Commit message (Collapse) | Author | Age | Files | Lines | |
---|---|---|---|---|---|
* | slang: Check return value from emit_instruction(). | Michal Krol | 2009-11-23 | 1 | -6/+67 |
| | |||||
* | tgsi: Clamp the source argument in micro_exp2() to avoid Inf. | Michal Krol | 2009-11-23 | 1 | -0/+18 |
| | |||||
* | tgsi: Remove code that actually had no effect. | Michal Krol | 2009-11-23 | 1 | -4/+0 |
| | |||||
* | softpipe: Initialise TGSI machine's Face. | Michal Krol | 2009-11-23 | 1 | -1/+7 |
| | |||||
* | tgsi: Fix POSITION and FACE fragment shader inputs. | Michal Krol | 2009-11-23 | 2 | -36/+42 |
| | |||||
* | tgsi: Bring BGNFOR/ENDFOR implementation up to spec. | Michal Krol | 2009-11-23 | 2 | -13/+35 |
| | |||||
* | tgsi: Fake TXD implementation in exec. | Michal Krol | 2009-11-23 | 1 | -1/+59 |
| | |||||
* | tgsi: Add execution debugging facilities to exec. | Michal Krol | 2009-11-23 | 1 | -4/+63 |
| | |||||
* | slang: Be more robust with memory in concat_shaders(). | Michal Krol | 2009-11-23 | 1 | -3/+19 |
| | |||||
* | r600 : add support for shader instruction trunc and discard. | Richard Li | 2009-11-22 | 2 | -29/+78 |
| | |||||
* | r600: hopefully fix segfault. | Dave Airlie | 2009-11-23 | 1 | -1/+2 |
| | |||||
* | r600: fix inline issues | Dave Airlie | 2009-11-23 | 2 | -4/+2 |
| | |||||
* | r600 : add stack depth calculation, enable CF pop. | Richard Li | 2009-11-22 | 3 | -161/+223 |
| | |||||
* | r600 : use cf for all pop now, left optimization for future. | Richard Li | 2009-11-22 | 2 | -16/+42 |
| | |||||
* | r300g: Build fix. | Corbin Simpson | 2009-11-21 | 1 | -1/+1 |
| | | | | Oops. | ||||
* | r300g: do not reallocate r300_vertex_info and r300_rs_block all the time | Marek Olšák | 2009-11-21 | 2 | -24/+20 |
| | |||||
* | r300g: clean up vs/fs tabs | Marek Olšák | 2009-11-21 | 2 | -94/+25 |
| | | | | | Instead of vs_tab, we use vs_output_tab and it's local now. fs_tab hasn't been used anywhere, so I removed it and r300_update_fs_tab too. | ||||
* | r300g: silence warnings | Marek Olšák | 2009-11-20 | 2 | -2/+4 |
| | |||||
* | r300g: fix typo in r300_reg.h to prevent the RS unit from doing random things | Marek Olšák | 2009-11-20 | 2 | -15/+13 |
| | | | | | | And reorder fragment shader inputs so that the colors are before texcoords, as is allocated by the shader compiler. This commit makes VS->FS attribute routing work on R500. | ||||
* | r300g: fix updating a vertex format | Marek Olšák | 2009-11-20 | 2 | -2/+5 |
| | | | | | We must update PSC when we change the vertex format, e.g. vertex colors from RGBA to BGRA. | ||||
* | r300g: add texture lod clamping | Marek Olšák | 2009-11-20 | 5 | -5/+23 |
| | | | | | | These now work: piglit/lodclamp piglit/levelclamp | ||||
* | r300g: Begin separating HW TCL and SW TCL state and setup. | Corbin Simpson | 2009-11-20 | 6 | -174/+110 |
| | | | | | | | This patch removes draw_context entirely from the HW TCL path and cleans up a few other things along the way. Hopefully, nothing got broken. Thanks to Marek Olšák for testing, review, and pointing out my bugs. :3 | ||||
* | r300g: emit R300_TEX_ENABLE to indicate there are no textures bound | Marek Olšák | 2009-11-20 | 2 | -2/+15 |
| | | | | Previously, this reg wasn't emitted at all if texture_count == 0. | ||||
* | r300g: remove variant states from emit_state_invariant | Marek Olšák | 2009-11-20 | 1 | -3/+1 |
| | |||||
* | r300g: Texrect factor cleanup. | Corbin Simpson | 2009-11-20 | 1 | -10/+14 |
| | | | | | (0, 0, 0, 1) is a much saner default value, and texrect factors only need to be (1/s, 1/t, 0, 1). | ||||
* | r300g: fix rectangle textures on r3xx | Marek Olšák | 2009-11-20 | 2 | -2/+28 |
