mesa.git
12 years agosvga: check for and skip null vertex buffer pointers
Brian Paul [Mon, 30 Apr 2012 20:44:26 +0000 (14:44 -0600)]
svga: check for and skip null vertex buffer pointers

Fixes regressions with google earth and other things.

Reviewed-by: José Fonseca <jfonseca@vmware.com>
12 years agosoftpipe: cast away const to silence warning
Brian Paul [Mon, 30 Apr 2012 20:37:12 +0000 (14:37 -0600)]
softpipe: cast away const to silence warning

12 years agosvga: cast away const to silence warning
Brian Paul [Mon, 30 Apr 2012 20:36:50 +0000 (14:36 -0600)]
svga: cast away const to silence warning

12 years agou_vbuf: set user buffer pointer for drivers which support user buffers
Marek Olšák [Mon, 30 Apr 2012 22:11:25 +0000 (00:11 +0200)]
u_vbuf: set user buffer pointer for drivers which support user buffers

It's not common to end up in u_vbuf and at the same time support user buffers
in a driver, but such a combination should work.

12 years agost/mesa: unmap upload buffer after uploading indices and constants
Marek Olšák [Mon, 30 Apr 2012 18:56:25 +0000 (20:56 +0200)]
st/mesa: unmap upload buffer after uploading indices and constants

12 years agost/xorg: don't use user_buffer_create
Marek Olšák [Sun, 29 Apr 2012 19:34:43 +0000 (21:34 +0200)]
st/xorg: don't use user_buffer_create

12 years agost/xa: don't use user_buffer_create
Marek Olšák [Sun, 29 Apr 2012 19:34:18 +0000 (21:34 +0200)]
st/xa: don't use user_buffer_create

12 years agost/vega: don't use user_buffer_create
Marek Olšák [Sun, 29 Apr 2012 19:33:37 +0000 (21:33 +0200)]
st/vega: don't use user_buffer_create

12 years agost/vega: use cso_draw_arrays
Marek Olšák [Sun, 29 Apr 2012 19:32:52 +0000 (21:32 +0200)]
st/vega: use cso_draw_arrays

12 years agocso: cso_context should install u_vbuf by itself and not st/mesa
Marek Olšák [Sun, 29 Apr 2012 19:24:39 +0000 (21:24 +0200)]
cso: cso_context should install u_vbuf by itself and not st/mesa

so that it's installed in the other state trackers too

12 years agost/xorg: fix compilation - wrong libkms include file
Marek Olšák [Sun, 29 Apr 2012 18:43:12 +0000 (20:43 +0200)]
st/xorg: fix compilation - wrong libkms include file

12 years agogallium/util: stop using user buffers in util_draw_texquad
Marek Olšák [Thu, 26 Apr 2012 11:45:29 +0000 (13:45 +0200)]
gallium/util: stop using user buffers in util_draw_texquad

12 years agogallium: remove pipe_resource::user_ptr
Marek Olšák [Thu, 26 Apr 2012 11:41:33 +0000 (13:41 +0200)]
gallium: remove pipe_resource::user_ptr

It's unused now.

12 years agoradeonsi: don't create temporary user buffer for r600_upload_const_buffer
Marek Olšák [Thu, 26 Apr 2012 11:39:19 +0000 (13:39 +0200)]
radeonsi: don't create temporary user buffer for r600_upload_const_buffer

12 years agogallium: add void *user_buffer to pipe_constant_buffer
Marek Olšák [Tue, 24 Apr 2012 20:53:05 +0000 (22:53 +0200)]
gallium: add void *user_buffer to pipe_constant_buffer

This reduces CPU overhead when updating constants.

12 years agost/mesa: reorder code in draw_vbo
Marek Olšák [Thu, 26 Apr 2012 09:19:35 +0000 (11:19 +0200)]
st/mesa: reorder code in draw_vbo

12 years agost/mesa: remove more unnecessary code in draw_vbo
Marek Olšák [Thu, 26 Apr 2012 09:14:51 +0000 (11:14 +0200)]
st/mesa: remove more unnecessary code in draw_vbo

The variables set here are not used anywhere.

