summaryrefslogtreecommitdiff
path: root/src
AgeCommit message (Collapse)Author
2010-01-25vega: adapt to blend changesRoland Scheidegger
2010-01-25xorg: adapt to blend changesRoland Scheidegger
2010-01-25gallium: fix cast for size calc of pipe_blend_stateRoland Scheidegger
2010-01-21docs: enhance blend documentationRoland Scheidegger
2010-01-21docs: clean up sampler descriptionRoland Scheidegger
2010-01-21gallium: adapt util code to blend changesRoland Scheidegger
2010-01-21gallium: new PIPE_CAP bits for the independent blend functionalityRoland Scheidegger
2010-01-20llvmpipe: adapt to per-rendertarget blend changesRoland Scheidegger
2010-01-20gallium: prepare for per-rendertarget blend enables, writemasks, blend funcsRoland Scheidegger
GL 3.0 (EXT_draw_buffers2) and other APIs allow independent blend enables and write masks per render target, ARB_draw_buffers_blend (and other APIs) also allow independent blend functions. Things like dithering, logic ops however are not extended to be per rendertarget, that might be conceptually possible however it doesn't look like any API wants to expose this.
2010-01-20r200: fix typo in 76cf2618327a7f008dcfd0d91d64d6d9e01f9a9cAlex Deucher
noticed by Maciej on IRC.
2010-01-20st/egl_g3d: Add support for driver probe.Chia-I Wu
Hook the drv->Probe callback to allow the drivers to probe displays. The loading of state trackers is delayed to avoid unnecessary loading.
2010-01-20st/egl_g3d: Add support for probe to native displays.Chia-I Wu
The functions can be used to check if a display is supported without creating a struct native_display. It uses a probe object that can be shared across drivers.
2010-01-20egl: Add a simple cache for driver probe.Chia-I Wu
In current design, multiple drivers will probe the same display and the best driver is determined. The cache can be used by the drivers to store and share the probed data.
2010-01-20egl: Improve driver matching.Chia-I Wu
Make drv->Probe return a score so that the matching can be done by finding the driver with the highest score.
2010-01-20egl: Rename _EGL_PLATFORM_X to _EGL_PLATFORM_POSIX.Chia-I Wu
The macro is used to determine if dlfcn.h or dirent.h is available. POSIX is a better name than X in such case.
2010-01-20egl: Remove unused driver and display functions.Chia-I Wu
Remove _eglPreloadDriver, _eglLookupDriver, and _eglSplitDisplayString.
2010-01-20egl: Use _eglPreloadDrivers.Chia-I Wu
Replace the use of _eglPreloadDriver by _eglPreloadDrivers. The latter supports EGL_DISPLAY which have a better chance to "just work".
2010-01-20egl: Add _eglPreloadDrivers.Chia-I Wu
It can be used to load the user driver specified by EGL_DRIVER, or a set of drivers specified by EGL_DISPLAY, or the default driver, and in that order.
2010-01-20st/egl_g3d: Do not reuse the pipe transfer in ximage.Chia-I Wu
A pipe transfer is supposed to be temporary. It should be created before X*PutImage and destroyed afterwards.
2010-01-19Remove hardcoded -Wall from Radeon DRI makefilesAlan Coopersmith
Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com> Signed-off-by: Brian Paul <brianp@vmware.com>
2010-01-19r300g: Move GB_AA_CONFIG to its eventual new home.Corbin Simpson
2010-01-19r300g: Make invariant state into an atom.Corbin Simpson
2010-01-20Merge remote branch 'origin/opengl-es-v2'Chia-I Wu
2010-01-19util/blit: document params to util_blit_pixels_writemask()Brian Paul
2010-01-19mesa: added comment about future FBO formatsBrian Paul
2010-01-19st/mesa: enable EXT_framebuffer_multisampleMarek Olšák
Signed-off-by: Brian Paul <brianp@vmware.com>
2010-01-19r100/r200: fix Y coord flipping in accelerated blitsAlex Deucher
2010-01-19r100/r200: fix dst pitch in blit codeAlex Deucher
2010-01-19r200: add missing symlink to radeon_tex_copy.cAlex Deucher
2010-01-20r100/r200: align to pitch updates in blit interfaceMaciej Cencora
2010-01-19r200: use common glCopyTex(Sub)Image codeMaciej Cencora
2010-01-19r100: use common glCopyTex(Sub)Image codeMaciej Cencora
2010-01-19radeon/r300/r600: share common glCopyTex(Sub)Image codeMaciej Cencora
2010-01-19r600: align to r300 changes in the blit codeMaciej Cencora
Pitch here means aligned width, not aligned width * bpp.
2010-01-19r300/r600: move some bo offsets checking to blit codeMaciej Cencora
In preperation for texcopy code sharing.
2010-01-19r600: prepare for some code sharingMaciej Cencora
2010-01-19r300: check if blitting for given format is supported earlierMaciej Cencora
Prevents failing assertions at later stage.
2010-01-19r300: use nearest texture filtering for accelerated blitsMaciej Cencora
2010-01-19r300: fix Y coord flipping in accelerated blitsMaciej Cencora
2010-01-19radeon: use mesa provided _mesa_tex_target_to_face functionMaciej Cencora
2010-01-19r300: prepare for texcopy code sharingMaciej Cencora
2010-01-19radeon: add blit function to vtblMaciej Cencora
2010-01-19intel: Remove dead note_fence vtbl hook.Eric Anholt
2010-01-19i965: Improve the hashing of brw_state_cache keys to include the cache_id.Eric Anholt
No measurable difference on cairoperf.
2010-01-19i965: Remove obsolete comment about the state atoms.Eric Anholt
2010-01-19i965: Upload as many VS constants as possible through the push constants.Eric Anholt
The pull constants require sending out to an overworked shared unit and waiting for a response, while push constants are nicely loaded in for us at thread dispatch time. By putting things we access in every VS invocation there, ETQW performance improved by 2.5% +/- 1.6% (n=6).
2010-01-19i965: Allow for variable-sized auxdata in the state cache.Eric Anholt
Everything has been constant-sized until now, but constant buffer handling changes will make us want some additional variable sized array.
2010-01-19intel: Use the new DRI2 flush invalidate entrypoint to signal frame done.Eric Anholt
Previously for frame throttling we would wait on the first batch after a swap before emitting another swap, because we had no hook after a swap was emitted. This meant that if an app managed to squeeze everything it for a frame had into one batch, it would lock-step with the GPU. With the swapbuffers changes, we now have the entrypoint we want. This takes the WoW intro screen from 25% GPU idle and visibly jerky to 4-5% GPU idle and rather smooth. Other apps such as OpenArena have run into this problem as well.
2010-01-19Sun compilers now support some gcc __attribute__ valuesAlan Coopersmith
Sun cc 5.9 and later (__SUNPRO_C >= 0x590) support __attribute__ calls for aligned, always_inline, noinline, pure, const, and malloc. This commit includes updates to files that were regenerated by gl_XML.py after adding the __SUNPRO_C checks to it Signed-off-by: Alan Coopersmith <alan.coopersmith@sun.com> Signed-off-by: Brian Paul <brianp@vmware.com>
2010-01-19softpipe: remove redundant assignmentsBrian Paul
We initialized width/height/depth from pt->width0/height0/depth0 above.
generated by cgit v1.2.3 (git 2.39.1) at 2024-11-18 18:53:00 +0000