profile/ivi/mesa.git
15 years agomain: fix some potential memory leaks
Brian Paul [Thu, 13 Aug 2009 19:48:36 +0000 (13:48 -0600)]
main: fix some potential memory leaks

Allocate dlist images after error checking.
Record GL_OUT_OF_MEMORY when we can't make a copy of an image.

15 years agomesa: fix some potential uninitialized memory references
Brian Paul [Thu, 13 Aug 2009 19:44:31 +0000 (13:44 -0600)]
mesa: fix some potential uninitialized memory references

15 years agoglsl: fix some uninitialized pointers
Brian Paul [Thu, 13 Aug 2009 19:44:01 +0000 (13:44 -0600)]
glsl: fix some uninitialized pointers

15 years agodraw: cope with more primitives in draw_pipeline_run
Keith Whitwell [Thu, 13 Aug 2009 14:54:10 +0000 (15:54 +0100)]
draw: cope with more primitives in draw_pipeline_run

This previously was used only for decomposed (POINT/LINE/TRI) primitives,
but for some time a full range of primitives could end up in here.

Fixes trivial/lineloop-clip on softpipe, among others.
(cherry picked from commit 87cd8a3b8a2407b30916be418ff2f95dfea5d2ad)

15 years agodri: Fix problems with unitialized values in dri screen object.
Pauli Nieminen [Fri, 7 Aug 2009 17:02:22 +0000 (10:02 -0700)]
dri: Fix problems with unitialized values in dri screen object.

This fixes crash in r200 KMS driver when pSAREA was set to 1 randomly because of memory wasn't cleared.

Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
15 years agoutil: fix incorrect assertion
Brian Paul [Thu, 6 Aug 2009 20:58:01 +0000 (14:58 -0600)]
util: fix incorrect assertion

Check that the dest surface/format is renderable.

15 years agoutil: include u_surface.h, added comment
Brian Paul [Thu, 6 Aug 2009 20:54:25 +0000 (14:54 -0600)]
util: include u_surface.h, added comment

15 years agoutil: use util_same_surface() to compare surface pointers
Brian Paul [Thu, 6 Aug 2009 15:02:36 +0000 (09:02 -0600)]
util: use util_same_surface() to compare surface pointers

15 years agoutil: added util_same_surface() helper function
Brian Paul [Thu, 6 Aug 2009 15:02:05 +0000 (09:02 -0600)]
util: added util_same_surface() helper function

15 years agoEnsure GL_EXT_blend_equation_separate is enabled when 2.0 is enabled.
Alan Hourihane [Thu, 6 Aug 2009 14:54:49 +0000 (15:54 +0100)]
Ensure GL_EXT_blend_equation_separate is enabled when 2.0 is enabled.

15 years agodocs:fix glBlitFramebuffer() for Gallium
Brian Paul [Wed, 5 Aug 2009 19:50:21 +0000 (13:50 -0600)]
docs:fix glBlitFramebuffer() for Gallium

15 years agost/mesa: implement BlitFramebuffer() for depth/stencil (incomplete)
Brian Paul [Wed, 5 Aug 2009 19:46:47 +0000 (13:46 -0600)]
st/mesa: implement BlitFramebuffer() for depth/stencil (incomplete)

We now handle the case of blitting Z+stencil to/from combined Z/stencil
surfaces.  But Z-only or stencil-only and separate depth/stencil surfaces
are not yet implemented.

15 years agomesa: generate GL_INVALID_OPERATION for missing z/stencil when blitting
Brian Paul [Wed, 5 Aug 2009 19:44:59 +0000 (13:44 -0600)]
mesa: generate GL_INVALID_OPERATION for missing z/stencil when blitting

If glBlitFramebuffer() is called with GL_DEPTH_BUFFER_BIT or
GL_STENCIL_BUFFER_BIT and the src/dst depth/stencil buffers are absent,
report an error.

15 years agost/mesa: fix Y inversion and optimize st_BlitFramebuffer()
Brian Paul [Wed, 5 Aug 2009 19:26:19 +0000 (13:26 -0600)]
st/mesa: fix Y inversion and optimize st_BlitFramebuffer()

Need to check for Y inversion separately for src/dest buffers.

If both the src and dest regions are upside down, make them right-side
up for a better chance at a fast path.

progs/tests/copypixrate -blit is much faster now.

