summaryrefslogtreecommitdiff
path: root/src/gallium
AgeCommit message (Collapse)Author
2010-06-04llvmpipe: Do unswizzling in parallel when flushing for transfers.José Fonseca
2010-06-04st/egl: Fix compiler warnings.Chia-I Wu
Forgot to stage this chunk in last commit.
2010-06-04st/egl: Move sw screen creation to native helper.Chia-I Wu
The code is shared by ximage and gdi backend.
2010-06-04r600g: implement clear_{render_target, depth_stencil}, resource_copy_regionMarek Olšák
resource_copy_region uses a software fallback because it relies on texture samplers which haven't been implemented yet.
2010-06-04r600g: make blit functions static (private)Marek Olšák
2010-06-04r600g: use cast wrappersMarek Olšák
2010-06-03r600g: Move declaration before code.Vinson Lee
Fixes SCons build.
2010-06-03util: Remove unnecessary header.Vinson Lee
2010-06-03nv50: Remove unnecessary header.Vinson Lee
2010-06-03svga: Remove unnecessary header.Vinson Lee
2010-06-03llvmpipe: Fix uninitialized variable on non-debug builds.Vinson Lee
2010-06-03r600g: added code for depthtestingBas Nieuwenhuizen
- added code to configure the depth buffer and to set up depth testing.
2010-06-03r600g: adapt to interface changesRoland Scheidegger
r600g should be able to handle separate depth stencil clears as well. Also adapt surface_fill/copy stubs to new interface (not that it matters).
2010-06-03Merge branch 'gallium-newclear'Roland Scheidegger
Conflicts: src/gallium/state_trackers/python/p_context.i
2010-06-03gallium: rename clearRT / clearDS to clear_render_target / clear_depth_stencilRoland Scheidegger
more consistent with rest of gallium naming conventions. Also rename driver-internal names for these the same.
2010-06-01gallium: Don't depend directly on staging dir in Xorg template makefileJakob Bornecrantz
2010-06-03tgsi: we don't support indirect input/output registers in SSE codegen yetBrian Paul
Extend the check for indirect addressing of temp regs to include input/output regs. Fixes failure with piglit glsl-texcoord-array.shader_test test when using SSE codegen.
2010-06-03tgsi: whitespace cleanupBrian Paul
2010-06-03gallium: add interpolation parameter to simple shader functionsBrian Paul
This lets us specify linear interpolation instead of perspective interpolation for blit operations. Might be a bit faster.
2010-06-01gallium: Don't always copy the dri driver to the staging dirJakob Bornecrantz
So something in the build keept updating the timestamp on the staging directory causing us to always copy the library to it.
2010-06-01gallium: Create a Xorg driver template MakefileJakob Bornecrantz
2010-06-03util/u_debug: use MAX2 macroBrian Paul
2010-06-03llvmpipe: Number shaders/variants for more reproducible function names in ↵José Fonseca
the IR.
2010-06-03gallivm: Register the JIT engine with oprofile on debug/profile builds.José Fonseca
2010-06-03gallivm: Factor out the quad derivative code into a single place. Fix ddy.José Fonseca
For ddy it should be (bottom - top).
2010-06-03r300g: use util/u_debugJoakim Sindholt
2010-06-03gallium: silence all debug_named_value related warningsJoakim Sindholt
2010-06-03util/u_debug: add description field to debug_named_valueJoakim Sindholt
2010-06-03scons: Handle Python OSError exception when missing libdrm_radeon.Vinson Lee
Instead of crashing when libdrm_radeon is unavailable, the SCons build just does not build r600g.
2010-06-03r300g: set the minimum texture stride to 64 bytes on RS600, RS690, and RS740Marek Olšák
I hope I've done it right. No way to test it here...
2010-06-02gallivm: bump up LP_BUILD_FLOW_MAX_VARIABLESBrian Paul
Fixes failed assertion with piglit fbo-drawbuffers-fragcolor.c See fd.o bug 28358. With 8 color buffers + Z we need to declare 33 vars, at least.
2010-06-02llvmpipe: Fix bad logic.José Fonseca
Courtesy of Vinson and coverity.
2010-06-02llvmpipe: Remove unnecessary header.Vinson Lee
2010-06-02llvmpipe: Fix typo in assert.José Fonseca
2010-06-02llvmpipe: Per quad interpolation.José Fonseca
First interpolate the 4 quads upper left corners, then sub-interpolate each quad pixel. Do the perspective divide once per quad. Saves some muls and reciprocates. But doesn't seem to make a noticeable improvement. It make the code simpler and more compact, so commiting anyway.
2010-06-02llvmpipe: Minor cleanup to the interpolator.José Fonseca
2010-06-02gallivm: Add a lp_build_const_elem().José Fonseca
2010-06-02llvmpipe: Store often used LLVM types in the lp_build_context.José Fonseca
2010-06-02gallivm: Zero min_lod and max_lod when only one view is selected and min/mag ↵José Fonseca
filter are equal.
2010-06-02gallivm: Make lp_build_sample_wrap_int usage more accurate.José Fonseca
Fixes mesa texwrap demo border with GL_CLAMP.
2010-06-02gallivm: Don't use reciprocate in lp_build_sample_wrap_linear().José Fonseca
Always clamp the scaled coordinates.
2010-06-02gallivm: Avoid fp arithmetic in lp_build_sample_wrap_nearest().José Fonseca
2010-06-02gallivm: Don't use lp_build_sample_wrap_nearest()José Fonseca
Always clamp with scaled coordinates.
2010-06-02llvmpipe: Centralize all position interpolation in lp_bld_interp.c.José Fonseca
2010-06-02r300g: use the stored GEM domain in add_buffer/write_relocMarek Olšák
I admit blitting to/from GTT is actually enforced in this commit, texture transfers are now damned fast!
2010-06-02r300g: store the GEM domain in buffer and texture structsMarek Olšák
2010-06-02r300g: use r300_buffer_domain everywhereMarek Olšák
2010-06-02r300g: blit directly to/from GTT when transfering texturesMarek Olšák
2010-06-02r300g: let the driver determine the GEM domain for buffer_createMarek Olšák
2010-06-02r300g: make texture transfers more pipelinedMarek Olšák
Mostly based on Michel Dänzer's patch.