Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | nv50: any cpu access to a texture is done on its backing images | Ben Skeggs | 2009-01-12 | 3 | -7/+72 |
| | | | | | | Still a little dodgy: - RTT will hit an assertion (hopefully!) and fail - 3D textures with depth >= 32 will cause bad things to happen | ||||
* | nv50: create buffers for each image that makes up a texture | Ben Skeggs | 2009-01-12 | 2 | -13/+38 |
| | |||||
* | nouveau: return buffer map to something sane. | Ben Skeggs | 2009-01-12 | 1 | -20/+0 |
| | | | | | | Sorry, but no, we're not doing this.. Correctness always takes precedence over speed. Implement this higher up where you know it's safe to do so, and doesn't break other things in the process. | ||||
* | nouveau: use usage, not uninitialised flags value... | Ben Skeggs | 2009-01-12 | 1 | -1/+1 |
| | |||||
* | cell: optimize unpack_colors() function, saving 12 cycles | Brian Paul | 2009-01-11 | 1 | -31/+34 |
| | |||||
* | cell: move color unpacking code into separate function | Brian Paul | 2009-01-11 | 1 | -76/+89 |
| | |||||
* | cell: re-order the z/stencil fetch/extract/convert instructions for better perf | Brian Paul | 2009-01-11 | 1 | -55/+51 |
| | | | | The new instruction order is 10 cycles faster. | ||||
* | cell: datatype clean-ups in SPE rtasm | Brian Paul | 2009-01-11 | 2 | -105/+99 |
| | |||||
* | cell: simplify the 'optional register' code | Brian Paul | 2009-01-11 | 1 | -62/+50 |
| | |||||
* | cell: asst datatype clean-ups | Brian Paul | 2009-01-11 | 1 | -73/+75 |
| | |||||
* | cell: move depth/stencil code into separate function | Brian Paul | 2009-01-11 | 1 | -179/+213 |
| | |||||
* | cell: clean-up, re-indent, comments | Brian Paul | 2009-01-11 | 1 | -95/+139 |
| | |||||
* | cell: use tgsi_dump_instruction() instead of spe_comment() | Brian Paul | 2009-01-10 | 1 | -101/+25 |
| | |||||
* | gallium: use tgsi_dump_instruction() instead of ppc_comment() | Brian Paul | 2009-01-10 | 1 | -84/+7 |
| | |||||
* | gallium: remove unused struct type | Brian Paul | 2009-01-10 | 1 | -8/+0 |
| | |||||
* | gallium: fix register clobber bug in TGSI->PPC codegen | Brian Paul | 2009-01-10 | 1 | -10/+41 |
| | | | | | | | | When negating a src vector that's stored in a altivec register, need to put negated value into a new register so we don't upset the original value. This solves the dark colors in the mandelbrot GLSL demo. Also, use new predicate functions to check if a TGSI temp is stored in an altivec register. | ||||
* | gallium: emit comments in TGSI->PPC codegen | Brian Paul | 2009-01-10 | 1 | -2/+90 |
| | |||||
* | gallium: code to dump/debug PPC code (disabled) | Brian Paul | 2009-01-10 | 1 | -0/+5 |
| | |||||
* | gallium: added comment/annotation support to PPC rtasm | Brian Paul | 2009-01-10 | 2 | -62/+187 |
| | |||||
* | cell: added rule to produce .s assembly files | Brian Paul | 2009-01-10 | 1 | -0/+3 |
| | |||||
* | gallium: s/false/FALSE/ | Brian Paul | 2009-01-10 | 1 | -1/+1 |
| | |||||
* | nouveau: Update nv30 swizzling. | Younes Manton | 2009-01-10 | 2 | -9/+10 |
| | |||||
* | nouveau: Factor out common winsys bits into libnouveaudrm.a | Younes Manton | 2009-01-10 | 44 | -922/+927 |
| | |||||
* | g3dvl: Use Gallium thread wrappers. | Younes Manton | 2009-01-10 | 1 | -4/+4 |
| | |||||
* | g3dvl: Use Gallium MALLOC wrappers. | Younes Manton | 2009-01-10 | 9 | -31/+34 |
| | |||||
* | g3dvl: Get rid of old unbuffered motion compensation code. | Younes Manton | 2009-01-10 | 4 | -2365/+1 |
| | |||||
