summaryrefslogtreecommitdiffstats
Commit message (Collapse)AuthorAgeFilesLines
* i965: Remove hint_gs_always and resulting dead codeIan Romanick2011-04-113-76/+13
| | | | | Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* intel: Fix ROUND_DOWN_TO macroIan Romanick2011-04-111-3/+27
| | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | | Previously the macro would (ALIGN(value - alignment - 1, alignment)). At the very least, this was missing parenthesis around "alignment - 1". As a result, if value was already aligned, it would be reduced by alignment. Condisder: x = ROUND_DOWN_TO(256, 128); This becomes: x = ALIGN(256 - 128 - 1, 128); Or: x = ALIGN(127, 128); Which becomes: x = 128; This macro is currently only used in brw_state_batch (brw_state_batch.c). It looks like the original version of this macro would just use too much space in the batch buffer. It's possible, but not at all clear to me from the code, that the original behavior is actually desired. In any case, this patch does not cause any piglit regressions on my Ironlake system. I also think that ALIGN_FLOOR would be a better name for this macro, but ROUND_DOWN_TO matches rounddown in the Linux kernel. Reviewed-by: Eric Anholt <[email protected]> Reviewed-by: Keith Whitwell <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* glsl: Make GL_ARB_shader_stencil_export enable block be similar to other blocksIan Romanick2011-04-111-7/+7
| | | | | Tested-by: Brian Paul <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* glsl: Only let a shader enable GL_ARB_draw_instanced if the driver supports itIan Romanick2011-04-111-6/+5
| | | | | | | | Also make the GL_ARB_draw_instanced block follow the same pattern as the other blocks. Tested-by: Brian Paul <[email protected]> Reviewed-by: Kenneth Graunke <[email protected]>
* i965: Move the SF VP from state caching to state streaming.Eric Anholt2011-04-113-8/+17
| | | | | | | | | This is a 49.6% +/- 2.0% (n=9, IPS outlier removed) performance improvement for the hacked-up-for-cache-misses scissor-many, and no statistically significant performance difference for the hacked-up-for-cache-hits version (n=9, IPS outlier removed). No statistically significant performance difference from ETQW (n=5) from these last two commits.
* i965: Change the SF unit from state caching to state streaming.Eric Anholt2011-04-113-107/+45
| | | | | | | | | This is a 28.1% +/- 1.4% (n=10) performance improvement for the hacked-up-for-cache-misses scissor-many (n=10), and no statistically significant wall-time performance difference for the hacked-up-for-cache-hits version (n=9, first outlier in each removed since IPS was warming up. User time increased by about 4.7%, but kernel time decreased equivalently).
* i965: Turn SF unit and viewport structs into pointers to prep for streaming.Eric Anholt2011-04-111-69/+70
| | | | I wanted to separate this mechanical change from the actual work.
* u_vbuf_mgr: fixed vbo max_index calculationpepp2011-04-111-1/+1
| | | | Signed-off-by: Marek Olšák <[email protected]>
* r600g: handle PIPE_CAP_FRAGMENT_COLOR_CLAMP_CONTROL in get_paramMarek Olšák2011-04-111-0/+1
|
* glx: Only remove the glx_display from the list after it's destroyed.Henri Verbeet2011-04-101-3/+8
| | | | Signed-off-by: Henri Verbeet <[email protected]>
* glx: Destroy dri2Hash on DRI2 display destruction.Henri Verbeet2011-04-101-0/+3
| | | | Signed-off-by: Henri Verbeet <[email protected]>
* mesa: Also update the color draw buffer if it's explicitly set to GL_NONE.Henri Verbeet2011-04-101-1/+1
| | | | | | NOTE: This is a candidate for the 7.10 branch. Signed-off-by: Henri Verbeet <[email protected]>
* docs: fix typos in wayland specNicolas Kaiser2011-04-101-3/+3
| | | | | Signed-off-by: Nicolas Kaiser <[email protected]> Signed-off-by: Brian Paul <[email protected]>
* docs: fix typos in Mesa specsNicolas Kaiser2011-04-1010-22/+22
| | | | | Signed-off-by: Nicolas Kaiser <[email protected]> Signed-off-by: Brian Paul <[email protected]>
* nv50,nvc0: silence shader debug outputChristoph Bumiller2011-04-1016-109/+85
|
* nvc0: make the card permit separate colour outputsChristoph Bumiller2011-04-102-14/+4
|
* nv50-nvc0: make use of COLOR_MASK,BLEND_ENABLE_COMMONChristoph Bumiller2011-04-109-31/+32
|
* nvc0: shrink CSOs a littleChristoph Bumiller2011-04-104-29/+25
|
* nv50,nvc0: identify some line rasterization stateChristoph Bumiller2011-04-106-6/+21
| | | | Fixes width of non-smooth (aliased) lines on nvc0.
