Age | Commit message (Collapse) | Author | |
---|---|---|---|
2009-08-25 | Wait after mapping buffer | Thomas White | |
2009-08-25 | Restore composite hooks | Thomas White | |
2009-08-25 | Fix DRI2 registration with recent X servers | Thomas White | |
2009-08-25 | Remove unnecessary variables | Thomas White | |
2009-08-25 | Fix headers for recent X.org | Thomas White | |
I think this breaks compatability with older versions. Some kind of conditional thing should be done. | |||
2009-08-22 | Fix GlamkKernelModesettingAvailable() so that it works | Thomas White | |
2009-08-21 | Don't unmap on FinishAccess, and return as quickly as possible on re-map | Thomas White | |
Combined with mixed mode pixmaps, this brings the speed up to about where it should be. | |||
2009-08-21 | Fix DRI2InfoRec problems | Thomas White | |
2009-08-21 | Typo | Thomas White | |
Thanks for Martin Jansa (JaMa) for spotting this. | |||
2009-08-21 | Strip out some RAC and loader stuff | Thomas White | |
This is necessary to keep up with the latest Xorg-server developments. It works with the latest Git master of the server, which we want to use for the great speedups arising from mixed mode EXA pixmap handling. Please let me know if it breaks on older versions... | |||
2009-08-21 | Handle the new DRI protocol as well | Thomas White | |
2009-08-21 | Use $PKG_CONFIG instead of pkg-config | Thomas White | |
This un-breaks cross-compilation in certain cases. | |||
2009-08-20 | EXA synchronisation stuff | Thomas White | |
2009-08-18 | Don't call xf86_reload_cursors() | Thomas White | |
It's not necessary for our software cursor (I think..?) | |||
2009-08-18 | Make the buffer size for very small pixmaps sane | Thomas White | |
2009-08-18 | Reallocate GEM objects when necessary | Thomas White | |
2009-08-12 | Tidy up headers, formatting and unused code | Thomas White | |
2009-08-09 | Use libDRM functions instead of bom->funcs etc | Thomas White | |
2009-08-09 | Remove upload/download EXA functions - can't accelerate these | Thomas White | |
2009-08-09 | Dispatch command buffer when cache is full | Thomas White | |
2009-08-09 | Small tidy up | Thomas White | |
2009-08-06 | Fix command queue submission | Thomas White | |
This reworks the command queue caching and submission so that it works, and makes it a lot more maintainable in the process. | |||
2009-07-30 | Use proper types for cmdq submission | Thomas White | |
2009-07-29 | Add a missing newline | Thomas White | |
2009-07-28 | Initialisation no longer segfaults... | Thomas White | |
2009-07-28 | De-crashify a few things | Thomas White | |
2009-07-26 | Cannot accelerate these | Thomas White | |
2009-07-26 | It all compiles now... | Thomas White | |
2009-07-26 | Make everything up to glamo-kms-exa.c compile | Thomas White | |
2009-07-26 | Remote crtc.c and output.c | Thomas White | |
2009-07-23 | Just backing up progress | Thomas White | |
2009-07-23 | Make glamo-kms-driver.c compile | Thomas White | |
2009-07-22 | ExaClose->GlamoKMSExaClose | Thomas White | |
2009-07-22 | Drop in most of the rest of the code | Thomas White | |
2009-07-22 | Just backing up progress | Thomas White | |
2009-07-21 | Makefile and initialisation stuff | Thomas White | |
2009-07-20 | Initial KMS stuff | Thomas White | |
2009-07-07 | remove useless symbol lists. | Lars-Peter Clausen | |
2009-07-05 | With the recent changes this should not be necessary anymore. | Lars-Peter Clausen | |
Revert "Don't wait when disabling the cmdq. it causes a wsod for some users." This reverts commit 25c4b0e80e93e04e6f7d4b8bca6d007fb9de6da8. | |||
2009-07-05 | Final engine ioctls fixup. | Lars-Peter Clausen | |
2009-07-05 | s,HAS_ENGINE_IOCTLS,HAVE_ENGINE_IOCTLS,g | Lars-Peter Clausen | |
2009-07-05 | Implement support for engine ioctls. The glamo Xorg driver should only touch | Lars-Peter Clausen | |
cmdq or 2d registers, so we enable/disable engines through framebuffer ioctls. This needs support from the kernel side. | |||
2009-07-04 | Don't disable the divider for all memory clocks when disabling the cmdq engine. | Lars-Peter Clausen | |
Other memory clocks might be in use and disabling the divider will cause a WSOD. | |||
2009-05-09 | Don't wait when disabling the cmdq. it causes a wsod for some users. | Lars-Peter Clausen | |
thanks to lindi for tracking it down. | |||
2009-05-09 | Only reset those bits which we actually change. | Lars-Peter Clausen | |
2009-05-05 | Only waiting for the cmdq engine to be finished when dispatching the queue | Lars-Peter Clausen | |
causes visual artifactes sometimes. Waiting for all engines to be finished seems to fix the issue. | |||
2009-05-05 | Sync graphics pipeline before uploading/downloading data to/from the graphics | Lars-Peter Clausen | |
memory. | |||
2009-05-04 | Cleanup: Replace pix->devKind with exaGetPixmapPitch(pix) | Lars-Peter Clausen | |
2009-05-04 | Don't try to accelerate 8bit solid fills, because in it's current shape it does | Lars-Peter Clausen | |
not work. | |||
2009-05-03 | Remove autogenerated file. | Lars-Peter Clausen | |