mesa.git
14 years agoARB_texture_rg: Allow RED and RG textures as FBO color buffer attachments
Ian Romanick [Thu, 30 Sep 2010 23:30:09 +0000 (16:30 -0700)]
ARB_texture_rg: Allow RED and RG textures as FBO color buffer attachments

14 years agoARB_texture_rg: Add R8, R16, RG88, and RG1616 internal formats
Ian Romanick [Wed, 29 Sep 2010 00:24:39 +0000 (17:24 -0700)]
ARB_texture_rg: Add R8, R16, RG88, and RG1616 internal formats

14 years agoARB_texture_rg: Handle RED and RG the same as RGB for tex env
Ian Romanick [Tue, 28 Sep 2010 20:58:28 +0000 (13:58 -0700)]
ARB_texture_rg: Handle RED and RG the same as RGB for tex env

14 years agoARB_texture_rg: Add GL_RED as a valid GL_DEPTH_TEXTURE_MODE
Ian Romanick [Tue, 28 Sep 2010 20:50:34 +0000 (13:50 -0700)]
ARB_texture_rg: Add GL_RED as a valid GL_DEPTH_TEXTURE_MODE

14 years agoARB_texture_rg: Add GL_TEXTURE_{RED,GREEN}_SIZE query support
Ian Romanick [Tue, 28 Sep 2010 18:17:27 +0000 (11:17 -0700)]
ARB_texture_rg: Add GL_TEXTURE_{RED,GREEN}_SIZE query support

14 years agoARB_texture_rg: Correct some errors in RED / RG internal format handling
Ian Romanick [Tue, 28 Sep 2010 18:11:43 +0000 (11:11 -0700)]
ARB_texture_rg: Correct some errors in RED / RG internal format handling

Fixes several problems:

The half-float, float, and integer internal formats depend on
ARB_texture_rg and other extensions.

RG_INTEGER is not a valid internal format.

Generic compressed formats depend on ARB_texture_rg, not
EXT_texture_compression_rgtc.

Use GL_RED instead of GL_R.

14 years agoARB_texture_rg: Add GLX protocol support
Ian Romanick [Fri, 20 Nov 2009 19:03:31 +0000 (11:03 -0800)]
ARB_texture_rg: Add GLX protocol support

14 years agoi965g: use Elements macro instead of manual sizeofs
Nicolas Kaiser [Fri, 1 Oct 2010 20:39:20 +0000 (13:39 -0700)]
i965g: use Elements macro instead of manual sizeofs

Signed-off-by: Nicolas Kaiser <nikai@nikai.net>
Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agoi965: Fix up copy'n'pasteo from moving coordinate setup around for gen4.
Eric Anholt [Fri, 1 Oct 2010 21:09:00 +0000 (14:09 -0700)]
i965: Fix up copy'n'pasteo from moving coordinate setup around for gen4.

14 years agoi965: Add real support for pre-gen5 texture sampling to the new FS.
Eric Anholt [Fri, 1 Oct 2010 20:37:40 +0000 (13:37 -0700)]
i965: Add real support for pre-gen5 texture sampling to the new FS.

Fixes 36 testcases, including glsl-fs-shadow2d*-bias which fail on the
Mesa IR backend.

14 years agoevergreen : fix z format setting, enable stencil.
richard [Fri, 1 Oct 2010 20:10:02 +0000 (16:10 -0400)]
evergreen : fix z format setting, enable stencil.

14 years agoi965: Pre-gen6, map VS outputs (not FS inputs) to URB setup in the new FS.
Eric Anholt [Fri, 1 Oct 2010 19:15:48 +0000 (12:15 -0700)]
i965: Pre-gen6, map VS outputs (not FS inputs) to URB setup in the new FS.

We should fix the SF to actually give us just the data we need, but
this fixes regressions in the new FS until then.

Fixes:
glsl-kwin-blur
glsl-routing

14 years agoi965: Also increment attribute location when skipping unused slots.
Eric Anholt [Fri, 1 Oct 2010 18:44:27 +0000 (11:44 -0700)]
i965: Also increment attribute location when skipping unused slots.

Fixes glsl1-texcoord varying.

