GL extensions are enabled
- * \param output  the post-process results
- * \param input  the input text
- * \param elog  log to record warnings, errors
- * \param extensions  out extension settings
- * \param pragmas  in/out #pragma settings
- * \return GL_TRUE for success, GL_FALSE for error
- */
-GLboolean
-_slang_preprocess_directives(slang_string *output,
-                             const char *input,
-                             slang_info_log *elog,
-                             const struct gl_extensions *extensions,
-                             struct gl_sl_pragmas *pragmas)
-{
-   grammar pid, eid;
-   GLboolean success;
-   slang_string without_backslashes;
-
-   pid = grammar_load_from_text ((const byte *) (slang_pp_directives_syn));
-   if (pid == 0) {
-      grammar_error_to_log (elog);
-      return GL_FALSE;
-   }
-   eid = grammar_load_from_text ((const byte *) (slang_pp_expression_syn));
-   if (eid == 0) {
-      grammar_error_to_log (elog);
-      grammar_destroy (pid);
-      return GL_FALSE;
-   }
-
-   slang_string_init(&without_backslashes);
-   success = _slang_preprocess_backslashes(&without_backslashes, input);
-
-   if (0) {
-      _mesa_printf("Pre-processed shader:\n");
-      _mesa_printf("%s", slang_string_cstr(&without_backslashes));
-      _mesa_printf("----------------------\n");
-   }
-
-   if (success) {
-      success = preprocess_source(output,
-                                  slang_string_cstr(&without_backslashes),
-                                  pid,
-                                  eid,
-                                  elog,
-                                  extensions,
-                                  pragmas);
-   }
-
-   slang_string_free(&without_backslashes);
-   grammar_destroy (eid);
-   grammar_destroy (pid);
-
-   if (0) {
-      _mesa_printf("Post-processed shader:\n");
-      _mesa_printf("%s", slang_string_cstr(output));
-      _mesa_printf("----------------------\n");
-   }
-
-   return success;
-}
-
diff --git a/src/mesa/shader/slang/slang_preprocess.h b/src/mesa/shader/slang/slang_preprocess.h
deleted file mode 100644 (file)
index f344820..0000000
+++ /dev/null
@@ -1,41 +0,0 @@
-/*
- * Mesa 3-D graphics library
- *
- * Copyright (C) 2005-2008  Brian Paul   All Rights Reserved.
- * Copyright (C) 2009 VMware, Inc.   All Rights Reserved.
- *
- * Permission is hereby granted, free of charge, to any person obtaining a
- * copy of this software and associated documentation files (the "Software"),
- * to deal in the Software without restriction, including without limitation
- * the rights to use, copy, modify, merge, publish, distribute, sublicense,
- * and/or sell copies of the Software, and to permit persons to whom the
- * Software is furnished to do so, subject to the following conditions:
- *
- * The above copyright notice and this permission notice shall be included
- * in all copies or substantial portions of the Software.
- *
- * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS
- * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
- * FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT.  IN NO EVENT SHALL
- * BRIAN PAUL BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN
- * AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
- * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE SOFTWARE.
- */
-
-#ifndef SLANG_PREPROCESS_H
-#define SLANG_PREPROCESS_H
-
-#include "slang_compile.h"
-#include "slang_log.h"
-
-
-extern GLboolean
-_slang_preprocess_version (const char *, GLuint *, GLuint *, slang_info_log *);
-
-extern GLboolean
-_slang_preprocess_directives(slang_string *output, const char *input,
-                             slang_info_log *,
-                             const struct gl_extensions *extensions,
-                             struct gl_sl_pragmas *pragmas);
-
-#endif /* SLANG_PREPROCESS_H */
index a0d7dbbace9eb7bd98471952b3c9f28e666f87fb..a7a3b9a7f97d1e65d71a800a80123fe5d21c25ad 100644 (file)
@@ -218,7 +218,6 @@ SHADER_SOURCES = \
        shader/arbprogparse.c \
        shader/arbprogram.c \
        shader/atifragshader.c \
-       shader/grammar/grammar_mesa.c \
        shader/hash_table.c \
        shader/lex.yy.c \
        shader/nvfragparse.c \
@@ -255,7 +254,6 @@ SLANG_SOURCES =     \
        shader/slang/slang_link.c       \
        shader/slang/slang_log.c        \
        shader/slang/slang_mem.c        \
-       shader/slang/slang_preprocess.c \
        shader/slang/slang_print.c      \
        shader/slang/slang_simplify.c   \
        shader/slang/slang_storage.c    \
@@ -365,6 +363,12 @@ GLAPI_OBJECTS = \
 COMMON_DRIVER_OBJECTS = $(COMMON_DRIVER_SOURCES:.c=.o)
 
 
+### Other archives/libraries
+
+GLSL_LIBS = \
+       $(TOP)/src/glsl/pp/libglslpp.a \
+       $(TOP)/src/glsl/cl/libglslcl.a
+
 
 ### Include directories
 
index 0e89a624c4e27ba6cb80371d6f7c6cc3adf26de7..73df44d1988ef8cf6e94fde0019193b700fecdbd 100644 (file)
@@ -46,7 +46,8 @@ static const struct st_tracked_state *atoms[] =
    &st_update_clip,
 
    &st_finalize_textures,
-   &st_update_shader,
+   &st_update_fp,
+   &st_update_vp,
 
    &st_update_rasterizer,
    &st_update_polygon_stipple,
index c7cffd85c8a89d03b2c729622108e29bce2077a8..f34b49203b226f4c8f2ea03417687d9c15aa4c15 100644 (file)
@@ -47,7 +47,8 @@ void st_validate_state( struct st_context *st );
 extern const struct st_tracked_state st_update_framebuffer;
 extern const struct st_tracked_state st_update_clip;
 extern const struct st_tracked_state st_update_depth_stencil_alpha;
-extern const struct st_tracked_state st_update_shader;
+extern const struct st_tracked_state st_update_fp;
+extern const struct st_tracked_state st_update_vp;
 extern const struct st_tracked_state st_update_rasterizer;
 extern const struct st_tracked_state st_update_polygon_stipple;
 extern const struct st_tracked_state st_update_viewport;
index e18c0f6e0a19b42c083600d7f2d1e1b1ac71a255..8ca4335e3307f2e90bf541a197635e6f417a57af 100644 (file)
@@ -40,6 +40,7 @@
 #include "pipe/p_inlines.h"
 #include "cso_cache/cso_context.h"
 #include "util/u_rect.h"
+#include "util/u_math.h"
 
 
 
@@ -64,8 +65,8 @@ update_renderbuffer_surface(struct st_context *st,
       GLuint level;
       /* find matching mipmap level size */
       for (level = 0; level <= texture->last_level; level++) {
-         if (texture->width[level] == rtt_width &&
-             texture->height[level] == rtt_height) {
+         if (u_minify(texture->width0, level) == rtt_width &&
+             u_minify(texture->height0, level) == rtt_height) {
 
             pipe_surface_reference(&strb->surface, NULL);
 
index babfcc87b4ab8e2081df7166e683faa290890b36..6a5854e9ba5b4d0a4d0122cd776afd228ec5f627 100644 (file)
@@ -145,7 +145,7 @@ load_color_map_texture(GLcontext *ctx, struct pipe_texture *pt)
    const GLuint gSize = ctx->PixelMaps.GtoG.Size;
    const GLuint bSize = ctx->PixelMaps.BtoB.Size;
    const GLuint aSize = ctx->PixelMaps.AtoA.Size;
-   const uint texSize = pt->width[0];
+   const uint texSize = pt->width0;
    uint *dest;
    uint i, j;
 
@@ -162,12 +162,14 @@ load_color_map_texture(GLcontext *ctx, struct pipe_texture *pt)
     */
    for (i = 0; i < texSize; i++) {
       for (j = 0; j < texSize; j++) {
+         union util_color uc;
          int k = (i * texSize + j);
          ubyte r = ctx->PixelMaps.RtoR.Map8[j * rSize / texSize];
          ubyte g = ctx->PixelMaps.GtoG.Map8[i * gSize / texSize];
          ubyte b = ctx->PixelMaps.BtoB.Map8[j * bSize / texSize];
          ubyte a = ctx->PixelMaps.AtoA.Map8[i * aSize / texSize];
-         util_pack_color_ub(r, g, b, a, pt->format, dest + k);
+         util_pack_color_ub(r, g, b, a, pt->format, &uc);
+         *(dest + k) = uc.ui;
       }
    }
 
index 6611956ae83ce77a9474a00633bd4c110770d102..d6e3a3e5610021f85b4bdc9f98ddc3446cd9ca57 100644 (file)
@@ -229,14 +229,23 @@ update_samplers(struct st_context *st)
 
          /*printf("%s su=%u non-null\n", __FUNCTION__, su);*/
          cso_single_sampler(st->cso_context, su, sampler);
+         if (su < st->ctx->Const.MaxVertexTextureImageUnits) {
+            cso_single_vertex_sampler(st->cso_context, su, sampler);
+         }
       }
       else {
          /*printf("%s su=%u null\n", __FUNCTION__, su);*/
          cso_single_sampler(st->cso_context, su, NULL);
+         if (su < st->ctx->Const.MaxVertexTextureImageUnits) {
+            cso_single_vertex_sampler(st->cso_context, su, NULL);
+         }
       }
    }
 
    cso_single_sampler_done(st->cso_context);
+   if (st->ctx->Const.MaxVertexTextureImageUnits > 0) {
+      cso_single_vertex_sampler_done(st->cso_context);
+   }
 }
 
 
index 9208f3f446fa36c6a70f8d185671d819a627606d..46c8cbb30987596c3269b7c4b39854eb97a9cb36 100644 (file)
 #include "st_mesa_to_tgsi.h"
 
 
-/**
- * This represents a vertex program, especially translated to match
- * the inputs of a particular fragment shader.
- */
-struct translated_vertex_program
-{
-   struct st_vertex_program *master;
-
-   /** The fragment shader "signature" this vertex shader is meant for: */
-   GLbitfield frag_inputs;
-
-   /** Compared against master vertex program's serialNo: */
-   GLuint serialNo;
-
-   /** Maps VERT_RESULT_x to slot */
-   GLuint output_to_slot[VERT_RESULT_MAX];
-   ubyte output_to_semantic_name[VERT_RESULT_MAX];
-   ubyte output_to_semantic_index[VERT_RESULT_MAX];
-
-   /** Pointer to the translated vertex program */
-   struct st_vertex_program *vp;
 
-   struct translated_vertex_program *next;  /**< next in linked list */
-};
-
-
-
-/**
- * Given a vertex program output attribute, return the corresponding
- * fragment program input attribute.
- * \return -1 for vertex outputs that have no corresponding fragment input
- */
-static GLint
-vp_out_to_fp_in(GLuint vertResult)
-{
-   if (vertResult >= VERT_RESULT_TEX0 &&
-       vertResult < VERT_RESULT_TEX0 + MAX_TEXTURE_COORD_UNITS)
-      return FRAG_ATTRIB_TEX0 + (vertResult - VERT_RESULT_TEX0);
-
-   if (vertResult >= VERT_RESULT_VAR0 &&
-       vertResult < VERT_RESULT_VAR0 + MAX_VARYING)
-      return FRAG_ATTRIB_VAR0 + (vertResult - VERT_RESULT_VAR0);
-
-   switch (vertResult) {
-   case VERT_RESULT_HPOS:
-      return FRAG_ATTRIB_WPOS;
-   case VERT_RESULT_COL0:
-      return FRAG_ATTRIB_COL0;
-   case VERT_RESULT_COL1:
-      return FRAG_ATTRIB_COL1;
-   case VERT_RESULT_FOGC:
-      return FRAG_ATTRIB_FOGC;
-   default:
-      /* Back-face colors, edge flags, etc */
-      return -1;
-   }
-}
 
 
-/**
- * Find a translated vertex program that corresponds to stvp and
- * has outputs matched to stfp's inputs.
- * This performs vertex and fragment translation (to TGSI) when needed.
+/*
+ * Translate fragment program if needed.
  */
-static struct translated_vertex_program *
-find_translated_vp(struct st_context *st,
-                   struct st_vertex_program *stvp,
-                   struct st_fragment_program *stfp)
+static void
+translate_fp(struct st_context *st,
+             struct st_fragment_program *stfp)
 {
-   static const GLuint UNUSED = ~0;
-   struct translated_vertex_program *xvp;
    const GLbitfield fragInputsRead = stfp->Base.Base.InputsRead;
 
-   /*
-    * Translate fragment program if needed.
-    */
    if (!stfp->state.tokens) {
       GLuint inAttr, numIn = 0;
 
@@ -141,7 +77,7 @@ find_translated_vp(struct st_context *st,
             numIn++;
          }
          else {
-            stfp->input_to_slot[inAttr] = UNUSED;
+            stfp->input_to_slot[inAttr] = -1;
          }
       }
 
@@ -151,168 +87,74 @@ find_translated_vp(struct st_context *st,
 
       st_translate_fragment_program(st, stfp, stfp->input_to_slot);
    }
+}
 
 
-   /* See if we've got a translated vertex program whose outputs match
-    * the fragment program's inputs.
-    * XXX This could be a hash lookup, using InputsRead as the key.
-    */
-   for (xvp = stfp->vertex_programs; xvp; xvp = xvp->next) {
-      if (xvp->master == stvp && xvp->frag_inputs == fragInputsRead) {
-         break;
-      }
-   }
-
-   /* No?  Allocate translated vp object now */
-   if (!xvp) {
-      xvp = ST_CALLOC_STRUCT(translated_vertex_program);
-      xvp->frag_inputs = fragInputsRead;
-      xvp->master = stvp;
-
-      xvp->next = stfp->vertex_programs;
-      stfp->vertex_programs = xvp;
-   }
 
-   /* See if we need to translate vertex program to TGSI form */
-   if (xvp->serialNo != stvp->serialNo) {
-      GLuint outAttr;
-      const GLbitfield64 outputsWritten = stvp->Base.Base.OutputsWritten;
-      GLuint numVpOuts = 0;
-      GLboolean emitPntSize = GL_FALSE, emitBFC0 = GL_FALSE, emitBFC1 = GL_FALSE;
-      GLbitfield usedGenerics = 0x0;
-      GLbitfield usedOutputSlots = 0x0;
-
-      /* Compute mapping of vertex program outputs to slots, which depends
-       * on the fragment program's input->slot mapping.
-       */
-      for (outAttr = 0; outAttr < VERT_RESULT_MAX; outAttr++) {
-         /* set defaults: */
-         xvp->output_to_slot[outAttr] = UNUSED;
-         xvp->output_to_semantic_name[outAttr] = TGSI_SEMANTIC_COUNT;
-         xvp->output_to_semantic_index[outAttr] = 99;
-
-         if (outAttr == VERT_RESULT_HPOS) {
-            /* always put xformed position into slot zero */
-            GLuint slot = 0;
-            xvp->output_to_slot[VERT_RESULT_HPOS] = slot;
-            xvp->output_to_semantic_name[outAttr] = TGSI_SEMANTIC_POSITION;
-            xvp->output_to_semantic_index[outAttr] = 0;
-            numVpOuts++;
-            usedOutputSlots |= (1 << slot);
-         }
-         else if (outputsWritten & (1 << outAttr)) {
-            /* see if the frag prog wants this vert output */
-            GLint fpInAttrib = vp_out_to_fp_in(outAttr);
-            if (fpInAttrib >= 0) {
-               GLuint fpInSlot = stfp->input_to_slot[fpInAttrib];
-               if (fpInSlot != ~0) {
-                  /* match this vp output to the fp input */
-                  GLuint vpOutSlot = stfp->input_map[fpInSlot];
-                  xvp->output_to_slot[outAttr] = vpOutSlot;
-                  xvp->output_to_semantic_name[outAttr] = stfp->input_semantic_name[fpInSlot];
-                  xvp->output_to_semantic_index[outAttr] = stfp->input_semantic_index[fpInSlot];
-                  numVpOuts++;
-                  usedOutputSlots |= (1 << vpOutSlot);
-               }
-               else {
-#if 0 /*debug*/
-                  printf("VP output %d not used by FP\n", outAttr);
-#endif
-               }
-            }
-            else if (outAttr == VERT_RESULT_PSIZ)
-               emitPntSize = GL_TRUE;
-            else if (outAttr == VERT_RESULT_BFC0)
-               emitBFC0 = GL_TRUE;
-            else if (outAttr == VERT_RESULT_BFC1)
-               emitBFC1 = GL_TRUE;
-         }
-#if 0 /*debug*/
-         printf("assign vp output_to_slot[%d] = %d\n", outAttr, 
-                xvp->output_to_slot[outAttr]);
-#endif
-      }
+/**
+ * Find a translated vertex program that corresponds to stvp and
+ * has outputs matched to stfp's inputs.
+ * This performs vertex and fragment translation (to TGSI) when needed.
+ */
+static struct st_vp_varient *
+find_translated_vp(struct st_context *st,
+                   struct st_vertex_program *stvp )
+{
+   struct st_vp_varient *vpv;
+   struct st_vp_varient_key key;
 
-      /* must do these last */
-      if (emitPntSize) {
-         GLuint slot = numVpOuts++;
-         xvp->output_to_slot[VERT_RESULT_PSIZ] = slot;
-         xvp->output_to_semantic_name[VERT_RESULT_PSIZ] = TGSI_SEMANTIC_PSIZE;
-         xvp->output_to_semantic_index[VERT_RESULT_PSIZ] = 0;
-         usedOutputSlots |= (1 << slot);
-      }
-      if (emitBFC0) {
-         GLuint slot = numVpOuts++;
-         xvp->output_to_slot[VERT_RESULT_BFC0] = slot;
-         xvp->output_to_semantic_name[VERT_RESULT_BFC0] = TGSI_SEMANTIC_COLOR;
-         xvp->output_to_semantic_index[VERT_RESULT_BFC0] = 0;
-         usedOutputSlots |= (1 << slot);
-      }
-      if (emitBFC1) {
-         GLuint slot = numVpOuts++;
-         xvp->output_to_slot[VERT_RESULT_BFC1] = slot;
-         xvp->output_to_semantic_name[VERT_RESULT_BFC1] = TGSI_SEMANTIC_COLOR;
-         xvp->output_to_semantic_index[VERT_RESULT_BFC1] = 1;
-         usedOutputSlots |= (1 << slot);
-      }
+   /* Nothing in our key yet.  This will change:
+    */
+   memset(&key, 0, sizeof key);
+
+   /* When this is true, we will add an extra input to the vertex
+    * shader translation (for edgeflags), an extra output with
+    * edgeflag semantics, and extend the vertex shader to pass through
+    * the input to the output.  We'll need to use similar logic to set
+    * up the extra vertex_element input for edgeflags.
+    * _NEW_POLYGON, ST_NEW_EDGEFLAGS_DATA
+    */
+   key.passthrough_edgeflags = (st->vertdata_edgeflags && (
+                                st->ctx->Polygon.FrontMode != GL_FILL ||
+                                st->ctx->Polygon.BackMode != GL_FILL));
 
-      /* build usedGenerics mask */
-      usedGenerics = 0x0;
-      for (outAttr = 0; outAttr < VERT_RESULT_MAX; outAttr++) {
-         if (xvp->output_to_semantic_name[outAttr] == TGSI_SEMANTIC_GENERIC) {
-            usedGenerics |= (1 << xvp->output_to_semantic_index[outAttr]);
-         }
-      }
 
-      /* For each vertex program output that doesn't match up to a fragment
-       * program input, map the vertex program output to a free slot and
-       * free generic attribute.
+   /* Do we need to throw away old translations after a change in the
+    * GL program string?
+    */
+   if (stvp->serialNo != stvp->lastSerialNo) {
+      /* These may have changed if the program string changed.
        */
-      for (outAttr = 0; outAttr < VERT_RESULT_MAX; outAttr++) {
-         if (outputsWritten & (1 << outAttr)) {
-            if (xvp->output_to_slot[outAttr] == UNUSED) {
-               GLint freeGeneric = _mesa_ffs(~usedGenerics) - 1;
-               GLint freeSlot = _mesa_ffs(~usedOutputSlots) - 1;
-               usedGenerics |= (1 << freeGeneric);
-               usedOutputSlots |= (1 << freeSlot);
-               xvp->output_to_slot[outAttr] = freeSlot;
-               xvp->output_to_semantic_name[outAttr] = TGSI_SEMANTIC_GENERIC;
-               xvp->output_to_semantic_index[outAttr] = freeGeneric;
-            }
-         }
+      st_prepare_vertex_program( st, stvp );
 
-#if 0 /*debug*/
-         printf("vp output_to_slot[%d] = %d\n", outAttr, 
-                xvp->output_to_slot[outAttr]);
-#endif
+      /* We are now up-to-date:
+       */
+      stvp->lastSerialNo = stvp->serialNo;
+   }
+   
+   /* See if we've got a translated vertex program whose outputs match
+    * the fragment program's inputs.
+    */
+   for (vpv = stvp->varients; vpv; vpv = vpv->next) {
+      if (memcmp(&vpv->key, &key, sizeof key) == 0) {
+         break;
       }
+   }
 
-      st_translate_vertex_program(st, stvp, xvp->output_to_slot,
-                                  xvp->output_to_semantic_name,
-                                  xvp->output_to_semantic_index);
-
-      xvp->vp = stvp;
-
-      /* translated VP is up to date now */
-      xvp->serialNo = stvp->serialNo;
+   /* No?  Perform new translation here. */
+   if (!vpv) {
+      vpv = st_translate_vertex_program(st, stvp, &key);
+      if (!vpv)
+         return NULL;
+      
+      vpv->next = stvp->varients;
+      stvp->varients = vpv;
    }
 
-   return xvp;
+   return vpv;
 }
 
 
-void
-st_free_translated_vertex_programs(struct st_context *st,
-                                   struct translated_vertex_program *xvp)
-{
-   struct translated_vertex_program *next;
-
-   while (xvp) {
-      next = xvp->next;
-      _mesa_free(xvp);
-      xvp = next;
-   }
-}
 
 
 static void *
@@ -326,32 +168,19 @@ get_passthrough_fs(struct st_context *st)
    return st->passthrough_fs;
 }
 
-
 static void
-update_linkage( struct st_context *st )
+update_fp( struct st_context *st )
 {
-   struct st_vertex_program *stvp;
    struct st_fragment_program *stfp;
-   struct translated_vertex_program *xvp;
-
-   /* find active shader and params -- Should be covered by
-    * ST_NEW_VERTEX_PROGRAM
-    */
-   assert(st->ctx->VertexProgram._Current);
-   stvp = st_vertex_program(st->ctx->VertexProgram._Current);
-   assert(stvp->Base.Base.Target == GL_VERTEX_PROGRAM_ARB);
 
    assert(st->ctx->FragmentProgram._Current);
    stfp = st_fragment_program(st->ctx->FragmentProgram._Current);
    assert(stfp->Base.Base.Target == GL_FRAGMENT_PROGRAM_ARB);
 
-   xvp = find_translated_vp(st, stvp, stfp);
+   translate_fp(st, stfp);
 