15 years agoutil: added comment/question about blit clipping
Brian Paul [Wed, 5 Aug 2009 19:22:26 +0000 (13:22 -0600)]
util: added comment/question about blit clipping

15 years agomesa: make _mesa_clip_blit() a shared function
Brian Paul [Wed, 5 Aug 2009 19:21:59 +0000 (13:21 -0600)]
mesa: make _mesa_clip_blit() a shared function

15 years agoutil: added util_blit_pixels() overlap test
Brian Paul [Wed, 5 Aug 2009 19:19:01 +0000 (13:19 -0600)]
util: added util_blit_pixels() overlap test

A comment alluded to this.  Now it's checked.

15 years agoutil: fix util_blit_pixels() test for surface_copy() path
Brian Paul [Wed, 5 Aug 2009 19:09:15 +0000 (13:09 -0600)]
util: fix util_blit_pixels() test for surface_copy() path

For the surface_copy() path require same format, no flipping and no stretching.

Fixes progs/tests/copypixrate -blit

15 years agoutil: reformatting and comments
Brian Paul [Wed, 5 Aug 2009 18:40:50 +0000 (12:40 -0600)]
util: reformatting and comments

15 years agodocs: 7.5.1 bug fixes
Brian Paul [Tue, 4 Aug 2009 15:27:49 +0000 (09:27 -0600)]
docs: 7.5.1 bug fixes

15 years agointel: Fix inverted test for disabling flushing of front buffer output.
Brian Paul [Tue, 4 Aug 2009 15:23:17 +0000 (09:23 -0600)]
intel: Fix inverted test for disabling flushing of front buffer output.

The comment disagreed with the code, and nicely drew my eyes to what was
going wrong.

Bug #21774 (blender)
Bug #21788 (readpix)

(cherry picked from master, commit fd65418f600874b05f902b622078b40bc1abb24a)

15 years agointel: Wait on the last swapbuffers to complete before queuing a new one.
Brian Paul [Tue, 4 Aug 2009 15:22:15 +0000 (09:22 -0600)]
intel: Wait on the last swapbuffers to complete before queuing a new one.

This fixes jerkiness in doom3 and other apps since the kernel change to
throttle less absurdly, which led to a thundering herd of frames.

Because this is a rather minimal fix, there is at least one downside: If
the whole scene completes in one batchbuffer, we'll end up stalling the GPU.

Thanks to Michel Dänzer for suggesting using glFlush to signal frame end
instead of going to all the effort of adding a new DRI2 extension.

(cherry picked from master, commit 0828579a658af01a64b5e699175dc9bbbedcd685)

15 years agotexenv: Use VP->Current, since _Current isn't updated at this point.
Brian Paul [Tue, 4 Aug 2009 15:15:33 +0000 (09:15 -0600)]
texenv: Use VP->Current, since _Current isn't updated at this point.

(cherry picked from master, commit a9ba1bfeb3a2852c6eda718e73c46c972a286648)

15 years agotexenv: Match state.c in deciding whether we'll be using a vertex shader.
Brian Paul [Tue, 4 Aug 2009 15:15:09 +0000 (09:15 -0600)]
texenv: Match state.c in deciding whether we'll be using a vertex shader.

(cherry picked from master, commit 40990d9dfb20b69585859b2a45596aa46c20140a)

15 years agotexenv: Add missing dependency on VP changes.
Brian Paul [Tue, 4 Aug 2009 15:14:33 +0000 (09:14 -0600)]
texenv: Add missing dependency on VP changes.

Funny thing is I annotated this dependency in
e5f63c403b767f9974e8eb5d412c012b8a69287f, but didn't actually use it.

(cherry picked from master, commit 03187571b63d97e3d1406d329c5e760e16ef3181)

Conflicts:
src/mesa/main/state.c

15 years agomklib: Ensure target directory exists for library
Dan Nicholson [Mon, 3 Aug 2009 15:35:43 +0000 (08:35 -0700)]
mklib: Ensure target directory exists for library

Instead of relying on the Makefile to always generate $(TOP)/$(LIB_DIR),
just have mklib handle creating the directory. This should fix any races
when using parallel make.