* nvc0: force perspective correction hint to nicestChristoph Bumiller2011-04-102-0/+7
| | | | | | | | | Fixes issues in e.g. nexuiz (desertfactory) or supertuxkart that look like lighting bugs. They're not visible with the software rasterizers because their notion of linear interpolation seems to be different from that of nv50/nvc0.
* nv50,nvc0: don't insert instructions after the block terminatorChristoph Bumiller2011-04-105-7/+25
|
* Make st_pipe_vertex_format return type in st_draw.h match st_draw.cAlan Coopersmith2011-04-081-1/+1
| | | | | | | | | Fixes compiler error from Sun compilers: "state_tracker/st_draw.c", line 185: identifier redeclared: st_pipe_vertex_format current : function(unsigned int, unsigned int, unsigned int, unsigned char) returning enum pipe_format previous: function(unsigned int, unsigned int, unsigned int, unsigned char) returning unsigned int : "state_tracker/st_draw.h", line 73 Signed-off-by: Alan Coopersmith <[email protected]>
* Fix GET_PROGRAM_NAME() on Solaris to not try to modify a read-only stringAlan Coopersmith2011-04-081-1/+19
| | | | Signed-off-by: Alan Coopersmith <[email protected]>
* Revert "r300g: handle DISCARD_WHOLE_RESOURCE for buffers"Marek Olšák2011-04-081-25/+10
| | | | | | | | | | | | | This reverts commit 437c748bf5072d2bded77a00c74c51cdb8b510e5. The commit is wrong for several reasons. One of them is when we grab a new buffer, we should update all the states it is bound in, including all parallel contexts. I don't think this is even doable. The correct solution would be upload data via a temporary buffer and do resource_copy_region to the original one. https://bugs.freedesktop.org/show_bug.cgi?id=36088
* r300g: fix another crash with hyperzMarek Olšák2011-04-082-22/+2
| | | | https://bugs.freedesktop.org/show_bug.cgi?id=36086
* Disable direct rendering on CygwinJon TURNEY2011-04-081-0/+14
| | | | | | | | | | | | Add Cygwin platform-specific settings and drivers to build for dri driver: - by default, disable direct rendering. - if direct rendering is enabled, the swrast dridriver is the only one it's sensible to try to build (this doesn't work at the moment as additional patches are required to build a libGL which can load just swrast without the DRM headers, even though there's no actual functional dependency) Signed-off-by: Jon TURNEY <[email protected]> Reviewed-by: Julien Cristau <[email protected]>
* Disable direct rendering on GNU/Hurdnobled2011-04-081-2/+11
| | | | | | | | The Hurd kernel doesn't have DRM yet. Signed-off-by: Jon TURNEY <[email protected]> Reviewed-by: Julien Cristau <[email protected]> Reviewed-By: Jakob Bornecrantz <[email protected]>
* Only require libdrm if direct rendering is actually enabled.Samuel Thibault2011-04-081-11/+18
| | | | | | | | | | | Fix build when configured --with-driver=dri --disable-driglx-direct on targets without drm e.g. GNU/Hurd and Cygwin Based on the Debian patch file '05_hurd-ftbfs.diff' by Samuel Thibault. Signed-off-by: Jon TURNEY <[email protected]> Reviewed-by: Julien Cristau <[email protected]> Reviewed-By: Jakob Bornecrantz <[email protected]>
* r300g: fix crash when adding a new relocationMarek Olšák2011-04-081-1/+2
| | | | | | | No idea why it's so hard to reproduce. Broken with: c35572352e3e92683988ee8d151b47f4190d62f9 Thanks to Toni Spets for assistance.