-   st_reference_vertprog(st, &st->vp, stvp);
    st_reference_fragprog(st, &st->fp, stfp);
 
-   cso_set_vertex_shader_handle(st->cso_context, stvp->driver_shader);
-
    if (st->missing_textures) {
       /* use a pass-through frag shader that uses no textures */
       void *fs = get_passthrough_fs(st);
@@ -360,16 +189,48 @@ update_linkage( struct st_context *st )
    else {
       cso_set_fragment_shader_handle(st->cso_context, stfp->driver_shader);
    }
+}
+
+const struct st_tracked_state st_update_fp = {
+   "st_update_fp",                                     /* name */
+   {                                                   /* dirty */
+      0,                                               /* mesa */
+      ST_NEW_FRAGMENT_PROGRAM                           /* st */
+   },
+   update_fp                                   /* update */
+};
+
+
+
+
+static void
+update_vp( struct st_context *st )
+{
+   struct st_vertex_program *stvp;
 
-   st->vertex_result_to_slot = xvp->output_to_slot;
+   /* find active shader and params -- Should be covered by
+    * ST_NEW_VERTEX_PROGRAM
+    */
+   assert(st->ctx->VertexProgram._Current);
+   stvp = st_vertex_program(st->ctx->VertexProgram._Current);
+   assert(stvp->Base.Base.Target == GL_VERTEX_PROGRAM_ARB);
+
+   st->vp_varient = find_translated_vp(st, stvp);
+
+   st_reference_vertprog(st, &st->vp, stvp);
+
+   cso_set_vertex_shader_handle(st->cso_context, 
+                                st->vp_varient->driver_shader);
+
+   st->vertex_result_to_slot = stvp->result_to_output;
 }
 
 
-const struct st_tracked_state st_update_shader = {
-   "st_update_shader",                                 /* name */
+const struct st_tracked_state st_update_vp = {
+   "st_update_vp",                                     /* name */
    {                                                   /* dirty */
-      0,                                               /* mesa */
-      ST_NEW_VERTEX_PROGRAM | ST_NEW_FRAGMENT_PROGRAM  /* st */
+      _NEW_POLYGON,                                    /* mesa */
+      ST_NEW_VERTEX_PROGRAM | ST_NEW_EDGEFLAGS_DATA    /* st */
    },
-   update_linkage                                      /* update */
+   update_vp                                   /* update */
 };
index 4d4f97da7ed37b87717ca854cb5bfc9bf925de47..0b68447d2122364daf4ab5402ddfac344faba5af 100644 (file)
@@ -32,6 +32,8 @@
   */
  
 
+#include "main/macros.h"
+
 #include "st_context.h"
 #include "st_atom.h"
 #include "st_texture.h"
@@ -99,6 +101,12 @@ update_textures(struct st_context *st)
    cso_set_sampler_textures(st->cso_context,
                             st->state.num_textures,
                             st->state.sampler_texture);
+   if (st->ctx->Const.MaxVertexTextureImageUnits > 0) {
+      cso_set_vertex_sampler_textures(st->cso_context,
+                                      MIN2(st->state.num_textures,
+                                           st->ctx->Const.MaxVertexTextureImageUnits),
+                                      st->state.sampler_texture);
+   }
 }
 
 
index 1960d171bf64ed12fe5763701e98d24cafb31ac2..1bdeaccda3e1d83ae4e6432a6c36166c01fb7b21 100644 (file)
@@ -169,11 +169,6 @@ make_bitmap_fragment_program(GLcontext *ctx, GLuint samplerIndex)
    stfp = (struct st_fragment_program *) p;
    stfp->Base.UsesKill = GL_TRUE;
 
-   /* No need to send this incomplete program down to hardware:
-    *
-    * st_translate_fragment_program(ctx->st, stfp, NULL);
-    */
-
    return stfp;
 }
 
index e83b6c92efb7ed340a99df357fefdc86190ca0dc..72b30e7c0435882b07d776b014ac30e30bf242bd 100644 (file)
@@ -52,6 +52,7 @@
 #include "pipe/p_inlines.h"
 #include "pipe/p_state.h"
 #include "pipe/p_defines.h"
+#include "util/u_format.h"
 #include "util/u_pack_color.h"
 #include "util/u_simple_shaders.h"
 #include "util/u_draw_quad.h"
@@ -341,7 +342,7 @@ static INLINE GLboolean
 check_clear_depth_with_quad(GLcontext *ctx, struct gl_renderbuffer *rb)
 {
    const struct st_renderbuffer *strb = st_renderbuffer(rb);
-   const GLboolean isDS = pf_is_depth_and_stencil(strb->surface->format);
+   const GLboolean isDS = util_format_is_depth_and_stencil(strb->surface->format);
 
    if (ctx->Scissor.Enabled &&
        (ctx->Scissor.X != 0 ||
@@ -365,7 +366,7 @@ static INLINE GLboolean
 check_clear_stencil_with_quad(GLcontext *ctx, struct gl_renderbuffer *rb)
 {
    const struct st_renderbuffer *strb = st_renderbuffer(rb);
-   const GLboolean isDS = pf_is_depth_and_stencil(strb->surface->format);
+   const GLboolean isDS = util_format_is_depth_and_stencil(strb->surface->format);
    const GLuint stencilMax = 0xff;
    const GLboolean maskStencil
       = (ctx->Stencil.WriteMask[0] & stencilMax) != stencilMax;
index 1d33e81c2c9c889efc0d82476c9c50097ca52322..7c664267d4e49d59936b4441f5689abe13bf8670 100644 (file)
 #include "pipe/p_context.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
+#include "tgsi/tgsi_ureg.h"
 #include "util/u_tile.h"
 #include "util/u_draw_quad.h"
+#include "util/u_format.h"
 #include "util/u_math.h"
 #include "util/u_rect.h"
 #include "shader/prog_instruction.h"
@@ -236,78 +238,41 @@ make_fragment_shader_z(struct st_context *st)
  * Create a simple vertex shader that just passes through the
  * vertex position and texcoord (and optionally, color).
  */
-static struct st_vertex_program *
-st_make_passthrough_vertex_shader(struct st_context *st, GLboolean passColor)
+static void *
+st_make_passthrough_vertex_shader(struct st_context *st, 
+                                  GLboolean passColor)
 {
-   GLcontext *ctx = st->ctx;
-   struct st_vertex_program *stvp;
-   struct gl_program *p;
-   GLuint ic = 0;
-
-   if (st->drawpix.vert_shaders[passColor])
-      return st->drawpix.vert_shaders[passColor];
-
-   /*
-    * Create shader now
-    */
-   p = ctx->Driver.NewProgram(ctx, GL_VERTEX_PROGRAM_ARB, 0);
-   if (!p)
-      return NULL;
-
-   if (passColor)
-      p->NumInstructions = 4;
-   else
-      p->NumInstructions = 3;
-
-   p->Instructions = _mesa_alloc_instructions(p->NumInstructions);
-   if (!p->Instructions) {
-      ctx->Driver.DeleteProgram(ctx, p);
-      return NULL;
-   }
-   _mesa_init_instructions(p->Instructions, p->NumInstructions);
-   /* MOV result.pos, vertex.pos; */
-   p->Instructions[0].Opcode = OPCODE_MOV;
-   p->Instructions[0].DstReg.File = PROGRAM_OUTPUT;
-   p->Instructions[0].DstReg.Index = VERT_RESULT_HPOS;
-   p->Instructions[0].SrcReg[0].File = PROGRAM_INPUT;
-   p->Instructions[0].SrcReg[0].Index = VERT_ATTRIB_POS;
-   /* MOV result.texcoord0, vertex.texcoord0; */
-   p->Instructions[1].Opcode = OPCODE_MOV;
-   p->Instructions[1].DstReg.File = PROGRAM_OUTPUT;
-   p->Instructions[1].DstReg.Index = VERT_RESULT_TEX0;
-   p->Instructions[1].SrcReg[0].File = PROGRAM_INPUT;
-   p->Instructions[1].SrcReg[0].Index = VERT_ATTRIB_TEX0;
-   ic = 2;
-   if (passColor) {
-      /* MOV result.color0, vertex.color0; */
-      p->Instructions[ic].Opcode = OPCODE_MOV;
-      p->Instructions[ic].DstReg.File = PROGRAM_OUTPUT;
-      p->Instructions[ic].DstReg.Index = VERT_RESULT_COL0;
-      p->Instructions[ic].SrcReg[0].File = PROGRAM_INPUT;
-      p->Instructions[ic].SrcReg[0].Index = VERT_ATTRIB_COLOR0;
-      ic++;
-   }
-
-   /* END; */
-   p->Instructions[ic].Opcode = OPCODE_END;
-   ic++;
-
-   assert(ic == p->NumInstructions);
+   if (!st->drawpix.vert_shaders[passColor]) {
+      struct ureg_program *ureg = 
+         ureg_create( TGSI_PROCESSOR_VERTEX );
+
+      if (ureg == NULL)
+         return NULL;
+
+      /* MOV result.pos, vertex.pos; */
+      ureg_MOV(ureg, 
+               ureg_DECL_output( ureg, TGSI_SEMANTIC_POSITION, 0 ),
+               ureg_DECL_vs_input( ureg, 0 ));
+      
+      /* MOV result.texcoord0, vertex.texcoord0; */
+      ureg_MOV(ureg, 
+               ureg_DECL_output( ureg, TGSI_SEMANTIC_GENERIC, 0 ),
+               ureg_DECL_vs_input( ureg, 1 ));
+      
+      if (passColor) {
+         /* MOV result.color0, vertex.color0; */
+         ureg_MOV(ureg, 
+                  ureg_DECL_output( ureg, TGSI_SEMANTIC_COLOR, 0 ),
+                  ureg_DECL_vs_input( ureg, 2 ));
+      }
 
-   p->InputsRead = VERT_BIT_POS | VERT_BIT_TEX0;
-   p->OutputsWritten = ((1 << VERT_RESULT_TEX0) |
-                        (1 << VERT_RESULT_HPOS));
-   if (passColor) {
-      p->InputsRead |= VERT_BIT_COLOR0;
-      p->OutputsWritten |= (1 << VERT_RESULT_COL0);
+      ureg_END( ureg );
+      
+      st->drawpix.vert_shaders[passColor] = 
+         ureg_create_shader_and_destroy( ureg, st->pipe );
    }
 
-   stvp = (struct st_vertex_program *) p;
-   st_translate_vertex_program(st, stvp, NULL, NULL, NULL);
-
-   st->drawpix.vert_shaders[passColor] = stvp;
-
-   return stvp;
+   return st->drawpix.vert_shaders[passColor];
 }
 
 
@@ -539,8 +504,8 @@ draw_textured_quad(GLcontext *ctx, GLint x, GLint y, GLfloat z,
                    GLsizei width, GLsizei height,
                    GLfloat zoomX, GLfloat zoomY,
                    struct pipe_texture *pt,
-                   struct st_vertex_program *stvp,
-                   struct st_fragment_program *stfp,
+                   void *driver_vp,
+                   void *driver_fp,
                    const GLfloat *color,
                    GLboolean invertTex)
 {
@@ -575,10 +540,10 @@ draw_textured_quad(GLcontext *ctx, GLint x, GLint y, GLfloat z,
    }
 
    /* fragment shader state: TEX lookup program */
-   cso_set_fragment_shader_handle(cso, stfp->driver_shader);
+   cso_set_fragment_shader_handle(cso, driver_fp);
 
    /* vertex shader state: position + texcoord pass-through */
-   cso_set_vertex_shader_handle(cso, stvp->driver_shader);
+   cso_set_vertex_shader_handle(cso, driver_vp);
 
 
    /* texture sampling state: */
@@ -621,10 +586,10 @@ draw_textured_quad(GLcontext *ctx, GLint x, GLint y, GLfloat z,
       struct pipe_texture *textures[2];
       textures[0] = pt;
       textures[1] = st->pixel_xfer.pixelmap_texture;
-      pipe->set_sampler_textures(pipe, 2, textures);
+      pipe->set_fragment_sampler_textures(pipe, 2, textures);
    }
    else {
-      pipe->set_sampler_textures(pipe, 1, &pt);
+      pipe->set_fragment_sampler_textures(pipe, 1, &pt);
    }
 
    /* Compute window coords (y=0=bottom) with pixel zoom.
@@ -637,8 +602,8 @@ draw_textured_quad(GLcontext *ctx, GLint x, GLint y, GLfloat z,
    y1 = y + height * ctx->Pixel.ZoomY;
 
    draw_quad(ctx, x0, y0, z, x1, y1, color, invertTex,
-            (GLfloat) width / pt->width[0],
-            (GLfloat) height / pt->height[0]);
+            (GLfloat) width / pt->width0,
+            (GLfloat) height / pt->height0);
 
    /* restore state */
    cso_restore_rasterizer(cso);
@@ -683,7 +648,7 @@ draw_stencil_pixels(GLcontext *ctx, GLint x, GLint y,
    }
 
    if(format != GL_DEPTH_STENCIL && 
-      pf_get_component_bits( strb->format, PIPE_FORMAT_COMP_Z ) != 0)
+      util_format_get_component_bits(strb->format, UTIL_FORMAT_COLORSPACE_ZS, 0) != 0)
       usage = PIPE_TRANSFER_READ_WRITE;
    else
       usage = PIPE_TRANSFER_WRITE;
@@ -737,7 +702,7 @@ draw_stencil_pixels(GLcontext *ctx, GLint x, GLint y,
             }
 
             /* now pack the stencil (and Z) values in the dest format */
-            switch (pt->format) {
+            switch (pt->texture->format) {
             case PIPE_FORMAT_S8_UNORM:
                {
                   ubyte *dest = stmap + spanY * pt->stride + spanX;
@@ -806,7 +771,7 @@ st_DrawPixels(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
               const struct gl_pixelstore_attrib *unpack, const GLvoid *pixels)
 {
    struct st_fragment_program *stfp;
-   struct st_vertex_program *stvp;
+   void *driver_vp;
    struct st_context *st = st_context(ctx);
    struct pipe_surface *ps;
    const GLfloat *color;
@@ -826,13 +791,13 @@ st_DrawPixels(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
    if (format == GL_DEPTH_COMPONENT) {
       ps = st->state.framebuffer.zsbuf;
       stfp = make_fragment_shader_z(st);
-      stvp = st_make_passthrough_vertex_shader(st, GL_TRUE);
+      driver_vp = st_make_passthrough_vertex_shader(st, GL_TRUE);
       color = ctx->Current.RasterColor;
    }
    else {
       ps = st->state.framebuffer.cbufs[0];
       stfp = combined_drawpix_fragment_program(ctx);
-      stvp = st_make_passthrough_vertex_shader(st, GL_FALSE);
+      driver_vp = st_make_passthrough_vertex_shader(st, GL_FALSE);
       color = NULL;
    }
 
@@ -843,7 +808,10 @@ st_DrawPixels(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
       if (pt) {
          draw_textured_quad(ctx, x, y, ctx->Current.RasterPos[2],
                             width, height, ctx->Pixel.ZoomX, ctx->Pixel.ZoomY,
-                            pt, stvp, stfp, color, GL_FALSE);
+                            pt, 
+                            driver_vp, 
+                            stfp->driver_shader,
+                            color, GL_FALSE);
          pipe_texture_reference(&pt, NULL);
       }
    }
@@ -875,7 +843,7 @@ copy_stencil_pixels(GLcontext *ctx, GLint srcx, GLint srcy,
                           GL_STENCIL_INDEX, GL_UNSIGNED_BYTE,
                           &ctx->DefaultPacking, buffer);
 
-   if(pf_get_component_bits( rbDraw->format, PIPE_FORMAT_COMP_Z ) != 0)
+   if(util_format_get_component_bits(rbDraw->format, UTIL_FORMAT_COLORSPACE_ZS, 0) != 0)
       usage = PIPE_TRANSFER_READ_WRITE;
    else
       usage = PIPE_TRANSFER_WRITE;
@@ -889,8 +857,8 @@ copy_stencil_pixels(GLcontext *ctx, GLint srcx, GLint srcy,
                                           usage, dstx, dsty,
                                           width, height);
 
-   assert(ptDraw->block.width == 1);
-   assert(ptDraw->block.height == 1);
+   assert(util_format_get_blockwidth(ptDraw->texture->format) == 1);
+   assert(util_format_get_blockheight(ptDraw->texture->format) == 1);
 
    /* map the stencil buffer */
    drawMap = screen->transfer_map(screen, ptDraw);
@@ -911,7 +879,7 @@ copy_stencil_pixels(GLcontext *ctx, GLint srcx, GLint srcy,
       dst = drawMap + y * ptDraw->stride;
       src = buffer + i * width;
 
-      switch (ptDraw->format) {
+      switch (ptDraw->texture->format) {
       case PIPE_FORMAT_S8Z24_UNORM:
          {
             uint *dst4 = (uint *) dst;
@@ -960,7 +928,7 @@ st_CopyPixels(GLcontext *ctx, GLint srcx, GLint srcy,
    struct pipe_context *pipe = st->pipe;
    struct pipe_screen *screen = pipe->screen;
    struct st_renderbuffer *rbRead;
-   struct st_vertex_program *stvp;
+   void *driver_vp;
    struct st_fragment_program *stfp;
    struct pipe_texture *pt;
    GLfloat *color;
@@ -1009,14 +977,14 @@ st_CopyPixels(GLcontext *ctx, GLint srcx, GLint srcy,
       rbRead = st_get_color_read_renderbuffer(ctx);
       color = NULL;
       stfp = combined_drawpix_fragment_program(ctx);
-      stvp = st_make_passthrough_vertex_shader(st, GL_FALSE);
+      driver_vp = st_make_passthrough_vertex_shader(st, GL_FALSE);
    }
    else {
       assert(type == GL_DEPTH);
       rbRead = st_renderbuffer(ctx->ReadBuffer->_DepthBuffer);
       color = ctx->Current.Attrib[VERT_ATTRIB_COLOR0];
       stfp = make_fragment_shader_z(st);
-      stvp = st_make_passthrough_vertex_shader(st, GL_TRUE);
+      driver_vp = st_make_passthrough_vertex_shader(st, GL_TRUE);
    }
 
    srcFormat = rbRead->texture->format;
@@ -1116,7 +1084,7 @@ st_CopyPixels(GLcontext *ctx, GLint srcx, GLint srcy,
       if (ST_DEBUG & DEBUG_FALLBACK)
          debug_printf("%s: fallback processing\n", __FUNCTION__);
 
-      if (type == GL_DEPTH && pf_is_depth_and_stencil(pt->format))
+      if (type == GL_DEPTH && util_format_is_depth_and_stencil(pt->format))
          transfer_usage = PIPE_TRANSFER_READ_WRITE;
       else
          transfer_usage = PIPE_TRANSFER_WRITE;
@@ -1148,7 +1116,10 @@ st_CopyPixels(GLcontext *ctx, GLint srcx, GLint srcy,
    /* draw textured quad */
    draw_textured_quad(ctx, dstx, dsty, ctx->Current.RasterPos[2],
                       width, height, ctx->Pixel.ZoomX, ctx->Pixel.ZoomY,
-                      pt, stvp, stfp, color, GL_TRUE);
+                      pt, 
+                      driver_vp, 
+                      stfp->driver_shader,
+                      color, GL_TRUE);
 
    pipe_texture_reference(&pt, NULL);
 }
index 65ce12ccd4f8081aabbf3462073a2c29564d2d35..45ce34a85f7e58ba46bc08f88459563dfff53cd6 100644 (file)
@@ -49,6 +49,7 @@
 #include "st_public.h"
 #include "st_texture.h"
 
+#include "util/u_format.h"
 #include "util/u_rect.h"
 
 
@@ -98,16 +99,14 @@ st_renderbuffer_alloc_storage(GLcontext * ctx, struct gl_renderbuffer *rb,
    strb->defined = GL_FALSE;  /* undefined contents now */
 
    if(strb->software) {
-      struct pipe_format_block block;
       size_t size;
       
       _mesa_free(strb->data);
 
       assert(strb->format != PIPE_FORMAT_NONE);
-      pf_get_block(strb->format, &block);
       
-      strb->stride = pf_get_stride(&block, width);
-      size = pf_get_2d_size(&block, strb->stride, height);
+      strb->stride = util_format_get_stride(strb->format, width);
+      size = util_format_get_2d_size(strb->format, strb->stride, height);
       
       strb->data = _mesa_malloc(size);
       
@@ -127,13 +126,12 @@ st_renderbuffer_alloc_storage(GLcontext * ctx, struct gl_renderbuffer *rb,
       memset(&template, 0, sizeof(template));
       template.target = PIPE_TEXTURE_2D;
       template.format = format;
-      pf_get_block(format, &template.block);
-      template.width[0] = width;
-      template.height[0] = height;
-      template.depth[0] = 1;
+      template.width0 = width;
+      template.height0 = height;
+      template.depth0 = 1;
       template.last_level = 0;
       template.nr_samples = rb->NumSamples;
-      if (pf_is_depth_stencil(format)) {
+      if (util_format_is_depth_or_stencil(format)) {
          template.tex_usage = PIPE_TEXTURE_USAGE_DEPTH_STENCIL;
       }
       else {
@@ -376,7 +374,7 @@ st_render_texture(GLcontext *ctx,
    rb->_BaseFormat = texImage->_BaseFormat;
    /*printf("***** render to texture level %d: %d x %d\n", att->TextureLevel, rb->Width, rb->Height);*/
 
-   /*printf("***** pipe texture %d x %d\n", pt->width[0], pt->height[0]);*/
+   /*printf("***** pipe texture %d x %d\n", pt->width0, pt->height0);*/
 
    pipe_texture_reference( &strb->texture, pt );
 
index b2d5c39a3a0d139676432de96616e482d7308e57..8c276f8128edb2046d2d02346a46112826ac803f 100644 (file)
@@ -138,24 +138,7 @@ st_delete_program(GLcontext *ctx, struct gl_program *prog)
    case GL_VERTEX_PROGRAM_ARB:
       {
          struct st_vertex_program *stvp = (struct st_vertex_program *) prog;
-
-         if (stvp->driver_shader) {
-            cso_delete_vertex_shader(st->cso_context, stvp->driver_shader);
-            stvp->driver_shader = NULL;
-         }
-
-         if (stvp->draw_shader) {
-#if FEATURE_feedback || FEATURE_drawpix
-            /* this would only have been allocated for the RasterPos path */
-            draw_delete_vertex_shader(st->draw, stvp->draw_shader);
-            stvp->draw_shader = NULL;
-#endif
-         }
-
-         if (stvp->state.tokens) {
-            st_free_tokens(stvp->state.tokens);
-            stvp->state.tokens = NULL;
-         }
+         st_vp_release_varients( st, stvp );
       }
       break;
    case GL_FRAGMENT_PROGRAM_ARB:
@@ -177,8 +160,6 @@ st_delete_program(GLcontext *ctx, struct gl_program *prog)
             _mesa_reference_program(ctx, &prg, NULL);
             stfp->bitmap_program = NULL;
          }
-
-         st_free_translated_vertex_programs(st, stfp->vertex_programs);
       }
       break;
    default:
@@ -219,8 +200,6 @@ static void st_program_string_notify( GLcontext *ctx,
          stfp->state.tokens = NULL;
       }
 