14 years agoi965: Fix the gen6 jump size for BREAK/CONT in new FS.
Eric Anholt [Fri, 1 Oct 2010 17:49:01 +0000 (10:49 -0700)]
i965: Fix the gen6 jump size for BREAK/CONT in new FS.

Since gen5, jumps are in increments of 64 bits instead of increments
of 128-bit instructions.

14 years agoi965: Add gen6 attribute interpolation to new FS backend.
Eric Anholt [Fri, 1 Oct 2010 17:45:26 +0000 (10:45 -0700)]
i965: Add gen6 attribute interpolation to new FS backend.

Untested, since my hardware is not booting at the moment.

14 years agor600g: indentation fixes
Jerome Glisse [Fri, 1 Oct 2010 14:26:58 +0000 (10:26 -0400)]
r600g: indentation fixes

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
14 years agor600g: setup basic loop consts on r600 + evergreen.
Dave Airlie [Fri, 1 Oct 2010 05:56:05 +0000 (15:56 +1000)]
r600g: setup basic loop consts on r600 + evergreen.

this sets up a single loop constant like r600c does.

14 years agor600g: only set the Z export if shader exports it.
Dave Airlie [Fri, 1 Oct 2010 05:16:47 +0000 (15:16 +1000)]
r600g: only set the Z export if shader exports it.

14 years agor600c: pull over 6xx/7xx vertex fixes for evergreen
Alex Deucher [Fri, 1 Oct 2010 03:13:50 +0000 (23:13 -0400)]
r600c: pull over 6xx/7xx vertex fixes for evergreen

14 years agor600g: flush SH cache on constant change on evergreen
Dave Airlie [Fri, 1 Oct 2010 04:43:02 +0000 (14:43 +1000)]
r600g: flush SH cache on constant change on evergreen

14 years agor600g: fix evergreen draw-buffers
Dave Airlie [Fri, 1 Oct 2010 04:24:14 +0000 (14:24 +1000)]
r600g: fix evergreen draw-buffers

just a typo in the register headers.

14 years agor600g: add cb flushing for extra buffers + depth buffer on r600/evergreen
Dave Airlie [Fri, 1 Oct 2010 04:00:27 +0000 (14:00 +1000)]
r600g: add cb flushing for extra buffers + depth buffer on r600/evergreen

14 years agor600g: sync vertex/texture cache on resources on evergreen
Dave Airlie [Fri, 1 Oct 2010 03:48:10 +0000 (13:48 +1000)]
r600g: sync vertex/texture cache on resources on evergreen

this gets rid of lots of the instability on evergreen,
which isn't surprising since it really broken not to flush caches.

14 years agor600g: fixup vertex format picking.
Dave Airlie [Fri, 1 Oct 2010 03:09:09 +0000 (13:09 +1000)]
r600g: fixup vertex format picking.

there are some vertex formats defined in r600c not in the docs.

14 years agor600g: add assembler support for other vtx fetch fields.
Dave Airlie [Fri, 1 Oct 2010 01:38:40 +0000 (11:38 +1000)]
r600g: add assembler support for other vtx fetch fields.

this shouldn't change behaviour, just push the choice of what
to do out to the shader.

14 years agoi965: Split the gen4 and gen5 sampler handling apart.
Eric Anholt [Fri, 1 Oct 2010 02:18:25 +0000 (19:18 -0700)]
i965: Split the gen4 and gen5 sampler handling apart.

Trying to track the insanity of the different argument layouts for
normal/shadow crossed with normal/lod/bias one generation at a time is
enough.

Fixes: glsl1-texture2D() with bias.
(first test passing in this code that doesn't pass without it!)

14 years agoi965: Use the lowering pass for texture projection.
Eric Anholt [Fri, 1 Oct 2010 03:09:53 +0000 (20:09 -0700)]
i965: Use the lowering pass for texture projection.

We should end up with the same code, but anyone else with this issue
could share the handling (which I got wrong for shadow comparisons in
the driver before).

14 years agoglsl: Add a lowering pass for texture projection.
Eric Anholt [Fri, 1 Oct 2010 03:07:27 +0000 (20:07 -0700)]
glsl: Add a lowering pass for texture projection.