| | | | | Adapted from Maciej Cencora's patch. | ||||
* | r300g: Clean up bitcounting function. | Corbin Simpson | 2009-11-20 | 1 | -5/+10 |
| | | | | | I didn't see this in u_math; surely somebody else has this wheel reinvented elsewhere. | ||||
* | r300g: Use MAX3 and MIN3. | Corbin Simpson | 2009-11-20 | 2 | -4/+1 |
| | |||||
* | util: Add MAX3 and MIN3. | Corbin Simpson | 2009-11-20 | 1 | -0/+3 |
| | |||||
* | r300g: set better values in the R300_VAP_CNTL register | Marek Olšák | 2009-11-20 | 1 | -5/+28 |
| | |||||
* | Fix comment to use /* */ rather than // | Alan Hourihane | 2009-11-20 | 1 | -16/+16 |
| | |||||
* | Fix indentation. | Alan Hourihane | 2009-11-20 | 1 | -3/+4 |
| | |||||
* | Fix vega compilation. | Alan Hourihane | 2009-11-20 | 3 | -5/+12 |
| | |||||
* | Fix memory leak. | Alan Hourihane | 2009-11-20 | 1 | -0/+1 |
| | |||||
* | r600 : eliminate Wondows line ending for test code. | Richard Li | 2009-11-20 | 1 | -28/+28 |
| | |||||
* | Merge remote branch 'origin/mesa_7_7_branch' | Dave Airlie | 2009-11-20 | 38 | -740/+743 |
|\ | |||||
| * | r100: fix texture_from_pixmap and compiz. | Dave Airlie | 2009-11-20 | 1 | -7/+10 |
| | | | | | | | | | | | | | | | | r100 state emission has separate rect and non-rect states, if we are doing TFP for a TEXTURE_2D we shouldn't use the rect states as they won't get emitted properly. Signed-off-by: Dave Airlie <[email protected]> | ||||
| * | st/xorg: Fix infinite loop in copy_packed_data. | Vinson Lee | 2009-11-19 | 1 | -1/+1 |
| | | |||||
| * | st/mesa: replace st_flush() with pipe->flush() | Brian Paul | 2009-11-19 | 1 | -1/+1 |
| | | | | | | | | | | We only need to flush the gallium driver in this case. Fixes a recursive state validation bug. | ||||
| * | vbo: added recursion check in vbo_exec_FlushVertices() | Brian Paul | 2009-11-19 | 1 | -1/+17 |
| | | |||||
| * | st/mesa: get rid of compressed_num_bytes() code | Brian Paul | 2009-11-19 | 1 | -36/+4 |
| | | |||||
| * | st/mesa: remove is_compressed_mesa_format() | Brian Paul | 2009-11-19 | 1 | -21/+2 |
| | | |||||
| * | st/mesa: remove pointless assertion | Brian Paul | 2009-11-19 | 1 | -2/+0 |
| | | |||||
| * | softpipe: whitespace/indentation fixes | Brian Paul | 2009-11-19 | 1 | -5/+3 |
| | | |||||
| * | softpipe: add missing check in softpipe_is_texture_referenced() | Brian Paul | 2009-11-19 | 1 | -2/+8 |
| | | | | | | | | Check if the named texture is referenced by the texture cache. | ||||
| * | st/xorg: Replace compile-time acceleration switch with Option "2DAccel". | Michel Dänzer | 2009-11-19 | 5 | -12/+25 |
| | | | | | | | | | | This option can be used to disable 2D acceleration. DRI2 and XVideo blits will still be accelerated, at least to some degree even with compositing. | ||||
| * | st/xorg: Miscellaneous XVideo fixes. | Michel Dänzer | 2009-11-19 | 1 | -3/+11 |
| | | | | | | | | | | | | * Make sure the destination pixmap has a texture to render into. * Fix damage reporting so the EXA migration code can do the right thing. * Fix destination coordinates for redirected windows. | ||||
| * | st/xorg: Try harder to ensure a shared texture has valid contents right away. | Michel Dänzer | 2009-11-19 | 1 | -0/+3 |
| | | |||||
| * | st/xorg: Don't complain about convolution filter being 'unknown'. | Michel Dänzer | 2009-11-19 | 1 | -1/+2 |
| | | | | | | | | | | Also add a newline to the complaint so it'll be a little less annoying if we actually encounter an unknown filter value again. | ||||
| * | st/xorg: Remove superfluous flushes from the EXA WaitMarker hook. | Michel Dänzer | 2009-11-19 | 1 | -9/+1 |
| | | | | | | | | Thanks to Thomas Hellström for pointing this out. |