Signed-off-by: Dan Nicholson <dbn.lists@gmail.com>
(cherry picked from commit 23671e5358ffc0abfec83aeea9a515b09a6b35f3)

15 years agoAdd missing X11_INCLUDES to egl/drivers/demo and egl/main.
Peter Hutterer [Mon, 20 Jul 2009 06:11:26 +0000 (16:11 +1000)]
Add missing X11_INCLUDES to egl/drivers/demo and egl/main.

Compiling mesa on a system with no X headers installed in the default
include paths fails due to missing X11 includes. The header includes are
picked up by configure but not applied.

Signed-off-by: Peter Hutterer <peter.hutterer@who-t.net>
Signed-off-by: Dave Airlie <airlied@redhat.com>
(cherry picked from commit 5358e54d1ae64ccfa81199b343a2931b415fcc0a)

15 years agoxdemo/glxswapcontrol: Move get_framge_usage after the swap.
Pauli Nieminen [Mon, 3 Aug 2009 19:05:33 +0000 (12:05 -0700)]
xdemo/glxswapcontrol: Move get_framge_usage after the swap.

This fixes the problem that first frame would report bogus usage values.
Problem was caused because get_frame_usage returned data from previous buffer
swap.

Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
15 years agoxdemos/glxcontexts: Don't leak visual info when choosing visual again.
Pauli Nieminen [Mon, 3 Aug 2009 17:43:38 +0000 (10:43 -0700)]
xdemos/glxcontexts: Don't leak visual info when choosing visual again.

Signed-off-by: Pauli Nieminen <suokkos@gmail.com>
15 years agost/mesa: we don't support GL_NV_point_sprite (see comment)
Brian Paul [Mon, 3 Aug 2009 17:16:23 +0000 (11:16 -0600)]
st/mesa: we don't support GL_NV_point_sprite (see comment)

15 years agomesa: move misplaced return statement
Brian Paul [Thu, 30 Jul 2009 15:02:27 +0000 (09:02 -0600)]
mesa: move misplaced return statement

Fixes regression from commit 7d93f817c9c5222e2bdbaa9fa9f03fc502bfe878

15 years agomesa: add missing return after catching program error
Brian Paul [Thu, 30 Jul 2009 14:20:01 +0000 (08:20 -0600)]
mesa: add missing return after catching program error

15 years agomesa/st: short-circuit glFinish calls on WIN32 only
Keith Whitwell [Wed, 29 Jul 2009 11:47:23 +0000 (12:47 +0100)]
mesa/st: short-circuit glFinish calls on WIN32 only

Windows opengl32.dll calls glFinish prior to every swapbuffers, which
makes it pretty hard to get decent performance...

Work around by mapping finish to flush on PIPE_OS_WINDOWS.  This is
conformant, though it might confuse poorly-written benchmarks which
attempt to measure a single event rather than figuring out the rate of
continuous processing.

15 years agomesa/st: recognize no-op scissor state when checking clear_with_quads
Keith Whitwell [Tue, 28 Jul 2009 17:45:22 +0000 (18:45 +0100)]
mesa/st: recognize no-op scissor state when checking clear_with_quads

Some apps enable scissor but set the rectangle to the dimensions of
the window.  Don't let this force us onto a slower clear path.

15 years agoutil: _debug_printf should print even when DEBUG is not defined
Keith Whitwell [Thu, 16 Jul 2009 16:51:02 +0000 (17:51 +0100)]
util: _debug_printf should print even when DEBUG is not defined

The leading underscore is meaningful...  This function is used by
_warning and _error functions as well as the more common
debug_printf().

debug_printf (without underscore) gets turned off when DEBUG is
disabled, but warning/error messages still use this function to get
their message out.
(cherry picked from commit 0ac879dca797360570543d5bd0fd64f8fb8e566e)

15 years agoglx: assign per screen driver configs (DRISW)
RALOVICH, Kristóf [Thu, 23 Jul 2009 15:05:59 +0000 (17:05 +0200)]
glx: assign per screen driver configs (DRISW)

15 years agoglx: assign per screen driver configs (DRI)
RALOVICH, Kristóf [Thu, 23 Jul 2009 15:05:50 +0000 (17:05 +0200)]
glx: assign per screen driver configs (DRI)

15 years agoglx: assign per screen driver configs (DRI2)
RALOVICH, Kristóf [Thu, 23 Jul 2009 15:05:43 +0000 (17:05 +0200)]
glx: assign per screen driver configs (DRI2)

15 years agoglx: properly release DRI configs
RALOVICH, Kristóf [Thu, 23 Jul 2009 15:04:52 +0000 (17:04 +0200)]
glx: properly release DRI configs

Release per screen DRI driver configs during screen destruction.

