mesa.git
15 years agoi915g: Disable vbuf fifo and minor commenting of vbuf code
Jakob Bornecrantz [Wed, 7 Oct 2009 00:40:37 +0000 (01:40 +0100)]
i915g: Disable vbuf fifo and minor commenting of vbuf code

The vbuf fifo doesn't appear to help once the libdrm
reuse flag has been set.

15 years agonv04-nv40: Fix swizzle transfers for NPOT sizes.
Younes Manton [Wed, 7 Oct 2009 02:07:47 +0000 (22:07 -0400)]
nv04-nv40: Fix swizzle transfers for NPOT sizes.

Workarounds not necessary, SIFM can handle NPOT, we just weren't setting
dst dimensions properly. SIFM can't handle odd w,h though, that still
needs fixing.

15 years agoprogs/tests: fix MSVC build.
Vinson Lee [Tue, 6 Oct 2009 22:02:47 +0000 (16:02 -0600)]
progs/tests: fix MSVC build.

15 years agoprogs/perf: Fix MSVC build.
Vinson Lee [Tue, 6 Oct 2009 01:42:58 +0000 (18:42 -0700)]
progs/perf: Fix MSVC build.

15 years agor600: r700PredictRenderSize can flush, so move index buffer setup after it
Andre Maasikas [Mon, 5 Oct 2009 06:38:52 +0000 (09:38 +0300)]
r600: r700PredictRenderSize can flush, so move index buffer setup after it

15 years agonv50: fix segfault when there's gaps in enabled texture units
Ben Skeggs [Tue, 6 Oct 2009 03:55:48 +0000 (13:55 +1000)]
nv50: fix segfault when there's gaps in enabled texture units

Tested with progs/demos/multiarb.

15 years agonouveau: fix buffer object leak
Ben Skeggs [Tue, 6 Oct 2009 03:18:09 +0000 (13:18 +1000)]
nouveau: fix buffer object leak

Very apparent with resizing windows on DRI2.

15 years agoi915g: Add more defines to tweek batchbuffer
Jakob Bornecrantz [Tue, 6 Oct 2009 01:30:43 +0000 (18:30 -0700)]
i915g: Add more defines to tweek batchbuffer

15 years agoi915g: Tweek vertexbuffer size
Jakob Bornecrantz [Tue, 6 Oct 2009 01:11:25 +0000 (18:11 -0700)]
i915g: Tweek vertexbuffer size

15 years agoi965g: remove 965/brw files from XLIB_WINSYS_SOURCES
Brian Paul [Mon, 5 Oct 2009 20:43:38 +0000 (14:43 -0600)]
i965g: remove 965/brw files from XLIB_WINSYS_SOURCES

15 years agoMerge branch 'mesa_7_6_branch'
Brian Paul [Mon, 5 Oct 2009 20:42:08 +0000 (14:42 -0600)]
Merge branch 'mesa_7_6_branch'

Conflicts:

src/gallium/auxiliary/util/u_cpu_detect.c

15 years agointel: use driReadDrawable in do_copy_texsubimage()
Brian Paul [Mon, 5 Oct 2009 20:26:16 +0000 (14:26 -0600)]
intel: use driReadDrawable in do_copy_texsubimage()

15 years agointel: use driReadDrawable, not driDrawable in do_blit_readpixels()
Brian Paul [Mon, 5 Oct 2009 20:25:36 +0000 (14:25 -0600)]
intel: use driReadDrawable, not driDrawable in do_blit_readpixels()

15 years agointel: remove a buffer equality test in _mesa_make_current()
Brian Paul [Mon, 5 Oct 2009 20:07:29 +0000 (14:07 -0600)]
intel: remove a buffer equality test in _mesa_make_current()

Before, if we called glXMakeCurrent() to change a context's window binding
while an FBO was bound, we weren't updating the intel->driDrawable and
intel->driReadDrawable fields.  This could cause us to dereference a null
pointer elsewhere.