12 years agogallium: add void *user_buffer in pipe_index_buffer
Marek Olšák [Tue, 24 Apr 2012 19:14:44 +0000 (21:14 +0200)]
gallium: add void *user_buffer in pipe_index_buffer

Adapted drivers: i915, llvmpipe, r300, r600, radeonsi, softpipe.

User index buffers have been disabled in nv30, nv50, nvc0 and svga to keep
things working.

12 years agogallium: remove pipe_context::redefine_user_buffer
Marek Olšák [Tue, 24 Apr 2012 19:12:37 +0000 (21:12 +0200)]
gallium: remove pipe_context::redefine_user_buffer

12 years agogallium: add void *user_buffer in pipe_vertex_buffer
Marek Olšák [Tue, 24 Apr 2012 18:58:47 +0000 (20:58 +0200)]
gallium: add void *user_buffer in pipe_vertex_buffer

This reduces CPU overhead in st_draw_vbo and removes a lot of unnecessary code
in that function which was required only to comply with the gallium interface,
but wasn't any useful really.

Adapted drivers: i915, llvmpipe, r300, softpipe.
No changes required in: r600, radeonsi.

User vertex buffers have been disabled in nv30, nv50, nvc0 and svga to keep
things working.

12 years agost/mesa: make user constant buffers optional
Marek Olšák [Tue, 24 Apr 2012 18:16:50 +0000 (20:16 +0200)]
st/mesa: make user constant buffers optional

12 years agogallium: change set_constant_buffer to be UBO-friendly
Marek Olšák [Tue, 24 Apr 2012 17:52:26 +0000 (19:52 +0200)]
gallium: change set_constant_buffer to be UBO-friendly

12 years agogallium: add PIPE_CAP_CONSTANT_BUFFER_OFFSET_ALIGNMENT
Marek Olšák [Tue, 24 Apr 2012 15:31:17 +0000 (17:31 +0200)]
gallium: add PIPE_CAP_CONSTANT_BUFFER_OFFSET_ALIGNMENT

This is required for any serious constant buffer support.
Constant buffer offsets on ATI and NVIDIA DX10 and DX11 GPUs must be
a multiple of 256.

In OpenGL, this can be queried via GL_UNIFORM_BUFFER_OFFSET_ALIGNMENT.

12 years agost/mesa: make user index buffers optional
Marek Olšák [Tue, 24 Apr 2012 15:04:18 +0000 (17:04 +0200)]
st/mesa: make user index buffers optional

v2: use a separate upload buffer for indices

12 years agost/mesa: only set index buffer when drawing is indexed
Marek Olšák [Tue, 24 Apr 2012 13:50:44 +0000 (15:50 +0200)]
st/mesa: only set index buffer when drawing is indexed

and restructure the code a bit

12 years agogallium: add PIPE_CAP_USER_INDEX_BUFFERS and PIPE_CAP_USER_CONSTANT_BUFFERS
Marek Olšák [Tue, 24 Apr 2012 13:19:31 +0000 (15:19 +0200)]
gallium: add PIPE_CAP_USER_INDEX_BUFFERS and PIPE_CAP_USER_CONSTANT_BUFFERS

12 years agoscons: Parse = operator in source lists too.
José Fonseca [Sun, 29 Apr 2012 20:44:05 +0000 (21:44 +0100)]
scons: Parse = operator in source lists too.

Should fix the scons build.

12 years agonv50,nvc0: fix depth/stencil resolve
Christoph Bumiller [Sun, 29 Apr 2012 13:38:20 +0000 (15:38 +0200)]
nv50,nvc0: fix depth/stencil resolve

Cannot sample depth/stencil with a single view, and needed to use
different shader code for nve4.

12 years agonvc0/ir/opt: INTERP does not support JOIN
Christoph Bumiller [Sun, 29 Apr 2012 15:43:46 +0000 (17:43 +0200)]
nvc0/ir/opt: INTERP does not support JOIN

12 years agonv50/ir/opt: try to convert ABS(SUB) to SAD
Christoph Bumiller [Sat, 28 Apr 2012 15:06:59 +0000 (17:06 +0200)]
nv50/ir/opt: try to convert ABS(SUB) to SAD

12 years agonvc0/ir: try to use the optimal texture op mode
Christoph Bumiller [Fri, 27 Apr 2012 21:27:34 +0000 (23:27 +0200)]
nvc0/ir: try to use the optimal texture op mode

Don't really know what they are yet but for groups of textures, the
last one should use mode "p" and the others "t".