-      stfp->param_state = stfp->Base.Base.Parameters->StateFlags;
-
       if (st->fp == stfp)
         st->dirty.st |= ST_NEW_FRAGMENT_PROGRAM;
    }
@@ -229,25 +208,7 @@ static void st_program_string_notify( GLcontext *ctx,
 
       stvp->serialNo++;
 
-      if (stvp->driver_shader) {
-         cso_delete_vertex_shader(st->cso_context, stvp->driver_shader);
-         stvp->driver_shader = NULL;
-      }
-
-      if (stvp->draw_shader) {
-#if FEATURE_feedback || FEATURE_drawpix
-         /* this would only have been allocated for the RasterPos path */
-         draw_delete_vertex_shader(st->draw, stvp->draw_shader);
-         stvp->draw_shader = NULL;
-#endif
-      }
-
-      if (stvp->state.tokens) {
-         st_free_tokens(stvp->state.tokens);
-         stvp->state.tokens = NULL;
-      }
-
-      stvp->param_state = stvp->Base.Base.Parameters->StateFlags;
+      st_vp_release_varients( st, stvp );
 
       if (st->vp == stvp)
         st->dirty.st |= ST_NEW_VERTEX_PROGRAM;
index 772bb3bb69e9184e320dcdcbe3e48ca788be4be9..6fa7bb64f2eb2cf16ce906ef58faebccae264329 100644 (file)
@@ -103,7 +103,7 @@ st_read_stencil_pixels(GLcontext *ctx, GLint x, GLint y,
       }
 
       /* get stencil (and Z) values */
-      switch (pt->format) {
+      switch (pt->texture->format) {
       case PIPE_FORMAT_S8_UNORM:
          {
             const ubyte *src = stmap + srcY * pt->stride;
@@ -243,7 +243,7 @@ st_fast_readpixels(GLcontext *ctx, struct st_renderbuffer *strb,
       GLint row, col, dy, dstStride;
 
       if (st_fb_orientation(ctx->ReadBuffer) == Y_0_TOP) {
-         y = strb->texture->height[0] - y - height;
+         y = strb->texture->height0 - y - height;
       }
 
       trans = st_cond_flush_get_tex_transfer(st_context(ctx), strb->texture,
@@ -431,8 +431,8 @@ st_readpixels(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
       const GLint dstStride = _mesa_image_row_stride(&clippedPacking, width,
                                                      format, type);
 
-      if (trans->format == PIPE_FORMAT_S8Z24_UNORM ||
-          trans->format == PIPE_FORMAT_X8Z24_UNORM) {
+      if (trans->texture->format == PIPE_FORMAT_S8Z24_UNORM ||
+          trans->texture->format == PIPE_FORMAT_X8Z24_UNORM) {
          if (format == GL_DEPTH_COMPONENT) {
             for (i = 0; i < height; i++) {
                GLuint ztemp[MAX_WIDTH];
@@ -463,8 +463,8 @@ st_readpixels(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
             }
          }
       }
-      else if (trans->format == PIPE_FORMAT_Z24S8_UNORM ||
-               trans->format == PIPE_FORMAT_Z24X8_UNORM) {
+      else if (trans->texture->format == PIPE_FORMAT_Z24S8_UNORM ||
+               trans->texture->format == PIPE_FORMAT_Z24X8_UNORM) {
          if (format == GL_DEPTH_COMPONENT) {
             for (i = 0; i < height; i++) {
                GLuint ztemp[MAX_WIDTH];
@@ -490,7 +490,7 @@ st_readpixels(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
             }
          }
       }
-      else if (trans->format == PIPE_FORMAT_Z16_UNORM) {
+      else if (trans->texture->format == PIPE_FORMAT_Z16_UNORM) {
          for (i = 0; i < height; i++) {
             GLushort ztemp[MAX_WIDTH];
             GLfloat zfloat[MAX_WIDTH];
@@ -505,7 +505,7 @@ st_readpixels(GLcontext *ctx, GLint x, GLint y, GLsizei width, GLsizei height,
             dst += dstStride;
          }
       }
-      else if (trans->format == PIPE_FORMAT_Z32_UNORM) {
+      else if (trans->texture->format == PIPE_FORMAT_Z32_UNORM) {
          for (i = 0; i < height; i++) {
             GLuint ztemp[MAX_WIDTH];
             GLfloat zfloat[MAX_WIDTH];
index d4630a514f8f2bf1bf4a00d2155c016a69712bfb..6e1ecb1c502887843cdbcbc8dd2ca7885ece4062 100644 (file)
@@ -62,6 +62,7 @@
 #include "pipe/p_shader_tokens.h"
 #include "util/u_tile.h"
 #include "util/u_blit.h"
+#include "util/u_format.h"
 #include "util/u_surface.h"
 #include "util/u_math.h"
 
@@ -204,7 +205,7 @@ static GLuint
 default_usage(enum pipe_format fmt)
 {
    GLuint usage = PIPE_TEXTURE_USAGE_SAMPLER;
-   if (pf_is_depth_stencil(fmt))
+   if (util_format_is_depth_or_stencil(fmt))
       usage |= PIPE_TEXTURE_USAGE_DEPTH_STENCIL;
    else
       usage |= PIPE_TEXTURE_USAGE_RENDER_TARGET;
@@ -405,10 +406,9 @@ compress_with_blit(GLcontext * ctx,
    memset(&templ, 0, sizeof(templ));
    templ.target = PIPE_TEXTURE_2D;
    templ.format = st_mesa_format_to_pipe_format(mesa_format);
-   pf_get_block(templ.format, &templ.block);
-   templ.width[0] = width;
-   templ.height[0] = height;
-   templ.depth[0] = 1;
+   templ.width0 = width;
+   templ.height0 = height;
+   templ.depth0 = 1;
    templ.last_level = 0;
    templ.tex_usage = PIPE_TEXTURE_USAGE_SAMPLER;
    src_tex = screen->texture_create(screen, &templ);
@@ -634,7 +634,7 @@ st_TexImage(GLcontext * ctx,
 
    if (stImage->pt) {
       if (format == GL_DEPTH_COMPONENT &&
-          pf_is_depth_and_stencil(stImage->pt->format))
+          util_format_is_depth_and_stencil(stImage->pt->format))
          transfer_usage = PIPE_TRANSFER_READ_WRITE;
       else
          transfer_usage = PIPE_TRANSFER_WRITE;
@@ -833,7 +833,7 @@ decompress_with_blit(GLcontext * ctx, GLenum target, GLint level,
    /* copy/pack data into user buffer */
    if (st_equal_formats(stImage->pt->format, format, type)) {
       /* memcpy */
-      const uint bytesPerRow = width * pf_get_size(stImage->pt->format);
+      const uint bytesPerRow = width * util_format_get_blocksize(stImage->pt->format);
       ubyte *map = screen->transfer_map(screen, tex_xfer);
       GLuint row;
       for (row = 0; row < height; row++) {
@@ -890,7 +890,7 @@ st_get_tex_image(GLcontext * ctx, GLenum target, GLint level,
    GLubyte *dest;
 
    if (stImage->pt &&
-       pf_is_compressed(stImage->pt->format) &&
+       util_format_is_compressed(stImage->pt->format) &&
        !compressed_dst) {
       /* Need to decompress the texture.
        * We'll do this by rendering a textured quad.
@@ -915,7 +915,7 @@ st_get_tex_image(GLcontext * ctx, GLenum target, GLint level,
                                             PIPE_TRANSFER_READ, 0, 0,
                                             stImage->base.Width,
                                             stImage->base.Height);
-      texImage->RowStride = stImage->transfer->stride / stImage->pt->block.size;
+      texImage->RowStride = stImage->transfer->stride / util_format_get_blocksize(stImage->pt->format);
    }
    else {
       /* Otherwise, the image should actually be stored in
@@ -1041,7 +1041,7 @@ st_TexSubimage(GLcontext *ctx, GLint dims, GLenum target, GLint level,
       unsigned face = _mesa_tex_target_to_face(target);
 
       if (format == GL_DEPTH_COMPONENT &&
-          pf_is_depth_and_stencil(stImage->pt->format))
+          util_format_is_depth_and_stencil(stImage->pt->format))
          transfer_usage = PIPE_TRANSFER_READ_WRITE;
       else
          transfer_usage = PIPE_TRANSFER_WRITE;
@@ -1163,10 +1163,10 @@ st_CompressedTexSubImage2D(GLcontext *ctx, GLenum target, GLint level,
                            struct gl_texture_image *texImage)
 {
    struct st_texture_image *stImage = st_texture_image(texImage);
-   struct pipe_format_block block;
    int srcBlockStride;
    int dstBlockStride;
    int y;
+   enum pipe_format pformat= stImage->pt->format;
 
    if (stImage->pt) {
       unsigned face = _mesa_tex_target_to_face(target);
@@ -1178,8 +1178,7 @@ st_CompressedTexSubImage2D(GLcontext *ctx, GLenum target, GLint level,
                                             xoffset, yoffset,
                                             width, height);
       
-      block = stImage->pt->block;
-      srcBlockStride = pf_get_stride(&block, width);
+      srcBlockStride = util_format_get_stride(pformat, width);
       dstBlockStride = stImage->transfer->stride;
    } else {
       assert(stImage->pt);
@@ -1193,16 +1192,16 @@ st_CompressedTexSubImage2D(GLcontext *ctx, GLenum target, GLint level,
       return;
    }
 
-   assert(xoffset % block.width == 0);
-   assert(yoffset % block.height == 0);
-   assert(width % block.width == 0);
-   assert(height % block.height == 0);
+   assert(xoffset % util_format_get_blockwidth(pformat) == 0);
+   assert(yoffset % util_format_get_blockheight(pformat) == 0);
+   assert(width % util_format_get_blockwidth(pformat) == 0);
+   assert(height % util_format_get_blockheight(pformat) == 0);
 
-   for (y = 0; y < height; y += block.height) {
+   for (y = 0; y < height; y += util_format_get_blockheight(pformat)) {
       /* don't need to adjust for xoffset and yoffset as st_texture_image_map does that */
-      const char *src = (const char*)data + srcBlockStride * pf_get_nblocksy(&block, y);
-      char *dst = (char*)texImage->Data + dstBlockStride * pf_get_nblocksy(&block, y);
-      memcpy(dst, src, pf_get_stride(&block, width));
+      const char *src = (const char*)data + srcBlockStride * util_format_get_nblocksy(pformat, y);
+      char *dst = (char*)texImage->Data + dstBlockStride * util_format_get_nblocksy(pformat, y);
+      memcpy(dst, src, util_format_get_stride(pformat, width));
    }
 
    if (stImage->pt) {
@@ -1266,7 +1265,7 @@ fallback_copy_texsubimage(GLcontext *ctx, GLenum target, GLint level,
 
    if ((baseFormat == GL_DEPTH_COMPONENT ||
         baseFormat == GL_DEPTH_STENCIL) &&
-       pf_is_depth_and_stencil(stImage->pt->format))
+       util_format_is_depth_and_stencil(stImage->pt->format))
       transfer_usage = PIPE_TRANSFER_READ_WRITE;
    else
       transfer_usage = PIPE_TRANSFER_WRITE;
@@ -1692,10 +1691,10 @@ copy_image_data_to_texture(struct st_context *st,
                             dstLevel,
                             stImage->base.Data,
                             stImage->base.RowStride * 
-                            stObj->pt->block.size,
+                            util_format_get_blocksize(stObj->pt->format),
                             stImage->base.RowStride *
                             stImage->base.Height *
-                            stObj->pt->block.size);
+                            util_format_get_blocksize(stObj->pt->format));
       _mesa_align_free(stImage->base.Data);
       stImage->base.Data = NULL;
    }
@@ -1761,10 +1760,9 @@ st_finalize_texture(GLcontext *ctx,
       if (stObj->pt->target != gl_target_to_pipe(stObj->base.Target) ||
           stObj->pt->format != fmt ||
           stObj->pt->last_level < stObj->lastLevel ||
-          stObj->pt->width[0] != firstImage->base.Width2 ||
-          stObj->pt->height[0] != firstImage->base.Height2 ||
-          stObj->pt->depth[0] != firstImage->base.Depth2 ||
-          stObj->pt->block.size != blockSize)
+          stObj->pt->width0 != firstImage->base.Width2 ||
+          stObj->pt->height0 != firstImage->base.Height2 ||
+          stObj->pt->depth0 != firstImage->base.Depth2)
       {
          pipe_texture_reference(&stObj->pt, NULL);
          ctx->st->dirty.st |= ST_NEW_FRAMEBUFFER;
index 18adb35e8721403b7217450823366ad73229cf09..831909a3f8fb6e07acd56239952cdb2a72eb6e6d 100644 (file)
@@ -55,6 +55,7 @@ struct bitmap_cache;
 #define ST_NEW_FRAGMENT_PROGRAM        0x2
 #define ST_NEW_VERTEX_PROGRAM          0x4
 #define ST_NEW_FRAMEBUFFER             0x8
+#define ST_NEW_EDGEFLAGS_DATA          0x10
 
 
 struct st_state_flags {
@@ -120,6 +121,7 @@ struct st_context
    struct st_state_flags dirty;
 
    GLboolean missing_textures;
+   GLboolean vertdata_edgeflags;
 
    /** Mapping from VERT_RESULT_x to post-transformed vertex slot */
    const GLuint *vertex_result_to_slot;
@@ -127,6 +129,8 @@ struct st_context
    struct st_vertex_program *vp;    /**< Currently bound vertex program */
    struct st_fragment_program *fp;  /**< Currently bound fragment program */
 
+   struct st_vp_varient *vp_varient;
+
    struct gl_texture_object *default_texture;
 
    struct {
index 3009cde9d5181c44c08546330121381790570a74..6e699ca5522477336456be0eacdeda12be9b2ef8 100644 (file)
@@ -86,7 +86,8 @@ st_print_current(void)
    }
 #endif
 
-   tgsi_dump( st->vp->state.tokens, 0 );
+   if (st->vp->varients)
+      tgsi_dump( st->vp->varients[0].state.tokens, 0 );
    if (st->vp->Base.Base.Parameters)
       _mesa_print_parameter_list(st->vp->Base.Base.Parameters);
 
index 5c6af1125a3da3988b8c89ab6064eb49d6af07db..e54f21be6009c7a1051127df9aeb68f1843c7a61 100644 (file)
@@ -217,59 +217,7 @@ st_pipe_vertex_format(GLenum type, GLuint size, GLenum format,
 }
 
 
-/*
- * If edge flags are needed, setup an bitvector of flags and call
- * pipe->set_edgeflags().
- * XXX memleak: need to free the returned pointer at some point
- */
-static void *
-setup_edgeflags(GLcontext *ctx, GLenum primMode, GLint start, GLint count,
-                const struct gl_client_array *array)
-{
-   struct pipe_context *pipe = ctx->st->pipe;
-
-   if ((primMode == GL_TRIANGLES ||
-        primMode == GL_QUADS ||
-        primMode == GL_POLYGON) &&
-       (ctx->Polygon.FrontMode != GL_FILL ||
-        ctx->Polygon.BackMode != GL_FILL)) {
-      /* need edge flags */
-      GLint i;
-      unsigned *vec;
-      struct st_buffer_object *stobj = st_buffer_object(array->BufferObj);
-      ubyte *map;
-
-      if (!stobj || stobj->Base.Name == 0) {
-         /* edge flags are not in a VBO */
-         return NULL;
-      }
-
-      vec = (unsigned *) _mesa_calloc(sizeof(unsigned) * ((count + 31) / 32));
-      if (!vec)
-         return NULL;
 
-      map = pipe_buffer_map(pipe->screen, stobj->buffer, PIPE_BUFFER_USAGE_CPU_READ);
-      map = ADD_POINTERS(map, array->Ptr);
-
-      for (i = 0; i < count; i++) {
-         if (*((float *) map))
-            vec[i/32] |= 1 << (i % 32);
-
-         map += array->StrideB;
-      }
-
-      pipe_buffer_unmap(pipe->screen, stobj->buffer);
-
-      pipe->set_edgeflags(pipe, vec);
-
-      return vec;
-   }
-   else {
-      /* edge flags not needed */
-      pipe->set_edgeflags(pipe, NULL);
-      return NULL;
-   }
-}
 
 
 /**
@@ -279,6 +227,7 @@ setup_edgeflags(GLcontext *ctx, GLenum primMode, GLint start, GLint count,
  */
 static GLboolean
 is_interleaved_arrays(const struct st_vertex_program *vp,
+                      const struct st_vp_varient *vpv,
                       const struct gl_client_array **arrays,
                       GLboolean *userSpace)
 {
@@ -288,7 +237,7 @@ is_interleaved_arrays(const struct st_vertex_program *vp,
    GLuint num_client_arrays = 0;
    const GLubyte *client_addr = NULL;
 
-   for (attr = 0; attr < vp->num_inputs; attr++) {
+   for (attr = 0; attr < vpv->num_inputs; attr++) {
       const GLuint mesaAttr = vp->index_to_input[attr];
       const struct gl_buffer_object *bufObj = arrays[mesaAttr]->BufferObj;
       const GLsizei stride = arrays[mesaAttr]->StrideB; /* in bytes */
@@ -321,7 +270,7 @@ is_interleaved_arrays(const struct st_vertex_program *vp,
       }
    }
 
-   *userSpace = (num_client_arrays == vp->num_inputs);
+   *userSpace = (num_client_arrays == vpv->num_inputs);
    /* printf("user space: %d (%d %d)\n", (int) *userSpace,num_client_arrays,vp->num_inputs); */
 
    return GL_TRUE;
@@ -333,15 +282,16 @@ is_interleaved_arrays(const struct st_vertex_program *vp,
  */
 static void
 get_arrays_bounds(const struct st_vertex_program *vp,
-                       const struct gl_client_array **arrays,
-                       GLuint max_index,
-                       const GLubyte **low, const GLubyte **high)
+                  const struct st_vp_varient *vpv,
+                  const struct gl_client_array **arrays,
+                  GLuint max_index,
+                  const GLubyte **low, const GLubyte **high)
 {
    const GLubyte *low_addr = NULL;
    const GLubyte *high_addr = NULL;
    GLuint attr;
 
-   for (attr = 0; attr < vp->num_inputs; attr++) {
+   for (attr = 0; attr < vpv->num_inputs; attr++) {
       const GLuint mesaAttr = vp->index_to_input[attr];
       const GLint stride = arrays[mesaAttr]->StrideB;
       const GLubyte *start = arrays[mesaAttr]->Ptr;
@@ -373,6 +323,7 @@ get_arrays_bounds(const struct st_vertex_program *vp,
 static void
 setup_interleaved_attribs(GLcontext *ctx,
                           const struct st_vertex_program *vp,
+                          const struct st_vp_varient *vpv,
                           const struct gl_client_array **arrays,
                           GLuint max_index,
                           GLboolean userSpace,
@@ -383,7 +334,7 @@ setup_interleaved_attribs(GLcontext *ctx,
    GLuint attr;
    const GLubyte *offset0 = NULL;
 
-   for (attr = 0; attr < vp->num_inputs; attr++) {
+   for (attr = 0; attr < vpv->num_inputs; attr++) {
       const GLuint mesaAttr = vp->index_to_input[attr];
       struct gl_buffer_object *bufobj = arrays[mesaAttr]->BufferObj;
       struct st_buffer_object *stobj = st_buffer_object(bufobj);
@@ -394,7 +345,7 @@ setup_interleaved_attribs(GLcontext *ctx,
       if (attr == 0) {
          const GLubyte *low, *high;
 
-         get_arrays_bounds(vp, arrays, max_index, &low, &high);
+         get_arrays_bounds(vp, vpv, arrays, max_index, &low, &high);
          /*printf("buffer range: %p %p  %d\n", low, high, high-low);*/
 
          offset0 = low;
@@ -435,6 +386,7 @@ setup_interleaved_attribs(GLcontext *ctx,
 static void
 setup_non_interleaved_attribs(GLcontext *ctx,
                               const struct st_vertex_program *vp,
+                              const struct st_vp_varient *vpv,
                               const struct gl_client_array **arrays,
                               GLuint max_index,
                               GLboolean *userSpace,
@@ -444,7 +396,7 @@ setup_non_interleaved_attribs(GLcontext *ctx,
    struct pipe_context *pipe = ctx->st->pipe;
    GLuint attr;
 
-   for (attr = 0; attr < vp->num_inputs; attr++) {
+   for (attr = 0; attr < vpv->num_inputs; attr++) {
       const GLuint mesaAttr = vp->index_to_input[attr];
       struct gl_buffer_object *bufobj = arrays[mesaAttr]->BufferObj;
       GLsizei stride = arrays[mesaAttr]->StrideB;
@@ -538,6 +490,20 @@ check_uniforms(GLcontext *ctx)
 }
 
 
+static unsigned translate_prim( GLcontext *ctx,
+                                unsigned prim )
+{
+   /* Avoid quadstrips if it's easy to do so:
+    */
+   if (prim == GL_QUAD_STRIP &&
+       ctx->Light.ShadeModel != GL_FLAT &&
+       ctx->Polygon.FrontMode == GL_FILL &&
+       ctx->Polygon.BackMode == GL_FILL)
+      prim = GL_TRIANGLE_STRIP;
+
+   return prim;
+}
+
 /**
  * This function gets plugged into the VBO module and is called when
  * we have something to render.
@@ -555,25 +521,36 @@ st_draw_vbo(GLcontext *ctx,
 {
    struct pipe_context *pipe = ctx->st->pipe;
    const struct st_vertex_program *vp;
+   const struct st_vp_varient *vpv;
    const struct pipe_shader_state *vs;
    struct pipe_vertex_buffer vbuffer[PIPE_MAX_SHADER_INPUTS];
    GLuint attr;
    struct pipe_vertex_element velements[PIPE_MAX_ATTRIBS];
    unsigned num_vbuffers, num_velements;
    GLboolean userSpace = GL_FALSE;
+   GLboolean vertDataEdgeFlags;
 
    /* Gallium probably doesn't want this in some cases. */
    if (!index_bounds_valid)
-      vbo_get_minmax_index(ctx, prims, ib, &min_index, &max_index);
+      if (!vbo_all_varyings_in_vbos(arrays))
+        vbo_get_minmax_index(ctx, prims, ib, &min_index, &max_index);
 
    /* sanity check for pointer arithmetic below */
    assert(sizeof(arrays[0]->Ptr[0]) == 1);
 
+   vertDataEdgeFlags = arrays[VERT_ATTRIB_EDGEFLAG]->BufferObj &&
+                       arrays[VERT_ATTRIB_EDGEFLAG]->BufferObj->Name;
+   if (vertDataEdgeFlags != ctx->st->vertdata_edgeflags) {
+      ctx->st->vertdata_edgeflags = vertDataEdgeFlags;
+      ctx->st->dirty.st |= ST_NEW_EDGEFLAGS_DATA;
+   }
+
    st_validate_state(ctx->st);
 
    /* must get these after state validation! */
    vp = ctx->st->vp;
-   vs = &ctx->st->vp->state;
+   vpv = ctx->st->vp_varient;
+   vs = &vpv->state;
 
