| Age | Commit message (Collapse) | Author | |
|---|---|---|---|
| 2010-01-13 | llvmpipe: improve empty-bin test further | Keith Whitwell | |
| Remove unused param, add comments. Thanks to Brian for review. | |||
| 2010-01-13 | llvmpipe: improve empty-bin test | Keith Whitwell | |
| We emit at most two clear packets (color and z respectively). | |||
| 2010-01-13 | llvmpipe: quick hack to short-circuit empty bins | Keith Whitwell | |
| 2010-01-13 | svga: Fix compile_vs error code. | José Fonseca | |
| It could erroneously return PIPE_OK in some circumstances. Make compile_fs code identical. | |||
| 2010-01-13 | svga: Fix call to translate::run(). | Michal Krol | |
| 2010-01-13 | r300g: Account for CS space used per atom. | Corbin Simpson | |
| Oh yeah, those atoms are startin' to pay off. The main obstacle now for OA playability is the absurdly low default mouse sensitivity, IMO. Not totally smooth yet, but getting there. | |||
| 2010-01-13 | r300g: Two dirty state optimizations. | Corbin Simpson | |
| 2010-01-13 | r300g: DCE in emit. | Corbin Simpson | |
| 2010-01-12 | r300g: Lower the immd mode threshold. | Corbin Simpson | |
| 2010-01-12 | r300g: Fix is_r400 status for rs4xx and rc4xx. | Corbin Simpson | |
| As suggested by agd5f. | |||
| 2010-01-12 | r300g: Switch to immediate mode for tiny renders. | Corbin Simpson | |
| Useful for e.g. blitter. | |||
| 2010-01-12 | r300g: emit blend LTE/GTE thresholds on >=RV350 | Marek Olšák | |
| As per classic r300. | |||
| 2010-01-12 | r300g: update the texture initialization so that it respects tiling | Marek Olšák | |
| This is still work-in-progress and tiling is not enabled by default. | |||
| 2010-01-12 | r300g: add emission of texture tiling bits | Marek Olšák | |
| 2010-01-12 | r300g: add is_r400 flag | Marek Olšák | |
| r4xx has some additional fragment shader registers compared to r3xx. | |||
| 2010-01-12 | r300g: silence a warning | Marek Olšák | |
| 2010-01-12 | r300g: only emit scissor when needed | Marek Olšák | |
| Reverting some bits from ce1c493ff8fad4b62e2b66f06636ac6560a6e0ad. Given the latest fixes, it's not needed to always emit scissor, really. | |||
| 2010-01-12 | Revert "r300g: Placate kernel checker by explicitly disabling depth test." | Marek Olšák | |
| This reverts commit e920ee23b32f6031a7b8527b540566e7ada6af8a. Assuming the FB state is not dirty and the DSA state is, then the depth test will be re-enabled, making the kernel checker angry. Reverting the commit fixes piglit/glsl-bug-22603. | |||
| 2010-01-12 | r300g: mark all states as dirty after flush | Marek Olšák | |
| It fixes almost all regressions introduced lately. | |||
| 2010-01-12 | llvmpipe: fix indentation, comment typo | Brian Paul | |
| 2010-01-12 | llvmpipe: debug checks: make sure scene is empty at key points | Brian Paul | |
| 2010-01-12 | llvmpipe: assert that we're putting data into a valid bin | Brian Paul | |
| 2010-01-12 | llvmpipe: clamp maxx,maxy to framebuffer size (in terms of tiles) | Brian Paul | |
| In some corner cases the right-most / bottom-most vertex can be right on the edge of the framebuffer. Because the maxx, maxy vals are computed with a series of float/int, pixel/tile transformations we can end up with maxx >= scene->x_tiles or maxy >= scene->y_tiles. This leads to putting data into bins that never get processed, or reset. This becomes stale data that can lead to segfaults. Clamping fixes this. | |||
| 2010-01-12 | llvmpipe: added lp_scene_is_empty() | Brian Paul | |
| 2010-01-12 | llvmpipe: silence unused var warnings | Brian Paul | |