15 years agoi915g: Drop the simple sufix
Jakob Bornecrantz [Tue, 29 Sep 2009 23:07:11 +0000 (16:07 -0700)]
i915g: Drop the simple sufix

None of the other driver have a silly sufix,
so just drop it. Nothing new added in this commit
or any other commit but this is better marketing.

15 years agoi965g: Drop i965simple
Jakob Bornecrantz [Tue, 29 Sep 2009 17:22:15 +0000 (10:22 -0700)]
i965g: Drop i965simple

The driver never work with real hardware and has
bitrotted for quite some time now, might as well
drop it. If somebody wants to look at it just use git.

15 years agor300g: fix scons build
Joakim Sindholt [Mon, 5 Oct 2009 17:25:04 +0000 (19:25 +0200)]
r300g: fix scons build

So I didn't touch r300compiler, but r300g now compiles after having
declarations and code untangled. As nha so gently points out, we shouldn't
have to do this just to comply with MSVC compilers.

15 years agomesa: fix return value when clipping {Read,Draw}Pixels height <= 0
Brian Paul [Mon, 5 Oct 2009 17:10:31 +0000 (11:10 -0600)]
mesa: fix return value when clipping {Read,Draw}Pixels height <= 0

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
(cherry picked from master, commit 7aeaca33c331f70d507fc83583b13b8d9fc3e847)

15 years agoFix building of GLSL demos which use M_PI
Jon TURNEY [Fri, 24 Jul 2009 19:33:25 +0000 (20:33 +0100)]
Fix building of GLSL demos which use M_PI

Some <math.h> files do not define M_PI, in which case, provide our own definition

Signed-off-by: Jon TURNEY <jon.turney@dronecode.org.uk>
15 years agoutil: Improve the cpuid assembly.
José Fonseca [Mon, 5 Oct 2009 15:45:38 +0000 (16:45 +0100)]
util: Improve the cpuid assembly.

No need to save ebx on 64bit. Use just xchgl. Refer to gcc's cpuid.h header.

Thanks to Uros Bizjak for pointing this out.

15 years agomesa: Copy textures' base format into wrapper renderbuffer.
José Fonseca [Mon, 5 Oct 2009 10:05:34 +0000 (11:05 +0100)]
mesa: Copy textures' base format into wrapper renderbuffer.

Otherwise st_copy_texsubimage will fallback to software blit due to
inconsistent base formats.

15 years agomesa/st: add ST_DEBUG environment variable
Keith Whitwell [Mon, 5 Oct 2009 14:50:11 +0000 (15:50 +0100)]
mesa/st: add ST_DEBUG environment variable

At last it's possible to turn on tgsi dumps and other debugging in the
state tracker without modifying sources...

15 years agoutil: add lost code to util_make_fragment_tex_shader_writemask()
Keith Whitwell [Mon, 5 Oct 2009 14:46:47 +0000 (15:46 +0100)]
util: add lost code to util_make_fragment_tex_shader_writemask()

This got ported to ureg at some point, but lost the code that
distinguishes it from regular util_make_fragment_tex_shader().

15 years agor600: update vertex program selection for draw path
Andre Maasikas [Mon, 5 Oct 2009 09:53:40 +0000 (12:53 +0300)]
r600: update vertex program selection for draw path

15 years agoUse _mesa_select_tex_image() rather than hardcoding face 0.
Michel Dänzer [Mon, 5 Oct 2009 10:31:51 +0000 (12:31 +0200)]
Use _mesa_select_tex_image() rather than hardcoding face 0.

Fixes crash loading a map in sauerbraten with

hwmipmap 1

in ~/.sauerbraten/config.cfg.

