[PATCH] dri2: Pass AsyncSwap [vblank_mode=0] requests to the drivers

[Eric Anholt]

On Wed, 8 Jun 2011 12:03:04 -0700, Jesse Barnes <jbarnes at virtuousgeek.org> wrote:
> On Wed,  8 Jun 2011 19:06:11 +0100
> Chris Wilson <chris at chris-wilson.co.uk> wrote:
> 
> > Currently, the midlayer dri2 code intercepts vblank_mode=0 SwapBuffers
> > and converts it to a CopyRegion request. This prevents the backend from
> > doing anything meaningful in this case and typically ends up being
> > vsync'ed since the drivers cannot distinguish it from a regular
> > CopyRegion request.
> > 
> > Signed-off-by: Chris Wilson <chris at chris-wilson.co.uk>
> > Cc: Jesse Barnes <jbarnes at virtuousgeek.org>
> > Cc: Kristian Høgsberg <krh at bitplanet.net>
> > ---
> > 
> > Sigh... I broke the patch with a last minute name change. Please pretend
> > you never saw the previous patch.
> 
> In hindsight, it would have been better if DRI2 never had a CopyRegion
> hook and just let the driver decide what to do given all the params
> (i.e. a very generic Swap hook).
> 
> But we don't have that and this looks like a nice addition.  It would
> be good to add a blurb to the header file about the hook though, since
> it could be easily confused with a hook that's supposed to schedule
> something and return immediately or something, rather than a "swap now,
> tearing be damned" hook.

\o/ making vblank handling more predictable.

So, I'm wondering if we want to just drop the no-tearing support in
CopyRegion (aka MESA_copy_sub_buffer) that we snuck in back when we
didn't really have other no-tearing support.  It sounds like the clutter
compositors are using another swap control to sync to vblank before
dispatching their subbuffer blits.  Do we know about compiz?

(The latency between vblank signal from OML_sync_control to client
wakeup to server wakeup to render engine wakeup might make us miss the
vblank.  If we do, it means dropping from 60fps to 30fps, though, which
is perhaps even nastier than a tear?)
-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 197 bytes
Desc: not available
URL: <http://lists.x.org/archives/xorg-devel/attachments/20110608/33b2ec61/attachment.pgp>