 #if 0
    if (MESA_VERBOSE & VERBOSE_GLSL) {
@@ -586,21 +563,21 @@ st_draw_vbo(GLcontext *ctx,
    /*
     * Setup the vbuffer[] and velements[] arrays.
     */
-   if (is_interleaved_arrays(vp, arrays, &userSpace)) {
+   if (is_interleaved_arrays(vp, vpv, arrays, &userSpace)) {
       /*printf("Draw interleaved\n");*/
-      setup_interleaved_attribs(ctx, vp, arrays, max_index, userSpace,
+      setup_interleaved_attribs(ctx, vp, vpv, arrays, max_index, userSpace,
                                 vbuffer, velements);
       num_vbuffers = 1;
-      num_velements = vp->num_inputs;
+      num_velements = vpv->num_inputs;
       if (num_velements == 0)
          num_vbuffers = 0;
    }
    else {
       /*printf("Draw non-interleaved\n");*/
-      setup_non_interleaved_attribs(ctx, vp, arrays, max_index,
+      setup_non_interleaved_attribs(ctx, vp, vpv, arrays, max_index,
                                     &userSpace, vbuffer, velements);
-      num_vbuffers = vp->num_inputs;
-      num_velements = vp->num_inputs;
+      num_vbuffers = vpv->num_inputs;
+      num_velements = vpv->num_inputs;
    }
 
 #if 0
@@ -633,6 +610,7 @@ st_draw_vbo(GLcontext *ctx,
       struct gl_buffer_object *bufobj = ib->obj;
       struct pipe_buffer *indexBuf = NULL;
       unsigned indexSize, indexOffset, i;
+      unsigned prim;
 
       switch (ib->type) {
       case GL_UNSIGNED_INT:
@@ -671,24 +649,20 @@ st_draw_vbo(GLcontext *ctx,
           * through to driver & draw module.  These interfaces still
           * need a bit of work...
           */
-         setup_edgeflags(ctx, prims[i].mode,
-                         prims[i].start + indexOffset, prims[i].count,
-                         arrays[VERT_ATTRIB_EDGEFLAG]);
+         prim = translate_prim( ctx, prims[i].mode );
 
          pipe->draw_range_elements(pipe, indexBuf, indexSize,
                                    min_index,
                                    max_index,
-                                   prims[i].mode,
+                                   prim,
                                    prims[i].start + indexOffset, prims[i].count);
       }
       else {
          for (i = 0; i < nr_prims; i++) {
-            setup_edgeflags(ctx, prims[i].mode,
-                            prims[i].start + indexOffset, prims[i].count,
-                            arrays[VERT_ATTRIB_EDGEFLAG]);
+            prim = translate_prim( ctx, prims[i].mode );
             
             pipe->draw_elements(pipe, indexBuf, indexSize,
-                                prims[i].mode,
+                                prim,
                                 prims[i].start + indexOffset, prims[i].count);
          }
       }
@@ -698,12 +672,12 @@ st_draw_vbo(GLcontext *ctx,
    else {
       /* non-indexed */
       GLuint i;
+      GLuint prim;
+
       for (i = 0; i < nr_prims; i++) {
-         setup_edgeflags(ctx, prims[i].mode,
-                         prims[i].start, prims[i].count,
-                         arrays[VERT_ATTRIB_EDGEFLAG]);
+         prim = translate_prim( ctx, prims[i].mode );
 
-         pipe->draw_arrays(pipe, prims[i].mode, prims[i].start, prims[i].count);
+         pipe->draw_arrays(pipe, prim, prims[i].start, prims[i].count);
       }
    }
 
index b2d682ef640c52cedad1b4fdd39714f0725de683..cfc0caac983e37fb96175718e16356442fd8cca1 100644 (file)
@@ -120,10 +120,10 @@ st_feedback_draw_vbo(GLcontext *ctx,
 
    /* must get these after state validation! */
    vp = ctx->st->vp;
-   vs = &st->vp->state;
+   vs = &st->vp_varient->state;
 
-   if (!st->vp->draw_shader) {
-      st->vp->draw_shader = draw_create_vertex_shader(draw, vs);
+   if (!st->vp_varient->draw_shader) {
+      st->vp_varient->draw_shader = draw_create_vertex_shader(draw, vs);
    }
 
    /*
@@ -136,7 +136,7 @@ st_feedback_draw_vbo(GLcontext *ctx,
    draw_set_viewport_state(draw, &st->state.viewport);
    draw_set_clip_state(draw, &st->state.clip);
    draw_set_rasterizer_state(draw, &st->state.rasterizer);
-   draw_bind_vertex_shader(draw, st->vp->draw_shader);
+   draw_bind_vertex_shader(draw, st->vp_varient->draw_shader);
    set_feedback_vertex_format(ctx);
 
    /* loop over TGSI shader inputs to determine vertex buffer
@@ -241,7 +241,8 @@ st_feedback_draw_vbo(GLcontext *ctx,
    mapped_constants = pipe_buffer_map(pipe->screen,
                                       st->state.constants[PIPE_SHADER_VERTEX].buffer,
                                       PIPE_BUFFER_USAGE_CPU_READ);
-   draw_set_mapped_constant_buffer(st->draw, mapped_constants,
+   draw_set_mapped_constant_buffer(st->draw, PIPE_SHADER_VERTEX,
+                                   mapped_constants,
                                    st->state.constants[PIPE_SHADER_VERTEX].buffer->size);
 
 
index 57fe72d76a7c19dd45fa8f06a93f29a4ddc703be..ef3cbc53ee0c85d401bf866aa5dfc365c797d3a9 100644 (file)
@@ -92,6 +92,10 @@ void st_init_limits(struct st_context *st)
       = _min(screen->get_param(screen, PIPE_CAP_MAX_VERTEX_TEXTURE_UNITS),
              MAX_VERTEX_TEXTURE_IMAGE_UNITS);
 
+   c->MaxCombinedTextureImageUnits
+      = _min(screen->get_param(screen, PIPE_CAP_MAX_COMBINED_SAMPLERS),
+             MAX_COMBINED_TEXTURE_IMAGE_UNITS);
+
    c->MaxTextureCoordUnits
       = _min(c->MaxTextureImageUnits, MAX_TEXTURE_COORD_UNITS);
 
index 091f68190e591f605c0b49b0c1d1dd15a68f124a..3e165c9febd4ef453cf929cc6b9778f514c9d4a3 100644 (file)
 #include "pipe/p_context.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_screen.h"
+#include "util/u_format.h"
 #include "st_context.h"
 #include "st_format.h"
 
-static GLuint
-format_bits(
-   pipe_format_rgbazs_t  info,
-   GLuint comp )
-{
-   return pf_get_component_bits( (enum pipe_format) info, comp );
-}
 
 static GLuint
-format_max_bits(
-   pipe_format_rgbazs_t  info )
+format_max_bits(enum pipe_format format)
 {
-   GLuint   size = format_bits( info, PIPE_FORMAT_COMP_R );
+   GLuint size = util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0);
 
-   size = MAX2( size, format_bits( info, PIPE_FORMAT_COMP_G ) );
-   size = MAX2( size, format_bits( info, PIPE_FORMAT_COMP_B ) );
-   size = MAX2( size, format_bits( info, PIPE_FORMAT_COMP_A ) );
-   size = MAX2( size, format_bits( info, PIPE_FORMAT_COMP_Z ) );
-   size = MAX2( size, format_bits( info, PIPE_FORMAT_COMP_S ) );
+   size = MAX2(size, util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 1));
+   size = MAX2(size, util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 2));
+   size = MAX2(size, util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 3));
+   size = MAX2(size, util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_ZS, 0));
+   size = MAX2(size, util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_ZS, 1));
    return size;
 }
 
 static GLuint
-format_size(
-   pipe_format_rgbazs_t  info )
+format_size(enum pipe_format format)
 {
    return
-      format_bits( info, PIPE_FORMAT_COMP_R ) +
-      format_bits( info, PIPE_FORMAT_COMP_G ) +
-      format_bits( info, PIPE_FORMAT_COMP_B ) +
-      format_bits( info, PIPE_FORMAT_COMP_A ) +
-      format_bits( info, PIPE_FORMAT_COMP_Z ) +
-      format_bits( info, PIPE_FORMAT_COMP_S );
+      util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0) +
+      util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 1) +
+      util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 2) +
+      util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 3) +
+      util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_ZS, 0) +
+      util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_ZS, 1);
 }
 
 /*
@@ -86,11 +78,13 @@ format_size(
 GLboolean
 st_get_format_info(enum pipe_format format, struct pipe_format_info *pinfo)
 {
-   if (pf_layout(format) == PIPE_FORMAT_LAYOUT_RGBAZS) {
-      pipe_format_rgbazs_t info;
+   const struct util_format_description *desc;
 
-      info = format;
+   desc = util_format_description(format);
+   assert(desc);
 
+   if (desc->layout == UTIL_FORMAT_LAYOUT_ARITH ||
+       desc->layout == UTIL_FORMAT_LAYOUT_ARRAY) {
 #if 0
       printf("%s\n", pf_name( format ) );
 #endif
@@ -103,22 +97,22 @@ st_get_format_info(enum pipe_format format, struct pipe_format_info *pinfo)
          pinfo->datatype = GL_UNSIGNED_INT_24_8;
       }
       else {
-         const GLuint size = format_max_bits( info );
+         const GLuint size = format_max_bits(format);
          if (size == 8) {
-            if (pf_type(info) == PIPE_FORMAT_TYPE_UNORM)
+            if (desc->channel[0].type == UTIL_FORMAT_TYPE_UNSIGNED)
                pinfo->datatype = GL_UNSIGNED_BYTE;
             else
                pinfo->datatype = GL_BYTE;
          }
          else if (size == 16) {
-            if (pf_type(info) == PIPE_FORMAT_TYPE_UNORM)
+            if (desc->channel[0].type == UTIL_FORMAT_TYPE_UNSIGNED)
                pinfo->datatype = GL_UNSIGNED_SHORT;
             else
                pinfo->datatype = GL_SHORT;
          }
          else {
             assert( size <= 32 );
-            if (pf_type(info) == PIPE_FORMAT_TYPE_UNORM)
+            if (desc->channel[0].type == UTIL_FORMAT_TYPE_UNSIGNED)
                pinfo->datatype = GL_UNSIGNED_INT;
             else
                pinfo->datatype = GL_INT;
@@ -126,23 +120,23 @@ st_get_format_info(enum pipe_format format, struct pipe_format_info *pinfo)
       }
 
       /* Component bits */
-      pinfo->red_bits = format_bits( info, PIPE_FORMAT_COMP_R );
-      pinfo->green_bits = format_bits( info, PIPE_FORMAT_COMP_G );
-      pinfo->blue_bits = format_bits( info, PIPE_FORMAT_COMP_B );
-      pinfo->alpha_bits = format_bits( info, PIPE_FORMAT_COMP_A );
-      pinfo->depth_bits = format_bits( info, PIPE_FORMAT_COMP_Z );
-      pinfo->stencil_bits = format_bits( info, PIPE_FORMAT_COMP_S );
+      pinfo->red_bits = util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 0);
+      pinfo->green_bits = util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 1);
+      pinfo->blue_bits = util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 2);
+      pinfo->alpha_bits = util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 3);
+      pinfo->depth_bits = util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_ZS, 0);
+      pinfo->stencil_bits = util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_ZS, 1);
       pinfo->luminance_bits = 0;
       pinfo->intensity_bits = 0;
 
       /* Format size */
-      pinfo->size = format_size( info ) / 8;
+      pinfo->size = format_size(format) / 8;
 
       /* Luminance & Intensity bits */
-      if( pf_swizzle_x(info) == PIPE_FORMAT_COMP_R &&
-          pf_swizzle_y(info) == PIPE_FORMAT_COMP_R &&
-          pf_swizzle_z(info) == PIPE_FORMAT_COMP_R ) {
-         if( pf_swizzle_w(info) == PIPE_FORMAT_COMP_R ) {
+      if (desc->swizzle[0] == UTIL_FORMAT_SWIZZLE_X &&
+          desc->swizzle[1] == UTIL_FORMAT_SWIZZLE_X &&
+          desc->swizzle[2] == UTIL_FORMAT_SWIZZLE_X) {
+         if (desc->swizzle[3] == UTIL_FORMAT_SWIZZLE_X) {
             pinfo->intensity_bits = pinfo->red_bits;
          }
          else {
@@ -153,7 +147,7 @@ st_get_format_info(enum pipe_format format, struct pipe_format_info *pinfo)
 
       pinfo->mesa_format = st_pipe_format_to_mesa_format(format);
    }
-   else if (pf_layout(format) == PIPE_FORMAT_LAYOUT_YCBCR) {
+   else if (desc->layout == UTIL_FORMAT_LAYOUT_YUV) {
       pinfo->mesa_format = MESA_FORMAT_YCBCR;
       pinfo->datatype = GL_UNSIGNED_SHORT;
       pinfo->size = 2; /* two bytes per "texel" */
@@ -390,6 +384,33 @@ default_rgba_format(struct pipe_screen *screen,
    return PIPE_FORMAT_NONE;
 }
 
+/**
+ * Find an RGB format supported by the context/winsys.
+ */
+static enum pipe_format
+default_rgb_format(struct pipe_screen *screen, 
+                   enum pipe_texture_target target,
+                   unsigned tex_usage, 
+                   unsigned geom_flags)
+{
+   static const enum pipe_format colorFormats[] = {
+      PIPE_FORMAT_X8R8G8B8_UNORM,
+      PIPE_FORMAT_B8G8R8X8_UNORM,
+      PIPE_FORMAT_R8G8B8X8_UNORM,
+      PIPE_FORMAT_A8R8G8B8_UNORM,
+      PIPE_FORMAT_B8G8R8A8_UNORM,
+      PIPE_FORMAT_R8G8B8A8_UNORM,
+      PIPE_FORMAT_R5G6B5_UNORM
+   };
+   uint i;
+   for (i = 0; i < Elements(colorFormats); i++) {
+      if (screen->is_format_supported( screen, colorFormats[i], target, tex_usage, geom_flags )) {
+         return colorFormats[i];
+      }
+   }
+   return PIPE_FORMAT_NONE;
+}
+
 /**
  * Find an sRGBA format supported by the context/winsys.
  */
@@ -474,13 +495,14 @@ st_choose_format(struct pipe_screen *screen, GLenum internalFormat,
    case 4:
    case GL_RGBA:
    case GL_COMPRESSED_RGBA:
-   case 3:
-   case GL_RGB:
-   case GL_COMPRESSED_RGB:
    case GL_RGBA8:
    case GL_RGB10_A2:
    case GL_RGBA12:
       return default_rgba_format( screen, target, tex_usage, geom_flags );
+   case 3:
+   case GL_RGB:
+   case GL_COMPRESSED_RGB:
+      return default_rgb_format( screen, target, tex_usage, geom_flags );
    case GL_RGBA16:
       if (tex_usage & PIPE_TEXTURE_USAGE_RENDER_TARGET)
          return default_deep_rgba_format( screen, target, tex_usage, geom_flags );
@@ -502,7 +524,7 @@ st_choose_format(struct pipe_screen *screen, GLenum internalFormat,
    case GL_RGB10:
    case GL_RGB12:
    case GL_RGB16:
-      return default_rgba_format( screen, target, tex_usage, geom_flags );
+      return default_rgb_format( screen, target, tex_usage, geom_flags );
 
    case GL_RGB5:
    case GL_RGB4:
index 2d404d5f4224002c325a1c092af289c44979b3f3..2c283d464ae40e63e216fbfb8ca7aa879419d6bb 100644 (file)
@@ -37,7 +37,9 @@
 #include "pipe/p_context.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_gen_mipmap.h"
+#include "util/u_math.h"
 
 #include "cso_cache/cso_cache.h"
 #include "cso_cache/cso_context.h"
@@ -133,29 +135,33 @@ fallback_generate_mipmap(GLcontext *ctx, GLenum target,
       srcTrans = st_cond_flush_get_tex_transfer(st_context(ctx), pt, face,
                                                srcLevel, zslice,
                                                PIPE_TRANSFER_READ, 0, 0,
-                                               pt->width[srcLevel],
-                                               pt->height[srcLevel]);
+                                               u_minify(pt->width0, srcLevel),
+                                               u_minify(pt->height0, srcLevel));
 
       dstTrans = st_cond_flush_get_tex_transfer(st_context(ctx), pt, face,
                                                dstLevel, zslice,
                                                PIPE_TRANSFER_WRITE, 0, 0,
-                                               pt->width[dstLevel],
-                                               pt->height[dstLevel]);
+                                               u_minify(pt->width0, dstLevel),
+                                               u_minify(pt->height0, dstLevel));
 
       srcData = (ubyte *) screen->transfer_map(screen, srcTrans);
       dstData = (ubyte *) screen->transfer_map(screen, dstTrans);
 
-      srcStride = srcTrans->stride / srcTrans->block.size;
-      dstStride = dstTrans->stride / dstTrans->block.size;
+      srcStride = srcTrans->stride / util_format_get_blocksize(srcTrans->texture->format);
+      dstStride = dstTrans->stride / util_format_get_blocksize(dstTrans->texture->format);
 
       _mesa_generate_mipmap_level(target, datatype, comps,
-                   0 /*border*/,
-                   pt->width[srcLevel], pt->height[srcLevel], pt->depth[srcLevel],
-                   srcData,
-                   srcStride, /* stride in texels */
-                   pt->width[dstLevel], pt->height[dstLevel], pt->depth[dstLevel],
-                   dstData,
-                   dstStride); /* stride in texels */
+                                  0 /*border*/,
+                                  u_minify(pt->width0, srcLevel),
+                                  u_minify(pt->height0, srcLevel),
+                                  u_minify(pt->depth0, srcLevel),
+                                  srcData,
+                                  srcStride, /* stride in texels */
+                                  u_minify(pt->width0, dstLevel),
+                                  u_minify(pt->height0, dstLevel),
+                                  u_minify(pt->depth0, dstLevel),
+                                  dstData,
+                                  dstStride); /* stride in texels */
 
       screen->transfer_unmap(screen, srcTrans);
       screen->transfer_unmap(screen, dstTrans);
@@ -235,9 +241,9 @@ st_generate_mipmap(GLcontext *ctx, GLenum target,
                                     oldTex->target,
                                     oldTex->format,
                                     lastLevel,
-                                    oldTex->width[0],
-                                    oldTex->height[0],
-                                    oldTex->depth[0],
+                                    oldTex->width0,
+                                    oldTex->height0,
+                                    oldTex->depth0,
                                     oldTex->tex_usage);
 
       /* The texture isn't in a "complete" state yet so set the expected
@@ -272,9 +278,9 @@ st_generate_mipmap(GLcontext *ctx, GLenum target,
          = _mesa_get_tex_image(ctx, texObj, target, srcLevel);
       struct gl_texture_image *dstImage;
       struct st_texture_image *stImage;
-      uint dstWidth = pt->width[dstLevel];
-      uint dstHeight = pt->height[dstLevel];
-      uint dstDepth = pt->depth[dstLevel];
+      uint dstWidth = u_minify(pt->width0, dstLevel);
+      uint dstHeight = u_minify(pt->height0, dstLevel);
+      uint dstDepth = u_minify(pt->depth0, dstLevel); 
       uint border = srcImage->Border;
 
       dstImage = _mesa_get_tex_image(ctx, texObj, target, dstLevel);
index bd94c9d79e8054afff6a9bede8f014ed49b80dbb..5c9be46a776fffe6fe02f320e73577b8742ba091 100644 (file)
@@ -396,6 +396,23 @@ static void emit_swz( struct st_translate *t,
 }
 
 
+/**
+ * Negate the value of DDY to match GL semantics where (0,0) is the
+ * lower-left corner of the window.
+ * Note that the GL_ARB_fragment_coord_conventions extension will
+ * effect this someday.
+ */
+static void emit_ddy( struct st_translate *t,
+                      struct ureg_dst dst,
+                      const struct prog_src_register *SrcReg )
+{
+   struct ureg_program *ureg = t->ureg;
+   struct ureg_src src = translate_src( t, SrcReg );
+   src = ureg_negate( src );
+   ureg_DDY( ureg, dst, src );
+}
+
+
 
 static unsigned
 translate_opcode( unsigned op )
@@ -619,7 +636,9 @@ compile_instruction(
       ureg_MOV( ureg, dst[0], ureg_imm1f(ureg, 0.5) );
       break;
                 
-
+   case OPCODE_DDY:
+      emit_ddy( t, dst[0], &inst->SrcReg[0] );
+      break;
 
    default:
       ureg_insn( ureg, 
@@ -675,6 +694,41 @@ emit_inverted_wpos( struct st_translate *t,
 }
 
 
+/**
+ * OpenGL's fragment gl_FrontFace input is 1 for front-facing, 0 for back.
+ * TGSI uses +1 for front, -1 for back.
+ * This function converts the TGSI value to the GL value.  Simply clamping/
+ * saturating the value to [0,1] does the job.
+ */
+static void
+emit_face_var( struct st_translate *t,
+               const struct gl_program *program )
+{
+   struct ureg_program *ureg = t->ureg;
+   struct ureg_dst face_temp = ureg_DECL_temporary( ureg );
+   struct ureg_src face_input = t->inputs[t->inputMapping[FRAG_ATTRIB_FACE]];
+
+   /* MOV_SAT face_temp, input[face]
+    */
+   face_temp = ureg_saturate( face_temp );
+   ureg_MOV( ureg, face_temp, face_input );
+
+   /* Use face_temp as face input from here on:
+    */
+   t->inputs[t->inputMapping[FRAG_ATTRIB_FACE]] = ureg_src(face_temp);
+}
+
+static void
+emit_edgeflags( struct st_translate *t,
+                 const struct gl_program *program )
+{
+   struct ureg_program *ureg = t->ureg;
+   struct ureg_dst edge_dst = t->outputs[t->outputMapping[VERT_RESULT_EDGE]];
+   struct ureg_src edge_src = t->inputs[t->inputMapping[VERT_ATTRIB_EDGEFLAG]];
+
+   ureg_MOV( ureg, edge_dst, edge_src );
+}
+
 /**
  * Translate Mesa program to TGSI format.
  * \param program  the program to translate
@@ -694,26 +748,24 @@ emit_inverted_wpos( struct st_translate *t,
  *
  * \return  array of translated tokens, caller's responsibility to free
  */
-const struct tgsi_token *
+enum pipe_error
 st_translate_mesa_program(
    GLcontext *ctx,
    uint procType,
+   struct ureg_program *ureg,
    const struct gl_program *program,
    GLuint numInputs,
    const GLuint inputMapping[],
    const ubyte inputSemanticName[],
    const ubyte inputSemanticIndex[],
    const GLuint interpMode[],
-   const GLbitfield inputFlags[],
    GLuint numOutputs,
    const GLuint outputMapping[],
    const ubyte outputSemanticName[],
    const ubyte outputSemanticIndex[],
-   const GLbitfield outputFlags[] )
+   boolean passthrough_edgeflags )
 {
    struct st_translate translate, *t;
-   struct ureg_program *ureg;
-   const struct tgsi_token *tokens = NULL;
    unsigned i;
 
    t = &translate;
@@ -722,11 +774,7 @@ st_translate_mesa_program(
    t->procType = procType;
    t->inputMapping = inputMapping;
    t->outputMapping = outputMapping;
-   t->ureg = ureg_create( procType );
-   if (t->ureg == NULL)
-      return NULL;
-
-   ureg = t->ureg;
+   t->ureg = ureg;
 