15 years agost/dri: avoid segfault if we can't get a pixmap's buffers
Ben Skeggs [Mon, 5 Oct 2009 05:51:47 +0000 (15:51 +1000)]
st/dri: avoid segfault if we can't get a pixmap's buffers

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
15 years agonv50: support PIPE_FORMAT_X8R8G8B8_UNORM
Ben Skeggs [Mon, 5 Oct 2009 03:41:33 +0000 (13:41 +1000)]
nv50: support PIPE_FORMAT_X8R8G8B8_UNORM

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
15 years agonouveau: return pitch value from local_handle_from_texture()
Ben Skeggs [Mon, 5 Oct 2009 03:39:34 +0000 (13:39 +1000)]
nouveau: return pitch value from local_handle_from_texture()

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
15 years agomesa: fix return value when clipping {Read,Draw}Pixels height <= 0
Ben Skeggs [Mon, 5 Oct 2009 04:42:45 +0000 (14:42 +1000)]
mesa: fix return value when clipping {Read,Draw}Pixels height <= 0

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
15 years agost/dri: no need to request fake front buffer, only handle it being returned
Ben Skeggs [Mon, 5 Oct 2009 05:32:55 +0000 (15:32 +1000)]
st/dri: no need to request fake front buffer, only handle it being returned

The previous behaviour was incorrect, and resulted in EXT_tfp being broken
for DDX drivers that implement the correct behaviour (intel/radeon/nouveau).

In the cases where a fake front buffer is required, the X server will
return one when requesting __DRI_BUFFER_FRONT_LEFT.

The Xorg state tracker (aka modesetting_drv) is likely broken now until
it's modified to match the other drivers.

Signed-off-by: Ben Skeggs <bskeggs@redhat.com>
15 years agor200: remove subpixel offset from viewport
Frederic Crozat [Sun, 4 Oct 2009 21:46:40 +0000 (17:46 -0400)]
r200: remove subpixel offset from viewport

Fixes bug fdo 20340 for r200.

15 years agoutil: Make assert a no-op on non-debug builds.
José Fonseca [Sun, 4 Oct 2009 20:59:24 +0000 (21:59 +0100)]
util: Make assert a no-op on non-debug builds.

This ensures that an assertion like

  assert(expensive_test());

won't have any penalty on release builds. It also implies that no vital
code should be in assert expressions.

15 years agollvmpipe: Ensure tile cache transfers are mapped before flushing it.
José Fonseca [Sun, 4 Oct 2009 12:25:24 +0000 (13:25 +0100)]
llvmpipe: Ensure tile cache transfers are mapped before flushing it.

15 years agollvmpipe: Autogenerate lp_tile_soa.c from u_format.csv.
José Fonseca [Sun, 4 Oct 2009 12:04:08 +0000 (13:04 +0100)]
llvmpipe: Autogenerate lp_tile_soa.c from u_format.csv.

This is just a temporary change until we code generate the tile read/write
functions in runtime. The new code avoids an extra memcpy that exists in
u_tile.c functions, from which lp_tile_soa.c was originally based.

This achieves up to 5% improvement, particularly in frames with
little geometry overlap.

15 years agoutil: Fix cpuid invocation for x86_64.
José Fonseca [Sun, 4 Oct 2009 11:49:31 +0000 (12:49 +0100)]
util: Fix cpuid invocation for x86_64.

15 years agollvmpipe: Remove loop testing from format testing.
José Fonseca [Sun, 4 Oct 2009 10:36:42 +0000 (11:36 +0100)]
llvmpipe: Remove loop testing from format testing.

Loop building will be rewritten.

15 years agollvmpipe: Match header's protection macro with filename.
José Fonseca [Sun, 4 Oct 2009 10:35:50 +0000 (11:35 +0100)]
llvmpipe: Match header's protection macro with filename.

15 years agollvmpipe: Adjust format assertion.
José Fonseca [Sat, 3 Oct 2009 21:15:17 +0000 (22:15 +0100)]
llvmpipe: Adjust format assertion.

We support array layout too -- if it has a single channel.

