summaryrefslogtreecommitdiffstats
path: root/src/gallium/drivers
Commit message (Collapse)AuthorAgeFilesLines
* llvmpipe: Move the determination of the number of threads to the screen.José Fonseca2010-04-248-25/+36
|
* gallivm: Centralize the cpu caps detection.José Fonseca2010-04-241-9/+0
|
* llvmpipe: Pass flatshade_first to setup.José Fonseca2010-04-241-0/+2
| | | | Fixes piglit provoking-vertex.
* svga: D3DCOLOR actually stands for B8G8R8A8.José Fonseca2010-04-241-1/+1
|
* svga: Translate recently added vertex formats.José Fonseca2010-04-241-6/+2
|
* llvmpipe: add initial autoconf support.Dave Airlie2010-04-241-2/+3
| | | | | | | | | allows the swrastg_dri.so to be built with llvmpipe, also links llvm to all dri drivers use --enable-gallium-llvm to use it. Signed-off-by: Dave Airlie <[email protected]>
* nouveau: Add to SCons build.Vinson Lee2010-04-241-0/+11
|
* llvmpipe: Remove unused variable.Vinson Lee2010-04-241-2/+0
|
* softpipe: compressed formats can't be render targetsBrian Paul2010-04-231-1/+4
|
* gallium/failover: include u_memory.h for MALLOC, FREEBrian Paul2010-04-231-0/+1
|
* gallium: s/free/FREE/ and same for friends.José Fonseca2010-04-239-28/+28
| | | | Based on Stephen Johnson's feedback.
* llvmpipe: rename texture refs to resource refsBrian Paul2010-04-233-26/+34
|
* llvmpipe: update for half float formatsAlan Hourihane2010-04-231-1/+1
|
* softpipe: YUV textures should be supported.José Fonseca2010-04-231-4/+0
|
* llvmpipe: increase max 2D texture / surface size to 4K x 4KBrian Paul2010-04-231-1/+1
|
* llvmpipe: rename MAXWIDTH, MAXHEIGHTBrian Paul2010-04-233-6/+6
|
* llvmpipe: rename lp_tile_size.h to lp_limits.hBrian Paul2010-04-235-7/+10
| | | | Collect the various driver limits in one place since.
* llvmpipe: move LP_MAX_TEXTURE_x_LEVELS, etc to lp_tile_size.hBrian Paul2010-04-233-8/+19
|
* llvmpipe: Remove unused variable.Vinson Lee2010-04-231-2/+0
|
* gallivm: implement indirect addressing over temporariesZack Rusin2010-04-221-1/+1
| | | | | | a bit more involved than indirect addressing over consts, but still fairly reasonable. we allocate an array instead of individual alloca's, and we do it only if the shader does indirect addressing.
* llvmpipe: Undo zs tiling changes.José Fonseca2010-04-221-0/+4
| | | | | | tile_w/tile_h have different meaning there. This partialy reverts commit ccdc6b5913885866e100780bfd661672c9a5d23c.
* llvmpipe: Portability fixes.José Fonseca2010-04-222-4/+4
|
* llvmpipe: Fix rendering to non 32bpp formats.José Fonseca2010-04-223-20/+22
| | | | | Tiled layout always used 32bpp, therefore linear/tiled strides are not related.
* llvmpipe: Use resource_is_texture() consistently.José Fonseca2010-04-222-25/+16
| | | | | | | | | Otherwise slightly difference order causes assertion failures. Also remove mentions of PIPE_BIND_SCANOUT/PIPE_BIND_SHARED. They are not propoer bind flags and will likely be deprecated. If surfaces should be passed to the winsys then they should have the DISPLAY_TARGET flag set, which is a proper bind flag.
* llvmpipe: Fix resource_is_texture.José Fonseca2010-04-221-8/+12
| | | | | | | | It was missing PIPE_BIND_RENDER_TARGET, causing assertion failures for pure render targets. Also bind flags are too variable and complex for a good assessment for whether the resource is a texture or not. Target is more concise.
* llvmpipe: Do not advertise S3TC rendering support.José Fonseca2010-04-221-10/+10
|
* llvmpipe: Fix typo in assert.José Fonseca2010-04-221-2/+2
|
* r300g: fix vertex unit setupAlex Deucher2010-04-221-3/+7
| | | | | | RV3xx is 2, RV560,RV570 is 8 Noticed by Tormod Volden.
* llvmpipe: add missing dependency in MakefileBrian Paul2010-04-211-1/+1
|
* r300g: raise the max index for generic varyings to 32Marek Olšák2010-04-211-1/+1
| | | | | This should fix an assertion failure in the game Heroes of Newearth. Yes, the game seems to generate semantic indices greater than 15.
* softpipe: call draw_wide_point_sprites()Brian Paul2010-04-211-0/+2
|
* llvmpipe: consolidate texture memory allocation codeBrian Paul2010-04-211-11/+48
| | | | | And fix incorrect allocation of linear memory for display targets. This fixes fd.o bugs 27761 and 27762.
* llvmpipe: Remove unused variable.Vinson Lee2010-04-201-3/+1
|
* i965g: s/inttypes.h/stdint.h/Vinson Lee2010-04-201-1/+1
| | | | MSVC doesn't have either but Mesa has stdint.h at include/c99.
* i965g: Remove unnecessary header.Vinson Lee2010-04-201-1/+0
|
* llvmpipe: Change return type of declaration to match code.Vinson Lee2010-04-201-1/+1
| | | | | | | | Commit 2142108e0e1cf1ed8d0142e9c41fe1947abe0907 changed the return type of llvmpipe_get_texture_image_address function but didn't change the declaration. Fixes build.
* llvmpipe: code consolidationBrian Paul2010-04-201-55/+18
|
* llvmpipe: remove debug code and simplifyBrian Paul2010-04-201-14/+4
|
* llvmpipe: fix cube face addressing bugBrian Paul2010-04-201-20/+60
| | | | Fixes fd.o bug 27760 (pigit fbo-cubemap).
* llvmpipe: whitespace fixBrian Paul2010-04-201-1/+1
|
* r300g: Spill a bit more info about implementation errors in surface_copy.Corbin Simpson2010-04-201-1/+7
| | | | compiz appears to hit this. Weird.
* nv50: make a working check for user memory vertex/index buffersChristoph Bumiller2010-04-202-2/+10
|
* nv50: fix FP result counting for depth outputChristoph Bumiller2010-04-201-1/+1
|
* nv50: support vertex index biasChristoph Bumiller2010-04-202-7/+41
|
* llvmpipe: fix depth+stencil logic errorBrian Paul2010-04-201-5/+18
| | | | | | If both Z-test and stencil-test were enabled, we were mis-computing the vector of updated Z buffer values. Fixes Z testing bug in progs/demos/fbotexture.c
* llvmpipe: fix comment/typoBrian Paul2010-04-201-1/+3
|
* llvmpipe: fix incorrect front-facing value for fragment shaderBrian Paul2010-04-201-1/+6
| | | | | The TGSI convention is +1 for front-facing, -1 for back-facing Fixes glean glsl1 gl_FrontFacing tests.
* llvmpipe: rename mode -> interpBrian Paul2010-04-202-16/+16
|
* llvmpipe: remove dead code/commentBrian Paul2010-04-201-2/+0
|
* nv50: move declaration before codeBrian Paul2010-04-201-1/+1
|