12 years agonvc0/ir: initial implementation of nve4 scheduling hints
Christoph Bumiller [Sun, 29 Apr 2012 15:59:06 +0000 (17:59 +0200)]
nvc0/ir: initial implementation of nve4 scheduling hints

12 years agonvc0/ir: implement better placement of texture barriers
Christoph Bumiller [Sun, 29 Apr 2012 15:56:57 +0000 (17:56 +0200)]
nvc0/ir: implement better placement of texture barriers

Put them before first uses instead of right after the texturing
instruction and cull unnecessary barriers.

12 years agonv50/ir/tgsi: fix handling of early RET
Christoph Bumiller [Sun, 29 Apr 2012 15:36:34 +0000 (17:36 +0200)]
nv50/ir/tgsi: fix handling of early RET

We have to actually emit RET, too, of course, not just the PRERET.

12 years agonvc0/ir/emit: fix emitTXQ 2nd src
Christoph Bumiller [Sat, 28 Apr 2012 16:04:05 +0000 (18:04 +0200)]
nvc0/ir/emit: fix emitTXQ 2nd src

12 years agonvc0/ir/target: integer ADD doesn't support ABS modifier
Christoph Bumiller [Sat, 28 Apr 2012 12:45:05 +0000 (14:45 +0200)]
nvc0/ir/target: integer ADD doesn't support ABS modifier

12 years agou_vbuf: unbind vertex buffers on destroy
Marek Olšák [Sun, 29 Apr 2012 15:41:37 +0000 (17:41 +0200)]
u_vbuf: unbind vertex buffers on destroy

12 years agou_blitter: fix resource leak
Marek Olšák [Sun, 29 Apr 2012 15:31:27 +0000 (17:31 +0200)]
u_blitter: fix resource leak

12 years agoradeonsi: make r600_buffer_transfer_unmap a no-op
Marek Olšák [Thu, 26 Apr 2012 10:12:33 +0000 (12:12 +0200)]
radeonsi: make r600_buffer_transfer_unmap a no-op

It's a no-op already in the winsys.

12 years agor600g: make r600_buffer_transfer_unmap a no-op
Marek Olšák [Thu, 26 Apr 2012 10:11:12 +0000 (12:11 +0200)]
r600g: make r600_buffer_transfer_unmap a no-op

It's a no-op already in the winsys.

12 years agor300g: make r300_buffer_transfer_unmap a no-op
Marek Olšák [Thu, 26 Apr 2012 10:10:06 +0000 (12:10 +0200)]
r300g: make r300_buffer_transfer_unmap a no-op

It's a no-op already in the winsys.

12 years agor300g: use u_default_transfer_inline_write
Marek Olšák [Thu, 26 Apr 2012 10:07:54 +0000 (12:07 +0200)]
r300g: use u_default_transfer_inline_write

12 years agoradeonsi: use u_default_transfer_inline_write
Marek Olšák [Thu, 26 Apr 2012 10:05:45 +0000 (12:05 +0200)]
radeonsi: use u_default_transfer_inline_write

12 years agowinsys/radeon: simplify buffer map/unmap functions
Marek Olšák [Thu, 26 Apr 2012 10:02:31 +0000 (12:02 +0200)]
winsys/radeon: simplify buffer map/unmap functions

The idea is not to use pb_map and pb_unmap wrappers, calling straight
into the winsys.

12 years agomesa: require GL_MAX_SAMPLES >= 4 for GL 3.0
Dylan Noblesmith [Sat, 21 Apr 2012 19:35:48 +0000 (19:35 +0000)]
mesa: require GL_MAX_SAMPLES >= 4 for GL 3.0

As noted in commit be4e46b21a60cfdc826bf89d1078df54966115b1,
this was missing before.

NOTE: This is a candidate for the 8.0 branch.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoautoconf: pass -Wall to automake
Dylan Noblesmith [Fri, 13 Apr 2012 20:36:39 +0000 (20:36 +0000)]
autoconf: pass -Wall to automake

And fix these warning that appear at autoreconf time:
"`:='-style assignments are not portable"

v2: Fix the recently-converted-to-automake r600.

