documentation on intel driver behavior?

[Michel Dänzer]

On Fri, 2008-02-29 at 10:18 -0800, Jesse Barnes wrote: 
> On Friday, February 29, 2008 12:59 am Michel Dänzer wrote:
> > On Wed, 2008-02-27 at 16:55 -0800, Eric Anholt wrote:
> > > For vblank syncing, we wait for vblank before emitting that copy. 
> > > Note that we're not doing anything about guaranteeing that that
> > > copy happens in any set amount of time.  We just assume that the
> > > hardware will get to it fast enough (and that we were woken up to
> > > do this soon enough) that no tearing occurs.
> >
> > Nice writeup Eric, I'd just like to add that with drm Git, the 3D
> > driver can also schedule buffer swaps to be executed by the DRM at an
> > arbitrary vertical blank period. This mechanism uses a tasklet to
> > mostly eliminate tearing even with several contexts running
> > concurrently.
> 
> Yeah, the tasklet stuff is really nice.  I wonder if it's something that 
> could be made into mostly generic code (ala vblank rework) so that 
> other drivers could use it?  I guess that would mean a generic drm 
> ioctl for it along with a chunk of generic code for handling the 
> tasklets; I'm not sure how much code it would save on the driver side, 
> but if it's significant the effort of porting it to the core may be 
> worth it.

Possibly, though I think it's probably not worth putting too much effort
into this before the DRI2 changes settle down.


-- 
Earthling Michel Dänzer           |          http://tungstengraphics.com
Libre software enthusiast         |          Debian, X and DRI developer