profile/ivi/mesa.git
15 years agogallium: fix texcoord loop for rasterpos attributes
Brian Paul [Fri, 2 Jan 2009 23:16:16 +0000 (16:16 -0700)]
gallium: fix texcoord loop for rasterpos attributes

15 years agogallium: clamp MaxVertexTextureImageUnits against Mesa limit
Brian Paul [Fri, 2 Jan 2009 23:15:43 +0000 (16:15 -0700)]
gallium: clamp MaxVertexTextureImageUnits against Mesa limit

15 years agomesa: remove redudant lines for libglapi.a
Brian Paul [Fri, 2 Jan 2009 23:15:16 +0000 (16:15 -0700)]
mesa: remove redudant lines for libglapi.a

15 years agopython: Pass a zero offset to util_draw_vertex_buffer.
José Fonseca [Wed, 31 Dec 2008 15:03:35 +0000 (15:03 +0000)]
python: Pass a zero offset to util_draw_vertex_buffer.

15 years agoutil: List new file in SConscript.
José Fonseca [Wed, 31 Dec 2008 13:29:09 +0000 (13:29 +0000)]
util: List new file in SConscript.

15 years agoMerge commit 'origin/gallium-0.1' into gallium-0.2
José Fonseca [Wed, 31 Dec 2008 13:05:39 +0000 (13:05 +0000)]
Merge commit 'origin/gallium-0.1' into gallium-0.2

15 years agoi915: Avoid ptr->int conversion.
José Fonseca [Tue, 30 Dec 2008 20:12:28 +0000 (20:12 +0000)]
i915: Avoid ptr->int conversion.

15 years agodraw: Avoid integer overflow converting pointers on 64bit archs.
José Fonseca [Tue, 30 Dec 2008 20:10:24 +0000 (20:10 +0000)]
draw: Avoid integer overflow converting pointers on 64bit archs.

Not really an error, as we only care for the lower 4 bits.

15 years agoMerge commit 'origin/master' into gallium-0.2
Brian Paul [Wed, 31 Dec 2008 01:01:15 +0000 (18:01 -0700)]
Merge commit 'origin/master' into gallium-0.2

Conflicts:

src/mesa/main/config.h

15 years agomesa: fix bug in evaluation of structure fields
Brian Paul [Wed, 31 Dec 2008 00:50:39 +0000 (17:50 -0700)]
mesa: fix bug in evaluation of structure fields

Fixes incorrect size information.  See bug 19273.

15 years agomesa: allow variable indexing into the predefined uniform variable arrays
Brian Paul [Wed, 31 Dec 2008 00:11:32 +0000 (17:11 -0700)]
mesa: allow variable indexing into the predefined uniform variable arrays

This allows code such as "vec4 a = gl_LightSource[i].ambient;" to work.

When a built-in uniform array is indexed with a variable index we need to
"unroll" the whole array into the parameter list (aka constant buffer) because
we don't know which elements may be accessed at compile-time.  In the case of
the gl_LightSource array of size [8], we emit 64 state references into the
parameter array (8 elements times 8 vec4s per gl_LightSourceParameters
struct).

Previously, we only allowed constant-indexed references to uniform arrays
(such as gl_LightSource[2].position) which resulted in a single state reference
being added to the parameter array, not 64.  We still optimize this case.

Users should be aware that using "gl_LightSource[i].ambient" in their shaders
is a bit expensive since state validation will involve updating all 64
light source entries in the parameter list.

15 years agomesa: better error message when running out of GLSL samplers
Brian Paul [Wed, 31 Dec 2008 00:03:09 +0000 (17:03 -0700)]
mesa: better error message when running out of GLSL samplers

15 years agomesa: comments for some state vars
Brian Paul [Wed, 31 Dec 2008 00:00:36 +0000 (17:00 -0700)]
mesa: comments for some state vars

15 years agomesa: increase max constants/uniforms to 256 (vec4 vectors)
Brian Paul [Wed, 31 Dec 2008 00:00:06 +0000 (17:00 -0700)]
mesa: increase max constants/uniforms to 256 (vec4 vectors)

15 years agogallium: Initialize var before use.
José Fonseca [Tue, 30 Dec 2008 19:21:30 +0000 (19:21 +0000)]
gallium: Initialize var before use.

