Re: suspend issue

[Pavel Machek <pavel@ucw.cz>]

On Fri 2008-04-11 00:05:52, Soeren Sonnenburg wrote:
> On Thu, 2008-04-10 at 12:02 +0200, Pavel Machek wrote:
> > On Wed 2008-04-09 12:25:30, Soeren Sonnenburg wrote:
> > > On Mon, 2008-04-07 at 01:28 +0200, Pavel Machek wrote:
> > > > On Sun 2008-04-06 13:55:04, Soeren Sonnenburg wrote:
> > > > > On Sun, 2008-04-06 at 04:03 +0000, Justin Mattock wrote:
> > > > > > Hello; I seem to have discovered something with echo mem > /sys/power/state.
> > > > > > I recompile the kernel and had forgotten to compile fglrx, when
> > > > > > performing a suspend without fglrx I was confronted with the dreaded
> > > > > > blank screen and a reboot. then as a test I compiled fglrx and loaded,
> > > > > > and to my amazement suspend reacted properly again.
> > > > > 
> > > > > Yeah, but I it does not even work without X - console only!
> > > > > 
> > > > > s2ram -f -p used to work since at least 2.6.20 ... and is now not
> > > > > anymore in 2.6.25-*
> > > > 
> > > > Can you do git bisect?
> > > 
> > > I just tried, and the problem is that 25-rc1 already fails to suspend -
> > > so I am not sure how I could get more insight into this failure *on
> > > resume*... any ideas?
> > 
> > Well, 2.6.20 worked, so bisect should be easy, right?
> 
> I don't quite understand... 2.6.24 worked yes, but 2.6.25-rc1 didn't
> even *suspend*. However the problem lies not in suspending but no
> display on *resume*. So the bug is hidden by the hang on suspend and I
> don't see how I can check when resume broke.

Hmm, you could bisect the resume regression, revert that, and then go
on to find the suspend regression.

...ok, not easy, I see.

Do you have kernel.org bugzilla # I could look at? Perhaps this can be
debugged using printks or something...
								Pavel
-- 
(english) http://www.livejournal.com/~pavelmachek
(cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html