    /*_mesa_print_program(program);*/
 
@@ -748,6 +796,10 @@ st_translate_mesa_program(
          emit_inverted_wpos( t, program );
       }
 
+      if (program->InputsRead & FRAG_BIT_FACE) {
+         emit_face_var( t, program );
+      }
+
       /*
        * Declare output attributes.
        */
@@ -782,6 +834,8 @@ st_translate_mesa_program(
                                            outputSemanticName[i],
                                            outputSemanticIndex[i] );
       }
+      if (passthrough_edgeflags)
+         emit_edgeflags( t, program );
    }
 
    /* Declare address register.
@@ -853,8 +907,7 @@ st_translate_mesa_program(
                         t->insn[t->labels[i].branch_target] );
    }
 
-   tokens = ureg_get_tokens( ureg, NULL );
-   ureg_destroy( ureg );
+   return PIPE_OK;
 
 out:
    FREE(t->insn);
@@ -863,17 +916,9 @@ out:
 
    if (t->error) {
       debug_printf("%s: translate error flag set\n", __FUNCTION__);
-      FREE((void *)tokens);
-      tokens = NULL;
-   }
-
-   if (!tokens) {
-      debug_printf("%s: failed to translate Mesa program:\n", __FUNCTION__);
-      _mesa_print_program(program);
-      debug_assert(0);
    }
 
-   return tokens;
+   return PIPE_ERROR_OUT_OF_MEMORY;
 }
 
 
index c0d1ff59e1f9f76c7c193ebb6e1c24d08c3f0afa..e3c5bd1d94db6e3d6f2e78cab0306e2146d9ac88 100644 (file)
@@ -30,6 +30,7 @@
 #define ST_MESA_TO_TGSI_H
 
 #include "main/mtypes.h"
+#include "tgsi/tgsi_ureg.h"
 
 
 #if defined __cplusplus
@@ -39,22 +40,22 @@ extern "C" {
 struct tgsi_token;
 struct gl_program;
 
-const struct tgsi_token *
+enum pipe_error
 st_translate_mesa_program(
    GLcontext *ctx,
    uint procType,
+   struct ureg_program *ureg,
    const struct gl_program *program,
    GLuint numInputs,
    const GLuint inputMapping[],
    const ubyte inputSemanticName[],
    const ubyte inputSemanticIndex[],
    const GLuint interpMode[],
-   const GLbitfield inputFlags[],
    GLuint numOutputs,
    const GLuint outputMapping[],
    const ubyte outputSemanticName[],
    const ubyte outputSemanticIndex[],
-   const GLbitfield outputFlags[] );
+   boolean passthrough_edgeflags );
 
 void
 st_free_tokens(const struct tgsi_token *tokens);
index cf19f8f1a8e48049df9fa94f546110d37616ed1b..6a869fae90443455d20b01c90ecc96b43838629e 100644 (file)
 #include "cso_cache/cso_context.h"
 
 
+
+/**
+ * Clean out any old compilations:
+ */
+void
+st_vp_release_varients( struct st_context *st,
+                        struct st_vertex_program *stvp )
+{
+   struct st_vp_varient *vpv;
+
+   for (vpv = stvp->varients; vpv; ) {
+      struct st_vp_varient *next = vpv->next;
+
+      if (vpv->driver_shader) 
+         cso_delete_vertex_shader(st->cso_context, vpv->driver_shader);
+      
+      if (vpv->draw_shader)
+         draw_delete_vertex_shader( st->draw, vpv->draw_shader );
+      
+      if (vpv->state.tokens)
+         st_free_tokens(vpv->state.tokens);
+      
+      FREE( vpv );
+
+      vpv = next;
+   }
+
+   stvp->varients = NULL;
+}
+
+
+
+
 /**
  * Translate a Mesa vertex shader into a TGSI shader.
  * \param outputMapping  to map vertex program output registers (VERT_RESULT_x)
  * \return  pointer to cached pipe_shader object.
  */
 void
-st_translate_vertex_program(struct st_context *st,
-                            struct st_vertex_program *stvp,
-                            const GLuint outputMapping[],
-                            const ubyte *outputSemanticName,
-                            const ubyte *outputSemanticIndex)
+st_prepare_vertex_program(struct st_context *st,
+                            struct st_vertex_program *stvp)
 {
-   struct pipe_context *pipe = st->pipe;
-   GLuint defaultOutputMapping[VERT_RESULT_MAX];
-   GLuint attr, i;
-   GLuint num_generic = 0;
-
-   ubyte vs_input_semantic_name[PIPE_MAX_SHADER_INPUTS];
-   ubyte vs_input_semantic_index[PIPE_MAX_SHADER_INPUTS];
-   uint vs_num_inputs = 0;
-
-   ubyte vs_output_semantic_name[PIPE_MAX_SHADER_OUTPUTS];
-   ubyte vs_output_semantic_index[PIPE_MAX_SHADER_OUTPUTS];
-   uint vs_num_outputs = 0;
-
-   GLbitfield input_flags[MAX_PROGRAM_INPUTS];
-   GLbitfield output_flags[MAX_PROGRAM_OUTPUTS];
+   GLuint attr;
 
-   /*memset(&vs, 0, sizeof(vs));*/
-   memset(input_flags, 0, sizeof(input_flags));
-   memset(output_flags, 0, sizeof(output_flags));
+   stvp->num_inputs = 0;
+   stvp->num_outputs = 0;
 
    if (stvp->Base.IsPositionInvariant)
       _mesa_insert_mvp_code(st->ctx, &stvp->Base);
@@ -95,162 +110,59 @@ st_translate_vertex_program(struct st_context *st,
     */
    for (attr = 0; attr < VERT_ATTRIB_MAX; attr++) {
       if (stvp->Base.Base.InputsRead & (1 << attr)) {
-         const GLuint slot = vs_num_inputs;
-
-         vs_num_inputs++;
-
-         stvp->input_to_index[attr] = slot;
-         stvp->index_to_input[slot] = attr;
-
-         switch (attr) {
-         case VERT_ATTRIB_POS:
-            vs_input_semantic_name[slot] = TGSI_SEMANTIC_POSITION;
-            vs_input_semantic_index[slot] = 0;
-            break;
-         case VERT_ATTRIB_WEIGHT:
-            /* fall-through */
-         case VERT_ATTRIB_NORMAL:
-            /* just label as a generic */
-            vs_input_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-            vs_input_semantic_index[slot] = 0;
-            break;
-         case VERT_ATTRIB_COLOR0:
-            vs_input_semantic_name[slot] = TGSI_SEMANTIC_COLOR;
-            vs_input_semantic_index[slot] = 0;
-            break;
-         case VERT_ATTRIB_COLOR1:
-            vs_input_semantic_name[slot] = TGSI_SEMANTIC_COLOR;
-            vs_input_semantic_index[slot] = 1;
-            break;
-         case VERT_ATTRIB_FOG:
-            vs_input_semantic_name[slot] = TGSI_SEMANTIC_FOG;
-            vs_input_semantic_index[slot] = 0;
-            break;
-         case VERT_ATTRIB_POINT_SIZE:
-            vs_input_semantic_name[slot] = TGSI_SEMANTIC_PSIZE;
-            vs_input_semantic_index[slot] = 0;
-            break;
-         case VERT_ATTRIB_TEX0:
-         case VERT_ATTRIB_TEX1:
-         case VERT_ATTRIB_TEX2:
-         case VERT_ATTRIB_TEX3:
-         case VERT_ATTRIB_TEX4:
-         case VERT_ATTRIB_TEX5:
-         case VERT_ATTRIB_TEX6:
-         case VERT_ATTRIB_TEX7:
-            assert(slot < Elements(vs_input_semantic_name));
-            vs_input_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-            vs_input_semantic_index[slot] = num_generic++;
-            break;
-         case VERT_ATTRIB_GENERIC0:
-         case VERT_ATTRIB_GENERIC1:
-         case VERT_ATTRIB_GENERIC2:
-         case VERT_ATTRIB_GENERIC3:
-         case VERT_ATTRIB_GENERIC4:
-         case VERT_ATTRIB_GENERIC5:
-         case VERT_ATTRIB_GENERIC6:
-         case VERT_ATTRIB_GENERIC7:
-         case VERT_ATTRIB_GENERIC8:
-         case VERT_ATTRIB_GENERIC9:
-         case VERT_ATTRIB_GENERIC10:
-         case VERT_ATTRIB_GENERIC11:
-         case VERT_ATTRIB_GENERIC12:
-         case VERT_ATTRIB_GENERIC13:
-         case VERT_ATTRIB_GENERIC14:
-         case VERT_ATTRIB_GENERIC15:
-            assert(attr < VERT_ATTRIB_MAX);
-            assert(slot < Elements(vs_input_semantic_name));
-            vs_input_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-            vs_input_semantic_index[slot] = num_generic++;
-            break;
-         default:
-            assert(0);
-         }
-
-         input_flags[slot] = stvp->Base.Base.InputFlags[attr];
+         stvp->input_to_index[attr] = stvp->num_inputs;
+         stvp->index_to_input[stvp->num_inputs] = attr;
+         stvp->num_inputs++;
       }
    }
+   /* bit of a hack, presetup potentially unused edgeflag input */
+   stvp->input_to_index[VERT_ATTRIB_EDGEFLAG] = stvp->num_inputs;
+   stvp->index_to_input[stvp->num_inputs] = VERT_ATTRIB_EDGEFLAG;
 
-#if 0
-   if (outputMapping && outputSemanticName) {
-      printf("VERT_RESULT  written  out_slot  semantic_name  semantic_index\n");
-      for (attr = 0; attr < VERT_RESULT_MAX; attr++) {
-         printf("    %-2d          %c       %3d          %2d              %2d\n",
-                attr, 
-                ((stvp->Base.Base.OutputsWritten & (1 << attr)) ? 'Y' : ' '),
-                outputMapping[attr],
-                outputSemanticName[attr],
-                outputSemanticIndex[attr]);
-      }
-   }
-#endif
-
-   /* initialize output semantics to defaults */
-   for (i = 0; i < PIPE_MAX_SHADER_OUTPUTS; i++) {
-      assert(i < Elements(vs_output_semantic_name));
-      vs_output_semantic_name[i] = TGSI_SEMANTIC_GENERIC;
-      vs_output_semantic_index[i] = 0;
-      output_flags[i] = 0x0;
-   }
-
-   num_generic = 0;
-   /*
-    * Determine number of outputs, the (default) output register
-    * mapping and the semantic information for each output.
+   /* Compute mapping of vertex program outputs to slots.
     */
    for (attr = 0; attr < VERT_RESULT_MAX; attr++) {
-      if (stvp->Base.Base.OutputsWritten & (1 << attr)) {
-         GLuint slot;
-
-         /* XXX
-          * Pass in the fragment program's input's semantic info.
-          * Use the generic semantic indexes from there, instead of
-          * guessing below.
-          */
-
-         if (outputMapping) {
-            slot = outputMapping[attr];
-            assert(slot != ~0);
-         }
-         else {
-            slot = vs_num_outputs;
-            vs_num_outputs++;
-            defaultOutputMapping[attr] = slot;
-         }
+      if ((stvp->Base.Base.OutputsWritten & (1 << attr)) == 0) {
+         stvp->result_to_output[attr] = ~0;
+      }
+      else {
+         unsigned slot = stvp->num_outputs++;
+
+         stvp->result_to_output[attr] = slot;
 
          switch (attr) {
          case VERT_RESULT_HPOS:
-            assert(slot == 0);
-            vs_output_semantic_name[slot] = TGSI_SEMANTIC_POSITION;
-            vs_output_semantic_index[slot] = 0;
+            stvp->output_semantic_name[slot] = TGSI_SEMANTIC_POSITION;
+            stvp->output_semantic_index[slot] = 0;
             break;
          case VERT_RESULT_COL0:
-            vs_output_semantic_name[slot] = TGSI_SEMANTIC_COLOR;
-            vs_output_semantic_index[slot] = 0;
+            stvp->output_semantic_name[slot] = TGSI_SEMANTIC_COLOR;
+            stvp->output_semantic_index[slot] = 0;
             break;
          case VERT_RESULT_COL1:
-            vs_output_semantic_name[slot] = TGSI_SEMANTIC_COLOR;
-            vs_output_semantic_index[slot] = 1;
+            stvp->output_semantic_name[slot] = TGSI_SEMANTIC_COLOR;
+            stvp->output_semantic_index[slot] = 1;
             break;
          case VERT_RESULT_BFC0:
-            vs_output_semantic_name[slot] = TGSI_SEMANTIC_BCOLOR;
-            vs_output_semantic_index[slot] = 0;
+            stvp->output_semantic_name[slot] = TGSI_SEMANTIC_BCOLOR;
+            stvp->output_semantic_index[slot] = 0;
             break;
          case VERT_RESULT_BFC1:
-            vs_output_semantic_name[slot] = TGSI_SEMANTIC_BCOLOR;
-            vs_output_semantic_index[slot] = 1;
+            stvp->output_semantic_name[slot] = TGSI_SEMANTIC_BCOLOR;
+            stvp->output_semantic_index[slot] = 1;
             break;
          case VERT_RESULT_FOGC:
-            vs_output_semantic_name[slot] = TGSI_SEMANTIC_FOG;
-            vs_output_semantic_index[slot] = 0;
+            stvp->output_semantic_name[slot] = TGSI_SEMANTIC_FOG;
+            stvp->output_semantic_index[slot] = 0;
             break;
          case VERT_RESULT_PSIZ:
-            vs_output_semantic_name[slot] = TGSI_SEMANTIC_PSIZE;
-            vs_output_semantic_index[slot] = 0;
+            stvp->output_semantic_name[slot] = TGSI_SEMANTIC_PSIZE;
+            stvp->output_semantic_index[slot] = 0;
             break;
          case VERT_RESULT_EDGE:
             assert(0);
             break;
+
          case VERT_RESULT_TEX0:
          case VERT_RESULT_TEX1:
          case VERT_RESULT_TEX2:
@@ -259,92 +171,79 @@ st_translate_vertex_program(struct st_context *st,
          case VERT_RESULT_TEX5:
          case VERT_RESULT_TEX6:
          case VERT_RESULT_TEX7:
-            /* fall-through */
+            stvp->output_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
+            stvp->output_semantic_index[slot] = attr - VERT_RESULT_TEX0;
+            break;
+
          case VERT_RESULT_VAR0:
-            /* fall-through */
          default:
-            assert(slot < Elements(vs_output_semantic_name));
-            if (outputSemanticName) {
-               /* use provided semantic into */
-               assert(outputSemanticName[attr] != TGSI_SEMANTIC_COUNT);
-               vs_output_semantic_name[slot] = outputSemanticName[attr];
-               vs_output_semantic_index[slot] = outputSemanticIndex[attr];
-            }
-            else {
-               /* use default semantic info */
-               vs_output_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-               vs_output_semantic_index[slot] = num_generic++;
-            }
+            assert(attr < VERT_RESULT_MAX);
+            stvp->output_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
+            stvp->output_semantic_index[slot] = (FRAG_ATTRIB_VAR0 - 
+                                                FRAG_ATTRIB_TEX0 +
+                                                attr - 
+                                                VERT_RESULT_VAR0);
+            break;
          }
-
-         assert(slot < Elements(output_flags));
-         output_flags[slot] = stvp->Base.Base.OutputFlags[attr];
       }
    }
+   /* similar hack to above, presetup potentially unused edgeflag output */
+   stvp->result_to_output[VERT_RESULT_EDGE] = stvp->num_outputs;
+   stvp->output_semantic_name[stvp->num_outputs] = TGSI_SEMANTIC_EDGEFLAG;
+   stvp->output_semantic_index[stvp->num_outputs] = 0;
+}
 
-   if (outputMapping) {
-      /* find max output slot referenced to compute vs_num_outputs */
-      GLuint maxSlot = 0;
-      for (attr = 0; attr < VERT_RESULT_MAX; attr++) {
-         if (outputMapping[attr] != ~0 && outputMapping[attr] > maxSlot)
-            maxSlot = outputMapping[attr];
-      }
-      vs_num_outputs = maxSlot + 1;
-   }
-   else {
-      outputMapping = defaultOutputMapping;
-   }
 
-#if 0 /* debug */
-   {
-      GLuint i;
-      printf("outputMapping? %d\n", outputMapping ? 1 : 0);
-      if (outputMapping) {
-         printf("attr -> slot\n");
-         for (i = 0; i < 16;  i++) {
-            printf(" %2d       %3d\n", i, outputMapping[i]);
-         }
-      }
-      printf("slot    sem_name  sem_index\n");
-      for (i = 0; i < vs_num_outputs; i++) {
-         printf(" %2d         %d         %d\n",
-                i,
-                vs_output_semantic_name[i],
-                vs_output_semantic_index[i]);
-      }
-   }
-#endif
-
-   /* free old shader state, if any */
-   if (stvp->state.tokens) {
-      st_free_tokens(stvp->state.tokens);
-      stvp->state.tokens = NULL;
-   }
-   if (stvp->driver_shader) {
-      cso_delete_vertex_shader(st->cso_context, stvp->driver_shader);
-      stvp->driver_shader = NULL;
+struct st_vp_varient *
+st_translate_vertex_program(struct st_context *st,
+                            struct st_vertex_program *stvp,
+                            const struct st_vp_varient_key *key)
+{
+   struct st_vp_varient *vpv = CALLOC_STRUCT(st_vp_varient);
+   struct pipe_context *pipe = st->pipe;
+   struct ureg_program *ureg;
+   enum pipe_error error;
+   unsigned num_outputs;
+
+   ureg = ureg_create( TGSI_PROCESSOR_VERTEX );
+   if (ureg == NULL)
+      return NULL;
+
+   vpv->num_inputs = stvp->num_inputs;
+   num_outputs = stvp->num_outputs;
+   if (key->passthrough_edgeflags) {
+      vpv->num_inputs++;
+      num_outputs++;
    }
 
-   stvp->state.tokens = 
+   error = 
       st_translate_mesa_program(st->ctx,
                                 TGSI_PROCESSOR_VERTEX,
+                                ureg,
                                 &stvp->Base.Base,
                                 /* inputs */
-                                vs_num_inputs,
+                                vpv->num_inputs,
                                 stvp->input_to_index,
-                                vs_input_semantic_name,
-                                vs_input_semantic_index,
+                                NULL, /* input semantic name */
+                                NULL, /* input semantic index */
                                 NULL,
-                                input_flags,
                                 /* outputs */
-                                vs_num_outputs,
-                                outputMapping,
-                                vs_output_semantic_name,
-                                vs_output_semantic_index,
-                                output_flags );
+                                num_outputs,
+                                stvp->result_to_output,
+                                stvp->output_semantic_name,
+                                stvp->output_semantic_index,
+                                key->passthrough_edgeflags );
+
+   if (error)
+      goto fail;
 
-   stvp->num_inputs = vs_num_inputs;
-   stvp->driver_shader = pipe->create_vs_state(pipe, &stvp->state);
+   vpv->state.tokens = ureg_get_tokens( ureg, NULL );
+   if (!vpv->state.tokens)
+      goto fail;
+
+   ureg_destroy( ureg );
+
+   vpv->driver_shader = pipe->create_vs_state(pipe, &vpv->state);
 
    if ((ST_DEBUG & DEBUG_TGSI) && (ST_DEBUG & DEBUG_MESA)) {
       _mesa_print_program(&stvp->Base.Base);
@@ -352,9 +251,19 @@ st_translate_vertex_program(struct st_context *st,
    }
 
    if (ST_DEBUG & DEBUG_TGSI) {
-      tgsi_dump( stvp->state.tokens, 0 );
+      tgsi_dump( vpv->state.tokens, 0 );
       debug_printf("\n");
    }
+
+   return vpv;
+
+fail:
+   debug_printf("%s: failed to translate Mesa program:\n", __FUNCTION__);
+   _mesa_print_program(&stvp->Base.Base);
+   debug_assert(0);
+
+   ureg_destroy( ureg );
+   return NULL;
 }
 
 
@@ -375,9 +284,10 @@ st_translate_fragment_program(struct st_context *st,
    GLuint defaultInputMapping[FRAG_ATTRIB_MAX];
    GLuint interpMode[16];  /* XXX size? */
    GLuint attr;
+   enum pipe_error error;
    const GLbitfield inputsRead = stfp->Base.Base.InputsRead;
+   struct ureg_program *ureg;
    GLuint vslot = 0;
-   GLuint num_generic = 0;
 
    uint fs_num_inputs = 0;
 
@@ -385,13 +295,6 @@ st_translate_fragment_program(struct st_context *st,
    ubyte fs_output_semantic_index[PIPE_MAX_SHADER_OUTPUTS];
    uint fs_num_outputs = 0;
 
-   GLbitfield input_flags[MAX_PROGRAM_INPUTS];
-   GLbitfield output_flags[MAX_PROGRAM_OUTPUTS];
-
-   /*memset(&fs, 0, sizeof(fs));*/
-   memset(input_flags, 0, sizeof(input_flags));
-   memset(output_flags, 0, sizeof(output_flags));
-
    /* which vertex output goes to the first fragment input: */
    if (inputsRead & FRAG_BIT_WPOS)
       vslot = 0;
@@ -434,14 +337,25 @@ st_translate_fragment_program(struct st_context *st,
             break;
          case FRAG_ATTRIB_FACE:
             stfp->input_semantic_name[slot] = TGSI_SEMANTIC_FACE;
-            stfp->input_semantic_index[slot] = num_generic++;
+            stfp->input_semantic_index[slot] = 0;
             interpMode[slot] = TGSI_INTERPOLATE_CONSTANT;
             break;
-         case FRAG_ATTRIB_PNTC:
-            stfp->input_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-            stfp->input_semantic_index[slot] = num_generic++;
-            interpMode[slot] = TGSI_INTERPOLATE_PERSPECTIVE;
-            break;
+
+            /* In most cases, there is nothing special about these
+             * inputs, so adopt a convention to use the generic
+             * semantic name and the mesa FRAG_ATTRIB_ number as the
+             * index. 
+             * 
+             * All that is required is that the vertex shader labels
+             * its own outputs similarly, and that the vertex shader
+             * generates at least every output required by the
+             * fragment shader plus fixed-function hardware (such as
+             * BFC).
+             * 
+             * There is no requirement that semantic indexes start at
+             * zero or be restricted to a particular range -- nobody
+             * should be building tables based on semantic index.
+             */
          case FRAG_ATTRIB_TEX0:
          case FRAG_ATTRIB_TEX1:
          case FRAG_ATTRIB_TEX2:
@@ -450,19 +364,18 @@ st_translate_fragment_program(struct st_context *st,
          case FRAG_ATTRIB_TEX5:
          case FRAG_ATTRIB_TEX6:
          case FRAG_ATTRIB_TEX7:
-            stfp->input_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-            stfp->input_semantic_index[slot] = num_generic++;
-            interpMode[slot] = TGSI_INTERPOLATE_PERSPECTIVE;
-            break;
+         case FRAG_ATTRIB_PNTC:
          case FRAG_ATTRIB_VAR0:
-            /* fall-through */
          default:
+            /* Actually, let's try and zero-base this just for
+             * readability of the generated TGSI.
+             */
+            assert(attr >= FRAG_ATTRIB_TEX0);
+            stfp->input_semantic_index[slot] = (attr - FRAG_ATTRIB_TEX0);
             stfp->input_semantic_name[slot] = TGSI_SEMANTIC_GENERIC;
-            stfp->input_semantic_index[slot] = num_generic++;
             interpMode[slot] = TGSI_INTERPOLATE_PERSPECTIVE;
+            break;
          }
-
-         input_flags[slot] = stfp->Base.Base.InputFlags[attr];
       }
    }
 