12 years agoglsl: Remove unused member predicate from ir_dead_functions_visitor.
Vinson Lee [Fri, 27 Apr 2012 04:27:37 +0000 (21:27 -0700)]
glsl: Remove unused member predicate from ir_dead_functions_visitor.

Fix uninitialized pointer field defect reported by Coverity.

Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Ian Romanick <ian.d.romanick@intel.com>
12 years agoi965/fs: Fix FB writes that tried to use the non-existent m16 register.
Kenneth Graunke [Tue, 24 Apr 2012 21:09:13 +0000 (14:09 -0700)]
i965/fs: Fix FB writes that tried to use the non-existent m16 register.

A little analysis shows that the worst-case value for "nr" is 17:
- base_mrf = 2                       ... 2
- header present (say gen == 5)      ... 4
- aa_dest_stencil_reg (stencil test) ... 5
- SIMD16 mode: += 4 * reg_width      ... 13
- source_depth_to_render_target      ... 15
- dest_depth_reg                     ... 17

This resulted in us setting base_mrf to 2 and mlen to 15.  In other
words, we'd try to use m2..m16.  But m16 doesn't exist pre-Gen6.  Also,
the instruction scheduler data structures use arrays of size 16, so this
would cause us to access them out of bounds.

While the debugger system routine may need m0 and m1, we don't use it
today, so the simplest solution is just to move base_mrf back to 1.
That way, our worst case message fits in m1..m15, which is legal.

An alternative would be to fail on SIMD16 in this case, but that seems
a bit unfortunate if there's no real need to reserve m0 and m1.

Fixes new piglit test shaders/depth-test-and-write on Ironlake.

Bugzilla: https://bugs.freedesktop.org/show_bug.cgi?id=48218
Signed-off-by: Kenneth Graunke <kenneth@whitecape.org>
Reviewed-by: Eric Anholt <eric@anholt.net>
12 years agoglsl: Remove unused member mem_ctx from ir_dead_functions_visitor.
Vinson Lee [Thu, 26 Apr 2012 05:49:12 +0000 (22:49 -0700)]
glsl: Remove unused member mem_ctx from ir_dead_functions_visitor.

Fix uninitialized pointer field defect reported by Coverity.

Signed-off-by: Vinson Lee <vlee@freedesktop.org>
Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agonv30: properly init window information
Ben Skeggs [Thu, 26 Apr 2012 23:26:13 +0000 (09:26 +1000)]
nv30: properly init window information

Should fix >2k rendering issues reported on nv4x.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
12 years agoradeonsi/llvm: Silence a warning
Tom Stellard [Wed, 25 Apr 2012 22:09:05 +0000 (18:09 -0400)]
radeonsi/llvm: Silence a warning

12 years agoradeon/llvm: Remove unused header files
Tom Stellard [Wed, 25 Apr 2012 16:20:51 +0000 (12:20 -0400)]
radeon/llvm: Remove unused header files

12 years agoradeon/llvm: Remove AMDILMachineFunctionInfo.cpp
Tom Stellard [Wed, 25 Apr 2012 16:16:27 +0000 (12:16 -0400)]
radeon/llvm: Remove AMDILMachineFunctionInfo.cpp

12 years agoradeon/llvm: Remove AMDILModuleInfo.cpp
Tom Stellard [Wed, 25 Apr 2012 15:55:01 +0000 (11:55 -0400)]
radeon/llvm: Remove AMDILModuleInfo.cpp

12 years agogallivm: Use lp_build_alloca instead of LLVMBuildAlloca on the loop limiter.
José Fonseca [Wed, 25 Apr 2012 15:30:06 +0000 (16:30 +0100)]
gallivm: Use lp_build_alloca instead of LLVMBuildAlloca on the loop limiter.

To ensure that the alloca is at the top of the function body, otherwise
LLVM will not eliminate them, causing stack misalignment on 32bits.