15 years agoglut: List new source file in sconscript.
José Fonseca [Tue, 30 Dec 2008 17:22:24 +0000 (17:22 +0000)]
glut: List new source file in sconscript.

15 years agomesa: Do not specify types in bitfields.
José Fonseca [Tue, 30 Dec 2008 17:21:25 +0000 (17:21 +0000)]
mesa: Do not specify types in bitfields.

As advised by gcc -pedantic.

15 years agodraw: Do not specify types in bitfields.
José Fonseca [Tue, 30 Dec 2008 17:21:15 +0000 (17:21 +0000)]
draw: Do not specify types in bitfields.

As advised by gcc -pedantic.

15 years agogallium: Remove unused variables.
José Fonseca [Tue, 30 Dec 2008 17:15:34 +0000 (17:15 +0000)]
gallium: Remove unused variables.

15 years agogallium: Don't redefine INLINE.
José Fonseca [Tue, 30 Dec 2008 17:14:48 +0000 (17:14 +0000)]
gallium: Don't redefine INLINE.

INLINE is commonly used in third-party headers.

15 years agoscons: Specify C99 throughout all the tree.
José Fonseca [Tue, 30 Dec 2008 17:13:38 +0000 (17:13 +0000)]
scons: Specify C99 throughout all the tree.

MSVC may not support full C99, but supports more than plain C90. And
-pedantic without -std=c99 generates too many spurious warnings
(specially C++ style comments) to be of any use.

Note that using certain C99 features in the cross-platform parts of Gallium
is still not possible; namely mid-of-scope variable declarations and named
structure initializers will break MSVC builds.

15 years agortasm: Remove spurious semi-colons after function bodies.
José Fonseca [Tue, 30 Dec 2008 17:06:51 +0000 (17:06 +0000)]
rtasm: Remove spurious semi-colons after function bodies.

15 years agogdi: Remove accidental keypresses.
José Fonseca [Tue, 30 Dec 2008 16:35:46 +0000 (16:35 +0000)]
gdi: Remove accidental keypresses.

15 years agomesa: updated compilation documentation
Brian Paul [Tue, 30 Dec 2008 14:57:16 +0000 (07:57 -0700)]
mesa: updated compilation documentation

15 years agodemos: minor fixes to twoside.c glsl demo
Brian Paul [Thu, 18 Dec 2008 23:08:42 +0000 (16:08 -0700)]
demos: minor fixes to twoside.c glsl demo

15 years agointel: disable ATI_texture_env_combine3 for i830( and related device).
Xiang, Haihao [Tue, 30 Dec 2008 03:25:45 +0000 (11:25 +0800)]
intel: disable ATI_texture_env_combine3 for i830( and related device).

Thanks to Eric for pointing it out.

15 years agodri: Fix driWaitForMSC32 when divisor >= 2 and msc < 0.
Eric Anholt [Wed, 24 Dec 2008 00:08:40 +0000 (16:08 -0800)]
dri: Fix driWaitForMSC32 when divisor >= 2 and msc < 0.

We'd come up with a negative remainder, while we were looking for the positive
version of it in the loop conditional.  And, since the "did we hit our target"
break was disabled for the target_msc == 0 ("Just make the divisor/remainder
work") path, we'd never exit.

Simplify the code by just using int64_t all over instead of trying to do it
in a u32 space.

15 years agoR300: missing semicolon
Alex Deucher [Mon, 29 Dec 2008 17:28:12 +0000 (12:28 -0500)]
R300: missing semicolon

15 years agointel: enable ATI_texture_env_combine3. Fixes #17707
Xiang, Haihao [Mon, 29 Dec 2008 01:30:41 +0000 (09:30 +0800)]
intel: enable ATI_texture_env_combine3. Fixes #17707

15 years agor300: remove the unknowns from the indx_buffer code
Dave Airlie [Sun, 28 Dec 2008 06:48:36 +0000 (16:48 +1000)]
r300: remove the unknowns from the indx_buffer code

15 years agoi915: separate the fog term from the specular color term.
Xiang, Haihao [Wed, 24 Dec 2008 01:26:46 +0000 (09:26 +0800)]
i915: separate the fog term from the specular color term.

