summaryrefslogtreecommitdiff
AgeCommit message (Collapse)Author
2008-06-29nv50: make TEX a halfieBen Skeggs
2008-06-29nv50: get tri-scissor-tri workingBen Skeggs
2008-06-29nv50: separate state validation and upload, similar to nv40Ben Skeggs
2008-06-29nv50: fix blend colourBen Skeggs
2008-06-29nv50: use stateobjs for sampler/image_control uploadsBen Skeggs
2008-06-29nv50: hack of a TEX opcodeBen Skeggs
2008-06-29nv50: hacks for stuff I don't really get yetBen Skeggs
2008-06-29nv50: abuse constbuf upload for program uploadBen Skeggs
2008-06-29nv50: a couple more bits'n'piecesBen Skeggs
2008-06-29nv50: disable inline IMMD for now, IMMD+pred == BANG!Ben Skeggs
fixes progs/fp/lit.txt
2008-06-29nv50: simplify interp crap a bit...Ben Skeggs
hopefully there wasn't a good reason I went the route I did.. can't recall..
2008-06-29nv50: do tsc/tic upload + stub out shader TEX stuffBen Skeggs
2008-06-29nv50: disable ztest for now - it doesn't work stillBen Skeggs
2008-06-29nv50: flag to indicate to winsys we want a surface for use as a zeta bufferBen Skeggs
NVIDIA love to make life difficult.. we need different flags in PTEs for zeta.. yay.. not.
2008-06-29nv50: comment on a so-far unseen bugBen Skeggs
2008-06-29nv50: another vbo formatBen Skeggs
2008-06-29nv50: fix blend csoBen Skeggs
2008-06-29nv50: valgrind complaintBen Skeggs
2008-06-29nv50: draw_elements() - inline only for the momentBen Skeggs
2008-06-29nv50: support a couple more common VBO formatsBen Skeggs
2008-06-29nv50: remove some cruft, don't upload program unless really neededBen Skeggs
2008-06-29nv50: more efficient const upload + fixes (fp/* works now!)Ben Skeggs
2008-06-29nv50: use constbuf segment 0 for everything - I can't make the others work..Ben Skeggs
2008-06-29nv50: allow relocating a shader's constants at upload timeBen Skeggs
2008-06-29nv50: carry instructions around in nv50_program_exec, not a flat arrayBen Skeggs
2008-06-29nv50: get vp working again, fp is broken regardless somehow..Ben Skeggs
g8x sucks :)
2008-06-29nv50: rename nv50_state.h to nv50_program.hBen Skeggs
2008-06-29nv50: move magics take 2Ben Skeggs
2008-06-29Revert "nv50: move some magics"Ben Skeggs
This reverts commit 0a38de30429d3075fc6dfc9ff3729c5ca11f0c2f.
2008-06-29nv50: move some magicsBen Skeggs
2008-06-29nv50: various fixes + SCSBen Skeggs
2008-06-29nv50: LRPBen Skeggs
2008-06-29nv50: quick hack so progs/fp/fp-tri works for the momentBen Skeggs
2008-06-29nv50: fix SGE/SLT when sources need swappingBen Skeggs
2008-06-29nv50: support the other TGSI_UTIL_SIGN modesBen Skeggs
2008-06-29nv50: note some things discovered during renouveau sessionBen Skeggs
2008-06-29nv50: use emit_flop() instead of building RCP manually on interpBen Skeggs
2008-06-29nv50: remove NV50_PROG_{VERTEX,FRAGMENT}Ben Skeggs
2008-06-29nv50: delayed write of fragprog result regs until end of programBen Skeggs
2008-06-29nv50: fix src1 & src2 == constBen Skeggs
2008-06-29nv50: note a critical bugBen Skeggs
2008-06-29nv50: hehe, damage from tex-surfaces.. surface_fill() reenabled now :)Ben Skeggs
2008-06-29nv50: obey writemask in a couple of placesBen Skeggs
2008-06-29nv50: LIT - sort-ofBen Skeggs
*somehow* we have the exact same bug here as on nv4x, the difference being on nv4x the hw actually has a LIT opcode.. NVIDIA doesn't have the bug on either arch FWIW.
2008-06-29nv50: obey per-source abs (TGSI_UTIL_SIGN_CLEAR)Ben Skeggs
2008-06-29nv50: create emit_pow() - emit_lit() will need to use itBen Skeggs
2008-06-29nv50: oops, copy+pastoBen Skeggs
2008-06-29nv50: POW!Ben Skeggs
2008-06-29nv50: fixes + untested _SAT modifierBen Skeggs
2008-06-29nv50: handle 0/1 SWZBen Skeggs
generated by cgit v1.2.3 (git 2.25.1) at 2024-11-10 08:40:52 +0000