Reviewed-by: James Benton <jbenton@vmware.com>
12 years agoradeon/llvm: Remove AMDILELFWriterInfo.cpp
Tom Stellard [Wed, 25 Apr 2012 02:04:43 +0000 (22:04 -0400)]
radeon/llvm: Remove AMDILELFWriterInfo.cpp

12 years agoradeon/llvm: Remove AMDILLiteralManager.cpp
Tom Stellard [Wed, 25 Apr 2012 01:57:14 +0000 (21:57 -0400)]
radeon/llvm: Remove AMDILLiteralManager.cpp

12 years agoradeon/llvm: Remove AMDILInliner.cpp
Tom Stellard [Wed, 25 Apr 2012 01:52:22 +0000 (21:52 -0400)]
radeon/llvm: Remove AMDILInliner.cpp

12 years agoradeon/llvm: Remove AMDILBarrierDetect.cpp
Tom Stellard [Wed, 25 Apr 2012 01:48:14 +0000 (21:48 -0400)]
radeon/llvm: Remove AMDILBarrierDetect.cpp

12 years agoradeon/llvm: Remove AMDILPrintfConvert.cpp
Tom Stellard [Wed, 25 Apr 2012 01:44:11 +0000 (21:44 -0400)]
radeon/llvm: Remove AMDILPrintfConvert.cpp

12 years agoradeon/llvm: Remove GlobalManager and KernelManager
Tom Stellard [Wed, 25 Apr 2012 01:39:11 +0000 (21:39 -0400)]
radeon/llvm: Remove GlobalManager and KernelManager

12 years agoradeon/llvm: Remove AsmPrinter files
Tom Stellard [Wed, 25 Apr 2012 01:20:26 +0000 (21:20 -0400)]
radeon/llvm: Remove AsmPrinter files

12 years agoradeon/llvm: Remove IOExpansion files
Tom Stellard [Wed, 25 Apr 2012 01:15:11 +0000 (21:15 -0400)]
radeon/llvm: Remove IOExpansion files

12 years agoradeon/llvm: Remove AMDILPointerManager.cpp
Tom Stellard [Tue, 24 Apr 2012 23:54:22 +0000 (19:54 -0400)]
radeon/llvm: Remove AMDILPointerManager.cpp

12 years agoradeonsi/llvm: Fix initialization of SIMachineFunctionInfo
Tom Stellard [Wed, 25 Apr 2012 13:03:21 +0000 (09:03 -0400)]
radeonsi/llvm: Fix initialization of SIMachineFunctionInfo

SIMachineFunctionInfo needs to be initialized before any of the
AMDIL passes.

12 years agomesa/st: Fix derreference after free.
José Fonseca [Wed, 25 Apr 2012 10:41:13 +0000 (11:41 +0100)]
mesa/st: Fix derreference after free.

12 years agogallium/docs: document the new vertex fetch CAPs
Marek Olšák [Tue, 24 Apr 2012 21:10:46 +0000 (23:10 +0200)]
gallium/docs: document the new vertex fetch CAPs

12 years agonv50: fix typo in nv50_fragprog_assign_slots
Christoph Bumiller [Tue, 24 Apr 2012 18:04:48 +0000 (20:04 +0200)]
nv50: fix typo in nv50_fragprog_assign_slots

12 years agoglsl: Add implementation of inverse() for mat2/3/4.
Eric Anholt [Wed, 18 Apr 2012 18:00:44 +0000 (11:00 -0700)]
glsl: Add implementation of inverse() for mat2/3/4.

This is taken from the ogl-math project, with Inverse renamed to adj
(since it's not actually the inverse), transposed, and our types
plugged in.  There are potential CSE opportunities in this code
(particularly for hardware with RCP but not DIV), but we should be
doing CSE anyway, so don't hand-optimize.

Fixes piglit inverse tests.

Acked-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoglsl: Add support for generating builtin code from GLSL instead of IR.
Eric Anholt [Wed, 18 Apr 2012 17:51:23 +0000 (10:51 -0700)]
glsl: Add support for generating builtin code from GLSL instead of IR.

This takes advantage of the builtin compiler to generate IR into a
string, the same way we read GLSL for function prototypes for our
profiles.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoglsl: Make ir_reader parse the "temporary" variable qualifier.
Eric Anholt [Wed, 18 Apr 2012 21:43:12 +0000 (14:43 -0700)]
glsl: Make ir_reader parse the "temporary" variable qualifier.