@@ -500,8 +413,6 @@ st_translate_fragment_program(struct st_context *st,
                break;
             }
 
-            output_flags[fs_num_outputs] = stfp->Base.Base.OutputFlags[attr];
-
             fs_num_outputs++;
          }
       }
@@ -510,9 +421,15 @@ st_translate_fragment_program(struct st_context *st,
    if (!inputMapping)
       inputMapping = defaultInputMapping;
 
-   stfp->state.tokens = 
+   ureg = ureg_create( TGSI_PROCESSOR_FRAGMENT );
+   if (ureg == NULL)
+      return;
+
+
+   error = 
       st_translate_mesa_program(st->ctx,
                                 TGSI_PROCESSOR_FRAGMENT,
+                                ureg,
                                 &stfp->Base.Base,
                                 /* inputs */
                                 fs_num_inputs,
@@ -520,14 +437,14 @@ st_translate_fragment_program(struct st_context *st,
                                 stfp->input_semantic_name,
                                 stfp->input_semantic_index,
                                 interpMode,
-                                input_flags,
                                 /* outputs */
                                 fs_num_outputs,
                                 outputMapping,
                                 fs_output_semantic_name,
-                                fs_output_semantic_index,
-                                output_flags );
+                                fs_output_semantic_index, FALSE );
 
+   stfp->state.tokens = ureg_get_tokens( ureg, NULL );
+   ureg_destroy( ureg );
    stfp->driver_shader = pipe->create_fs_state(pipe, &stfp->state);
 
    if ((ST_DEBUG & DEBUG_TGSI) && (ST_DEBUG & DEBUG_MESA)) {
index e2e5eddef22c170bc0e86de2dcca2d928d42a012..6b9a9226df583b0de95b340d1ab59a5dacafd384 100644 (file)
@@ -64,41 +64,73 @@ struct st_fragment_program
    struct pipe_shader_state state;
    void *driver_shader;
 
-   GLuint param_state;
-
-   /** List of vertex programs which have been translated such that their
-    * outputs match this fragment program's inputs.
-    */
-   struct translated_vertex_program *vertex_programs;
-
    /** Program prefixed with glBitmap prologue */
    struct st_fragment_program *bitmap_program;
    uint bitmap_sampler;
 };
 
 
+
+struct st_vp_varient_key
+{
+   boolean passthrough_edgeflags;
+};
+
+
+/**
+ * This represents a vertex program, especially translated to match
+ * the inputs of a particular fragment shader.
+ */
+struct st_vp_varient
+{
+   /* Parameters which generated this translated version of a vertex
+    * shader:
+    */
+   struct st_vp_varient_key key;
+
+   /** TGSI tokens -- why?
+    */
+   struct pipe_shader_state state;
+
+   /** Driver's compiled shader */
+   void *driver_shader;
+
+   /** For using our private draw module (glRasterPos) */
+   struct draw_vertex_shader *draw_shader;
+
+   /** Next in linked list */
+   struct st_vp_varient *next;  
+
+   /** similar to that in st_vertex_program, but with information about edgeflags too */
+   GLuint num_inputs;
+};
+
+
+
+
 /**
  * Derived from Mesa gl_fragment_program:
  */
 struct st_vertex_program
 {
    struct gl_vertex_program Base;  /**< The Mesa vertex program */
-   GLuint serialNo;
+   GLuint serialNo, lastSerialNo;
 
    /** maps a Mesa VERT_ATTRIB_x to a packed TGSI input index */
    GLuint input_to_index[VERT_ATTRIB_MAX];
    /** maps a TGSI input index back to a Mesa VERT_ATTRIB_x */
    GLuint index_to_input[PIPE_MAX_SHADER_INPUTS];
-
    GLuint num_inputs;
 
-   struct pipe_shader_state state;
-   void *driver_shader;
+   /** Maps VERT_RESULT_x to slot */
+   GLuint result_to_output[VERT_RESULT_MAX];
+   ubyte output_semantic_name[VERT_RESULT_MAX];
+   ubyte output_semantic_index[VERT_RESULT_MAX];
+   GLuint num_outputs;
 
-   /** For using our private draw module (glRasterPos) */
-   struct draw_vertex_shader *draw_shader;
-
-   GLuint param_state;
+   /** List of translated varients of this vertex program.
+    */
+   struct st_vp_varient *varients;
 };
 
 
@@ -143,13 +175,21 @@ st_translate_fragment_program(struct st_context *st,
                               const GLuint inputMapping[]);
 
 
+/* Called after program string change, discard all previous
+ * compilation results.
+ */
 extern void
+st_prepare_vertex_program(struct st_context *st,
+                          struct st_vertex_program *stvp);
+
+extern struct st_vp_varient *
 st_translate_vertex_program(struct st_context *st,
-                            struct st_vertex_program *vp,
-                            const GLuint vert_output_to_slot[],
-                            const ubyte *fs_input_semantic_name,
-                            const ubyte *fs_input_semantic_index);
+                            struct st_vertex_program *stvp,
+                            const struct st_vp_varient_key *key);
 
+void
+st_vp_release_varients( struct st_context *st,
+                        struct st_vertex_program *stvp );
 
 extern void
 st_print_shaders(GLcontext *ctx);
index 10f13512830f4d7fd7f4fd69b3855b17e34b0676..8a3e4cd3ac3d3807e6b67f0b83ee5e3c7d14e720 100644 (file)
@@ -43,7 +43,9 @@
 #include "pipe/p_context.h"
 #include "pipe/p_defines.h"
 #include "pipe/p_inlines.h"
+#include "util/u_format.h"
 #include "util/u_rect.h"
+#include "util/u_math.h"
 
 
 #define DBG if(0) printf
@@ -100,10 +102,9 @@ st_texture_create(struct st_context *st,
    pt.target = target;
    pt.format = format;
    pt.last_level = last_level;
-   pt.width[0] = width0;
-   pt.height[0] = height0;
-   pt.depth[0] = depth0;
-   pf_get_block(format, &pt.block);
+   pt.width0 = width0;
+   pt.height0 = height0;
+   pt.depth0 = depth0;
    pt.tex_usage = usage;
 
    newtex = screen->texture_create(screen, &pt);
@@ -135,9 +136,9 @@ st_texture_match_image(const struct pipe_texture *pt,
    /* Test if this image's size matches what's expected in the
     * established texture.
     */
-   if (image->Width != pt->width[level] ||
-       image->Height != pt->height[level] ||
-       image->Depth != pt->depth[level])
+   if (image->Width != u_minify(pt->width0, level) ||
+       image->Height != u_minify(pt->height0, level) ||
+       image->Depth != u_minify(pt->depth0, level))
       return GL_FALSE;
 
    return GL_TRUE;
@@ -241,8 +242,9 @@ st_surface_data(struct pipe_context *pipe,
    struct pipe_screen *screen = pipe->screen;
    void *map = screen->transfer_map(screen, dst);
 
+   assert(dst->texture);
    util_copy_rect(map,
-                  &dst->block,
+                  dst->texture->format,
                   dst->stride,
                   dstx, dsty, 
                   width, height, 
@@ -265,7 +267,7 @@ st_texture_image_data(struct st_context *st,
 {
    struct pipe_context *pipe = st->pipe;
    struct pipe_screen *screen = pipe->screen;
-   GLuint depth = dst->depth[level];
+   GLuint depth = u_minify(dst->depth0, level);
    GLuint i;
    const GLubyte *srcUB = src;
    struct pipe_transfer *dst_transfer;
@@ -275,15 +277,16 @@ st_texture_image_data(struct st_context *st,
    for (i = 0; i < depth; i++) {
       dst_transfer = st_no_flush_get_tex_transfer(st, dst, face, level, i,
                                                  PIPE_TRANSFER_WRITE, 0, 0,
-                                                 dst->width[level],
-                                                 dst->height[level]);
+                                                 u_minify(dst->width0, level),
+                                                  u_minify(dst->height0, level));
 
       st_surface_data(pipe, dst_transfer,
                      0, 0,                             /* dstx, dsty */
                      srcUB,
                      src_row_stride,
                      0, 0,                             /* source x, y */
-                     dst->width[level], dst->height[level]);       /* width, height */
+                     u_minify(dst->width0, level),
+                      u_minify(dst->height0, level));      /* width, height */
 
       screen->tex_transfer_destroy(dst_transfer);
 
@@ -301,9 +304,9 @@ st_texture_image_copy(struct pipe_context *pipe,
                       GLuint face)
 {
    struct pipe_screen *screen = pipe->screen;
-   GLuint width = dst->width[dstLevel];
-   GLuint height = dst->height[dstLevel];
-   GLuint depth = dst->depth[dstLevel];
+   GLuint width = u_minify(dst->width0, dstLevel); 
+   GLuint height = u_minify(dst->height0, dstLevel); 
+   GLuint depth = u_minify(dst->depth0, dstLevel); 
    struct pipe_surface *src_surface;
    struct pipe_surface *dst_surface;
    GLuint i;
@@ -313,13 +316,13 @@ st_texture_image_copy(struct pipe_context *pipe,
 
       /* find src texture level of needed size */
       for (srcLevel = 0; srcLevel <= src->last_level; srcLevel++) {
-         if (src->width[srcLevel] == width &&
-             src->height[srcLevel] == height) {
+         if (u_minify(src->width0, srcLevel) == width &&
+             u_minify(src->height0, srcLevel) == height) {
             break;
          }
       }
-      assert(src->width[srcLevel] == width);
-      assert(src->height[srcLevel] == height);
+      assert(u_minify(src->width0, srcLevel) == width);
+      assert(u_minify(src->height0, srcLevel) == height);
 
 #if 0
       {
@@ -403,7 +406,7 @@ st_bind_texture_surface(struct pipe_surface *ps, int target, int level,
    }
 
    /* map pipe format to base format for now */
-   if (pf_get_component_bits(format, PIPE_FORMAT_COMP_A) > 0)
+   if (util_format_get_component_bits(format, UTIL_FORMAT_COLORSPACE_RGB, 3) > 0)
       internalFormat = GL_RGBA;
    else
       internalFormat = GL_RGB;
index bd20a8d972f2a36fdee0407ddec062be02edbb49..8484aab5a93fd7f86f7438e54c0e685aad59fa07 100644 (file)
@@ -67,8 +67,8 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
         if (facing == 1) {
            if (IND & SS_TWOSIDE_BIT) {
               if (IND & SS_RGBA_BIT) {
-                  if (VB->ColorPtr[1]) {
-                     GLfloat (*vbcolor)[4] = VB->ColorPtr[1]->data;
+                  if (VB->BackfaceColorPtr) {
+                     GLfloat (*vbcolor)[4] = VB->BackfaceColorPtr->data;
 
                      if (swsetup->intColors) {
                         COPY_CHAN4(saved_color[0], v[0]->color);
@@ -81,7 +81,7 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
                         COPY_4V(saved_col0[2], v[2]->attrib[FRAG_ATTRIB_COL0]);
                      }
 
-                     if (VB->ColorPtr[1]->stride) {
+                     if (VB->BackfaceColorPtr->stride) {
                         if (swsetup->intColors) {
                            SS_COLOR(v[0]->color, vbcolor[e0]);
                            SS_COLOR(v[1]->color, vbcolor[e1]);
@@ -108,14 +108,14 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
                      }
                   }
 
-                 if (VB->SecondaryColorPtr[1]) {
-                    GLfloat (*vbspec)[4] = VB->SecondaryColorPtr[1]->data;
+                 if (VB->BackfaceSecondaryColorPtr) {
+                    GLfloat (*vbspec)[4] = VB->BackfaceSecondaryColorPtr->data;
 
                     COPY_4V(saved_spec[0], v[0]->attrib[FRAG_ATTRIB_COL1]);
                     COPY_4V(saved_spec[1], v[1]->attrib[FRAG_ATTRIB_COL1]);
                     COPY_4V(saved_spec[2], v[2]->attrib[FRAG_ATTRIB_COL1]);
 
-                    if (VB->SecondaryColorPtr[1]->stride) {
+                    if (VB->BackfaceSecondaryColorPtr->stride) {
                        SS_SPEC(v[0]->attrib[FRAG_ATTRIB_COL1], vbspec[e0]);
                        SS_SPEC(v[1]->attrib[FRAG_ATTRIB_COL1], vbspec[e1]);
                        SS_SPEC(v[2]->attrib[FRAG_ATTRIB_COL1], vbspec[e2]);
@@ -127,7 +127,7 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
                     }
                  }
               } else {
-                 GLfloat *vbindex = (GLfloat *)VB->IndexPtr[1]->data;
+                 GLfloat *vbindex = (GLfloat *)VB->BackfaceIndexPtr->data;
                  saved_index[0] = v[0]->attrib[FRAG_ATTRIB_CI][0];
                  saved_index[1] = v[1]->attrib[FRAG_ATTRIB_CI][0];
                  saved_index[2] = v[2]->attrib[FRAG_ATTRIB_CI][0];
@@ -200,7 +200,7 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
    if (IND & SS_TWOSIDE_BIT) {
       if (facing == 1) {
         if (IND & SS_RGBA_BIT) {
-            if (VB->ColorPtr[1]) {
+            if (VB->BackfaceColorPtr) {
                if (swsetup->intColors) {
                   COPY_CHAN4(v[0]->color, saved_color[0]);
                   COPY_CHAN4(v[1]->color, saved_color[1]);
@@ -213,7 +213,7 @@ static void TAG(triangle)(GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
                }
             }
 
-           if (VB->SecondaryColorPtr[1]) {
+           if (VB->BackfaceSecondaryColorPtr) {
               COPY_4V(v[0]->attrib[FRAG_ATTRIB_COL1], saved_spec[0]);
               COPY_4V(v[1]->attrib[FRAG_ATTRIB_COL1], saved_spec[1]);
               COPY_4V(v[2]->attrib[FRAG_ATTRIB_COL1], saved_spec[2]);
index 6137c2d2fe52220f6c9e862e5f5a282f79264664..ebaae6335b942d906bacce86ae0dc9ac08d21a50 100644 (file)
@@ -198,26 +198,23 @@ struct vertex_buffer
     */
    GLuint Count;  /**< Number of vertices currently in buffer */
 
-   /* Pointers to current data.
-    * XXX some of these fields alias AttribPtr below and should be removed
-    * such as NormalPtr, TexCoordPtr, FogCoordPtr, etc.
+   /* Pointers to current data.  Most of the data is in AttribPtr -- all of
+    * it that is one of VERT_ATTRIB_X.  For things only produced by TNL,
+    * such as backface color or eye-space coordinates, they are stored
+    * here.
     */
    GLuint      *Elts;                          
-   GLvector4f  *ObjPtr;                                /* _TNL_BIT_POS */
    GLvector4f  *EyePtr;                                /* _TNL_BIT_POS */
    GLvector4f  *ClipPtr;                       /* _TNL_BIT_POS */
    GLvector4f  *NdcPtr;                         /* _TNL_BIT_POS */
    GLubyte     ClipOrMask;                     /* _TNL_BIT_POS */
    GLubyte     ClipAndMask;                    /* _TNL_BIT_POS */
    GLubyte     *ClipMask;                      /* _TNL_BIT_POS */
-   GLvector4f  *NormalPtr;                     /* _TNL_BIT_NORMAL */
    GLfloat     *NormalLengthPtr;               /* _TNL_BIT_NORMAL */
    GLboolean   *EdgeFlag;                      /* _TNL_BIT_EDGEFLAG */
-   GLvector4f  *TexCoordPtr[MAX_TEXTURE_COORD_UNITS]; /* VERT_TEX_0..n */
-   GLvector4f  *IndexPtr[2];                   /* _TNL_BIT_INDEX */
-   GLvector4f  *ColorPtr[2];                   /* _TNL_BIT_COLOR0 */
-   GLvector4f  *SecondaryColorPtr[2];           /* _TNL_BIT_COLOR1 */
-   GLvector4f  *FogCoordPtr;                   /* _TNL_BIT_FOG */
+   GLvector4f  *BackfaceIndexPtr;
+   GLvector4f  *BackfaceColorPtr;
+   GLvector4f  *BackfaceSecondaryColorPtr;
 
    const struct _mesa_prim  *Primitive;                      
    GLuint      PrimitiveCount;       
@@ -402,11 +399,6 @@ struct tnl_device_driver
    /* Alert tnl-aware drivers of changes to material.
     */
 
-   void (*NotifyInputChanges)(GLcontext *ctx, GLuint bitmask);
-   /* Alert tnl-aware drivers of changes to size and stride of input
-    * arrays.
-    */
-
    /***
     *** Rendering -- These functions called only from t_vb_render.c
     ***/
index 04fa106300351cf97e52f420621b91275041a3e0..1c7c733883c1c56630e6f3db14be8b5f75efc690 100644 (file)
@@ -251,22 +251,10 @@ static void bind_inputs( GLcontext *ctx,
     */
    VB->Count = count;
 
-
-   /* Legacy pointers -- remove one day.
-    */
-   VB->ObjPtr = VB->AttribPtr[_TNL_ATTRIB_POS];
-   VB->NormalPtr = VB->AttribPtr[_TNL_ATTRIB_NORMAL];
-   VB->ColorPtr[0] = VB->AttribPtr[_TNL_ATTRIB_COLOR0];
-   VB->ColorPtr[1] = NULL;
-   VB->IndexPtr[0] = VB->AttribPtr[_TNL_ATTRIB_COLOR_INDEX];
-   VB->IndexPtr[1] = NULL;
-   VB->SecondaryColorPtr[0] = VB->AttribPtr[_TNL_ATTRIB_COLOR1];
-   VB->SecondaryColorPtr[1] = NULL;
-   VB->FogCoordPtr = VB->AttribPtr[_TNL_ATTRIB_FOG];
-
-   for (i = 0; i < ctx->Const.MaxTextureCoordUnits; i++) {
-      VB->TexCoordPtr[i] = VB->AttribPtr[_TNL_ATTRIB_TEX0 + i];
-   }
+   /* These should perhaps be part of _TNL_ATTRIB_* */
+   VB->BackfaceColorPtr = NULL;
+   VB->BackfaceIndexPtr = NULL;
+   VB->BackfaceSecondaryColorPtr = NULL;
 
    /* Clipping and drawing code still requires this to be a packed
     * array of ubytes which can be written into.  TODO: Fix and
index 357ef1e24b5bf4914fe2b3e3b4e31d86eb314ae3..01b30babb4828b4c8ebcdee8babea9dfba790086 100644 (file)
@@ -86,10 +86,6 @@ static GLuint check_input_changes( GLcontext *ctx )
       }
    }
 
-   if (tnl->pipeline.input_changes &&
-      tnl->Driver.NotifyInputChanges) 
-      tnl->Driver.NotifyInputChanges( ctx, tnl->pipeline.input_changes );
-
    return tnl->pipeline.input_changes;
 }
 
index f3a7bd49f40189a4ca81ebc92e0365a9d315138a..4a0e6ad4f997cad65803b2e6663f6339d48bf25e 100644 (file)
@@ -156,7 +156,7 @@ run_fog_stage(GLcontext *ctx, struct tnl_pipeline_stage *stage)
       GLuint i;
       GLfloat *coord;
       /* Fog is computed from vertex or fragment Z values */
-      /* source = VB->ObjPtr or VB->EyePtr coords */
+      /* source = VB->AttribPtr[_TNL_ATTRIB_POS] or VB->EyePtr coords */
       /* dest = VB->AttribPtr[_TNL_ATTRIB_FOG] = fog stage private storage */
       VB->AttribPtr[_TNL_ATTRIB_FOG] = &store->fogcoord;
 
@@ -176,11 +176,12 @@ run_fog_stage(GLcontext *ctx, struct tnl_pipeline_stage *stage)
         /* Full eye coords weren't required, just calculate the
          * eye Z values.
          */
-        _mesa_dotprod_tab[VB->ObjPtr->size]( (GLfloat *) input->data,
-                                             4 * sizeof(GLfloat),
-                                             VB->ObjPtr, plane );
+        _mesa_dotprod_tab[VB->AttribPtr[_TNL_ATTRIB_POS]->size]
+           ( (GLfloat *) input->data,
+             4 * sizeof(GLfloat),
+             VB->AttribPtr[_TNL_ATTRIB_POS], plane );
 
-        input->count = VB->ObjPtr->count;
+        input->count = VB->AttribPtr[_TNL_ATTRIB_POS]->count;
 
         /* make sure coords are really positive
            NOTE should avoid going through array twice */
@@ -213,7 +214,7 @@ run_fog_stage(GLcontext *ctx, struct tnl_pipeline_stage *stage)
       /* input->count may be one if glFogCoord was only called once
        * before glBegin.  But we need to compute fog for all vertices.
        */
-      input->count = VB->ObjPtr->count;
+      input->count = VB->AttribPtr[_TNL_ATTRIB_POS]->count;
 
       VB->AttribPtr[_TNL_ATTRIB_FOG] = &store->fogcoord;  /* dest data */
    }
@@ -227,7 +228,6 @@ run_fog_stage(GLcontext *ctx, struct tnl_pipeline_stage *stage)
       VB->AttribPtr[_TNL_ATTRIB_FOG] = input;
    }
 
-   VB->FogCoordPtr = VB->AttribPtr[_TNL_ATTRIB_FOG];
    return GL_TRUE;
 }
 
index f47f99397c9ad29bf59664884f66325f1d25d879..8a0fe63fd8b75ab60feaa3042c883ae093aa79b8 100644 (file)
@@ -127,7 +127,7 @@ prepare_materials(GLcontext *ctx,
       const GLuint bitmask = ctx->Light.ColorMaterialBitmask;
       for (i = 0 ; i < MAT_ATTRIB_MAX ; i++)
         if (bitmask & (1<<i))
-           VB->AttribPtr[_TNL_ATTRIB_MAT_FRONT_AMBIENT + i] = VB->ColorPtr[0];
+           VB->AttribPtr[_TNL_ATTRIB_MAT_FRONT_AMBIENT + i] = VB->AttribPtr[_TNL_ATTRIB_COLOR0];
    }
 
