summaryrefslogtreecommitdiff
path: root/src/mesa/drivers/dri
AgeCommit message (Collapse)Author
2009-05-10Merge commit 'origin/master' into radeon-rewriteJerome Glisse
Conflicts: src/mesa/drivers/dri/r200/r200_state.c src/mesa/drivers/dri/r300/r300_context.h src/mesa/drivers/dri/r300/r300_fragprog.c src/mesa/drivers/dri/r300/r300_state.c src/mesa/drivers/dri/r300/r300_texmem.c src/mesa/drivers/dri/r300/r300_texstate.c src/mesa/drivers/dri/r300/r500_fragprog.c src/mesa/drivers/dri/radeon/radeon_screen.c src/mesa/drivers/dri/radeon/radeon_state.c
2009-05-09radeon: don't include cs uncondionalyJerome Glisse
2009-05-08i965: improve debug loggingRobert Ellison
Looking for memory leaks that were causing crashes in my environment in a situation where valgrind would not work, I ended up improving the i965 debug traces so I could better see where the memory was being allocated and where it was going, in the regions and miptrees code, and in the state caches. These traces were specific enough that external scripts could determine what elements were not being released, and where the memory leaks were. I also ended up creating my own backtrace code in intel_regions.c, to determine exactly where regions were being allocated and for what, since valgrind wasn't working. Because it was useful, I left it in, but disabled and compiled out. It can be activated by changing a flag at the top of the file.
2009-05-08i965: fix memory leak in context/renderbuffer region managementRobert Ellison
A temporary change to the intelMakeCurrent() function to make it work with frame buffer objects causes the static regions associated with the context (the front_region, back_region, and depth_region) to take on an additional reference, with no corresponding release. This causes a memory leak if a program repeatedly creates and destroys contexts. The fix is the corresponding hack, to unreference these regions when the context is deleted, but only if the framebuffer objects are still present and the same regions are still referenced within. Both sets of code have comment blocks referring to each other.
2009-05-08i965: fix segfault on low memory conditionsRobert Ellison
When out of memory (in at least one case, triggered by a longrunning memory leak), this code will segfault and crash. By checking for the out-of-memory condition, the system can continue, and will report the out-of-memory error later, a much preferable outcome.
2009-05-08intel: Add a metaops version of glGenerateMipmapEXT/SGIS_generate_mipmaps.Eric Anholt
In addition to being HW accelerated, it avoids the incorrect (black) rendering of the mipmaps that SW was doing in fbo-generatemipmap. Improves the performance of the mipmap generation and drawing in fbo-generatemipmap by 30%.
2009-05-08intel: Put the constant texcoords used in metaops into a vbo.Eric Anholt
Make this be its own function for setup/teardown of the binding of these texcoords. No performance difference in the engine demo (I just felt dirty not using a VBO for this), and I think it should be more resilient to interference from current GL state.
2009-05-08i965: const qualifiersBrian Paul
2009-05-08i965: don't use GRF regs 126,127 for WM programsBrian Paul
They seem to be used for something else and using them for shader temps seems to lead to GPU lock-ups. Call _mesa_warning() when we run out of temps. Also, clean up some debug code.
2009-05-08radeon: IRQ always enabled in DRI2 path no need to query kernel for itJerome Glisse
2009-05-07r300: add missing texture formatMaciej Cencora
Fixes piglit/texdepth
2009-05-07r300: moar cleanupMaciej Cencora
- remove unused r300UpdateClipPlanes function - move reg definition to r300_reg.h - remove incorrect forward definition of tnl_UpdateFixedFunctionProgram and add proper #include - remove unreachable code
2009-05-07r300: fix compiler warningsMaciej Cencora
2009-05-07i965: relAddr local var (to make debug/test a little easier)Brian Paul
2009-05-06i965: Remove bad constant buffer constant-reg-already-loaded optimization.Eric Anholt
Thanks to branching, the state of c->current_const[i].index at the point of emitting constant loads for this instruction may not match the actual constant currently loaded in the reg at runtime. Fixes a regression in my GLSL program for idr's class since b58b3a786aa38dcc9d72144c2cc691151e46e3d5.
2009-05-06intel: Unmap buffers if needed at DeleteBuffer time.Eric Anholt
This fixes a crash in glean's pbo test, which tripped over the assert when a context was destroyed while a buffer was still mapped (Mesa doesn't call UnmapBuffer in that case). Regression in c6bde8873fbda6d8467600b7491d8543c75b0509
2009-05-06When clearing the stencil buffer, don't use a two-sided stencilOwen W. Taylor
In radeon_clear_tris(), when clearing the stencil buffer, pass GL_FRONT_AND_BACK to _mesa_StencilFuncSeparate(), to avoid triggering a software fallback on r300 and below. https://bugs.freedesktop.org/show_bug.cgi?id=21601
2009-05-06i965: Remove the forced lack of caching for renderbuffer surface state.Eric Anholt
This snuck in with the multi-draw-buffers commit, and is a major penalty to performance. It doesn't appear to be required, as the only dependency the surface BO has is on the state key (and if there's some other dependency, it should just be in the key). This brings openarena performance up to almost 2% faster than Mesa 7.4.
2009-05-06i965: Remove _NEW_PROGRAM from brw_wm_surfaces setup dependencies.Eric Anholt
This was a leftover from the brw_wm_constant_buffer change.
2009-05-06i965: Split WM constant buffer update from other WM surfaces.Eric Anholt
This can avoid re-uploading constant data when it isn't necessary, and is a step towards not updating other surfaces just because constants change. It also brings the upload of the constant buffer next to the creation. This brings openarena performance up another 4%, to 91% of the Mesa 7.4 branch.