14 years agor600g: realign evergreen code with r600 code.
Dave Airlie [Fri, 1 Oct 2010 01:13:02 +0000 (11:13 +1000)]
r600g: realign evergreen code with r600 code.

fixes segfault in depth-tex-modes-glsl and OA startup.

14 years agor600c: add reloc for CB_COLOR0_ATTRIB
Alex Deucher [Fri, 1 Oct 2010 00:54:26 +0000 (20:54 -0400)]
r600c: add reloc for CB_COLOR0_ATTRIB

We'll need a reloc for tiling eventually,
so add it now.

14 years agor600g: add reloc for evergreen color attrib
Dave Airlie [Fri, 1 Oct 2010 00:52:09 +0000 (10:52 +1000)]
r600g: add reloc for evergreen color attrib

we'll need this for color tiling on evergreen.

14 years agor600g: drop depth quirk on evergreen
Dave Airlie [Fri, 1 Oct 2010 00:19:39 +0000 (10:19 +1000)]
r600g: drop depth quirk on evergreen

none of the EG cards need the quirk.

14 years agor600g: add winsys support for CTL constants.
Dave Airlie [Thu, 30 Sep 2010 23:43:14 +0000 (09:43 +1000)]
r600g: add winsys support for CTL constants.

These need to be emitted, we also need them to do proper vtx start,
instead of abusing index offset.

14 years agor600g: fix evergreen depth flushing.
Dave Airlie [Fri, 1 Oct 2010 00:13:04 +0000 (10:13 +1000)]
r600g: fix evergreen depth flushing.

although evergreen can apparantly sample direct from 24-bit,
just make it work with the current method for now.

14 years agor600g: use Elements macro instead of manual sizeofs
Dave Airlie [Thu, 30 Sep 2010 23:39:35 +0000 (09:39 +1000)]
r600g: use Elements macro instead of manual sizeofs

14 years agodraw: check for null sampler pointers
Brian Paul [Thu, 30 Sep 2010 22:41:31 +0000 (16:41 -0600)]
draw: check for null sampler pointers

http://bugs.freedesktop.org/show_bug.cgi?id=30516

14 years agogallivm: added some comments
Brian Paul [Thu, 30 Sep 2010 17:01:09 +0000 (11:01 -0600)]
gallivm: added some comments

14 years agor600g: keep a mapping around for each bo
John Doe [Thu, 30 Sep 2010 21:53:36 +0000 (17:53 -0400)]
r600g: keep a mapping around for each bo

Save a lot of call into the kernel and thus improve performances.

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
14 years agor600g: don't double count dirty block
John Doe [Thu, 30 Sep 2010 21:30:25 +0000 (17:30 -0400)]
r600g: don't double count dirty block

This avoid to overcount the number of dwords we need and
thus avoid maximazation of cs buffer use.

Signed-off-by: Jerome Glisse <jglisse@redhat.com
14 years agoevergreeng: avoid overlapping border color btw VS & PS
Jerome Glisse [Thu, 30 Sep 2010 21:06:29 +0000 (17:06 -0400)]
evergreeng: avoid overlapping border color btw VS & PS

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
14 years agoi965: Fix new FS handling of builtin uniforms with packed scalars in structs.
Eric Anholt [Thu, 30 Sep 2010 20:26:38 +0000 (13:26 -0700)]
i965: Fix new FS handling of builtin uniforms with packed scalars in structs.

We were pointing each element at the .x channel of the
ParameterValues.

Fixes glsl1-linear fog.

14 years agomesa: Don't reference a W component in setting up a vec3 uniform component.
Eric Anholt [Thu, 30 Sep 2010 20:40:22 +0000 (13:40 -0700)]
mesa: Don't reference a W component in setting up a vec3 uniform component.

The 965 driver would try to set up storage for the W component, and
the offsets would get mixed up.

14 years agoi965: Fix whole-structure/array assignment in new FS.
Eric Anholt [Thu, 30 Sep 2010 18:46:24 +0000 (11:46 -0700)]
i965: Fix whole-structure/array assignment in new FS.

We need to walk the type tree to get the right register types for
structure components.  Fixes glsl-fs-statevar-call.

14 years agoRevert "Prefer intrinsics to handrolled atomic ops."
Tom Fogal [Thu, 30 Sep 2010 20:39:14 +0000 (14:39 -0600)]
Revert "Prefer intrinsics to handrolled atomic ops."