Previously fog parameter and specular color are packed into the
same dword. Note specular color should be packed in BGRA for device,
so if fog parameter and specular color all are present, fog parameter
will dirty the alpha term of specular color. This fixes rendering
issue when playing 'Yo Frankie' on 915/945.

15 years agointel: Fix glBitmap clipping for DRI1.
Eric Anholt [Tue, 23 Dec 2008 22:45:37 +0000 (14:45 -0800)]
intel: Fix glBitmap clipping for DRI1.

15 years agoRemove third buffer support from Mesa.
Dave Airlie [Sat, 20 Dec 2008 01:38:23 +0000 (17:38 -0800)]
Remove third buffer support from Mesa.

This is part of the deprecated pageflipping infrastructure.

15 years agoegl: fix egl closure
Alan Hourihane [Tue, 23 Dec 2008 15:33:46 +0000 (15:33 +0000)]
egl: fix egl closure

15 years agoegl: fix startup query version
Alan Hourihane [Tue, 23 Dec 2008 14:18:36 +0000 (14:18 +0000)]
egl: fix startup query version

15 years agoegl: support GLXFBConfigs, pbuffers and pixmaps.
Alan Hourihane [Tue, 23 Dec 2008 13:20:18 +0000 (13:20 +0000)]
egl: support GLXFBConfigs, pbuffers and pixmaps.

15 years agosoftpipe: Don't fill surfaces's winsys fields.
José Fonseca [Mon, 22 Dec 2008 20:23:59 +0000 (20:23 +0000)]
softpipe: Don't fill surfaces's winsys fields.

This is sometimes checked to distinguish between texture views and
(deprecated) standalone surfaces.

15 years agosoftpipe: Call surface_alloc_storage to get the pipebuffer for display targets.
José Fonseca [Mon, 22 Dec 2008 20:20:58 +0000 (20:20 +0000)]
softpipe: Call surface_alloc_storage to get the pipebuffer for display targets.

Otherwise blitting from display target surfaces to front screen fails in
several platforms.

15 years agoMerge commit 'origin/gallium-0.1' into gallium-0.2
José Fonseca [Mon, 22 Dec 2008 19:31:52 +0000 (19:31 +0000)]
Merge commit 'origin/gallium-0.1' into gallium-0.2

Conflicts:

src/gallium/auxiliary/util/Makefile

15 years agogdi: Cleanup sconsfile.
José Fonseca [Thu, 11 Dec 2008 05:42:25 +0000 (14:42 +0900)]
gdi: Cleanup sconsfile.

15 years agosoftpipe: initialize refcount and winsys
Jerome Glisse [Thu, 18 Dec 2008 18:56:45 +0000 (19:56 +0100)]
softpipe: initialize refcount and winsys

15 years agosoftpipe: convert to use texture instead of surface
Jerome Glisse [Thu, 18 Dec 2008 12:34:27 +0000 (13:34 +0100)]
softpipe: convert to use texture instead of surface

15 years agogallium: const correctness.
José Fonseca [Mon, 22 Dec 2008 16:55:27 +0000 (16:55 +0000)]
gallium: const correctness.

15 years agoIgnore new tests executables.
José Fonseca [Sun, 21 Dec 2008 13:48:43 +0000 (13:48 +0000)]
Ignore new tests executables.

15 years agogallium: Fix typo in define name.
José Fonseca [Sat, 20 Dec 2008 12:59:51 +0000 (12:59 +0000)]
gallium: Fix typo in define name.

15 years agoMerge commit 'origin/master' into gallium-0.2
Brian Paul [Fri, 19 Dec 2008 23:56:56 +0000 (16:56 -0700)]
Merge commit 'origin/master' into gallium-0.2

15 years agogallium: begin adapting Ian's 3D mipmap gen code to gallium utility lib
Brian Paul [Fri, 19 Dec 2008 23:37:06 +0000 (16:37 -0700)]
gallium: begin adapting Ian's 3D mipmap gen code to gallium utility lib

Unfinished, a big no-op for now.

