fdo-mirrors/mesa
1
0
Fork 0
mirror of https://gitlab.freedesktop.org/mesa/mesa.git synced 2026-05-09 13:18:04 +02:00
Code Issues Projects Releases Packages Wiki Activity Actions 1
29289 commits 66 branches 934 tags 961 MiB
Commit graph

29289 commits

This branch
This branch
All branches
Author SHA1 Message Date
Eric Anholt
99077e7792 i965: Flag BRW_NEW_CONTEXT on some context state. 
Fixing this is a prereq for avoiding flagging all state at new
batch time.  Eliminating that still causes problems, though (notably
glean logicOp fails on my GM965).
 2009-11-13 14:39:09 -08:00
Eric Anholt
1ffd0a7789 intel: Remove some dead context structure fields. 2009-11-13 13:18:57 -08:00
Eric Anholt
3727858ceb i965: Remove an unused cache_item field. 2009-11-13 13:18:57 -08:00
Eric Anholt
b54f8eeae8 i965: Remove long dead structures for ffvertex_prog.c. 2009-11-13 13:18:57 -08:00
Eric Anholt
ded0ec1ea5 i965: Use bo_map instead of subdata to upload the bits of constant buffer. 
Saves CPU time, resulting in a 2.5% FPS win on ETQW.
 2009-11-13 13:18:57 -08:00
Eric Anholt
8e8586e626 i965: Validate the number of URB entries selected for the VS. 2009-11-13 13:18:56 -08:00
Eric Anholt
3c05c1eb63 intel: When subdataing a busy buffer, use a temporary and blit in. 
This cuts a massive number of waits in ET:QW, which uses a VBO ringbuffer.
Unfortunately it doesn't BufferData when wrapping back to 0, so we can't
be clever with tracking what's been initialized.
 2009-11-13 13:18:56 -08:00
Eric Anholt
4e50ce35ee i965: Clean up Ironlake sampler type definitions. 
They're the same regardless of execution width for 8, 4x2, and 16.
 2009-11-13 13:18:56 -08:00
Eric Anholt
91bd593109 i965: Avoid moving the current value back into the accumulator for MAD. 
This is a 2.9% (+/-.3%) performance win for my GL demo, which hits MAD
sequences for matrix transforms.
 2009-11-13 13:18:56 -08:00
Eric Anholt
d6690ce15f mesa: Improve the eliminate-move-use to work across multiple instructions. 
This shaves more instructions off of the VS of my GL demo, but no
performance difference this time at n=6.  This also fixes a regression
that was in this path, which is now piglit's glsl-vs-mov-after-deref.
 2009-11-13 13:18:22 -08:00
Zack Rusin
654122ba7b st/xorg: try to fix non-uniform transforms 2009-11-13 14:58:35 -05:00
Michal Krol
eaa34c2dea glsl/apps: Report syntax parser errors. 2009-11-13 19:51:49 +01:00
Michal Krol
3f147c71ed slang: Report syntax parser errors. 2009-11-13 19:51:24 +01:00
Michal Krol
99c89ebdb0 glsl/cl: Add simple error reporting. 2009-11-13 19:51:02 +01:00
Michal Krol
b385312bc7 slang: Regenerate .gc files. 2009-11-13 06:03:56 +01:00
Michal Krol
7593562a61 slang: Get rid of the old syntax file and utilities. 2009-11-13 05:58:46 +01:00
Michal Krol
cd5553b457 slang: Plug in the new syntax parser. 2009-11-13 05:52:39 +01:00
Michal Krol
f5b6e06390 gdi: Link to glslcl. 2009-11-13 05:49:25 +01:00
Michal Krol
38a1f0b5d1 glsl/apps: Add GLSL compiler that translates source text into binary stream. 
Should be used in place of gc_to_bin utility to precompile builtin library.
 2009-11-13 03:02:14 +01:00
Michal Krol
08e90bdea1 glsl/cl: Add a hard-coded syntax parser. 2009-11-13 03:00:45 +01:00
Vinson Lee
43080e40aa progs/glsl: Add missing break statement in multinoise.c. 2009-11-12 16:21:49 -08:00
Ian Romanick
b6b753f727 intel: Don't check for context pointer to be NULL during extension init 
Thanks to Chia-I Wu's changes to the extension function
infrastructure, we no longer have to tell the loader which extensions
the driver might enable.  This means that intelInitExtensions will
never be called with a NULL context pointer.  Remove all the NULL checks.

Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
Reviewed-by: Eric Anholt <eric@anholt.net>
 2009-11-12 15:43:09 -08:00
Ian Romanick
a736d3f439 intel: Remove unused enable_imaging parameter to intelInitExtensions 2009-11-12 15:43:09 -08:00
Eric Anholt
514544f373 i965: Fix Ironlake shadow comparisons. 
The cube map array index arg is always present.
 2009-11-12 11:41:39 -08:00
Eric Anholt
5f305b1db9 i965: Fix VBO last-valid-offset setup on Ironlake. 
Instead of doing math based on the (broken for VBO && offset != 0)
input->count number, just use the BO size.  Fixes assertion failure in ETQW.
 2009-11-12 11:41:39 -08:00
Keith Whitwell
2f54d02d20 i965g: consult fs inputs when laying out vs output regs 
Vertex shader now emits just the FS inputs, in the positions and order
expected by the fragment shader.

This means potentially regenerating the vertex shader to match
different fragment shader's input layouts.
 2009-11-11 18:51:58 -08:00
Keith Whitwell
0c547d63c4 i965g: skip over vertex position output when preallocating FS inputs 2009-11-11 18:51:57 -08:00
Keith Whitwell
a010307e0a i965g: some more cases where file_max[] is used without obligatory +1 2009-11-11 18:51:57 -08:00
Keith Whitwell
e4d174ea1b i965g: smaller upload buffer for index data 2009-11-11 18:51:57 -08:00
Keith Whitwell
d86e9079a8 i965g: fix off-by-one on curbe upload 2009-11-11 18:51:57 -08:00
Keith Whitwell
e243279a48 i965g: avoid use of internally generated immediates 
Currently not working, so don't generate more of them.
 2009-11-11 18:51:57 -08:00
Keith Whitwell
98c2b53381 trivial: make tri-orig more closely match the original version of tri... 2009-11-11 18:51:57 -08:00
Zack Rusin
4d72f8f520 st/xorg: use an immediate instead of a full blown const 2009-11-11 21:47:28 -05:00
Zack Rusin
72b4a7d67f st/xorg: don't use flow control 2009-11-11 21:43:40 -05:00
Brian Paul
0a39620d6d swrast: handle additional Z24 formats in read_depth_pixels() 2009-11-11 19:38:19 -07:00
Roland Scheidegger
ab12e764ba i965: fix EXT_provoking_vertex support 
This didn't work for quad/quadstrips at all, and for all other primitive types
it only worked when they were unclipped.
Fix up the former in gs stage (could probably do without these changes and
instead set QuadsFollowProvokingVertexConvention to false), and the rest in
clip stage.
 2009-11-11 17:57:56 -08:00
Vinson Lee
7dfea5c072 demos/glsl: Add missing break statement to noise test. 2009-11-11 17:39:58 -08:00
Zack Rusin
2cfbbc76e4 st/xorg: implement repeatnone and make some code smell less like ass 2009-11-11 19:55:50 -05:00
Zack Rusin
493d599af4 st/xorg: fallback until daddy can implement you properly 2009-11-11 19:55:50 -05:00
Brian Paul
2ff4f5a7fe progs/tests: add alpha/blend testing to packedpixels.c 2009-11-11 12:25:08 -07:00
Corbin Simpson
cbee31a1f8 r300, r300g: Add missing registers. 2009-11-11 10:24:19 -08:00
Corbin Simpson
d030ce6a84 dri-st: Add some required GL 2.0 extensions. 
Two-sided stencil and NPOT textures.
 2009-11-11 10:24:06 -08:00
Eric Anholt
1220aba99b i965: Fix VS constant buffer value loading. 
Previously, we'd load linearly from ParameterValues[0] for the constants,
though ParameterValues[1] may not equal ParameterValues[0] + 4.  Additionally,
the STATE_VAL type paramters didn't get updated.

Fixes piglit vp-constant-array-huge.vpfp and ET:QW object locations.

Bug #23226.
 2009-11-10 14:44:40 -08:00
Eric Anholt
e5ffb9f5ea i965: Unalias src/dst registers for SGE and friends. 
Fixes piglit vp-sge-alias test, and the googleearth ground shader.  \o/

Bug #22228
(cherry picked from commit 56ab92bad8)
 2009-11-10 14:44:39 -08:00
Eric Anholt
23a4a6727e i965: Allow use of PROGRAM_LOCAL constants in ARB_vp. 
Fixes piglit arl.vp.
(cherry picked from commit d52d78b4bc)
 2009-11-10 14:44:39 -08:00
Michal Krol
d44cebd1ee glsl/pp: Add sl_pp_purify_getc(). 2009-11-10 20:49:45 +01:00
Jakob Bornecrantz
e4a19ffb13 st/xorg: Fallback if picture format doesn't match texture format 2009-11-10 10:05:40 +01:00
Jakob Bornecrantz
3201c655e4 st/xorg: Don't segfault when debug printing 2009-11-10 10:00:04 +01:00
Zack Rusin
fca8b2c3ae st/xorg: wrap to border color 2009-11-10 20:30:48 -05:00
brian
b81f213157 swrast: update renderbuffer format assertions 2009-11-10 18:23:59 -07:00