This reverts commit 5f66b340aa49c6bc8d0acb2d1a6f8e9a7ef2cb2e, quickly
fixing 30514.

14 years agor600g: fix constant & literal src splitting, also fix mplayer gl2 shader
Jerome Glisse [Thu, 30 Sep 2010 20:26:33 +0000 (16:26 -0400)]
r600g: fix constant & literal src splitting, also fix mplayer gl2 shader

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
14 years agoPrefer intrinsics to handrolled atomic ops.
Tom Fogal [Mon, 27 Sep 2010 04:32:15 +0000 (22:32 -0600)]
Prefer intrinsics to handrolled atomic ops.

14 years agoImplement x86_64 atomics for compilers w/o intrinsics.
Tom Fogal [Mon, 27 Sep 2010 00:57:59 +0000 (18:57 -0600)]
Implement x86_64 atomics for compilers w/o intrinsics.

Really old gcc's (3.3, at least) don't have support for the
intrinsics we need.  This implements a fallback for that case.

14 years agoi965: Update renderer strings for sandybridge
Adam Jackson [Thu, 30 Sep 2010 18:08:35 +0000 (14:08 -0400)]
i965: Update renderer strings for sandybridge

Signed-off-by: Adam Jackson <ajax@redhat.com>
14 years agor600g: use constant buffer instead of register for constant
Jerome Glisse [Thu, 30 Sep 2010 14:43:26 +0000 (10:43 -0400)]
r600g: use constant buffer instead of register for constant

Signed-off-by: Jerome Glisse <jglisse@redhat.com>
14 years agogallivm: check for level=0 case in lp_build_minify()
Brian Paul [Thu, 30 Sep 2010 16:52:26 +0000 (10:52 -0600)]
gallivm: check for level=0 case in lp_build_minify()

This lets us avoid the shift and max() operations.

14 years agogallivm: More comprehensive border usage logic.
José Fonseca [Thu, 30 Sep 2010 16:39:17 +0000 (17:39 +0100)]
gallivm: More comprehensive border usage logic.

14 years agost/egl: Drop context argument from egl_g3d_get_egl_image.
Chia-I Wu [Thu, 30 Sep 2010 15:43:58 +0000 (23:43 +0800)]
st/egl: Drop context argument from egl_g3d_get_egl_image.

Fix a regression since 17eace581d25a626a7d75d9d1205d012cbb14a6e.

14 years agost: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:53:13 +0000 (07:53 -0700)]
st: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agomath: remove duplicated includes
Nicolas Kaiser [Thu, 30 Sep 2010 14:52:02 +0000 (07:52 -0700)]
math: remove duplicated includes

Remove duplicated includes.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agomain: remove duplicated includes
Nicolas Kaiser [Thu, 30 Sep 2010 14:50:43 +0000 (07:50 -0700)]
main: remove duplicated includes

Remove duplicated includes.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/savage: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:48:29 +0000 (07:48 -0700)]
dri/savage: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/radeon: remove duplicated includes
Nicolas Kaiser [Thu, 30 Sep 2010 14:47:22 +0000 (07:47 -0700)]
dri/radeon: remove duplicated includes

Remove duplicated includes.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/r600: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:46:06 +0000 (07:46 -0700)]
dri/r600: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/r300: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:44:55 +0000 (07:44 -0700)]
dri/r300: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/r128: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:43:48 +0000 (07:43 -0700)]
dri/r128: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/mga: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:42:41 +0000 (07:42 -0700)]
dri/mga: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/intel: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:41:26 +0000 (07:41 -0700)]
dri/intel: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/i965: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:39:59 +0000 (07:39 -0700)]
dri/i965: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/i915: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:38:53 +0000 (07:38 -0700)]
dri/i915: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/i810: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:37:26 +0000 (07:37 -0700)]
dri/i810: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/common: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:36:12 +0000 (07:36 -0700)]
dri/common: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agoglx: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:34:53 +0000 (07:34 -0700)]
glx: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agogallium/winsys: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:33:43 +0000 (07:33 -0700)]
gallium/winsys: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agogallium/st: remove duplicated includes
Nicolas Kaiser [Thu, 30 Sep 2010 14:32:40 +0000 (07:32 -0700)]
gallium/st: remove duplicated includes