* r300g: emit US_FORMAT on R520 onlyMarek Olšák2011-04-082-2/+2
|
* i965/fs: Remove broken optimization for live intervals in loops.Eric Anholt2011-04-081-4/+2
| | | | | | | | | The theory here was to detect a temporary variable used within a loop, and avoid considering it live across the entire loop. However, it was overeager and failed when the first definition of the variable appeared within the loop but was only conditionally defined. Fixes glsl-fs-loop-redundant-condition.
* u_vbuf_mgr: fix crash with unsupported vertex format in hw VBOsMarek Olšák2011-04-082-2/+6
|
* r300g: fix crash with hyperzMarek Olšák2011-04-081-6/+4
| | | | r300_update_hyperz_state is no longer required to be called here.
* r300g: Properly take the sampler view's first_level into account when sampling.Henri Verbeet2011-04-081-15/+21
| | | | | Signed-off-by: Brian Paul <[email protected]> Signed-off-by: Marek Olšák <[email protected]>
* llvmpipe: Take the sampler view's first_level into account when sampling.Fabian Bieler2011-04-0814-33/+78
|
* softpipe: Take the sampler view's first_level into account when sampling.Henri Verbeet2011-04-081-16/+16
| | | | Signed-off-by: Brian Paul <[email protected]>
* st/mesa: Clamp min_lod to the max miplevel as well.Henri Verbeet2011-04-081-1/+3
| | | | | | | | Otherwise min_lod can potentially be larger than the clamped max_lod. The code that follows will swap min_lod and max_lod in that case, resulting in a max_lod larger than MAX_LEVEL. Signed-off-by: Brian Paul <[email protected]>
* st/mesa: Set samplers views' first_level.Henri Verbeet2011-04-082-7/+6
| | | | | | | | Base level and min LOD aren't equivalent. In particular, min LOD has no effect on image array selection for magnification and non-mipmapped minification. Signed-off-by: Brian Paul <[email protected]>
* r300g: do not advertise R16F and RG16F on DRM < 2.8.0Marek Olšák2011-04-081-2/+5
|
* r300g: fix texturing with sizes > 2048 on r500Marek Olšák2011-04-087-8/+52
|
* r300g: add 'radeon: ' prefix to error messages in winsysMarek Olšák2011-04-082-5/+5
|
* mesa: fix dstRowDiff computation in RGTC texstore functionsMarek Olšák2011-04-081-4/+4
| | | | | Copied from libtxc_dxtn, this fixes NPOT RGTC1 textures with r300g. I also did the same for RGTC2.
* egl_dri2.c: Open drm device with O_CLOEXECKristian Høgsberg2011-04-071-1/+1
| | | | Not a good fd to leak.
* r600g: Unreference framebuffer state on context destruction.Henri Verbeet2011-04-071-0/+1
| | | | Signed-off-by: Henri Verbeet <[email protected]>
* r600g: Use symbolic names for NUMBER_TYPE.Henri Verbeet2011-04-072-5/+5
| | | | Signed-off-by: Henri Verbeet <[email protected]>
* docs: replace llvmpipe/README with docs/llvmpipe.htmlBrian Paul2011-04-072-34/+101
|
* llvmpipe: document issue with LLVM 2.8 and earlier with AVXBrian Paul2011-04-071-0/+4
|
* util: add cpu detection for sse 4.2 and avxBrian Paul2011-04-072-0/+6
|
* util: simplify bit shifting in util_cpu_detect()Brian Paul2011-04-071-11/+11
|
* svga: another tweak to adjust_x term (use -0.5)Brian Paul2011-04-071-1/+1
|
generated by cgit v1.2.3 (git 2.39.1) at 2024-12-24 13:19:51 +0000