15 years agogallium: Fix typeo in mipmap filter for GL_UNSIGNED_SHORT_1_5_5_5_REV
Brian Paul [Fri, 19 Dec 2008 23:14:48 +0000 (16:14 -0700)]
gallium: Fix typeo in mipmap filter for GL_UNSIGNED_SHORT_1_5_5_5_REV

This is copied from Ian's commit a330933bb75c38148668637cd22b90d75d39506f

15 years agoAdd do_row_3d for mipmapping 3D textures
Ian Romanick [Fri, 19 Dec 2008 21:03:51 +0000 (13:03 -0800)]
Add do_row_3d for mipmapping 3D textures

Previously 3D textures were mipmapped using multiple passed through
the 2D mipmap generation code.  This had 3 disadvantages.  First, the
extra passes were slow.  Second, this required the allocation of a
temporary buffer to hold intermediate data.  Third, and most
important, the extra passes caused loss of additional bits due to
integer division / bit-shifting.

With this change, our mipmapgen conformance test passes for
non-compressed texture formats.

15 years agoFix typeo in mipmap filter for GL_UNSIGNED_SHORT_1_5_5_5_REV
Ian Romanick [Fri, 19 Dec 2008 20:58:01 +0000 (12:58 -0800)]
Fix typeo in mipmap filter for GL_UNSIGNED_SHORT_1_5_5_5_REV

15 years ago965 / GLSL: Use full precision for EXP instruction
Ian Romanick [Thu, 18 Dec 2008 22:13:26 +0000 (14:13 -0800)]
965 / GLSL: Use full precision for EXP instruction

The partial precision mode doesn't have quite enough bits of precision
to pass conformance tests.

15 years agoGLSL: The LOG2 macro doesn't have enough precision
Ian Romanick [Thu, 18 Dec 2008 22:11:06 +0000 (14:11 -0800)]
GLSL: The LOG2 macro doesn't have enough precision

It looks like the LOG2 macro only has 8 or 9 bits of precission, but
the ARB_vertex_program spec says "accurate to at least 10 bits".

15 years agointel: Fix glBitmap top/bottom clipping.
Eric Anholt [Fri, 19 Dec 2008 00:00:09 +0000 (16:00 -0800)]
intel: Fix glBitmap top/bottom clipping.

Bug #19139.

15 years agointel: Don't forget the source bitmap size when clipping the size we draw.
Eric Anholt [Thu, 18 Dec 2008 06:10:57 +0000 (22:10 -0800)]
intel: Don't forget the source bitmap size when clipping the size we draw.

15 years agointel: Update mesa state in blit operations that want post-scissor draw bounds.
Eric Anholt [Thu, 18 Dec 2008 05:18:20 +0000 (21:18 -0800)]
intel: Update mesa state in blit operations that want post-scissor draw bounds.

15 years agointel: don't clip to scissor-clipped read framebuffer bounds in copypixels.
Eric Anholt [Thu, 18 Dec 2008 05:18:00 +0000 (21:18 -0800)]
intel: don't clip to scissor-clipped read framebuffer bounds in copypixels.

15 years agogallium: Simple and efficient cache.
José Fonseca [Fri, 19 Dec 2008 20:06:11 +0000 (20:06 +0000)]
gallium: Simple and efficient cache.

Fixed size hash table. Collisions are handled by simply destroying the
previous entry.

It hasn't received much testing yet.

15 years agocell: fix build breakage
Brian Paul [Fri, 19 Dec 2008 15:01:03 +0000 (08:01 -0700)]
cell: fix build breakage

15 years agogallium: set ST_NEW_FRAMEBUFFER when setting new surfaces
Alan Hourihane [Fri, 19 Dec 2008 14:59:43 +0000 (14:59 +0000)]
gallium: set ST_NEW_FRAMEBUFFER when setting new surfaces

15 years agomesa: s/CC/APP_CC/ in progs/glsl/Makefile
Brian Paul [Fri, 19 Dec 2008 14:57:09 +0000 (07:57 -0700)]
mesa: s/CC/APP_CC/ in progs/glsl/Makefile

15 years agogallium: replace #elif with #else
Brian Paul [Fri, 19 Dec 2008 14:33:17 +0000 (07:33 -0700)]
gallium: replace #elif with #else

15 years agogallium: replace #elif with #else
Brian Paul [Fri, 19 Dec 2008 14:32:52 +0000 (07:32 -0700)]
gallium: replace #elif with #else

15 years agointel: Move copyteximage source clipping out of copytexsubimage.
Eric Anholt [Fri, 19 Dec 2008 02:42:06 +0000 (18:42 -0800)]
intel: Move copyteximage source clipping out of copytexsubimage.

glCopyTexSubImage already gets the (correct) clipping for us, so it doesn't
need the path.  While moving the clipping out, replace the code with the mesa
path to do the same job.