Remove duplicated includes.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agogallium/softpipe: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:31:08 +0000 (07:31 -0700)]
gallium/softpipe: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agogallium/llvmpipe: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:29:49 +0000 (07:29 -0700)]
gallium/llvmpipe: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agogallium/i915: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:28:18 +0000 (07:28 -0700)]
gallium/i915: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agogallium/util: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:27:09 +0000 (07:27 -0700)]
gallium/util: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agogallium/rtasm: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:25:36 +0000 (07:25 -0700)]
gallium/rtasm: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agoegl: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:25:31 +0000 (07:25 -0700)]
egl: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agoswrast: remove duplicated include
Nicolas Kaiser [Thu, 30 Sep 2010 14:25:21 +0000 (07:25 -0700)]
swrast: remove duplicated include

Remove duplicated include.

Signed-off-by: Brian Paul <brianp@vmware.com>
14 years agodri/nv10: Use fast Z clears.
Francisco Jerez [Sat, 25 Sep 2010 13:29:02 +0000 (15:29 +0200)]
dri/nv10: Use fast Z clears.

14 years agodri/nouveau: Remove unnecessary flush.
Francisco Jerez [Tue, 28 Sep 2010 20:24:05 +0000 (22:24 +0200)]
dri/nouveau: Remove unnecessary flush.

14 years agodri/nouveau: Have a smaller amount of larger scratch buffers.
Francisco Jerez [Tue, 28 Sep 2010 20:20:12 +0000 (22:20 +0200)]
dri/nouveau: Have a smaller amount of larger scratch buffers.

Larger VBOs avoid many kernel trips to get them in sync with the GPU.

14 years agomapi: Fix compiler warnings.
Chia-I Wu [Thu, 30 Sep 2010 09:09:59 +0000 (17:09 +0800)]
mapi: Fix compiler warnings.

Do not use "void *" in arithmetics.

14 years agost/egl: Skip single-buffered configs in EGL.
Chia-I Wu [Thu, 30 Sep 2010 08:53:33 +0000 (16:53 +0800)]
st/egl: Skip single-buffered configs in EGL.

Let DRI2 report single-buffered configs and skip them in EGL.  This is
based on the patch by Luca Barbieri.

14 years agoegl: Check extensions.
Chia-I Wu [Thu, 23 Sep 2010 18:42:15 +0000 (02:42 +0800)]
egl: Check extensions.

Do not call into the driver if the extension for the called function is
not enabled.

14 years agoi965: always set tiling for fbo depth buffer on sandybridge
Zhenyu Wang [Thu, 30 Sep 2010 02:49:47 +0000 (10:49 +0800)]
i965: always set tiling for fbo depth buffer on sandybridge

Sandybridge requires depth buffer must be tiling.

Fix 'fbo_firecube' demo.

14 years agor300g: fix conditional rendering in non-wait path
Marek Olšák [Wed, 29 Sep 2010 23:08:58 +0000 (01:08 +0200)]
r300g: fix conditional rendering in non-wait path

NOTE: This is a candidate for the 7.9 branch.

14 years agoi965: Remove my "safety counter" code from loops.
Eric Anholt [Wed, 29 Sep 2010 23:20:48 +0000 (16:20 -0700)]
i965: Remove my "safety counter" code from loops.

I've screwed this up enough times that I don't think it's worth it.
This time, it was that I was doing it once per top-level body
instruction instead of just once at the end of the loop body.

14 years agoi965: Add live interval analysis and hook it up to the register allocator.
Eric Anholt [Wed, 29 Sep 2010 22:20:12 +0000 (15:20 -0700)]
i965: Add live interval analysis and hook it up to the register allocator.

Fixes 13 piglit cases that failed at register allocation before.

14 years agoi965: First cut at register allocation using graph coloring.
Eric Anholt [Wed, 29 Sep 2010 19:08:11 +0000 (12:08 -0700)]
i965: First cut at register allocation using graph coloring.

The interference is totally bogus (maximal), so this is equivalent to
our trivial register assignment before.  As in, passes the same set of
piglit tests.