This lets ir_reader eat the output of builtin_compiler on actual
function definitions.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoglsl: Add doxygen explaining what main.cpp is for.
Eric Anholt [Wed, 18 Apr 2012 21:40:10 +0000 (14:40 -0700)]
glsl: Add doxygen explaining what main.cpp is for.

I keep getting lost in the Makefile trying to figure out what to edit
to work on builtin_compiler or glsl_compiler.

Reviewed-by: Kenneth Graunke <kenneth@whitecape.org>
12 years agoi965/vs: Fix texelFetchOffset()
Eric Anholt [Tue, 17 Apr 2012 18:56:26 +0000 (11:56 -0700)]
i965/vs: Fix texelFetchOffset()

It appears that when using 'ld' with the offset bits, address bounds
checking happens before the offset is applied, so parts of the drawing
in piglit texelFetchOffset() with a negative texcoord go black.

12 years agoi965/fs: Fix texelFetchOffset()
Eric Anholt [Tue, 17 Apr 2012 01:43:48 +0000 (18:43 -0700)]
i965/fs: Fix texelFetchOffset()

It appears that when using 'ld' with the offset bits, address bounds
checking happens before the offset is applied, so parts of the drawing
in piglit texelFetchOffset() with a negative texcoord go black.

12 years agonv50,nvc0: handle new vertex fetch caps
Christoph Bumiller [Tue, 24 Apr 2012 11:34:36 +0000 (13:34 +0200)]
nv50,nvc0: handle new vertex fetch caps

12 years agonv50,nvc0: don't initialize the draw module, we don't use it
Christoph Bumiller [Tue, 24 Apr 2012 11:35:44 +0000 (13:35 +0200)]
nv50,nvc0: don't initialize the draw module, we don't use it

But some day we might (e.g. for blending 16 bpp formats on nv50).

12 years agonv50: enable shader watchdog timer
Christoph Bumiller [Mon, 23 Apr 2012 18:08:54 +0000 (20:08 +0200)]
nv50: enable shader watchdog timer

12 years agodarwin: Eliminate a pthread mutex leak
Jeremy Huddleston [Mon, 23 Apr 2012 23:43:22 +0000 (16:43 -0700)]
darwin: Eliminate a pthread mutex leak

Signed-off-by: Jeremy Huddleston <jeremyhu@apple.com>
Tested-by: Charles Davis <cdavis@mines.edu>
12 years agoradeonsi: stop using u_vbuf and adapt to gallium interface changes
Marek Olšák [Mon, 23 Apr 2012 23:23:33 +0000 (01:23 +0200)]
radeonsi: stop using u_vbuf and adapt to gallium interface changes

12 years agou_vbuf: don't call u_vbuf_upload_buffers if there's nothing to upload
Marek Olšák [Thu, 19 Apr 2012 17:54:23 +0000 (19:54 +0200)]
u_vbuf: don't call u_vbuf_upload_buffers if there's nothing to upload

12 years agou_vbuf: don't upload user buffers if the driver supports them
Marek Olšák [Thu, 19 Apr 2012 03:28:27 +0000 (05:28 +0200)]
u_vbuf: don't upload user buffers if the driver supports them

12 years agou_vbuf: optimize u_vbuf_translate_find_free_vb_slots
Marek Olšák [Thu, 19 Apr 2012 01:47:03 +0000 (03:47 +0200)]
u_vbuf: optimize u_vbuf_translate_find_free_vb_slots

12 years agou_vbuf: turn bool arrays into bitmasks and loops into bitwise operations
Marek Olšák [Wed, 18 Apr 2012 22:56:44 +0000 (00:56 +0200)]
u_vbuf: turn bool arrays into bitmasks and loops into bitwise operations

Optimized functions:
- u_vbuf_need_minmax_index
- u_vbuf_mapping_vertex_buffer_blocks

12 years agou_vbuf: take advantage of all new vertex fetch caps
Marek Olšák [Wed, 18 Apr 2012 23:41:46 +0000 (01:41 +0200)]
u_vbuf: take advantage of all new vertex fetch caps

12 years agou_vbuf: simplify u_vbuf_set_vertex_buffers
Marek Olšák [Wed, 18 Apr 2012 23:24:32 +0000 (01:24 +0200)]
u_vbuf: simplify u_vbuf_set_vertex_buffers

