On Wednesday, November 14, 2007 10:47 pm Janne wrote:
> On Wednesday 14 November 2007, Jesse Barnes wrote:
> > On Wednesday, November 14, 2007 11:57 am Janne wrote:
> > > Top causes for wakeups:
> > >   71.5% ( 50.2)       <interrupt> : [EMAIL PROTECTED]:0000:01:00.0
> > >    9.1% (  6.4)             artsd : schedule_timeout (process_timeout)
> > >    6.0% (  4.2)       <interrupt> : acpi
> > >    2.0% (  1.4)              kwin : schedule_timeout (process_timeout)
> > >    1.4% (  1.0)              Xorg : do_setitimer (it_real_fn)
> > >    1.4% (  1.0)            dhcdbd : schedule_timeout (process_timeout)
> > >    1.4% (  1.0)             artsd : do_setitimer (it_real_fn)
> > >
> > > - There's Radeon there at the top.
> > > => ???
> >
> > Sounds like the radeon driver is enabling vblank interrupts after resume.
> > Does the same thing happen if you VT switch to a console and back to X?
>
> Yes, indeed it happens.
>
> > This problem is fixed properly in the vblank-rework branch of the DRM
> > tree. Those changes should be making their way into upstream Linux in the
> > 2.6.25 timeframe...
>
> Ok, cool. I'll see if I can make enough sense out of all bits and pieces so
> that I can make that work. I not, then I'll wait until it
> becomes "official" :)

The radeon X driver also has some code for selectively enabling vblank events.  
A similar approach would probably work in its EnterVT routine, so it may be 
worth filing a bug for this problem or raising it on [EMAIL PROTECTED]

Jesse

_______________________________________________
Discuss mailing list
[email protected]
http://mail.lesswatts.org/mailman/listinfo/discuss

Reply via email to