15 years agomesa: Clip copytexsubimage to read framebuffer bounds, not scissor region.
Eric Anholt [Fri, 19 Dec 2008 02:31:25 +0000 (18:31 -0800)]
mesa: Clip copytexsubimage to read framebuffer bounds, not scissor region.

15 years agomesa: Correct _mesa_clip_to_region() off-by-one.
Eric Anholt [Fri, 19 Dec 2008 02:23:51 +0000 (18:23 -0800)]
mesa: Correct _mesa_clip_to_region() off-by-one.

Note how if:
x + width == xmax + 0: width -= 0
x + width == xmax + 1: width -= 0
x + width == xmax + 2: width -= 1

So, the function was clipping to [xmin, xmax+1), not [xmin, xmax) like it was
supposed to.  Same for ymax.

15 years agotgsi: scan for additional info: uses_fogcoord, uses_frontfacing
Brian Paul [Fri, 19 Dec 2008 01:06:38 +0000 (18:06 -0700)]
tgsi: scan for additional info: uses_fogcoord, uses_frontfacing

15 years agotgsi: scan for additional info: uses_fogcoord, uses_frontfacing
Brian Paul [Fri, 19 Dec 2008 01:06:38 +0000 (18:06 -0700)]
tgsi: scan for additional info: uses_fogcoord, uses_frontfacing

15 years agogallium: fix two-sided lighting test in state tracker
Brian Paul [Thu, 18 Dec 2008 23:00:20 +0000 (16:00 -0700)]
gallium: fix two-sided lighting test in state tracker

This fixes two-sided lighting for vertex shaders.

15 years agogallium: fix two-sided lighting test in state tracker
Brian Paul [Thu, 18 Dec 2008 23:00:20 +0000 (16:00 -0700)]
gallium: fix two-sided lighting test in state tracker

This fixes two-sided lighting for vertex shaders.

15 years agoglsl: Fix handling of nested parens in macro actual arguments.
Michal Krol [Thu, 18 Dec 2008 17:45:32 +0000 (18:45 +0100)]
glsl: Fix handling of nested parens in macro actual arguments.

15 years agodemos: all glutDestroyWindow() upon exit to test query object clean-up
Brian Paul [Thu, 18 Dec 2008 16:48:20 +0000 (09:48 -0700)]
demos: all glutDestroyWindow() upon exit to test query object clean-up

15 years agoi915: check WRAP_T instead of WRAP_R for cube map texture.
Xiang, Haihao [Thu, 18 Dec 2008 04:57:41 +0000 (12:57 +0800)]
i915: check WRAP_T instead of WRAP_R for cube map texture.

15 years agoi915: fix abort issue. (bug #19147)
Xiang, Haihao [Thu, 18 Dec 2008 02:07:45 +0000 (10:07 +0800)]
i915: fix abort issue. (bug #19147)

15 years agomesa: remove unneeded _mesa_reference_fragprog() call
Brian Paul [Thu, 18 Dec 2008 01:05:03 +0000 (18:05 -0700)]
mesa: remove unneeded _mesa_reference_fragprog() call

The subsequent if/else cases always call _mesa_reference_fragprog() anyway.

15 years agodri: fix for Cygwin compilation, bug 19144
Jon Turney [Thu, 18 Dec 2008 01:01:16 +0000 (18:01 -0700)]
dri: fix for Cygwin compilation, bug 19144

15 years agomesa: updated comments
Brian Paul [Wed, 17 Dec 2008 21:54:46 +0000 (14:54 -0700)]
mesa: updated comments

15 years agodemos: add test of vertex-only and fragment-only shader programs
Brian Paul [Wed, 17 Dec 2008 20:58:31 +0000 (13:58 -0700)]
demos: add test of vertex-only and fragment-only shader programs

15 years agoutil: check for frag/vertShader=0 before attaching
Brian Paul [Wed, 17 Dec 2008 20:58:05 +0000 (13:58 -0700)]
util: check for frag/vertShader=0 before attaching

15 years agosoftpipe: initialize refcount and winsys
Jerome Glisse [Thu, 18 Dec 2008 18:56:45 +0000 (19:56 +0100)]
softpipe: initialize refcount and winsys

15 years agoGallium: fix for conform test
Robert Ellison [Thu, 18 Dec 2008 16:46:53 +0000 (09:46 -0700)]
Gallium: fix for conform test

The following construction in util_surface_copy() in
gallium/auxiliary/util/u_rect.c, introduced in commit
d177c9ddda2c452cf7d6696d89cf4458ef986f98, incorrectly inverts
the Y coordinate in the last parameter to pipe_copy_rect().