15 years agoglx: cache DRI configs in __GLXscreenConfigsRec
RALOVICH, Kristóf [Thu, 23 Jul 2009 15:03:54 +0000 (17:03 +0200)]
glx: cache DRI configs in __GLXscreenConfigsRec

15 years agoprogs/trivial: add missing files to Makefile, .gitignore
Vinson Lee [Tue, 28 Jul 2009 14:54:14 +0000 (08:54 -0600)]
progs/trivial: add missing files to Makefile, .gitignore

15 years agointel: Fix leak of DRI option info due to using the wrong free routine.
Brian Paul [Mon, 27 Jul 2009 21:32:50 +0000 (15:32 -0600)]
intel: Fix leak of DRI option info due to using the wrong free routine.

(cherry picked from commit 6d66f23c50ebe8f973757b6fd1b81c9b7920c447)

15 years agointel: Clean up leak of driver context structure on context destroy.
Brian Paul [Mon, 27 Jul 2009 21:28:49 +0000 (15:28 -0600)]
intel: Clean up leak of driver context structure on context destroy.

(cherry picked from commit ddef7dc87b2001fbe117ee5f24a0c645ee95a03c)

15 years agost/mesa: silence warning
Brian Paul [Mon, 27 Jul 2009 21:10:28 +0000 (15:10 -0600)]
st/mesa: silence warning

15 years agomesa: separate some finite/pragma Watcom stuff
Brian Paul [Mon, 27 Jul 2009 21:09:56 +0000 (15:09 -0600)]
mesa: separate some finite/pragma Watcom stuff

15 years agosoftpipe: include sp_winsys.h to silence function prototype warning
Brian Paul [Mon, 27 Jul 2009 21:08:44 +0000 (15:08 -0600)]
softpipe: include sp_winsys.h to silence function prototype warning

15 years agointel: Use _mesa_warning() to report GEM warnings
Brian Paul [Mon, 27 Jul 2009 20:35:18 +0000 (14:35 -0600)]
intel: Use _mesa_warning() to report GEM warnings

15 years agowindows: updated VC8 project files
Karl Schultz [Mon, 27 Jul 2009 19:38:35 +0000 (13:38 -0600)]
windows: updated VC8 project files

See bug 22882.

15 years agodocs: 7.5 tarball md5sums
Brian Paul [Fri, 17 Jul 2009 22:58:27 +0000 (16:58 -0600)]
docs: 7.5 tarball md5sums

15 years agoutil: fix typo.
José Fonseca [Sun, 26 Jul 2009 19:31:11 +0000 (20:31 +0100)]
util: fix typo.

15 years agomesa: include glew headers in MesaDemos tarballs
Brian Paul [Fri, 24 Jul 2009 02:20:39 +0000 (20:20 -0600)]
mesa: include glew headers in MesaDemos tarballs

15 years agodocs: initial release notes for 7.5.1
Brian Paul [Wed, 22 Jul 2009 18:49:34 +0000 (12:49 -0600)]
docs: initial release notes for 7.5.1

15 years agomesa: bump version to 7.5.1
Brian Paul [Wed, 22 Jul 2009 18:46:53 +0000 (12:46 -0600)]
mesa: bump version to 7.5.1

15 years agomesa: include GLEW sources in MesaDemos tarball
Brian Paul [Wed, 22 Jul 2009 18:34:07 +0000 (12:34 -0600)]
mesa: include GLEW sources in MesaDemos tarball

15 years agointel: Fall back on glBitmap with fog enabled.
Eric Anholt [Fri, 3 Jul 2009 02:21:22 +0000 (19:21 -0700)]
intel: Fall back on glBitmap with fog enabled.

We would have to build the program with the appropriate fog mode, and
also supply the fog coordinate if appropriate.

Bug #19413.
(cherry picked from commit 8ae02a3919bf31bd33f86208472e100eedb58497)

15 years agoi965: Don't clip everything if FRONT_AND_BACK culling while culling disabled.
Eric Anholt [Tue, 21 Jul 2009 00:58:12 +0000 (17:58 -0700)]
i965: Don't clip everything if FRONT_AND_BACK culling while culling disabled.