2009-05-06i965: Disentangle VS constant surface state from WM surface state.Eric Anholt
Also, only create VS surface state if there's a VS constant buffer to be uploaded, and set the contents of the buffer at the same time as creation.
2009-05-06i965: Don't create constant buffers if they won't be used.Eric Anholt
Really, the creation and upload of constants should be in the same place, since they should only happen together, and a state flag should be triggered by them so that we don't thrash state around so much for just updating constants. But this still recovers openarena performance by another 19%, leaving us 16% behind Mesa 7.4 branch.
2009-05-06radeon: hopefully fixup radeon cube state emission for kmsDave Airlie
2009-05-06r100/r200: try and allocate miptree correct for hw.Dave Airlie
This doesn't make things worse but according to sroland it is how the GPU hw expects things on the r100/r200
2009-05-06r200: fix cubic emission.Dave Airlie
Still doesn't fix cubemaps, I really missed the whole drmsupports thing when testing this all originally
2009-05-05r200: fix some cube map issuesRoland Scheidegger
remove the r100-ism of swapping cube faces which doesn't apply to r200, and also use precalculated offsets. Note that cube textures will still not work on r100 and r200 since mipmap layout is level-first order (for r300) whereas r100/r200 require face-first (and possibly also 2k alignment for face at least with tiling).
2009-05-05r200: fix another section size mismatchRoland Scheidegger
2009-05-05r200: fix CS section size mismatch (bug 21565)Roland Scheidegger
2009-05-05radeon/r200: enable all the optional drm support bitsDave Airlie
2009-05-04r300: set proper texture row alignment for IGP chipsMaciej Cencora
Looks like r400 based IGP chips require 64 byte alignment
2009-05-01mesa: in glReadBufer() set _NEW_BUFFERS, not _NEW_PIXELBrian Paul
Since GL_READ_BUFFER is historically part of the gl_pixel_attrib group it made sense to signal changes with _NEW_PIXEL. But now with FBOs it's also part of the framebuffer state. Now _NEW_PIXEL strictly indicates pixels transfer state changes. This change avoids framebuffer state validation when any random bit of pixel-transfer state is set. DRI drivers updated too: don't check _NEW_COLOR when updating framebuffer state. I think that was just copied from the Xlib driver because we care about dither enable/disable state there.
2009-05-01Test either GL_FRONT_LEFT or GL_FRONT for front-buffer renderingIan Romanick
For non-stereo visuals, which is all we support, we treat GL_FRONT_LEFT as GL_FRONT. However, they are technically different, and they have different enum values. Test for either one to determine if we're in front-buffer rendering mode. This fix was suggested by Pierre Willenbrock. Signed-off-by: Ian Romanick <ian.d.romanick@intel.com>
2009-05-01Merge branch 'const-buffer-changes'Brian Paul
Conflicts: src/mesa/drivers/dri/i965/brw_curbe.c src/mesa/drivers/dri/i965/brw_vs_emit.c src/mesa/drivers/dri/i965/brw_wm_glsl.c
2009-04-30r300: Increase reference count of texture objects referenced by current state.Michel Dänzer
Fixes a use-after-free reported in http://bugs.freedesktop.org/show_bug.cgi?id=20539, so this possibly fixes that bug. It has been confirmed to fix http://bugs.freedesktop.org/show_bug.cgi?id=17895 .
2009-04-28R300: add quadpipe overridesAlex Deucher
RV410 SE chips only have 1 quadpipe. Also, handle other R300 chip with quadpipe override.
2009-04-28R300: add quadpipe overridesAlex Deucher
RV410 SE chips only have 1 quadpipe. Also, handle other R300 chip with quadpipe override
2009-04-28i965: avoid segfault in intel_update_renderbuffers() if using DRI1Brian Paul
2009-04-28radeno: cleanup the startup path furtherDave Airlie
2009-04-28radeon: further cleanup dri1 screen initDave Airlie
2009-04-28r300: remove unused debugging in set tex buffer pathsDave Airlie
2009-04-28radeon: remove kernel mm, dri2 path takes care of itDave Airlie
2009-04-27r300: fix fragment program limitsMaciej Cencora
2009-04-27r300: fallback only if stencil test is enabledMaciej Cencora
2009-04-27r300: do front/back color selection in HW for software TCL pathMaciej Cencora
2009-04-27r300: fix valgrind warningsMaciej Cencora
2009-04-27r300: fix point size clamping when point is not attenuatedMaciej Cencora
2009-04-27i965: #include prog_print.h to silence warningBrian Paul
2009-04-27i965: only upload constant buffer data when we actually need the const bufferBrian Paul
Make the use_const_buffer field per-program and only call the code which updates the constant buffer's data if the flag is set. This should undo the perf regression from 20f3497e4b6756e330f7b3f54e8acaa1d6c92052 (cherry picked from master, commit dc9705d12d162ba6d087eb762e315de9f97bc456)
2009-04-27i965: only upload constant buffer data when we actually need the const bufferBrian Paul
Make the use_const_buffer field per-program and only call the code which updates the constant buffer's data if the flag is set. This should undo the perf regression from 20f3497e4b6756e330f7b3f54e8acaa1d6c92052
2009-04-27r300: fix performance regressionMaciej Cencora
This performance regression on non TCL hw was introduced by ed4c6cbe017b4e8bacb7e012d4baaf77a20a2c33. This patch depends on "r300: always route 4 texcoord components to RS" and "r300: add point attenuation stage for TCL fallbacks".