    /* Now, for each material attribute that's tracking vertex color, save
@@ -200,7 +200,7 @@ static GLboolean run_lighting( GLcontext *ctx,
    struct light_stage_data *store = LIGHT_STAGE_DATA(stage);
    TNLcontext *tnl = TNL_CONTEXT(ctx);
    struct vertex_buffer *VB = &tnl->vb;
-   GLvector4f *input = ctx->_NeedEyeCoords ? VB->EyePtr : VB->ObjPtr;
+   GLvector4f *input = ctx->_NeedEyeCoords ? VB->EyePtr : VB->AttribPtr[_TNL_ATTRIB_POS];
    GLuint idx;
 
    if (!ctx->Light.Enabled || ctx->VertexProgram._Current)
@@ -208,13 +208,13 @@ static GLboolean run_lighting( GLcontext *ctx,
 
    /* Make sure we can talk about position x,y and z:
     */
-   if (input->size <= 2 && input == VB->ObjPtr) {
+   if (input->size <= 2 && input == VB->AttribPtr[_TNL_ATTRIB_POS]) {
 
       _math_trans_4f( store->Input.data,
-                     VB->ObjPtr->data,
-                     VB->ObjPtr->stride,
+                     VB->AttribPtr[_TNL_ATTRIB_POS]->data,
+                     VB->AttribPtr[_TNL_ATTRIB_POS]->stride,
                      GL_FLOAT,
-                     VB->ObjPtr->size,
+                     VB->AttribPtr[_TNL_ATTRIB_POS]->size,
                      0,
                      VB->Count );
 
@@ -246,10 +246,6 @@ static GLboolean run_lighting( GLcontext *ctx,
     */
    store->light_func_tab[idx]( ctx, VB, stage, input );
 
-   VB->AttribPtr[_TNL_ATTRIB_COLOR0] = VB->ColorPtr[0];
-   VB->AttribPtr[_TNL_ATTRIB_COLOR1] = VB->SecondaryColorPtr[0];
-   VB->AttribPtr[_TNL_ATTRIB_COLOR_INDEX] = VB->IndexPtr[0];
-
    return GL_TRUE;
 }
 
index 124ca3c74fe258b21616263979db681910ffc267..4ebef2356f50e7ca459b9fff1423999778b5c553 100644 (file)
@@ -72,13 +72,13 @@ static void TAG(light_rgba_spec)( GLcontext *ctx,
    fprintf(stderr, "%s\n", __FUNCTION__ );
 #endif
 
-   VB->ColorPtr[0] = &store->LitColor[0];
-   VB->SecondaryColorPtr[0] = &store->LitSecondary[0];
+   VB->AttribPtr[_TNL_ATTRIB_COLOR0] = &store->LitColor[0];
+   VB->AttribPtr[_TNL_ATTRIB_COLOR1] = &store->LitSecondary[0];
    sumA[0] = ctx->Light.Material.Attrib[MAT_ATTRIB_FRONT_DIFFUSE][3];
 
 #if IDX & LIGHT_TWOSIDE
-   VB->ColorPtr[1] = &store->LitColor[1];
-   VB->SecondaryColorPtr[1] = &store->LitSecondary[1];
+   VB->BackfaceColorPtr = &store->LitColor[1];
+   VB->BackfaceSecondaryColorPtr = &store->LitSecondary[1];
    sumA[1] = ctx->Light.Material.Attrib[MAT_ATTRIB_BACK_DIFFUSE][3];
 #endif
 
@@ -259,11 +259,11 @@ static void TAG(light_rgba)( GLcontext *ctx,
    fprintf(stderr, "%s\n", __FUNCTION__ );
 #endif
 
-   VB->ColorPtr[0] = &store->LitColor[0];
+   VB->AttribPtr[_TNL_ATTRIB_COLOR0] = &store->LitColor[0];
    sumA[0] = ctx->Light.Material.Attrib[MAT_ATTRIB_FRONT_DIFFUSE][3];
 
 #if IDX & LIGHT_TWOSIDE
-   VB->ColorPtr[1] = &store->LitColor[1];
+   VB->BackfaceColorPtr = &store->LitColor[1];
    sumA[1] = ctx->Light.Material.Attrib[MAT_ATTRIB_BACK_DIFFUSE][3];
 #endif
 
@@ -449,9 +449,9 @@ static void TAG(light_fast_rgba_single)( GLcontext *ctx,
 
    (void) input;               /* doesn't refer to Eye or Obj */
 
-   VB->ColorPtr[0] = &store->LitColor[0];
+   VB->AttribPtr[_TNL_ATTRIB_COLOR0] = &store->LitColor[0];
 #if IDX & LIGHT_TWOSIDE
-   VB->ColorPtr[1] = &store->LitColor[1];
+   VB->BackfaceColorPtr = &store->LitColor[1];
 #endif
 
    if (nr > 1) {
@@ -559,9 +559,9 @@ static void TAG(light_fast_rgba)( GLcontext *ctx,
    sumA[0] = ctx->Light.Material.Attrib[MAT_ATTRIB_FRONT_DIFFUSE][3];
    sumA[1] = ctx->Light.Material.Attrib[MAT_ATTRIB_BACK_DIFFUSE][3];
 
-   VB->ColorPtr[0] = &store->LitColor[0];
+   VB->AttribPtr[_TNL_ATTRIB_COLOR0] = &store->LitColor[0];
 #if IDX & LIGHT_TWOSIDE
-   VB->ColorPtr[1] = &store->LitColor[1];
+   VB->BackfaceColorPtr = &store->LitColor[1];
 #endif
 
    if (nr > 1) {
@@ -665,14 +665,14 @@ static void TAG(light_ci)( GLcontext *ctx,
    fprintf(stderr, "%s\n", __FUNCTION__ );
 #endif
 
-   VB->IndexPtr[0] = &store->LitIndex[0];
+   VB->AttribPtr[_TNL_ATTRIB_COLOR_INDEX] = &store->LitIndex[0];
 #if IDX & LIGHT_TWOSIDE
-   VB->IndexPtr[1] = &store->LitIndex[1];
+   VB->BackfaceIndexPtr = &store->LitIndex[1];
 #endif
 
-   indexResult[0] = (GLfloat *)VB->IndexPtr[0]->data;
+   indexResult[0] = (GLfloat *)VB->AttribPtr[_TNL_ATTRIB_COLOR_INDEX]->data;
 #if IDX & LIGHT_TWOSIDE
-   indexResult[1] = (GLfloat *)VB->IndexPtr[1]->data;
+   indexResult[1] = (GLfloat *)VB->BackfaceIndexPtr->data;
 #endif
 
    /* loop over vertices */
index a4821cc1cccf7230a53752d6012b7c591ad84032..693d3dc118d17d806c816b676fb84581fd939868 100644 (file)
@@ -79,7 +79,6 @@ run_normal_stage(GLcontext *ctx, struct tnl_pipeline_stage *stage)
    }
 
    VB->AttribPtr[_TNL_ATTRIB_NORMAL] = &store->normal;
-   VB->NormalPtr = &store->normal;
 
    VB->NormalLengthPtr = NULL; /* no longer valid */
    return GL_TRUE;
index e69f7d5766466908f93e89200d97bd7f8939989c..c289cdfbaac84379d8305739979f31e8f26e8070 100644 (file)
@@ -454,19 +454,14 @@ run_vp( GLcontext *ctx, struct tnl_pipeline_stage *stage )
       VB->ClipPtr->count = VB->Count;
    }
 
-   VB->ColorPtr[0] = &store->results[VERT_RESULT_COL0];
-   VB->ColorPtr[1] = &store->results[VERT_RESULT_BFC0];
-   VB->SecondaryColorPtr[0] = &store->results[VERT_RESULT_COL1];
-   VB->SecondaryColorPtr[1] = &store->results[VERT_RESULT_BFC1];
-   VB->FogCoordPtr = &store->results[VERT_RESULT_FOGC];
-
    VB->AttribPtr[VERT_ATTRIB_COLOR0] = &store->results[VERT_RESULT_COL0];
    VB->AttribPtr[VERT_ATTRIB_COLOR1] = &store->results[VERT_RESULT_COL1];
    VB->AttribPtr[VERT_ATTRIB_FOG] = &store->results[VERT_RESULT_FOGC];
    VB->AttribPtr[_TNL_ATTRIB_POINTSIZE] = &store->results[VERT_RESULT_PSIZ];
+   VB->BackfaceColorPtr = &store->results[VERT_RESULT_BFC0];
+   VB->BackfaceSecondaryColorPtr = &store->results[VERT_RESULT_BFC1];
 
    for (i = 0; i < ctx->Const.MaxTextureCoordUnits; i++) {
-      VB->TexCoordPtr[i] = 
       VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]
          = &store->results[VERT_RESULT_TEX0 + i];
    }
index 7c1819b223cc7d6d9f100d1d507df8aff67fe153..9ef13bc96d87cadb66f2f604779b30dbf77b9c3b 100644 (file)
@@ -341,7 +341,7 @@ static void texgen( GLcontext *ctx,
    GLvector4f *in = VB->AttribPtr[VERT_ATTRIB_TEX0 + unit];
    GLvector4f *out = &store->texcoord[unit];
    const struct gl_texture_unit *texUnit = &ctx->Texture.Unit[unit];
-   const GLvector4f *obj = VB->ObjPtr;
+   const GLvector4f *obj = VB->AttribPtr[_TNL_ATTRIB_POS];
    const GLvector4f *eye = VB->EyePtr;
    const GLvector4f *normal = VB->AttribPtr[_TNL_ATTRIB_NORMAL];
    const GLfloat *m = store->tmp_m;
@@ -498,7 +498,6 @@ static GLboolean run_texgen_stage( GLcontext *ctx,
 
         store->TexgenFunc[i]( ctx, store, i );
 
-         VB->TexCoordPtr[i] =
          VB->AttribPtr[VERT_ATTRIB_TEX0 + i] = &store->texcoord[i];
       }
    }
index 0abe8cc35dd94d39f81ec6643c2d9c917f25dc6a..83688290e59fcbe59ce6ae7028f38227c82713a1 100644 (file)
@@ -73,7 +73,6 @@ static GLboolean run_texmat_stage( GLcontext *ctx,
                              ctx->TextureMatrixStack[i].Top,
                              VB->AttribPtr[_TNL_ATTRIB_TEX0 + i]);
 
-         VB->TexCoordPtr[i] = 
         VB->AttribPtr[VERT_ATTRIB_TEX0+i] = &store->texcoord[i];
       }
    }
index 4734754ea4016dd470fdf9466a109230387f29d8..bc7e0951ec8c93c47bf011d586ceda5c6816df62 100644 (file)
@@ -152,16 +152,16 @@ static GLboolean run_vertex_stage( GLcontext *ctx,
        * Use combined ModelProject to avoid some depth artifacts
        */
       if (ctx->ModelviewMatrixStack.Top->type == MATRIX_IDENTITY)
-        VB->EyePtr = VB->ObjPtr;
+        VB->EyePtr = VB->AttribPtr[_TNL_ATTRIB_POS];
       else
         VB->EyePtr = TransformRaw( &store->eye,
                                    ctx->ModelviewMatrixStack.Top,
-                                   VB->ObjPtr);
+                                   VB->AttribPtr[_TNL_ATTRIB_POS]);
    }
 
    VB->ClipPtr = TransformRaw( &store->clip,
                               &ctx->_ModelProjectMatrix,
-                              VB->ObjPtr );
+                              VB->AttribPtr[_TNL_ATTRIB_POS] );
 
    /* Drivers expect this to be clean to element 4...
     */
index 9812f8c8081aed83eac539a36f07fa389a4fbc9b..99ddace73d13f1c39d2e06bea72dd71f16d3d62c 100644 (file)
@@ -210,7 +210,7 @@ static INLINE void insert_3f_xyw_err( const struct tnl_clipspace_attr *a, GLubyt
 {
    (void) a; (void) v; (void) in;
    DEBUG_INSERT;
-   _mesa_exit(1);
+   exit(1);
 }
 
 static INLINE void insert_3f_3( const struct tnl_clipspace_attr *a, GLubyte *v, const GLfloat *in )
@@ -1092,33 +1092,33 @@ void _tnl_generic_interp_extras( GLcontext *ctx,
 {
    struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
-   /* If stride is zero, ColorPtr[1] is constant across the VB, so
+   /* If stride is zero, BackfaceColorPtr is constant across the VB, so
     * there is no point interpolating between two values as they will
     * be identical.  In all other cases, this value is generated by
     * t_vb_lighttmp.h and has a stride of 4 dwords.
     */
-   if (VB->ColorPtr[1] && VB->ColorPtr[1]->stride) {
-      assert(VB->ColorPtr[1]->stride == 4 * sizeof(GLfloat));
+   if (VB->BackfaceColorPtr && VB->BackfaceColorPtr->stride) {
+      assert(VB->BackfaceColorPtr->stride == 4 * sizeof(GLfloat));
 
       INTERP_4F( t,
-                VB->ColorPtr[1]->data[dst],
-                VB->ColorPtr[1]->data[out],
-                VB->ColorPtr[1]->data[in] );
+                VB->BackfaceColorPtr->data[dst],
+                VB->BackfaceColorPtr->data[out],
+                VB->BackfaceColorPtr->data[in] );
    }
 
-   if (VB->SecondaryColorPtr[1]) {
-      assert(VB->SecondaryColorPtr[1]->stride == 4 * sizeof(GLfloat));
+   if (VB->BackfaceSecondaryColorPtr) {
+      assert(VB->BackfaceSecondaryColorPtr->stride == 4 * sizeof(GLfloat));
       
       INTERP_3F( t,
-                VB->SecondaryColorPtr[1]->data[dst],
-                VB->SecondaryColorPtr[1]->data[out],
-                VB->SecondaryColorPtr[1]->data[in] );
+                VB->BackfaceSecondaryColorPtr->data[dst],
+                VB->BackfaceSecondaryColorPtr->data[out],
+                VB->BackfaceSecondaryColorPtr->data[in] );
    }
    
-   if (VB->IndexPtr[1]) {
-      VB->IndexPtr[1]->data[dst][0] = LINTERP( t,
-                                              VB->IndexPtr[1]->data[out][0],
-                                              VB->IndexPtr[1]->data[in][0] );
+   if (VB->BackfaceIndexPtr) {
+      VB->BackfaceIndexPtr->data[dst][0] = LINTERP( t,
+                                              VB->BackfaceIndexPtr->data[out][0],
+                                              VB->BackfaceIndexPtr->data[in][0] );
    }
 
    if (VB->EdgeFlag) {
@@ -1135,18 +1135,18 @@ void _tnl_generic_copy_pv_extras( GLcontext *ctx,
 
    /* See above comment:
     */
-   if (VB->ColorPtr[1] && VB->ColorPtr[1]->stride) {
-      COPY_4FV( VB->ColorPtr[1]->data[dst], 
-               VB->ColorPtr[1]->data[src] );
+   if (VB->BackfaceColorPtr && VB->BackfaceColorPtr->stride) {
+      COPY_4FV( VB->BackfaceColorPtr->data[dst],
+               VB->BackfaceColorPtr->data[src] );
    }
 
-   if (VB->SecondaryColorPtr[1]) {
-      COPY_4FV( VB->SecondaryColorPtr[1]->data[dst], 
-               VB->SecondaryColorPtr[1]->data[src] );
+   if (VB->BackfaceSecondaryColorPtr) {
+      COPY_4FV( VB->BackfaceSecondaryColorPtr->data[dst],
+               VB->BackfaceSecondaryColorPtr->data[src] );
    }
 
-   if (VB->IndexPtr[1]) {
-      VB->IndexPtr[1]->data[dst][0] = VB->IndexPtr[1]->data[src][0];
+   if (VB->BackfaceIndexPtr) {
+      VB->BackfaceIndexPtr->data[dst][0] = VB->BackfaceIndexPtr->data[src][0];
    }
 
    _tnl_generic_copy_pv(ctx, dst, src);
index d568bfdb5841fa8c19eaff8cc57e5f0376aa2e85..2424204b886f4f8e08f9fa4c8dfc4daa319967ff 100644 (file)
@@ -447,7 +447,7 @@ static void TAG(render_quad_strip_verts)( GLcontext *ctx,
 
    } else if (HAVE_TRI_STRIPS && 
              ctx->Light.ShadeModel == GL_FLAT &&
-             TNL_CONTEXT(ctx)->vb.ColorPtr[0]->stride) {
+             TNL_CONTEXT(ctx)->vb.AttribPtr[_TNL_ATTRIB_COLOR0]->stride) {
       if (HAVE_ELTS) {
         LOCAL_VARS;
         int dmasz = GET_SUBSEQUENT_VB_MAX_ELTS();
@@ -1230,7 +1230,7 @@ static GLboolean TAG(validate_render)( GLcontext *ctx,
            ok = GL_TRUE;
         } else if (HAVE_TRI_STRIPS && 
                    ctx->Light.ShadeModel == GL_FLAT &&
-                   VB->ColorPtr[0]->stride != 0) {
+                   VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride != 0) {
            if (HAVE_ELTS) {
               ok = (GLint) count < GET_SUBSEQUENT_VB_MAX_ELTS();
            }
index c3ba8514c8ab87116dd4a7e838d4c3dc70d4784d..6acd837317f5dd870d2e9d613c54275f6ee0c635 100644 (file)
@@ -195,7 +195,7 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
                  }
               }
               else {
-                 GLfloat (*vbcolor)[4] = VB->ColorPtr[1]->data;
+                 GLfloat (*vbcolor)[4] = VB->BackfaceColorPtr->data;
                  (void) vbcolor;
 
                  if (!DO_FLAT) {
@@ -204,8 +204,8 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
                  }
                  VERT_SAVE_RGBA( 2 );
 
-                 if (VB->ColorPtr[1]->stride) {
-                    ASSERT(VB->ColorPtr[1]->stride == 4*sizeof(GLfloat));
+                 if (VB->BackfaceColorPtr->stride) {
+                    ASSERT(VB->BackfaceColorPtr->stride == 4*sizeof(GLfloat));
 
                     if (!DO_FLAT) {              
                        VERT_SET_RGBA( v[0], vbcolor[e0] );
@@ -221,9 +221,9 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
                     VERT_SET_RGBA( v[2], vbcolor[0] );
                  }
 
-                 if (HAVE_SPEC && VB->SecondaryColorPtr[1]) {
-                    GLfloat (*vbspec)[4] = VB->SecondaryColorPtr[1]->data;
-                    ASSERT(VB->SecondaryColorPtr[1]->stride == 4*sizeof(GLfloat));
+                 if (HAVE_SPEC && VB->BackfaceSecondaryColorPtr) {
+                    GLfloat (*vbspec)[4] = VB->BackfaceSecondaryColorPtr->data;
+                    ASSERT(VB->BackfaceSecondaryColorPtr->stride == 4*sizeof(GLfloat));
 
                     if (!DO_FLAT) {
                        VERT_SAVE_SPEC( 0 );
@@ -237,7 +237,7 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
               }
            }
            else {
-              GLfloat (*vbindex) = (GLfloat *)VB->IndexPtr[1]->data;
+              GLfloat (*vbindex) = (GLfloat *)VB->BackfaceIndexPtr->data;
               if (!DO_FLAT) {
                  VERT_SAVE_IND( 0 );
                  VERT_SAVE_IND( 1 );
@@ -279,7 +279,7 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
         VERT_SAVE_RGBA( 1 );
         VERT_COPY_RGBA( v[0], v[2] );
         VERT_COPY_RGBA( v[1], v[2] );
-        if (HAVE_SPEC && VB->SecondaryColorPtr[0]) {
+        if (HAVE_SPEC && VB->AttribPtr[_TNL_ATTRIB_COLOR1]) {
            VERT_SAVE_SPEC( 0 );
            VERT_SAVE_SPEC( 1 );
            VERT_COPY_SPEC( v[0], v[2] );
@@ -374,7 +374,7 @@ static void TAG(triangle)( GLcontext *ctx, GLuint e0, GLuint e1, GLuint e2 )
       if (HAVE_RGBA) {
         VERT_RESTORE_RGBA( 0 );
         VERT_RESTORE_RGBA( 1 );
-        if (HAVE_SPEC && VB->SecondaryColorPtr[0]) {
+        if (HAVE_SPEC && VB->AttribPtr[_TNL_ATTRIB_COLOR1]) {
            VERT_RESTORE_SPEC( 0 );
            VERT_RESTORE_SPEC( 1 );
         }
@@ -436,7 +436,7 @@ static void TAG(quadr)( GLcontext *ctx,
         if (DO_TWOSIDE && facing == 1)
         {
            if (HAVE_RGBA) {
-              GLfloat (*vbcolor)[4] = VB->ColorPtr[1]->data;
+              GLfloat (*vbcolor)[4] = VB->BackfaceColorPtr->data;
               (void)vbcolor;
 
               if (HAVE_BACK_COLORS) {
@@ -471,7 +471,7 @@ static void TAG(quadr)( GLcontext *ctx,
                  }
                  VERT_SAVE_RGBA( 3 );
 
-                 if (VB->ColorPtr[1]->stride) {
+                 if (VB->BackfaceColorPtr->stride) {
                     if (!DO_FLAT) {
                        VERT_SET_RGBA( v[0], vbcolor[e0] );
                        VERT_SET_RGBA( v[1], vbcolor[e1] );
@@ -488,9 +488,9 @@ static void TAG(quadr)( GLcontext *ctx,
                     VERT_SET_RGBA( v[3], vbcolor[0] );
                  }
 
-                 if (HAVE_SPEC && VB->SecondaryColorPtr[1]) {
-                    GLfloat (*vbspec)[4] = VB->SecondaryColorPtr[1]->data;
-                    ASSERT(VB->SecondaryColorPtr[1]->stride==4*sizeof(GLfloat));
+                 if (HAVE_SPEC && VB->BackfaceSecondaryColorPtr) {
+                    GLfloat (*vbspec)[4] = VB->BackfaceSecondaryColorPtr->data;
+                    ASSERT(VB->BackfaceSecondaryColorPtr->stride==4*sizeof(GLfloat));
 
                     if (!DO_FLAT) {
                        VERT_SAVE_SPEC( 0 );
@@ -506,7 +506,7 @@ static void TAG(quadr)( GLcontext *ctx,
               }
            }
            else {
-              GLfloat *vbindex = (GLfloat *)VB->IndexPtr[1]->data;
+              GLfloat *vbindex = (GLfloat *)VB->BackfaceIndexPtr->data;
               if (!DO_FLAT) {
                  VERT_SAVE_IND( 0 );
                  VERT_SAVE_IND( 1 );
@@ -553,7 +553,7 @@ static void TAG(quadr)( GLcontext *ctx,
         VERT_COPY_RGBA( v[0], v[3] );
         VERT_COPY_RGBA( v[1], v[3] );
         VERT_COPY_RGBA( v[2], v[3] );
-        if (HAVE_SPEC && VB->SecondaryColorPtr[0]) {
+        if (HAVE_SPEC && VB->AttribPtr[_TNL_ATTRIB_COLOR1]) {
            VERT_SAVE_SPEC( 0 );
            VERT_SAVE_SPEC( 1 );
            VERT_SAVE_SPEC( 2 );
@@ -659,7 +659,7 @@ static void TAG(quadr)( GLcontext *ctx,
         VERT_RESTORE_RGBA( 0 );
         VERT_RESTORE_RGBA( 1 );
         VERT_RESTORE_RGBA( 2 );
-        if (HAVE_SPEC && VB->SecondaryColorPtr[0]) {
+        if (HAVE_SPEC && VB->AttribPtr[_TNL_ATTRIB_COLOR1]) {
            VERT_RESTORE_SPEC( 0 );
            VERT_RESTORE_SPEC( 1 );
            VERT_RESTORE_SPEC( 2 );
@@ -708,7 +708,7 @@ static void TAG(line)( GLcontext *ctx, GLuint e0, GLuint e1 )
       if (HAVE_RGBA) {
         VERT_SAVE_RGBA( 0 );
         VERT_COPY_RGBA( v[0], v[1] );
-        if (HAVE_SPEC && VB->SecondaryColorPtr[0]) {
+        if (HAVE_SPEC && VB->AttribPtr[_TNL_ATTRIB_COLOR1]) {
            VERT_SAVE_SPEC( 0 );
            VERT_COPY_SPEC( v[0], v[1] );
         }
@@ -725,7 +725,7 @@ static void TAG(line)( GLcontext *ctx, GLuint e0, GLuint e1 )
       if (HAVE_RGBA) {
         VERT_RESTORE_RGBA( 0 );
 