| 2010-01-12 | gallium: draw_arrays/elements_instanced() are of type void. | Michal Krol | |
| 2010-01-12 | llvmpipe: remove unused #define | Brian Paul | |
| 2010-01-12 | nv50: fix 2 off by one memory leaks (nv50_miptree_level->image_offset) | Marcin Slusarz | |
| 2010-01-12 | nv50: fix memory leak on nv50_pc free | Marcin Slusarz | |
| 2010-01-12 | gallium: Rename ALIGN_STACK -> PIPE_ALIGN_STACK for consistency. | José Fonseca | |
| 2010-01-12 | gallium: Simplify PIPE_ALIGN_VAR. | José Fonseca | |
| gcc allows pre-fix variable attributes. Suggested by Ian Romanick. | |||
| 2010-01-12 | gallium: Generalize the alignment macros to other compilers and any alignment. | José Fonseca | |
| 2010-01-12 | r300g: Always emit scissors. | Corbin Simpson | |
| Save some code, improve FPS, and fix piglit tests. Everybody wins. | |||
| 2010-01-11 | r300g: Add back dirty state check. | Corbin Simpson | |
| From 07ea7e6c80. This is the only questionable part of that commit, AFAICT... | |||
| 2010-01-11 | r300g: Placate kernel checker by explicitly disabling depth test. | Corbin Simpson | |
| This probably should disable stencil, too, if the kernel cares enough. Note: When atomized, framebuffer setup should go towards the end anyway, but it *must* follow these test setups anyway. | |||
| 2010-01-11 | r300g: Avoid segfault on binding null RS state. | Corbin Simpson | |
| 2010-01-11 | llvmpipe: disable the all in/out test code for now | Brian Paul | |
| It's still faster not to try to special case the "all pixels are known to be inside the triangle" case. | |||
| 2010-01-11 | llvmpipe: move, update comments | Brian Paul | |
| 2010-01-11 | llvmpipe: refactor generate_fragment() code | Brian Paul | |
| This will make it easier to generate multiple versions of the fragment code per variant. | |||
| 2010-01-11 | llvmpipe: do the all-in test on the scalar c0 instead of vector c0 | Brian Paul | |
| This still isn't faster, but committing it for posterity. | |||
| 2010-01-11 | gallium: remove const qualifier from pipe_buffer argument in set_constant_buffer | Roland Scheidegger | |
| 2010-01-11 | llvmpipe: force constant interpolation of flatshade colors | Keith Whitwell | |
| Nice speedup for gears. | |||
| 2010-01-11 | llvmpipe: remove opencoded constant | Keith Whitwell | |
| 2010-01-11 | llvmpipe: remove scissor cliprect derived state | Keith Whitwell | |
| Was previously calculating the intersection of the scissor rectangle and the framebuffer dimensions. Rendering is already restricted to framebuffer dimensions by other means, so scissor testing (when implemented) can just use the scissor state directly. | |||
| 2010-01-10 | r300g: Don't bother testing for empty scissors. | Corbin Simpson | |
| If somebody goes through that much effort, they probably intended it. So humor them. :3 | |||
| 2010-01-10 | r300g: Atomize scissors. | Corbin Simpson | |
| Argfl. Some of this code is so questionable. | |||
| 2010-01-10 | r300g: Move TCL bypass switch to main context. | Corbin Simpson | |
| 2010-01-10 | r300g: Atomize viewport. | Corbin Simpson | |
| Goddammit, some of these hax are really annoying. | |||
| 2010-01-10 | r300g: Move ROPCNTL to the top of the emit order. | Corbin Simpson | |
| According to the docs, this decreases stalls, and indeed we get a tiny bit more glxgears from it. | |||
| 2010-01-10 | r300g: Don't avoid rewriting ZTOP. | Corbin Simpson | |
| Simplify the code. Added a comment to keep me from doing it again. | |||
 |
index : android-x86-mesa.git | |
| Androïd/x86 port of Mesa drivers | Hugues Hiegel |
| summaryrefslogtreecommitdiff |