On Friday, October 26, 2007 1:14 am Michel Dänzer wrote:
> On Thu, 2007-10-25 at 10:25 -0700, Jesse Barnes wrote:
> > On Thursday, October 25, 2007 2:02 am Michel Dänzer wrote:
> > > > It still has some bugs.  When moving windows between screens,
> > > > Mesa seems to lose track of the right vblank count to sync
> > > > against sometimes, so my test app's calls to
> > > > glXWaitVideoSyncSGI return immediately.  Moving the window back
> > > > and forth between the screens a few times fixes this problem
> > > > though, maybe there's a race in the update of base_msc and
> > > > which pipe to sync against?
> > >
> > > There can't really be a race with a single context... If this is
> > > still an issue with your updated patch, it might be related to
> > > some of the issues below.
> >
> > I was thinking the internal state may be updated by the
> > intelWindowMoved function after the client calls
> > glXWaitVideoSyncSGI but before it returned...  So not really a race
> > between two threads, but possible confusion anyway.  Or is that not
> > possible?
>
> Shouldn't be - the driver only calls WindowMoved after taking the DRM
> lock (generally during rendering operations) or when the window is
> bound to a context.
>
> There could probably be races between several threads though, some
> kind of mutex might be necessary to protect against those.

Yeah, it should only be called from makecurrent and the contendedlock 
functions, right?  I was thinking it would be called asynchronously 
somehow, but I guess not, so I must have a problem in the WaitMSC 
function somewhere.

As for actual multithreaded programs, how is that dealt with in the rest 
of Mesa?

Jesse

-------------------------------------------------------------------------
This SF.net email is sponsored by: Splunk Inc.
Still grepping through log files to find problems?  Stop.
Now Search log events and configuration files using AJAX and a browser.
Download your FREE copy of Splunk now >> http://get.splunk.com/
--
_______________________________________________
Dri-devel mailing list
Dri-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/dri-devel

Reply via email to