summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2007-01-25fix dstelt memory allocation to avoid hash corruptionRoland Scheidegger
2007-01-25remove tnl/t_vtx_x86_gcc.S file from sources as it no longer existsRoland Scheidegger
2007-01-25re-fix bug 9062 in vbo codeRoland Scheidegger
2007-01-25nouveau: unbreak NOUVEAU_RING_DEBUGBen Skeggs
2007-01-25nouveau: add result scaling to shader backend, use it in RSQ emul for NV40.Ben Skeggs
2007-01-25nouveau: and here's the "oops, I suck" commit :)Ben Skeggs
2007-01-25nouveau: some extra debug outputBen Skeggs
2007-01-24remove some parenthesisBrian
2007-01-24use FRAG_BIT_TEX()Brian
2007-01-24fix varying interpolationBrian
2007-01-24only setup varying vars that will be usedBrian
2007-01-24added FRAG_BIT_TEX() and FRAG_BIT_VAR() macrosBrian
2007-01-24965 glxswapcontrol fixZou Nan hai
2007-01-24 965 ARB_Occlusion_query fixZou Nan hai
2007-01-23updated commentBrian
2007-01-23indicate which opcodes are used by Mesa GLSL, updated commentsBrian
2007-01-23fix g++ warnings/errorsBrian
2007-01-23silence C++ warningsBrian
2007-01-23fixes for C++ warnings/errorsBrian
2007-01-23i915tex: Relocation fixes:Thomas Hellstrom
Increase the number of allowed relocations per batchbuffer. Fix an assert to avoid an array index overflow. (Reported by Steve Wilkins)
2007-01-23i915tex: Fix randr resizing. Rotation still broken.Thomas Hellstrom
2007-01-23nouveau: DPH and CMP for NV40 which doesn't do it natively.Ben Skeggs
2007-01-23nouveau: allow for card-specific shader infos to be keptBen Skeggs
NV30/40 fragprog: build FP_CONTROL per-shader, still some hardcoded bits for this reg.. It looks like it has to do with the number of temps used, but needs more looking at. NV40 vtxprog : build VP_IN_REG/VP_OUT_REG during shader compile
2007-01-23nouveau: fill in condition info for instructionsBen Skeggs
2007-01-23nouveau: unbreak nv40Ben Skeggs
2007-01-23nouveau: reindent shader pass0/pass2Ben Skeggs
if this gets rejected by the commit list, just ignore it.. nothing interesting to see here :)
2007-01-23nouveau: start converting shaders into a tree format again..Ben Skeggs
No branching stuff implemented yet. Works enough for gears, probably other stuff broken.
2007-01-21Merge branch 'master' of git+ssh://git.freedesktop.org/git/mesa/mesaJeremy Kolb
2007-01-21nouveau: Fixes for nv30.Jeremy Kolb
2007-01-22nouveau: fix some bugs in the nv10 swtcl.Stephane Marchesin
2007-01-21nouveau: add nv04 state support, and small nv04 fixes.Stephane Marchesin
2007-01-21nouveau: rename pass0_arb to pass0.Ben Skeggs
I was expecting to have 2 frontends for the shader code (asm, glsl). With Brian's work on GLSL this is unnecessary :)
2007-01-21nouveau: kill off shader pass1.Ben Skeggs
It sucks, and we have someone who can do a much better job than I can starting work on it soon. alloc_temp/free_temp is left in pass2 to workaround fragprog temps/outputs overlapping, but this all belongs in the optimiser.
2007-01-21nouveau: shader backend branching support for all cards that support it.Ben Skeggs
2007-01-20Add a simple mechanism for annotating instructions for easier debugging.Brian
2007-01-20optimization: emit MAD instructions when possibleBrian
2007-01-20remove unused swizzle parameter to -storage_to_src_reg()Brian
2007-01-20Initial implementation of OPCODE_IF/ELSE/ENDIF instructions.Brian
2007-01-20Reimplement && and || to do short-circuit evaluation.Brian
Improved shader error handling.
2007-01-20rewrite more __postIncr functionsBrian
2007-01-19Implement do/while loops. Replace IR_CJUMP with IR_CJUMP0 and IR_CJUMP1 soBrian
we can either jump on zero, or non-zero predicate.
2007-01-19Rewrite normalize(vec3/vec4) to use one less register.Brian
2007-01-19report error message when something failsBrian
2007-01-19print error msg when there's a problemBrian
2007-01-19remove stray tabBrian
2007-01-19Implement fragment discard/kill.Brian
2007-01-19change while-loop to create new scope for loop body, per specBrian
2007-01-19nouveau: bump drm patchlevelBen Skeggs
2007-01-18Implement constant sharing so that 4 float constants can share a singleBrian
float[4] register slot.
2007-01-18_mesa_add_unnamed_constant() now tries to re-use constants already inBrian
the parameter list.