      /* If do_flip, invert src_y position and pass negative src stride
*/
      pipe_copy_rect(dst_map,
                     &dst->block,
                     dst->stride,
                     dst_x, dst_y,
                     w, h,
                     src_map,
                     do_flip ? -(int) src->stride : src->stride,
                     src_x,
                     do_flip ? w - src_y : src_y);

The intention is to start at the last Y coordinate line and move
backwards, in the case of a flip; in that case, the correct
calculation is "src_y + h - 1", not "w - src_y".

This fixes a Gallium assertion failure in the conformance tests:

      u_rect.c:65:pipe_copy_rect: Assertion `src_y >= 0' failed.
      debug_get_bool_option: GALLIUM_ABORT_ON_ASSERT = TRUE
      Trace/breakpoint trap

15 years agoglsl: Fix handling of nested parens in macro actual arguments.
Michal Krol [Thu, 18 Dec 2008 17:45:32 +0000 (18:45 +0100)]
glsl: Fix handling of nested parens in macro actual arguments.

15 years agoglsl: Fix handling of nested parens in macro actual arguments.
Michal Krol [Thu, 18 Dec 2008 17:45:32 +0000 (18:45 +0100)]
glsl: Fix handling of nested parens in macro actual arguments.

15 years agodemos: all glutDestroyWindow() upon exit to test query object clean-up
Brian Paul [Thu, 18 Dec 2008 16:48:20 +0000 (09:48 -0700)]
demos: all glutDestroyWindow() upon exit to test query object clean-up

15 years agosoftpipe: convert to use texture instead of surface
Jerome Glisse [Thu, 18 Dec 2008 12:34:27 +0000 (13:34 +0100)]
softpipe: convert to use texture instead of surface

15 years agoGallium: fix for conform test
Robert Ellison [Thu, 18 Dec 2008 16:46:53 +0000 (09:46 -0700)]
Gallium: fix for conform test

The following construction in util_surface_copy() in
gallium/auxiliary/util/u_rect.c, introduced in commit
d177c9ddda2c452cf7d6696d89cf4458ef986f98, incorrectly inverts
the Y coordinate in the last parameter to pipe_copy_rect().

      /* If do_flip, invert src_y position and pass negative src stride
*/
      pipe_copy_rect(dst_map,
                     &dst->block,
                     dst->stride,
                     dst_x, dst_y,
                     w, h,
                     src_map,
                     do_flip ? -(int) src->stride : src->stride,
                     src_x,
                     do_flip ? w - src_y : src_y);

The intention is to start at the last Y coordinate line and move
backwards, in the case of a flip; in that case, the correct
calculation is "src_y + h - 1", not "w - src_y".

This fixes a Gallium assertion failure in the conformance tests:

      u_rect.c:65:pipe_copy_rect: Assertion `src_y >= 0' failed.
      debug_get_bool_option: GALLIUM_ABORT_ON_ASSERT = TRUE
      Trace/breakpoint trap

15 years agomesa: Pass the context to query object delete cb to avoid null dereference.
Eric Anholt [Wed, 8 Oct 2008 01:47:31 +0000 (18:47 -0700)]
mesa: Pass the context to query object delete cb to avoid null dereference.

15 years agogallium: Enable memory debugging on all windows platforms.
José Fonseca [Thu, 18 Dec 2008 14:53:24 +0000 (14:53 +0000)]
gallium: Enable memory debugging on all windows platforms.

15 years agoi915: check WRAP_T instead of WRAP_R for cube map texture.
Xiang, Haihao [Thu, 18 Dec 2008 04:57:41 +0000 (12:57 +0800)]
i915: check WRAP_T instead of WRAP_R for cube map texture.