-        if (HAVE_SPEC && VB->SecondaryColorPtr[0]) {
+        if (HAVE_SPEC && VB->AttribPtr[_TNL_ATTRIB_COLOR1]) {
            VERT_RESTORE_SPEC( 0 );
         }
       }
index b3937c29a000596a0fcd960c2a4caad5bf704b6d..a8a0a69768d3b7a304002ac6b650704c0420b919 100644 (file)
@@ -297,19 +297,19 @@ INTERP_QUALIFIER void TAG(interp_extras)( GLcontext *ctx,
    LOCALVARS
    struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
-   if (VB->ColorPtr[1]) {
-      assert(VB->ColorPtr[1]->stride == 4 * sizeof(GLfloat));
+   if (VB->BackfaceColorPtr) {
+      assert(VB->BackfaceColorPtr->stride == 4 * sizeof(GLfloat));
 
       INTERP_4F( t,
-                   GET_COLOR(VB->ColorPtr[1], dst),
-                   GET_COLOR(VB->ColorPtr[1], out),
-                   GET_COLOR(VB->ColorPtr[1], in) );
+                   GET_COLOR(VB->BackfaceColorPtr, dst),
+                   GET_COLOR(VB->BackfaceColorPtr, out),
+                   GET_COLOR(VB->BackfaceColorPtr, in) );
 
-      if (VB->SecondaryColorPtr[1]) {
+      if (VB->BackfaceSecondaryColorPtr) {
         INTERP_3F( t,
-                      GET_COLOR(VB->SecondaryColorPtr[1], dst),
-                      GET_COLOR(VB->SecondaryColorPtr[1], out),
-                      GET_COLOR(VB->SecondaryColorPtr[1], in) );
+                      GET_COLOR(VB->BackfaceSecondaryColorPtr, dst),
+                      GET_COLOR(VB->BackfaceSecondaryColorPtr, out),
+                      GET_COLOR(VB->BackfaceSecondaryColorPtr, in) );
       }
    }
 
@@ -326,13 +326,13 @@ INTERP_QUALIFIER void TAG(copy_pv_extras)( GLcontext *ctx,
    LOCALVARS
       struct vertex_buffer *VB = &TNL_CONTEXT(ctx)->vb;
 
-   if (VB->ColorPtr[1]) {
-      COPY_4FV( GET_COLOR(VB->ColorPtr[1], dst), 
-               GET_COLOR(VB->ColorPtr[1], src) );
+   if (VB->BackfaceColorPtr) {
+      COPY_4FV( GET_COLOR(VB->BackfaceColorPtr, dst),
+               GET_COLOR(VB->BackfaceColorPtr, src) );
 
-      if (VB->SecondaryColorPtr[1]) {
-        COPY_4FV( GET_COLOR(VB->SecondaryColorPtr[1], dst), 
-                  GET_COLOR(VB->SecondaryColorPtr[1], src) );
+      if (VB->BackfaceSecondaryColorPtr) {
+        COPY_4FV( GET_COLOR(VB->BackfaceSecondaryColorPtr, dst),
+                  GET_COLOR(VB->BackfaceSecondaryColorPtr, src) );
       }
    }
 
index 92dd8931c33b710e3e0e9e58222fd5560a4bc0b9..85101b9ceb8eca6f1e7041d8b8d7898c9ec68144 100644 (file)
@@ -153,46 +153,46 @@ static void TAG(emit)( GLcontext *ctx,
 
    if (DO_TEX3) {
       const GLuint t3 = GET_TEXSOURCE(3);
-      tc3 = VB->TexCoordPtr[t3]->data;
-      tc3_stride = VB->TexCoordPtr[t3]->stride;
+      tc3 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t3]->data;
+      tc3_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t3]->stride;
       if (DO_PTEX)
-        tc3_size = VB->TexCoordPtr[t3]->size;
+        tc3_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t3]->size;
    }
 
    if (DO_TEX2) {
       const GLuint t2 = GET_TEXSOURCE(2);
-      tc2 = VB->TexCoordPtr[t2]->data;
-      tc2_stride = VB->TexCoordPtr[t2]->stride;
+      tc2 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t2]->data;
+      tc2_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t2]->stride;
       if (DO_PTEX)
-        tc2_size = VB->TexCoordPtr[t2]->size;
+        tc2_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t2]->size;
    }
 
    if (DO_TEX1) {
       const GLuint t1 = GET_TEXSOURCE(1);
-      tc1 = VB->TexCoordPtr[t1]->data;
-      tc1_stride = VB->TexCoordPtr[t1]->stride;
+      tc1 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->data;
+      tc1_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->stride;
       if (DO_PTEX)
-        tc1_size = VB->TexCoordPtr[t1]->size;
+        tc1_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t1]->size;
    }
 
    if (DO_TEX0) {
       const GLuint t0 = GET_TEXSOURCE(0);
-      tc0_stride = VB->TexCoordPtr[t0]->stride;
-      tc0 = VB->TexCoordPtr[t0]->data;
+      tc0_stride = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->stride;
+      tc0 = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->data;
       if (DO_PTEX) 
-        tc0_size = VB->TexCoordPtr[t0]->size;
+        tc0_size = VB->AttribPtr[_TNL_ATTRIB_TEX0 + t0]->size;
    }
 
    if (DO_RGBA) {
-      col_stride = VB->ColorPtr[0]->stride;
-      col = VB->ColorPtr[0]->data;
-      col_size = VB->ColorPtr[0]->size;
+      col_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
+      col = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+      col_size = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->size;
    }
 
    if (DO_SPEC) {
-      if (VB->SecondaryColorPtr[0]) {
-        spec_stride = VB->SecondaryColorPtr[0]->stride;
-        spec = VB->SecondaryColorPtr[0]->data;
+      if (VB->AttribPtr[_TNL_ATTRIB_COLOR1]) {
+        spec_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR1]->stride;
+        spec = VB->AttribPtr[_TNL_ATTRIB_COLOR1]->data;
       } else {
         spec = (GLfloat (*)[4])ctx->Current.Attrib[VERT_ATTRIB_COLOR1];
         spec_stride = 0;
@@ -200,9 +200,9 @@ static void TAG(emit)( GLcontext *ctx,
    }
 
    if (DO_FOG) {
-      if (VB->FogCoordPtr) {
-        fog = VB->FogCoordPtr->data;
-        fog_stride = VB->FogCoordPtr->stride;
+      if (VB->AttribPtr[_TNL_ATTRIB_FOG]) {
+        fog = VB->AttribPtr[_TNL_ATTRIB_FOG]->data;
+        fog_stride = VB->AttribPtr[_TNL_ATTRIB_FOG]->stride;
       }
       else {
         static GLfloat tmp[4] = {0, 0, 0, 0};
@@ -356,9 +356,9 @@ static void TAG(emit)( GLcontext *ctx, GLuint start, GLuint end,
 
    ASSERT(stride == 4);
 
-   col = VB->ColorPtr[0]->data;
-   col_stride = VB->ColorPtr[0]->stride;
-   col_size = VB->ColorPtr[0]->size;
+   col = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->data;
+   col_stride = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->stride;
+   col_size = VB->AttribPtr[_TNL_ATTRIB_COLOR0]->size;
 
 /*     fprintf(stderr, "%s(small) importable %x\n",  */
 /*        __FUNCTION__, VB->importable_data); */
@@ -410,22 +410,22 @@ static GLboolean TAG(check_tex_sizes)( GLcontext *ctx )
 
    /* Force 'missing' texcoords to something valid.
     */
-   if (DO_TEX3 && VB->TexCoordPtr[2] == 0)
-      VB->TexCoordPtr[2] = VB->TexCoordPtr[3];
+   if (DO_TEX3 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + 2] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX0 + 2] = VB->AttribPtr[_TNL_ATTRIB_TEX0 + 3];
 
-   if (DO_TEX2 && VB->TexCoordPtr[1] == 0)
-      VB->TexCoordPtr[1] = VB->TexCoordPtr[2];
+   if (DO_TEX2 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + 1] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX0 + 1] = VB->AttribPtr[_TNL_ATTRIB_TEX0 + 2];
 
-   if (DO_TEX1 && VB->TexCoordPtr[0] == 0)
-      VB->TexCoordPtr[0] = VB->TexCoordPtr[1];
+   if (DO_TEX1 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + 0] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX0 + 0] = VB->AttribPtr[_TNL_ATTRIB_TEX0 + 1];
 
    if (DO_PTEX)
       return GL_TRUE;
    
-   if ((DO_TEX3 && VB->TexCoordPtr[GET_TEXSOURCE(3)]->size == 4) ||
-       (DO_TEX2 && VB->TexCoordPtr[GET_TEXSOURCE(2)]->size == 4) ||
-       (DO_TEX1 && VB->TexCoordPtr[GET_TEXSOURCE(1)]->size == 4) ||
-       (DO_TEX0 && VB->TexCoordPtr[GET_TEXSOURCE(0)]->size == 4))
+   if ((DO_TEX3 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(3)]->size == 4) ||
+       (DO_TEX2 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(2)]->size == 4) ||
+       (DO_TEX1 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(1)]->size == 4) ||
+       (DO_TEX0 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(0)]->size == 4))
       return GL_FALSE;
 
    return GL_TRUE;
@@ -438,14 +438,14 @@ static GLboolean TAG(check_tex_sizes)( GLcontext *ctx )
 
    /* Force 'missing' texcoords to something valid.
     */
-   if (DO_TEX3 && VB->TexCoordPtr[2] == 0)
-      VB->TexCoordPtr[2] = VB->TexCoordPtr[3];
+   if (DO_TEX3 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + 2] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX0 + 2] = VB->AttribPtr[_TNL_ATTRIB_TEX0 + 3];
 
-   if (DO_TEX2 && VB->TexCoordPtr[1] == 0)
-      VB->TexCoordPtr[1] = VB->TexCoordPtr[2];
+   if (DO_TEX2 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + 1] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX0 + 1] = VB->AttribPtr[_TNL_ATTRIB_TEX0 + 2];
 
-   if (DO_TEX1 && VB->TexCoordPtr[0] == 0)
-      VB->TexCoordPtr[0] = VB->TexCoordPtr[1];
+   if (DO_TEX1 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + 0] == 0)
+      VB->AttribPtr[_TNL_ATTRIB_TEX0 + 0] = VB->AttribPtr[_TNL_ATTRIB_TEX0 + 1];
 
    if (DO_PTEX)
       return GL_TRUE;
@@ -453,14 +453,14 @@ static GLboolean TAG(check_tex_sizes)( GLcontext *ctx )
    /* No hardware support for projective texture.  Can fake it for
     * TEX0 only.
     */
-   if ((DO_TEX3 && VB->TexCoordPtr[GET_TEXSOURCE(3)]->size == 4) ||
-       (DO_TEX2 && VB->TexCoordPtr[GET_TEXSOURCE(2)]->size == 4) ||
-       (DO_TEX1 && VB->TexCoordPtr[GET_TEXSOURCE(1)]->size == 4)) {
+   if ((DO_TEX3 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(3)]->size == 4) ||
+       (DO_TEX2 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(2)]->size == 4) ||
+       (DO_TEX1 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(1)]->size == 4)) {
       PTEX_FALLBACK();
       return GL_FALSE;
    }
 
-   if (DO_TEX0 && VB->TexCoordPtr[GET_TEXSOURCE(0)]->size == 4) {
+   if (DO_TEX0 && VB->AttribPtr[_TNL_ATTRIB_TEX0 + GET_TEXSOURCE(0)]->size == 4) {
       if (DO_TEX1 || DO_TEX2 || DO_TEX3) {
         PTEX_FALLBACK();
       }
index baa00af29a63b4ca1941cfe59380c788f555a69f..0af8893c30224902bac17138d877c8934af3114e 100644 (file)
@@ -211,7 +211,7 @@ static INLINE void insert_3f_xyw_4( const struct vf_attr *a, GLubyte *v, const G
 static INLINE void insert_3f_xyw_err( const struct vf_attr *a, GLubyte *v, const GLfloat *in )
 {
    (void) a; (void) v; (void) in;
-   _mesa_exit(1);
+   exit(1);
 }
 
 static INLINE void insert_3f_3( const struct vf_attr *a, GLubyte *v, const GLfloat *in )
index d56b701aa8c1f1a189558aa2fd76c657909fe862..0d7e0f1f98b85f58beae62e003b0a36c958d9dee 100644 (file)
@@ -120,22 +120,22 @@ int main( int argc, char **argv )
    OFFSET( "VB_COUNT               ", struct vertex_buffer, Count );
    printf( "\n" );
    OFFSET( "VB_ELTS                ", struct vertex_buffer, Elts );
-   OFFSET( "VB_OBJ_PTR             ", struct vertex_buffer, ObjPtr );
+   OFFSET( "VB_OBJ_PTR             ", struct vertex_buffer, AttribPtr[_TNL_ATTRIB_POS] );
    OFFSET( "VB_EYE_PTR             ", struct vertex_buffer, EyePtr );
    OFFSET( "VB_CLIP_PTR            ", struct vertex_buffer, ClipPtr );
    OFFSET( "VB_PROJ_CLIP_PTR       ", struct vertex_buffer, NdcPtr );
    OFFSET( "VB_CLIP_OR_MASK        ", struct vertex_buffer, ClipOrMask );
    OFFSET( "VB_CLIP_MASK           ", struct vertex_buffer, ClipMask );
-   OFFSET( "VB_NORMAL_PTR          ", struct vertex_buffer, NormalPtr );
+   OFFSET( "VB_NORMAL_PTR          ", struct vertex_buffer, AttribPtr[_TNL_ATTRIB_NORMAL] );
    OFFSET( "VB_EDGE_FLAG           ", struct vertex_buffer, EdgeFlag );
-   OFFSET( "VB_TEX0_COORD_PTR      ", struct vertex_buffer, TexCoordPtr[0] );
-   OFFSET( "VB_TEX1_COORD_PTR      ", struct vertex_buffer, TexCoordPtr[1] );
-   OFFSET( "VB_TEX2_COORD_PTR      ", struct vertex_buffer, TexCoordPtr[2] );
-   OFFSET( "VB_TEX3_COORD_PTR      ", struct vertex_buffer, TexCoordPtr[3] );
-   OFFSET( "VB_INDEX_PTR           ", struct vertex_buffer, IndexPtr );
-   OFFSET( "VB_COLOR_PTR           ", struct vertex_buffer, ColorPtr );
-   OFFSET( "VB_SECONDARY_COLOR_PTR ", struct vertex_buffer, SecondaryColorPtr );
-   OFFSET( "VB_FOG_COORD_PTR       ", struct vertex_buffer, FogCoordPtr );
+   OFFSET( "VB_TEX0_COORD_PTR      ", struct vertex_buffer, AttribPtr[_TNL_ATTRIB_TEX0] );
+   OFFSET( "VB_TEX1_COORD_PTR      ", struct vertex_buffer, AttribPtr[_TNL_ATTRIB_TEX1] );
+   OFFSET( "VB_TEX2_COORD_PTR      ", struct vertex_buffer, AttribPtr[_TNL_ATTRIB_TEX2] );
+   OFFSET( "VB_TEX3_COORD_PTR      ", struct vertex_buffer, AttribPtr[_TNL_ATTRIB_TEX3] );
+   OFFSET( "VB_INDEX_PTR           ", struct vertex_buffer, AttribPtr[_TNL_ATTRIB_COLOR_INDEX] );
+   OFFSET( "VB_COLOR_PTR           ", struct vertex_buffer, AttribPtr[_TNL_ATTRIB_COLOR0] );
+   OFFSET( "VB_SECONDARY_COLOR_PTR ", struct vertex_buffer, AttribPtr[_TNL_ATTRIB_COLOR1] );
+   OFFSET( "VB_FOG_COORD_PTR       ", struct vertex_buffer, AttribPtr[_TNL_ATTRIB_FOG] );
    OFFSET( "VB_PRIMITIVE           ", struct vertex_buffer, Primitive );
    printf( "\n" );
 
index 3a93544b03cf3898eadd7db11c3cca11521a41ca..caee6c0ca6e2b60624ea904590f0bc47f649e558 100644 (file)
@@ -22,7 +22,7 @@
                                Name="VCCLCompilerTool"\r
                                AdditionalOptions="/Zm1000 "\r
                                InlineFunctionExpansion="1"\r
-                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang,../../../../src/mesa/shader/grammar"\r
+                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang"\r
                                PreprocessorDefinitions="NDEBUG,WIN32,_LIB,_DLL,BUILD_GL32,MESA_MINWARN"\r
                                StringPooling="TRUE"\r
                                RuntimeLibrary="4"\r
@@ -74,7 +74,7 @@
                                Name="VCCLCompilerTool"\r
                                AdditionalOptions="/Zm1000 "\r
                                Optimization="0"\r
-                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang,../../../../src/mesa/shader/grammar"\r
+                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang"\r
                                PreprocessorDefinitions="_DEBUG,WIN32,_LIB,_DLL,BUILD_GL32,MESA_MINWARN"\r
                                BasicRuntimeChecks="3"\r
                                RuntimeLibrary="5"\r
                        <File\r
                                RelativePath="..\..\..\..\src\mesa\glapi\glthread.c">\r
                        </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar.c">\r
-                               <FileConfiguration\r
-                                       Name="Release|Win32"\r
-                                       ExcludedFromBuild="TRUE">\r
-                                       <Tool\r
-                                               Name="VCCLCompilerTool"/>\r
-                               </FileConfiguration>\r
-                               <FileConfiguration\r
-                                       Name="Debug|Win32"\r
-                                       ExcludedFromBuild="TRUE">\r
-                                       <Tool\r
-                                               Name="VCCLCompilerTool"/>\r
-                               </FileConfiguration>\r
-                       </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar_crt.c">\r
-                               <FileConfiguration\r
-                                       Name="Debug|Win32"\r
-                                       ExcludedFromBuild="TRUE">\r
-                                       <Tool\r
-                                               Name="VCCLCompilerTool"/>\r
-                               </FileConfiguration>\r
-                       </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar_mesa.c">\r
-                       </File>\r
                        <File\r
                                RelativePath="..\..\..\..\src\mesa\main\hash.c">\r
                        </File>\r
                        <File\r
                                RelativePath="..\..\..\..\src\mesa\glapi\glthread.h">\r
                        </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar.h">\r
-                       </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar_crt.h">\r
-                       </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar_mesa.h">\r
-                       </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar_syn.h">\r
-                       </File>\r
                        <File\r
                                RelativePath="..\..\..\..\src\mesa\main\hash.h">\r
                        </File>\r
index 993c28ddc1728f838257562916f18f0dd4150ac9..05bf7d2ea0f0e1676447c1917e4522955bd69f07 100644 (file)
@@ -43,7 +43,7 @@
                                Name="VCCLCompilerTool"\r
                                AdditionalOptions="/Zm1000 "\r
                                InlineFunctionExpansion="1"\r
-                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang,../../../../src/mesa/shader/grammar"\r
+                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang"\r
                                PreprocessorDefinitions="NDEBUG;WIN32;_LIB;_DLL;BUILD_GL32;MESA_MINWARN;_CRT_SECURE_NO_DEPRECATE"\r
                                StringPooling="true"\r
                                RuntimeLibrary="2"\r
                                Name="VCCLCompilerTool"\r
                                AdditionalOptions="/Zm1000 "\r
                                Optimization="0"\r
-                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang,../../../../src/mesa/shader/grammar"\r
+                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang"\r
                                PreprocessorDefinitions="_DEBUG;WIN32;_LIB;_DLL;BUILD_GL32;MESA_MINWARN;_CRT_SECURE_NO_DEPRECATE"\r
                                BasicRuntimeChecks="3"\r
                                RuntimeLibrary="3"\r
                                Name="VCCLCompilerTool"\r
                                AdditionalOptions="/Zm1000 "\r
                                Optimization="0"\r
-                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang,../../../../src/mesa/shader/grammar"\r
+                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang"\r
                                PreprocessorDefinitions="_DEBUG;WIN32;_LIB;BUILD_GL32;MESA_MINWARN;_CRT_SECURE_NO_DEPRECATE"\r
                                BasicRuntimeChecks="3"\r
                                RuntimeLibrary="1"\r
                                Name="VCCLCompilerTool"\r
                                AdditionalOptions="/Zm1000 "\r
                                InlineFunctionExpansion="1"\r
-                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang,../../../../src/mesa/shader/grammar"\r
+                               AdditionalIncludeDirectories="../../../../include,../../../../src/mesa,../../../../src/mesa/glapi,../../../../src/mesa/main,../../../../src/mesa/shader,../../../../src/mesa/shader/slang"\r
                                PreprocessorDefinitions="NDEBUG;WIN32;_LIB;BUILD_GL32;MESA_MINWARN;_CRT_SECURE_NO_DEPRECATE"\r
                                StringPooling="true"\r
                                RuntimeLibrary="0"\r
                                RelativePath="..\..\..\..\src\mesa\glapi\glthread.c"\r
                                >\r
                        </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar_mesa.c"\r
-                               >\r
-                       </File>\r
                        <File\r
                                RelativePath="..\..\..\..\src\mesa\main\hash.c"\r
                                >\r
                                RelativePath="..\..\..\..\src\mesa\glapi\glthread.h"\r
                                >\r
                        </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar.h"\r
-                               >\r
-                       </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar_crt.h"\r
-                               >\r
-                       </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar_mesa.h"\r
-                               >\r
-                       </File>\r
-                       <File\r
-                               RelativePath="..\..\..\..\src\mesa\shader\grammar\grammar_syn.h"\r
-                               >\r
-                       </File>\r
                        <File\r
                                RelativePath="..\..\..\..\src\mesa\main\hash.h"\r
                                >\r