15 years agor300g: xRGB and RGBx formats.
Corbin Simpson [Thu, 1 Oct 2009 02:44:38 +0000 (19:44 -0700)]
r300g: xRGB and RGBx formats.

We now have 48 GLX visuals. Pretty soon, we'll have 90+ visuals,
only five of which ever get tested. :3

15 years agointel: Suppress a compiler warning for an pointer->int cast
Chris Wilson [Sat, 3 Oct 2009 22:08:39 +0000 (23:08 +0100)]
intel: Suppress a compiler warning for an pointer->int cast

intel_pixel_read.c: In function ‘do_blit_readpixels’:
intel_pixel_read.c:221: warning: cast from pointer to integer of
different size

Cast via an intermediate (GLintptr) instead and hope the result fits
within GLuint... [It should as we simply do not support textures *that*
large!]

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15 years agointel: Assert that relocation offsets are within the target
Chris Wilson [Fri, 2 Oct 2009 18:46:03 +0000 (19:46 +0100)]
intel: Assert that relocation offsets are within the target

This should catch the common programming error where we attempt to
emit a relocation to beyond the end of the target buffer.

Signed-off-by: Chris Wilson <chris@chris-wilson.co.uk>
15 years agollvmpipe: Fetch tile only if a color buffer is bound.
José Fonseca [Sat, 3 Oct 2009 18:42:22 +0000 (19:42 +0100)]
llvmpipe: Fetch tile only if a color buffer is bound.

15 years agoprog_parameter: Document the fact that Size may be > 4
Nicolai Hähnle [Sat, 3 Oct 2009 14:18:57 +0000 (16:18 +0200)]
prog_parameter: Document the fact that Size may be > 4

Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
15 years agor300g: Do not abort on fragment program compiler error
Nicolai Hähnle [Sat, 3 Oct 2009 17:20:31 +0000 (19:20 +0200)]
r300g: Do not abort on fragment program compiler error

Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
15 years agor300g: Fix memory leak in radeon_texture_from_shared_handle
Nicolai Hähnle [Sat, 3 Oct 2009 15:56:51 +0000 (17:56 +0200)]
r300g: Fix memory leak in radeon_texture_from_shared_handle

Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
15 years agor300g: Build in the trace and softpipe driver for xorg state tracker
Sedat Dilek [Sat, 3 Oct 2009 16:01:58 +0000 (18:01 +0200)]
r300g: Build in the trace and softpipe driver for xorg state tracker

Same as in src/gallium/winsys/drm/intel/xorg/Makefile

Thanks MrCooper for explanations on IRC

[ Summary amended by Michel Dänzer to clarify that this is related to the xorg
  state tracker ]

15 years agoradeon: Cope better with texture images with no miptrees.
Michel Dänzer [Sat, 3 Oct 2009 16:01:58 +0000 (18:01 +0200)]
radeon: Cope better with texture images with no miptrees.

Fixes crash with compiz magnifier plugin.

15 years agoswrast: Move up state validation in _swrast_ReadPixels.
Michel Dänzer [Sat, 3 Oct 2009 16:01:58 +0000 (18:01 +0200)]
swrast: Move up state validation in _swrast_ReadPixels.

This ensures the driver won't map the wrong set of textures.

15 years agometa: Make sure texImage->TexFormat is valid for CopyTex(Sub)Image.
Michel Dänzer [Sat, 3 Oct 2009 16:01:57 +0000 (18:01 +0200)]
meta: Make sure texImage->TexFormat is valid for CopyTex(Sub)Image.

15 years agor300g: Remove an unnecessarily created pipe buffer (and thus fix a leak)
Nicolai Hähnle [Sat, 3 Oct 2009 15:49:16 +0000 (17:49 +0200)]
r300g: Remove an unnecessarily created pipe buffer (and thus fix a leak)

Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
15 years agost/dri: Install APPLE_vertex_array_object functions
Nicolai Hähnle [Sat, 3 Oct 2009 15:39:32 +0000 (17:39 +0200)]
st/dri: Install APPLE_vertex_array_object functions

Besides from being necessary to use that extension, it also fixes a crash
when deleting the currently bound vertex array object.

Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
15 years agor300g: Reset vbo_offset after allocation of a new buffer
Nicolai Hähnle [Sat, 3 Oct 2009 15:24:04 +0000 (17:24 +0200)]
r300g: Reset vbo_offset after allocation of a new buffer

This fixes the glxgears bug, among other things.

Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
15 years agost/dri: Install ARB_vertex_array_object functions
Nicolai Hähnle [Sat, 3 Oct 2009 00:11:02 +0000 (02:11 +0200)]
st/dri: Install ARB_vertex_array_object functions

Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
15 years agor300: Workaround problem on R500 with very large fragment programs
Nicolai Hähnle [Fri, 2 Oct 2009 23:26:38 +0000 (01:26 +0200)]
r300: Workaround problem on R500 with very large fragment programs

The non-KMS interface is to blame here. In theory, a proper fix
could be produced that works for the KMS interface only, but it
require cleaning a lot of mess. Easier to just do it right in r300g.

Signed-off-by: Nicolai Hähnle <nhaehnle@gmail.com>
15 years agomesa: optimized _mesa_meta_BlitFramebuffer() for src=texture case
Brian Paul [Fri, 2 Oct 2009 19:59:41 +0000 (13:59 -0600)]
mesa: optimized _mesa_meta_BlitFramebuffer() for src=texture case

If the src renderbuffer is actually a texture, we can directly use that
texture as the src and avoid a copy.

15 years agoi965: Use a little stack space to avoid a malloc in wm_get_binding_table.
Eric Anholt [Fri, 2 Oct 2009 17:53:56 +0000 (10:53 -0700)]
i965: Use a little stack space to avoid a malloc in wm_get_binding_table.

15 years agomesa: Remove another unexplained Flush call, this time from BindFramebuffer.
Eric Anholt [Fri, 2 Oct 2009 00:59:05 +0000 (17:59 -0700)]
mesa: Remove another unexplained Flush call, this time from BindFramebuffer.

Combined with the previous fix, it takes cairo-gl firefox-talos-gfx time
from 120 seconds to 90 seconds on my GM45.

15 years agoRevert "Flush driver, not just tnl module."
Eric Anholt [Thu, 1 Oct 2009 23:53:12 +0000 (16:53 -0700)]
Revert "Flush driver, not just tnl module."

This reverts commit df058298e1570eea8712f9bb051f674fab2eaf24.  It didn't
explain why it was required, doesnt appear to be required, and is a
significant performance penalty for cairo-gl firefox.

Conflicts:

src/mesa/main/fbobject.c

15 years agointel: Remove an unexplained flush from intelClearWithBlit.
Eric Anholt [Fri, 2 Oct 2009 01:07:57 +0000 (18:07 -0700)]
intel: Remove an unexplained flush from intelClearWithBlit.

15 years agor600: remove support for host-based ibs
Alex Deucher [Fri, 2 Oct 2009 18:25:52 +0000 (14:25 -0400)]
r600: remove support for host-based ibs

no longer used now that the hw supports this natively.

Also, clean up some formatting.

15 years agomesa/st: don't reuse vertex buffers for bitmap, clear quads
Keith Whitwell [Fri, 2 Oct 2009 14:36:47 +0000 (15:36 +0100)]
mesa/st: don't reuse vertex buffers for bitmap, clear quads

Currently using max_slots > 1 will cause synchronous rendering
if the driver flushes its command buffers between one bitmap and the
next.

Need to improve buffer_write to allow NO_WAIT (as well as no_flush)
updates to buffers where we know there is no conflict with previous
data.

15 years agost/xorg: Use PIPE_TRANSFER_MAP_DIRECTLY flag in EXA PrepareAccess hook.
Michel Dänzer [Fri, 2 Oct 2009 16:13:26 +0000 (18:13 +0200)]
st/xorg: Use PIPE_TRANSFER_MAP_DIRECTLY flag in EXA PrepareAccess hook.