15 years agoi915: fix abort issue. (bug #19147)
Xiang, Haihao [Thu, 18 Dec 2008 02:07:45 +0000 (10:07 +0800)]
i915: fix abort issue. (bug #19147)

15 years agomesa: choose GLSL vertex shader over ARB/internal vertex program in get_fp_input_mask()
Brian Paul [Thu, 18 Dec 2008 02:01:34 +0000 (19:01 -0700)]
mesa: choose GLSL vertex shader over ARB/internal vertex program in get_fp_input_mask()

This is a work-around the for the fact that we do fragment shader state
validation before vertex shader validation (see comments in state.c) so in
get_fp_input_mask() we can't rely on ctx->VertexProgram._Current being up to
date yet.

This fixes a glean glsl1 test failure.

15 years agogallium: fix memory corruption in u_gen_mipmap.c
Brian Paul [Thu, 18 Dec 2008 01:52:10 +0000 (18:52 -0700)]
gallium: fix memory corruption in u_gen_mipmap.c

Remove the old/initial vbuf allocation in util_create_gen_mipmap().
We were allocating a small vbuf at this point so get_next_slot() didn't have
as large of buffer as it expected.  So all but the first set_vertex_data()
was writing out of bounds.

Also added some comments.

15 years agomesa: choose GLSL vertex shader over ARB/internal vertex program in get_fp_input_mask()
Brian Paul [Thu, 18 Dec 2008 02:01:34 +0000 (19:01 -0700)]
mesa: choose GLSL vertex shader over ARB/internal vertex program in get_fp_input_mask()

This is a work-around the for the fact that we do fragment shader state
validation before vertex shader validation (see comments in state.c) so in
get_fp_input_mask() we can't rely on ctx->VertexProgram._Current being up to
date yet.

This fixes a glean glsl1 test failure.

15 years agosoftpipe: fix vertex shader texture sampling
Brian Paul [Thu, 18 Dec 2008 01:59:58 +0000 (18:59 -0700)]
softpipe: fix vertex shader texture sampling

Need to disable/bypass lambda calculation since derivatives of texcoords
are meaningless for adjacent vertices.

15 years agogallium: fix memory corruption in u_gen_mipmap.c
Brian Paul [Thu, 18 Dec 2008 01:52:10 +0000 (18:52 -0700)]
gallium: fix memory corruption in u_gen_mipmap.c

Remove the old/initial vbuf allocation in util_create_gen_mipmap().
We were allocating a small vbuf at this point so get_next_slot() didn't have
as large of buffer as it expected.  So all but the first set_vertex_data()
was writing out of bounds.

Also added some comments.

15 years agomesa: remove unneeded _mesa_reference_fragprog() call
Brian Paul [Thu, 18 Dec 2008 01:05:03 +0000 (18:05 -0700)]
mesa: remove unneeded _mesa_reference_fragprog() call

The subsequent if/else cases always call _mesa_reference_fragprog() anyway.

15 years agodri: fix for Cygwin compilation, bug 19144
Jon Turney [Thu, 18 Dec 2008 01:01:16 +0000 (18:01 -0700)]
dri: fix for Cygwin compilation, bug 19144

15 years agomesa: updated comments
Brian Paul [Wed, 17 Dec 2008 21:54:46 +0000 (14:54 -0700)]
mesa: updated comments

15 years agomesa: fix fixed-function test in get_fp_input_mask() - again.
Brian Paul [Wed, 17 Dec 2008 21:04:03 +0000 (14:04 -0700)]
mesa: fix fixed-function test in get_fp_input_mask() - again.

The problem we're solving only occured when there was a user-defined
vertex shader but no fragment shader.  Check for that case now.
Fixes glean api2 vertex array failure.

15 years agomesa: add missing cases for texture array targets
Brian Paul [Wed, 17 Dec 2008 20:17:15 +0000 (13:17 -0700)]
mesa: add missing cases for texture array targets

15 years agomesa: fix fixed-function test in get_fp_input_mask() - again.
Brian Paul [Wed, 17 Dec 2008 21:04:03 +0000 (14:04 -0700)]
mesa: fix fixed-function test in get_fp_input_mask() - again.

The problem we're solving only occured when there was a user-defined
vertex shader but no fragment shader.  Check for that case now.
Fixes glean api2 vertex array failure.