Fixes everything-black with meta_clear_tris on quake4-mpdemo and doom3-demo.

Bug #18844, 22077.
(cherry picked from commit 81d555068408d4343d7627c8bedda5675f09bd21)

15 years agoradeon: With DRI1, if we have HW stencil, only expose fbconfigs with stencil.
Michel Dänzer [Mon, 20 Jul 2009 18:26:28 +0000 (20:26 +0200)]
radeon: With DRI1, if we have HW stencil, only expose fbconfigs with stencil.

Otherwise simple apps like glxgears pick up a DirectColor visual since the X
server mixes the depth 32 visual in with the other GLX visuals, and this seems
to result in a (mostly) black screen due to a bad ColorMap for a lot of people.

The bad ColorMap may be a bug in the apps, the X server or X driver, and
regardless of that I think the X server should ideally make the depth 32 GLX
visual separate from the rest again, but in the meantime this makes us cope.

(depth_bits is either 16 or 24, never 0)

15 years agoprogs/util: remove extfuncs.h (we use GLEW instead)
Brian Paul [Fri, 17 Jul 2009 20:43:29 +0000 (14:43 -0600)]
progs/util: remove extfuncs.h (we use GLEW instead)

15 years agoprogs/glsl: finish conversion to GLEW
Brian Paul [Fri, 17 Jul 2009 19:23:11 +0000 (13:23 -0600)]
progs/glsl: finish conversion to GLEW

This is a follow-on to b799af91d5ffbee1481161fec29eb4c92b161272.
Remove _func function suffix and unneeded #includes.

15 years agomesa: set version to 7.5
Brian Paul [Fri, 17 Jul 2009 18:40:48 +0000 (12:40 -0600)]
mesa: set version to 7.5

15 years agomesa: set version to 7.5
Brian Paul [Fri, 17 Jul 2009 18:40:44 +0000 (12:40 -0600)]
mesa: set version to 7.5

15 years agodocs: set date for 7.5 release
Brian Paul [Fri, 17 Jul 2009 18:40:24 +0000 (12:40 -0600)]
docs: set date for 7.5 release

15 years agodocs: news entry for 7.5 release
Brian Paul [Fri, 17 Jul 2009 15:40:45 +0000 (09:40 -0600)]
docs: news entry for 7.5 release

15 years agodocs: update links to Gallium wiki page on freedesktop.org
Brian Paul [Fri, 17 Jul 2009 15:40:28 +0000 (09:40 -0600)]
docs: update links to Gallium wiki page on freedesktop.org

15 years agopython/retrace: Process the call no passed to --to option inclusively.
José Fonseca [Thu, 16 Jul 2009 18:34:44 +0000 (19:34 +0100)]
python/retrace: Process the call no passed to --to option inclusively.

15 years agopython/retrace: Dump the surface copy contents.
José Fonseca [Thu, 16 Jul 2009 18:32:40 +0000 (19:32 +0100)]
python/retrace: Dump the surface copy contents.

15 years agopython/retrace: Flush stdout before calling the pipe driver.
José Fonseca [Thu, 16 Jul 2009 18:31:36 +0000 (19:31 +0100)]
python/retrace: Flush stdout before calling the pipe driver.

So that messages are in sync with stderr.

15 years agomesa: Fix logbase2.
José Fonseca [Thu, 16 Jul 2009 18:20:25 +0000 (19:20 +0100)]
mesa: Fix logbase2.

It was providing 1 too many for non power two values.

15 years agopython: Hack to prevent segmentation faults when python exits.
José Fonseca [Thu, 16 Jul 2009 10:21:12 +0000 (11:21 +0100)]
python: Hack to prevent segmentation faults when python exits.

15 years agowgl: Expose pipe_screen/pipe_context via an extension.
José Fonseca [Thu, 16 Jul 2009 10:20:12 +0000 (11:20 +0100)]
wgl: Expose pipe_screen/pipe_context via an extension.

15 years agopython: Obtain pipe_screen/pipe_context from the system's OpenGL driver.
José Fonseca [Thu, 16 Jul 2009 10:19:06 +0000 (11:19 +0100)]
python: Obtain pipe_screen/pipe_context from the system's OpenGL driver.