Propagate NULL return value.

This also allows removing the DRM_MODE_FEATURE_DIRTYFB specific pixmap
management hacks.

15 years agogallium: Add PIPE_TRANSFER_MAP_DIRECTLY usage flag.
Michel Dänzer [Fri, 2 Oct 2009 16:13:26 +0000 (18:13 +0200)]
gallium: Add PIPE_TRANSFER_MAP_DIRECTLY usage flag.

Asks the driver to map the texture storage directly or return NULL if that's
not possible.

15 years agogallium: Preparations for adding more PIPE_TRANSFER_* usage flags.
Michel Dänzer [Fri, 2 Oct 2009 16:13:26 +0000 (18:13 +0200)]
gallium: Preparations for adding more PIPE_TRANSFER_* usage flags.

Always test for PIPE_TRANSFER_READ/WRITE using the bit-wise and operator, and
add a pipe_transfer_buffer_flags() helper for getting the buffer usage flags
corresponding to them.

15 years agodocs: document default texture binding fix
Brian Paul [Fri, 2 Oct 2009 15:58:16 +0000 (09:58 -0600)]
docs: document default texture binding fix

15 years agomesa: fix incorrect default texture binding in unbind_texobj_from_texunits()
Brian Paul [Fri, 2 Oct 2009 15:55:07 +0000 (09:55 -0600)]
mesa: fix incorrect default texture binding in unbind_texobj_from_texunits()

If we deleted a currently bound texture, we were always reverting the texture
binding to the default 1D texture rather than the proper default texture.

15 years agointel: wrap _mesa_meta_GenerateMipmap()
Brian Paul [Fri, 2 Oct 2009 14:55:25 +0000 (08:55 -0600)]
intel: wrap _mesa_meta_GenerateMipmap()

Need to check if we'll take the software path so which requires mapping the
src texture image.

Fixes crash in piglit gen-compressed-teximage, bug 24219.  However, the
test still does not pass (it may never have).

15 years agomesa: added _mesa_meta_check_generate_mipmap_fallback()
Brian Paul [Fri, 2 Oct 2009 14:54:55 +0000 (08:54 -0600)]
mesa: added _mesa_meta_check_generate_mipmap_fallback()

15 years agogallium/xlib: use bitwise-and to test GLX_RGBA_BIT in choose_visual()
Brian Paul [Fri, 2 Oct 2009 13:52:27 +0000 (07:52 -0600)]
gallium/xlib: use bitwise-and to test GLX_RGBA_BIT in choose_visual()

The parameter is a bitmask.

15 years agoxlib: use bitwise-and to test GLX_RGBA_BIT in choose_visual()
Brian Paul [Fri, 2 Oct 2009 13:51:15 +0000 (07:51 -0600)]
xlib: use bitwise-and to test GLX_RGBA_BIT in choose_visual()

The parameter is a bitmask.

15 years agogallium/xlib: return GLX_RGBA_BIT or GLX_COLOR_INDEX_BIT in get_config()
Brian Paul [Fri, 2 Oct 2009 13:47:49 +0000 (07:47 -0600)]
gallium/xlib: return GLX_RGBA_BIT or GLX_COLOR_INDEX_BIT in get_config()

This reverts part of commit a6b84aef4ad3a7bac40704146a98977c62bfb6e8

15 years agogallium: replace // comments with /* */
Brian Paul [Fri, 2 Oct 2009 13:31:42 +0000 (07:31 -0600)]
gallium: replace // comments with /* */

15 years agogallium: remove // comment and extra whitespace
Brian Paul [Fri, 2 Oct 2009 13:26:28 +0000 (07:26 -0600)]
gallium: remove // comment and extra whitespace

15 years agog3dvl: Formatting.
Younes Manton [Fri, 2 Oct 2009 02:25:46 +0000 (22:25 -0400)]
g3dvl: Formatting.