12 years agou_vbuf: rename native_attribs -> driver_attribs
Marek Olšák [Wed, 18 Apr 2012 23:12:55 +0000 (01:12 +0200)]
u_vbuf: rename native_attribs -> driver_attribs

12 years agor600g: move updating vertex buffer state into set_vertex_buffers
Marek Olšák [Sun, 1 Apr 2012 21:54:44 +0000 (23:54 +0200)]
r600g: move updating vertex buffer state into set_vertex_buffers

12 years agost/mesa: always expose ARB_ES2_compatibility
Marek Olšák [Wed, 11 Apr 2012 10:55:23 +0000 (12:55 +0200)]
st/mesa: always expose ARB_ES2_compatibility

u_vbuf translates GL_FIXED too if needed.

12 years agost/mesa: always expose ARB_half_float_vertex
Marek Olšák [Wed, 11 Apr 2012 10:54:10 +0000 (12:54 +0200)]
st/mesa: always expose ARB_half_float_vertex

u_vbuf kicks in and translates it to float if it's unsupported.

12 years agogallium: make user vertex buffers optional
Marek Olšák [Tue, 10 Apr 2012 04:00:17 +0000 (06:00 +0200)]
gallium: make user vertex buffers optional

This couldn't be split because it would break bisecting.

Summary:
* r300g,r600g: stop using u_vbuf
* r300g,r600g: also report that the FIXED vertex type is unsupported
* u_vbuf: refactor for use in the state tracker
* cso: wire up u_vbuf with cso_context
* st/mesa: conditionally install u_vbuf

12 years agoi915g: report that all vertex formats are supported
Marek Olšák [Wed, 11 Apr 2012 13:16:09 +0000 (15:16 +0200)]
i915g: report that all vertex formats are supported

So that u_vbuf isn't enabled.

12 years agor600g: don't share u_upload_mgr with u_vbuf, create its own
Marek Olšák [Wed, 11 Apr 2012 14:00:09 +0000 (16:00 +0200)]
r600g: don't share u_upload_mgr with u_vbuf, create its own

12 years agor300g: don't share u_upload_mgr with u_vbuf, create its own
Marek Olšák [Wed, 11 Apr 2012 14:00:09 +0000 (16:00 +0200)]
r300g: don't share u_upload_mgr with u_vbuf, create its own

12 years agou_vbuf: pull u_vbuf_draw_max_vertex_count into r300g
Marek Olšák [Wed, 11 Apr 2012 13:56:59 +0000 (15:56 +0200)]
u_vbuf: pull u_vbuf_draw_max_vertex_count into r300g

12 years agou_vbuf: make use of the new CAPs to determine what to do
Marek Olšák [Tue, 10 Apr 2012 03:14:26 +0000 (05:14 +0200)]
u_vbuf: make use of the new CAPs to determine what to do

This adds the ability to initialize u_vbuf_caps before creating u_vbuf itself.
It will be useful for determining if u_vbuf should be used or not.

Also adapt r300g and r600g.

12 years agogallium drivers: report that user vertex buffers are supported
Marek Olšák [Wed, 11 Apr 2012 13:40:00 +0000 (15:40 +0200)]
gallium drivers: report that user vertex buffers are supported

12 years agogallium: add CAPs for vertex fetcher
Marek Olšák [Tue, 10 Apr 2012 03:12:22 +0000 (05:12 +0200)]
gallium: add CAPs for vertex fetcher

Supported vertex formats will be queried using
is_format_supported(.., PIPE_BIND_VERTEX_BUFFER, ..).

12 years agou_vbuf: remove u_vbuf_resource
Marek Olšák [Tue, 10 Apr 2012 19:29:06 +0000 (21:29 +0200)]
u_vbuf: remove u_vbuf_resource

12 years agou_vbuf: use user_ptr from pipe_resource
Marek Olšák [Tue, 10 Apr 2012 02:37:52 +0000 (04:37 +0200)]
u_vbuf: use user_ptr from pipe_resource

12 years agogallium: add user_ptr in pipe_resource
Marek Olšák [Tue, 10 Apr 2012 02:33:23 +0000 (04:33 +0200)]
gallium: add user_ptr in pipe_resource

I need to access the pointer in st/mesa when I only have pipe_resource.