15 years agosoftpipe: limit blend results to [0,1]
Luca Barbieri [Wed, 15 Jul 2009 17:34:36 +0000 (11:34 -0600)]
softpipe: limit blend results to [0,1]

15 years agoFix state flag dependencies for fixed function fragment program updates.
Brian Paul [Wed, 15 Jul 2009 17:12:05 +0000 (11:12 -0600)]
Fix state flag dependencies for fixed function fragment program updates.

I started looking into why _NEW_ARRAY punishes us, and while annotating
dependencies noticed that a bunch of dependencies were missing.

(cherry picked from master, commit e5f63c403b767f9974e8eb5d412c012b8a69287f)

15 years agopython/retrace: Interpret surface_copy.
José Fonseca [Wed, 15 Jul 2009 14:37:27 +0000 (15:37 +0100)]
python/retrace: Interpret surface_copy.

15 years agopython/samples: Use PIPE_FORMAT_Z16_UNORM instead of PIPE_FORMAT_Z32_UNORM.
José Fonseca [Wed, 15 Jul 2009 14:37:04 +0000 (15:37 +0100)]
python/samples: Use PIPE_FORMAT_Z16_UNORM instead of PIPE_FORMAT_Z32_UNORM.

More common. True fix would be to use whatever the screen supports though.

15 years agomesa: recognize and eliminate repeated error messages
Keith Whitwell [Wed, 15 Jul 2009 12:29:58 +0000 (13:29 +0100)]
mesa: recognize and eliminate repeated error messages

15 years agomesa: split out errorstring switch from _mesa_error
Keith Whitwell [Tue, 14 Jul 2009 15:28:08 +0000 (16:28 +0100)]
mesa: split out errorstring switch from _mesa_error

Move a chunk of code out of _mesa_error()

15 years agomesa: remove dead code in _mesa_error
Keith Whitwell [Tue, 14 Jul 2009 15:25:59 +0000 (16:25 +0100)]
mesa: remove dead code in _mesa_error

Remove early and unused snprintf and where[] string.

15 years agomesa: don't call getenv every time _mesa_error is called
Keith Whitwell [Tue, 14 Jul 2009 15:23:04 +0000 (16:23 +0100)]
mesa: don't call getenv every time _mesa_error is called

Buggy apps can generate thousands of mesa_error calls.  Don't need
to keep calling getenv to retreive the same MESA_DEBUG string each
time.

15 years agomesa: fix texture border color code for glPopAttrib()
Brian Paul [Tue, 14 Jul 2009 20:26:42 +0000 (14:26 -0600)]
mesa: fix texture border color code for glPopAttrib()

The texture object's border color used to be stored as GLchan but it's
been GLfloat for a while now.

15 years agoprogs/tests: Use compressed texture in mipmap_comp_tests
Jakob Bornecrantz [Sat, 4 Jul 2009 14:01:45 +0000 (16:01 +0200)]
progs/tests: Use compressed texture in mipmap_comp_tests

15 years agoprogs/tests: Tests more views in mipmap_comp_tests
Jakob Bornecrantz [Sat, 4 Jul 2009 05:40:34 +0000 (07:40 +0200)]
progs/tests: Tests more views in mipmap_comp_tests

15 years agoprogs/tests: Add yet another mipmap test
Jakob Bornecrantz [Sat, 13 Jun 2009 23:04:00 +0000 (01:04 +0200)]
progs/tests: Add yet another mipmap test

15 years agomesa: regenerated enums.c file
Brian Paul [Tue, 14 Jul 2009 13:58:22 +0000 (07:58 -0600)]
mesa: regenerated enums.c file

15 years agoscons: Monkey patch os.spawnve on Windows to become thread safe.
José Fonseca [Tue, 14 Jul 2009 10:09:23 +0000 (11:09 +0100)]
scons: Monkey patch os.spawnve on Windows to become thread safe.

See also:
- http://bugs.python.org/issue6476
- http://scons.tigris.org/issues/show_bug.cgi?id=2449

15 years agomesa: Report the true face number when flushing.
José Fonseca [Tue, 7 Jul 2009 16:56:12 +0000 (17:56 +0100)]
mesa: Report the true face number when flushing.

