Improvements to Intel DDX for implementation of DRI2 OML_sync_control extensions.

Mario Kleiner mario.kleiner at
Wed Mar 3 20:24:38 PST 2010

They are against Jesse Barnes xf86-video-intel tree:

git:// master branch

The tree is in the same state as when i wrote the patches - they are  
following commit


Ideas to fix? I'm a git newbie.


On Mar 4, 2010, at 1:28 AM, Eric Anholt wrote:

> On Sun, 21 Feb 2010 18:45:46 +0100, Mario Kleiner  
> <mario.kleiner at> wrote:
>> This series of 2 patches fixes issues in the Intel DDX implementation
>> of I830DRI2ScheduleWaitMSC() and I830DRI2ScheduleSwap().
>> The previous implementation should mostly work in the simple
>> glXSwapBuffers() case, although it may reduce the maximum
>> swaprate to half the video refresh interval, e.g., max 30 fps
>> at a refresh of 60 Hz. It won't work correctly for the
>> glXSwapBuffersMscOML(target_msc, divisor, remainder) case if
>> divisor or remainder are non-zero or target_msc is not > 2
>> frames ahead of the current vblank count. It doesn't handle
>> the (msc % divisor) == remainder case correctly.
>> A similar problem exists for calls to glXWaitForMscOML().
>> These patches should fix above issues, or at least make
>> handling more robust.
>> Jesse Barnes already had a look at these patches, but due to
>> lack of suitable hardware, i haven't tested them myself, not
>> even if they compile cleanly.
>> Could somebody please give them some good testing and merge
>> them if appropriate?
> These don't seem to apply, and I can't quite work out what code  
> they're
> against

Mario Kleiner
Max Planck Institute for Biological Cybernetics
Spemannstr. 38
72076 Tuebingen

e-mail: mario.kleiner at
office: +49 (0)7071/601-1623
fax:    +49 (0)7071/601-616
"For a successful technology, reality must take precedence
over public relations, for Nature cannot be fooled."
(Richard Feynman)

More information about the xorg-devel mailing list