* | g3dvl: Map vertex bufs once per frame/flush. | Younes Manton | 2009-01-10 | 1 | -36/+48 |
| | |||||
* | nouveau: Catch some more leaks. | Younes Manton | 2009-01-10 | 2 | -0/+3 |
| | |||||
* | g3dvl: Fix some memory leaks. | Younes Manton | 2009-01-10 | 2 | -0/+21 |
| | |||||
* | g3dvl: Allocate one set of bufs, let winsys rename them as necessary. | Younes Manton | 2009-01-10 | 2 | -158/+178 |
| | |||||
* | nouveau: Add busy() query, determines if BOs can be mapped immediately. | Younes Manton | 2009-01-10 | 2 | -0/+37 |
| | |||||
* | nouveau: Use PIPE_BUFFER_USAGE_CPU_* instead of custom. | Younes Manton | 2009-01-10 | 4 | -5/+4 |
| | |||||
* | g3dvl: Expand YCbCr to full RGB range by default. | Younes Manton | 2009-01-10 | 1 | -1/+1 |
| | |||||
* | g3dvl: Work around mplayer weirdness in XvMCPutSurface(). | Younes Manton | 2009-01-10 | 1 | -2/+8 |
| | |||||
* | g3dvl: Flag textures holding incoming data as dynamic. | Younes Manton | 2009-01-10 | 1 | -1/+1 |
| | |||||
* | nouveau: Put dynamic textures in GART for CPU access and don't swizzle. | Younes Manton | 2009-01-10 | 5 | -8/+17 |
| | | | | Also flag shadows as dynamic since they're for CPU access as well. | ||||
* | gallium: Define PIPE_TEXTURE_USAGE_DYNAMIC. | Younes Manton | 2009-01-10 | 1 | -0/+1 |
| | | | | | Knowing how the client intends to use the texture will give the driver the opportunity to optimize for such cases. | ||||
* | nouveau: Swizzle textures larger than nv04 SIFM limit in parts. | Younes Manton | 2009-01-10 | 2 | -24/+69 |
| | | | | Limit of SIFM on nv40 is 1024x1024, not sure about others. | ||||
* | draw: Add missing include. | José Fonseca | 2009-01-09 | 1 | -0/+2 |
| | |||||
* | st: build fix | Alan Hourihane | 2009-01-09 | 1 | -1/+0 |
| | |||||
* | glsl: fix typo in the vec2 += operator function | Brian Paul | 2009-01-09 | 2 | -2/+2 |
| | |||||
* | mesa: set version string to 7.3-rc1 | Brian Paul | 2009-01-09 | 1 | -1/+1 |
| | |||||
* | glsl: fix broken +=, -=, *=, /= operators | Brian Paul | 2009-01-09 | 2 | -737/+786 |
| | | | | | | | These functions need to return the final computed value. Now expressions such as a = (b += c) work properly. Also, no need to use __asm intrinsics in these functions. The resulting code is the same when using ordinary arithmetic operators and is more legible. | ||||
* | mesa: fix off-by-one bug in _mesa_delete_instructions() | Brian Paul | 2009-01-09 | 1 | -1/+1 |
| | |||||
* | i965: Fix GLSL FS DPH to return the right value instead of src0.w * src1.w. | Eric Anholt | 2009-01-09 | 1 | -1/+1 |
| | |||||
* | i965: Remove worrisome comment about _NEW_PROGRAM signaling fp change. | Eric Anholt | 2009-01-09 | 1 | -4/+0 |
| | | | | | Everything now depends on either BRW_NEW_FRAGMENT_PROGRAM or BRW_NEW_VERTEX_PROGRAM. | ||||
* | mesa: Remove _Active and _UseTexEnvProgram flags from fragment programs. | Eric Anholt | 2009-01-09 | 8 | -41/+5 |
| | | | | | | | | There was a note in state.c about _Active deserving to die, and there were potential issues with it due to i965 forgetting to set _UseTexEnvProgram. Removing both simplifies things. Reviewed-by: Brian Paul <[email protected]> | ||||
* | i965: Remove dead brw_vs_tnl.c | Eric Anholt | 2009-01-09 | 3 | -61/+0 |
| | |||||
* | i965: allow gl_FragData[0] usage when there's only one color buffer | Brian Paul | 2009-01-09 | 1 | -2/+9 |
| | | | | If gl_FragData[0] is written but not gl_FragCOlor, use the former. | ||||
* | mesa: additional case in file_string() | Brian Paul | 2009-01-09 | 1 | -0/+2 |
| |