14 years agora: First cut at a graph-coloring register allocator for mesa.
Eric Anholt [Mon, 27 Sep 2010 19:34:33 +0000 (12:34 -0700)]
ra: First cut at a graph-coloring register allocator for mesa.

Notably missing is choice of registers to spill.

14 years agoi965: Clean up the virtual GRF handling.
Eric Anholt [Wed, 29 Sep 2010 17:43:46 +0000 (10:43 -0700)]
i965: Clean up the virtual GRF handling.

Now, virtual GRFs are consecutive integers, rather than offsetting the
next one by the size.  We need the size information to still be around
for real register allocation, anyway.

14 years agor600g: return string for chip family
Dave Airlie [Wed, 29 Sep 2010 23:17:20 +0000 (09:17 +1000)]
r600g: return string for chip family

use same strings as r600c.

14 years agor600g: clean up some code from move to new paths.
Dave Airlie [Wed, 29 Sep 2010 23:07:07 +0000 (09:07 +1000)]
r600g: clean up some code from move to new paths.

mainly remove 2 suffix from function names

14 years agor600g: add L8A8 unorm.
Dave Airlie [Wed, 29 Sep 2010 23:04:04 +0000 (09:04 +1000)]
r600g: add L8A8 unorm.

fixes texEnv warnings.

14 years agor600g: port r300g fix for X* formats in texformat code
Dave Airlie [Wed, 29 Sep 2010 22:56:37 +0000 (08:56 +1000)]
r600g: port r300g fix for X* formats in texformat code

14 years agoi956: Make new FS discard do its work in a temp, not the null reg!
Eric Anholt [Wed, 29 Sep 2010 22:48:21 +0000 (15:48 -0700)]
i956: Make new FS discard do its work in a temp, not the null reg!

Fixes:
glsl-fs-discard-02 (GPU hang)
glsl1-discard statement (2)

14 years agoi965: Fix use of undefined mem_ctx in vector splitting.
Eric Anholt [Wed, 29 Sep 2010 22:19:52 +0000 (15:19 -0700)]
i965: Fix use of undefined mem_ctx in vector splitting.

14 years agogallivm: Use SSE4.1's ROUNDSS/ROUNDSD for scalar rounding.
José Fonseca [Wed, 29 Sep 2010 21:27:38 +0000 (22:27 +0100)]
gallivm: Use SSE4.1's ROUNDSS/ROUNDSD for scalar rounding.

14 years agopython/retrace: Handle set_index_buffer and draw_vbo.
José Fonseca [Sat, 25 Sep 2010 13:55:23 +0000 (14:55 +0100)]
python/retrace: Handle set_index_buffer and draw_vbo.

14 years agotrace: Fix set_index_buffer and draw_vbo tracing.
José Fonseca [Sat, 25 Sep 2010 13:55:00 +0000 (14:55 +0100)]
trace: Fix set_index_buffer and draw_vbo tracing.

14 years agor300/compiler: Move declaration before code.
Vinson Lee [Wed, 29 Sep 2010 21:22:20 +0000 (14:22 -0700)]
r300/compiler: Move declaration before code.

Fixes this GCC warning on linux-x86 build.
r3xx_vertprog.c: In function ‘ei_if’:
r3xx_vertprog.c:396: warning: ISO C90 forbids mixed declarations and code

14 years agor300/compiler: Move declaration before code.
Vinson Lee [Wed, 29 Sep 2010 21:13:49 +0000 (14:13 -0700)]
r300/compiler: Move declaration before code.

Fixes these GCC warnings on linux-x86 build.
r500_fragprog_emit.c: In function ‘emit_paired’:
r500_fragprog_emit.c:237: warning: ISO C90 forbids mixed declarations and code
r500_fragprog_emit.c: In function ‘emit_tex’:
r500_fragprog_emit.c:367: warning: ISO C90 forbids mixed declarations and code
r500_fragprog_emit.c: In function ‘emit_flowcontrol’:
r500_fragprog_emit.c:415: warning: ISO C90 forbids mixed declarations and code
r500_fragprog_emit.c: In function ‘r500BuildFragmentProgramHwCode’:
r500_fragprog_emit.c:633: warning: ISO C90 forbids mixed declarations and code