15 years agog3dvl: Delete state_trackers/g3dvl, other unused files.
Younes Manton [Fri, 2 Oct 2009 02:17:47 +0000 (22:17 -0400)]
g3dvl: Delete state_trackers/g3dvl, other unused files.

15 years agog3dvl: Copyright blocks.
Younes Manton [Fri, 2 Oct 2009 02:16:10 +0000 (22:16 -0400)]
g3dvl: Copyright blocks.

15 years agost/xvmc: Set default CSC matrix to BT.601, no ProcAmp, full range RGB.
Younes Manton [Fri, 2 Oct 2009 02:01:18 +0000 (22:01 -0400)]
st/xvmc: Set default CSC matrix to BT.601, no ProcAmp, full range RGB.

15 years agog3dvl: Color space conv interface & vl impl.
Younes Manton [Fri, 2 Oct 2009 01:53:17 +0000 (21:53 -0400)]
g3dvl: Color space conv interface & vl impl.

Interface is pipe_video_context::set_csc_matrix().

vl_csc.h defines some helpers to generate CSC matrices based on one of
the color standard and a user defined ProcAmp (brightness, contrast,
saturation, hue).

15 years agoi915: Add stub ARB_occlusion_query support under a driconf debug option.
Eric Anholt [Wed, 29 Jul 2009 20:07:49 +0000 (13:07 -0700)]
i915: Add stub ARB_occlusion_query support under a driconf debug option.

This is useful for enabling our GLSL testcases using the 2.0 entrypoints
even though we don't have full GL 2.0.

15 years agoi915: Add optional support for ARB_fragment_shader under a driconf option.
Eric Anholt [Wed, 29 Jul 2009 20:00:09 +0000 (13:00 -0700)]
i915: Add optional support for ARB_fragment_shader under a driconf option.

Other vendors have enabled ARB_fragment_shader as part of OpenGL 2.0
enablement even on hardware like the 915 with no dynamic branching or
dFdx/dFdy support.  But for now we'll leave it disabled because we don't
do any flattening of ifs or loops, which is rather restrictive.

This support is not complete, and may be unstable depending on your shaders.
It passes 10/15 of the piglit glsl tests, but hangs on glean glsl1.

15 years agoi915: Add support or fallbacks for GLSL fragment shader opcodes.
Eric Anholt [Thu, 30 Jul 2009 03:44:39 +0000 (20:44 -0700)]
i915: Add support or fallbacks for GLSL fragment shader opcodes.

15 years agoi915: Add support for varying inputs.
Eric Anholt [Thu, 1 Oct 2009 21:00:28 +0000 (14:00 -0700)]
i915: Add support for varying inputs.

15 years agoi915: Enable ARB_vertex_shader for both i915 and i830.
Eric Anholt [Wed, 29 Jul 2009 19:15:14 +0000 (12:15 -0700)]
i915: Enable ARB_vertex_shader for both i915 and i830.

Since the TNL is all done in software anyway, it should be the same to
the user who's probably using ARB_vertex_program otherwise, but gives them
a nicer programming environment.

15 years agoi915: Increase maximum program size to the hardware limits.
Eric Anholt [Thu, 30 Jul 2009 07:03:21 +0000 (00:03 -0700)]
i915: Increase maximum program size to the hardware limits.

This fixes potential heap trashing if the program of choice exceeds limits,
and fixes the native instructions limit being lower than what can be
used by valid programs.

15 years agoi915: Update and translate the fragment program along with state updates.
Eric Anholt [Thu, 30 Jul 2009 06:37:04 +0000 (23:37 -0700)]
i915: Update and translate the fragment program along with state updates.

Previously, we were doing it in the midst of the pipeline run, which gave
an opportunity to enable/disable fallbacks, which is certainly the wrong
time to be doing so.  This manifested itself in a NULL dereference for PutRow
after transitioning out of a fallback during a run_pipeline in glean glsl1.

