summaryrefslogtreecommitdiff
path: root/src/mesa
AgeCommit message (Collapse)Author
2008-11-28i965: Add a new state flag BRW_NEW_NR_SURFACES instead of CACHE_NEW_SURFACEEric Anholt
The CACHE_NEW_SURFACE bit always gets spammed since we get many different surface BOs per state emit, but the only consumer of it wanted to just know how many surfaces were enabled.
2008-11-28i965: Fix clashing enums for BRW_NEW_INDICES/VERTICES versus BATCH/DEPTH_BUFFER.Eric Anholt
Fixes upload of large amounts of state for every new primitive emit.
2008-11-28i965: Remove BRW_WM_LOCK dirty bit, introduced to work around lack of relocs.Eric Anholt
This was causing a prepare of wm state at every primitive emit.
2008-11-28i965: Add debug code for dumping how frequently different dirty bits are set.Eric Anholt
2008-11-28i915: Remove dead early z enable bit which was always on.Eric Anholt
2008-11-28i965: Reduce fast-pathiness of brw_try_draw_prims, bringing in important checks.Eric Anholt
Later primitives, even if they caused a full state validate, wouldn't check that there was enough space in the batchbuffer, occasionally triggering the sanity check. We also skipped the aperture space check, even if it would mean bringing in new programs and associated state.
2008-11-28mesa: add missing break statementsBrian Paul
2008-11-28mesa: enable texture compression extensions for software drivers when possibleBrian Paul
2008-11-28mesa: remove unneeded compressed texure size checksPhillip Klaus Krause
2008-11-28mesa: fix default switch case in append_token(), see bug 18734Brian Paul
2008-11-27mesa: remove an assertion (see bug 18734)Brian Paul
2008-11-26windows: fix visual object memleakThomas Henn
2008-11-26windows: fix _mesa_unreference_framebuffer() to pass ** type.Thomas Henn
2008-11-26mesa: add missing type check for function callsBrian Paul
2008-11-26mesa: add missing size check for assignment optimizationBrian Paul
2008-11-26mesa: remove debug codeBrian Paul
2008-11-26mesa: add some missing switch cases for generating state var stringsBrian Paul
2008-11-24mesa: added support for GLSL 1.20 array.length() methodBrian Paul
This is the only method supported in GLSL 1.20 so we take a few short-cuts.
2008-11-24mesa: support for GLSL 1.20 array typesBrian Paul
This allows syntax like "float[8] foo, bar;"
2008-11-24mesa: add gl_program::Input/OutputFlags[] arrayBrian Paul
These arrays will indicate per-input or per-output options for vertex/fragment programs such as centroid-sampling and invariance.
2008-11-24windows: replace free() with _mesa_unreference_framebuffer()Brian Paul
Fixes invalid memory reference bug when exiting.
2008-11-24mesa: copy centroid/invariance/precision info in parse_init_declarator()Brian Paul
2008-11-24mesa: check that varying variable qualifiers agreeBrian Paul
2008-11-24mesa: copy precision/variant/centroid info in slang_fully_specified_type_copy()Brian Paul
2008-11-24mesa: set flags for varying varsBrian Paul
2008-11-24mesa: rename program parameter flags to match other Mesa conventionsBrian Paul
2008-11-24mesa: copy Flags in _mesa_clone_parameter_list()Brian Paul
2008-11-24mesa: dump/debug varying vars listBrian Paul
2008-11-24mesa: added PROG_PARAM_ bits for invariant, flat/linear interpolationBrian Paul
Plus, update the print/debug code.
2008-11-24mesa: add Flags field to gl_program_parameterBrian Paul
Only one flag defined so far: PROG_PARAM_CENTROID_BIT
2008-11-21mesa: issue error, don't crash, when calling a prototyped, but undefined ↵Brian Paul
function Bug #18659.
2008-11-21mesa: better variable name: s/aux/store/Brian Paul
2008-11-21i965: Add support for accelerated CopyTexSubImage.Eric Anholt
There were hacks in EmitCopyBlit before to adjust offsets so that y=0 after the offsets had been adjusted for a negative pitch. It appears that those hacks were due to an unclear and surprising aspect of the hardware: inverting the pitch results in the blit into the specified rectangle being inverted, without the user needing to adjust y and base offset. Tested with piglit copytexsubimage test on 915GM and GM965. Should fix serious performance issues with ETQW and other applications.
2008-11-21intel: Don't glBitmap fallback with scissoring enabled.Eric Anholt
The blit bitmap code already handles scissoring. This is a 15-100% speedup on blender benchmark.blend thanks to avoiding fallbacks. Bug #17951.
2008-11-21i915: Don't overwrite i915's Viewport function from generic code.Eric Anholt
Instead, have i965 and i915 both call the generic function from their Viewport.
2008-11-20mesa: fix shadow sampling unit issue.Xiang, Haihao
texture comparison logic is bypassed if the currently bound texture is not a depth/depth_stencil texture.
2008-11-20intel: fix i830 comment + backwards VB offsets.airlied
According to Keith the docs have these offsets the other way around
2008-11-20intel: fix i8xx vbo enable bitairlied
2008-11-20intel: add lots of i830 engine to intel_decode debugairlied
2008-11-19mesa: minor comment reformatttingBrian Paul
2008-11-19mesa: glsl compiler debug codeBrian Paul
RETURN0 macro reports file/line before returning zero.
2008-11-19mesa: rework GLSL array code generationBrian Paul
We now express arrays in terms of indirect addressing. For example: dst = a[i]; becomes: MOV dst, TEMP[1 + TEMP[2].y]; At instruction-emit time indirect addressing is converted into ARL/ ADDR-relative form: ARL ADDR.x, TEMP[2].y; MOV dst, TEMP[1 + ADDR.x]; This fixes a number of array-related issues. Arrays of arrays and complex array/struct nesting works now. There may be some regressions, but more work is coming.
2008-11-19mesa: don't realloc instruction buffer so oftenBrian Paul
2008-11-19mesa: updated commentBrian Paul
2008-11-19mesa: clamp luminance if needed.Xiang, Haihao
This fixes glReadPixels(GL_LUMINANCE, GL_FLOAT)/glGetTexImage(GL_LUMINANCE, GL_FLOAT) issue on fixed-point color buffers.
2008-11-13mesa: no longer need Writemask field in GLSL IR nodesBrian Paul
The Swizzle and Size fields carry all the info we need now.
2008-11-13mesa: revamp GLSL instruction emit codeBrian Paul
This is a step toward better array handling code. In particular, when more than one operand of an instruction uses indirect addressing, we'll need some temporary instructions and registers. By converting IR storage to instruction operands all in one place (emit_instruction()) we can be smarter about this. Also, somewhat better handling of dst register swizzle/writemask handling. This results in tighter writemasks on some instructions which is good for SOA execution. And, cleaner instruction commenting with inst_comment(). Next: remove some more dead code and additional clean-ups...
2008-11-13mesa: make writemask_string() non-staticBrian Paul
2008-11-13mesa: remove some do-nothing GLSL codeBrian Paul
2008-11-13mesa: fix accidental regression in GLSL built-in texture matrix lookupBrian Paul
Was broken by commit 9aca9a4b72b2a7b378e50bd88f9c3324d07375ec.