Re: New sleep code for ATI M6, M7 and M9

[Michel Dänzer <daenzer@debian.org>]

On Son, 2003-03-23 at 10:18, Benjamin Herrenschmidt wrote: 
> On Wed, 2003-03-19 at 02:54, Michel Dänzer wrote:
> > On Die, 2003-03-18 at 23:26, Jamie Wilkinson wrote:
> > > This one time, at band camp, Brett Carter wrote:
> > > > 
> > > >> As you can tell from the versions though, I haven't come around to
> > > >> updating the woody packages in a while. It's possible that something in
> > > >> the current packages for sid makes a difference. Would be great if
> > > >> someone could verify that.
> > > >
> > > >Ok, I'm verifying that the sid packages don't work for me either.
> > > >xserver-xfree86-dri-trunk
> > > >Version: 2003.03.11-1
> > > >
> > > >drm-trunk-module-2.4.20-ben8
> > > >Version: 2003.03.11-1+lucy+22+lowlat+sleep
> > > >
> > > >If I disable dri, I can sleep just fine, otherwise, the screen never comes
> > > >back on.
> 
> Can you all try with DRI and PCI GART instead of AGP GART ?
> 
> Also, Michel, what is X doing ? Is it always re-enabling AGP on wakeup ?

No, only with Option "DRIReinit", and even then only if a client isn't
keeping the DRM open during sleep.

> Is it making sure all ring processing is completely finished before
> sleep ?

The X server basically emulates a VT switch for sleep, so the radeon
driver stops the CP, which is supposed to flush and idle it, but I guess
it's possible that this doesn't work perfectly yet.


-- 
Earthling Michel Dänzer (MrCooper)/ Debian GNU/Linux (powerpc) developer
XFree86 and DRI project member   /  CS student, Free Software enthusiast