15 years agoRevert "i915: don't validate PS program when falling back to software"
Eric Anholt [Thu, 30 Jul 2009 05:46:14 +0000 (22:46 -0700)]
Revert "i915: don't validate PS program when falling back to software"

This reverts commit e7044d552c6d16389447880b8744a51de1cf0199.  It
prevented the driver from ever recovering from a software fallback due
to a program error.  The original bug it claimed to fix doesn't appear to
exist post-revert.

15 years agoi915: Bail when the fragment program has too many total instructions.
Eric Anholt [Thu, 30 Jul 2009 05:39:15 +0000 (22:39 -0700)]
i915: Bail when the fragment program has too many total instructions.

Previously, we'd go trashing the heap.

15 years agoi915: Let i915_program_error take a format string, and don't use _mesa_problem.
Eric Anholt [Thu, 30 Jul 2009 19:32:40 +0000 (12:32 -0700)]
i915: Let i915_program_error take a format string, and don't use _mesa_problem.

It's misleading to report things like the program having too many native
instructions as a Mesa implementation error, when the program may just be
too big for the hardware.

15 years agomesa: fix memory leak when generating mipmaps for compressed textures
Brian Paul [Thu, 1 Oct 2009 20:55:13 +0000 (14:55 -0600)]
mesa: fix memory leak when generating mipmaps for compressed textures

15 years agomesa: fix mem leaks
Brian Paul [Thu, 1 Oct 2009 20:52:28 +0000 (14:52 -0600)]
mesa: fix mem leaks

15 years agogallium/util: silence uninitialized var warning
Brian Paul [Thu, 1 Oct 2009 20:52:10 +0000 (14:52 -0600)]
gallium/util: silence uninitialized var warning

15 years agomesa: add missing return when out of memory
Brian Paul [Thu, 1 Oct 2009 20:51:43 +0000 (14:51 -0600)]
mesa: add missing return when out of memory

15 years agomesa: fix potential uninitialized memory reads
Brian Paul [Thu, 1 Oct 2009 20:51:26 +0000 (14:51 -0600)]
mesa: fix potential uninitialized memory reads

15 years agoglsl: fix mem leak
Brian Paul [Thu, 1 Oct 2009 20:50:45 +0000 (14:50 -0600)]
glsl: fix mem leak

15 years agost/mesa: check for null before asserts, fix possible mem leak
Brian Paul [Thu, 1 Oct 2009 20:49:34 +0000 (14:49 -0600)]
st/mesa: check for null before asserts, fix possible mem leak

15 years agost/mesa: silence hidden parameter warning
Brian Paul [Thu, 1 Oct 2009 20:49:04 +0000 (14:49 -0600)]
st/mesa: silence hidden parameter warning

15 years agoMerge branch 'mesa_7_6_branch'
Brian Paul [Thu, 1 Oct 2009 19:35:42 +0000 (13:35 -0600)]
Merge branch 'mesa_7_6_branch'

15 years agost/mesa: fix non-mipmap lastLevel calculation.
Brian Paul [Thu, 1 Oct 2009 19:34:49 +0000 (13:34 -0600)]
st/mesa: fix non-mipmap lastLevel calculation.

reviewed by Brian Paul.
(cherry picked from master, commit ae2daacbac7242938cffe0e2409071e030e00863)

15 years agomesa: Return -FLT_MAX instead of 0 for LG2(0).
Vinson Lee [Thu, 1 Oct 2009 19:33:20 +0000 (13:33 -0600)]
mesa: Return -FLT_MAX instead of 0 for LG2(0).

lim x->0 log(x) = -inf so -FLT_MAX is a better approximation than 0
for LG2(0).

15 years agoobjviewer: remove disabled code
Brian Paul [Thu, 1 Oct 2009 19:14:16 +0000 (13:14 -0600)]
objviewer: remove disabled code