15 years agor128: fix two-sided lighting segfault seen in GLUT's olight demo
Peteri Andras [Mon, 13 Jul 2009 14:03:16 +0000 (08:03 -0600)]
r128: fix two-sided lighting segfault seen in GLUT's olight demo

15 years agointel: Bump driver data, add RC3 tag
Ian Romanick [Mon, 13 Jul 2009 04:07:38 +0000 (21:07 -0700)]
intel: Bump driver data, add RC3 tag

15 years agogallium: compare the actual register, not all the inputs
Zack Rusin [Sat, 11 Jul 2009 17:48:41 +0000 (13:48 -0400)]
gallium: compare the actual register, not all the inputs

otherwise we decrement indexes for all registers

15 years agotgsi: update some assertions
Brian Paul [Fri, 10 Jul 2009 22:26:09 +0000 (16:26 -0600)]
tgsi: update some assertions

15 years agotgsi: tgsi: add semantic_names[] string for TGSI_SEMANTIC_FACE
Brian Paul [Fri, 10 Jul 2009 21:44:48 +0000 (15:44 -0600)]
tgsi: tgsi: add semantic_names[] string for TGSI_SEMANTIC_FACE

Same story as in the tgsi_dump.c code (see prev commit).

15 years agotgsi: add semantic_names[] string for TGSI_SEMANTIC_FACE
Brian Paul [Fri, 10 Jul 2009 21:41:26 +0000 (15:41 -0600)]
tgsi: add semantic_names[] string for TGSI_SEMANTIC_FACE

Fixes TGSI dump output when front/back-face register is declared.

Also, add some assertions to make sure the semantic/interpolate string
arrays have as many elements as there are tokens in the p_shader_tokens.h
file.  That should catch problems like this in the future.

15 years agost/mesa: implement indirect addressing for destination registers
Brian Paul [Fri, 10 Jul 2009 19:07:45 +0000 (13:07 -0600)]
st/mesa: implement indirect addressing for destination registers

15 years agotgis: implement indirect addressing for destination registers
Brian Paul [Fri, 10 Jul 2009 19:07:16 +0000 (13:07 -0600)]
tgis: implement indirect addressing for destination registers

Includes the TGSI interpreter, but not the SSE/PPC/etc code generators.

15 years agovbo: fix vbo/dlist memory leak
Brian Paul [Fri, 10 Jul 2009 18:55:30 +0000 (12:55 -0600)]
vbo: fix vbo/dlist memory leak

Based on a patch by kristof.ralovich@gmail.com

15 years agodemos: set 4th component of texcoord to 1.0
Brian Paul [Thu, 9 Jul 2009 19:59:03 +0000 (13:59 -0600)]
demos: set 4th component of texcoord to 1.0

Avoid potential randomness in resulting texcoords.

15 years agoRemove stale reference to non-Gallium nouveau driver from configure.ac.
Michel Dänzer [Fri, 10 Jul 2009 12:49:46 +0000 (14:49 +0200)]
Remove stale reference to non-Gallium nouveau driver from configure.ac.

15 years agomesa: From float type modifier from values to large for singles
Ian Romanick [Tue, 23 Jun 2009 01:22:51 +0000 (18:22 -0700)]
mesa: From float type modifier from values to large for singles

The values 2147483648.0 and 4294967294.0 are too larget to be stored in single
precision floats.  Forcing these to be singles causes bits to be lost, which
results in errors in some pixel transfer tests.

This fixes bug #22344.
(cherry picked from commit 70e72070fce6aa1e0918dcc62c1949465cee69f7)

15 years agodocs: document gl_TextureMatrix[i][j] array indexing bug fix
Brian Paul [Thu, 9 Jul 2009 14:04:07 +0000 (08:04 -0600)]
docs: document gl_TextureMatrix[i][j] array indexing bug fix

15 years agoglsl: do const parameter optimization for array element actual parameters
Brian Paul [Thu, 9 Jul 2009 13:58:50 +0000 (07:58 -0600)]
glsl: do const parameter optimization for array element actual parameters

When a function parameter is const-qualified we can avoid making a copy
of the actual parameter (we basically do a search/replace when inlining).
This is now done for array element params too, resulting in better code
(fewer MOV instructions).

We should allow some other types of function arguments here but let's be
conservative for the moment.