Hi Sorry, I've been physically away from the notebook since early august till last weekend.
On Dienstag, 26. August 2008, Brice Goglin wrote:
> On Sun, Jun 01, 2008 at 08:49:18PM +0200, Stefan Lippers-Hollmann wrote:
> > Yes, it is completely dead, open ssh connections trying to gather last bits
> > from tail -f /var/log/messages stall immediately (unfortunately the
> > notebook has no serial connectors for a serial console), pings get no
> > answer - a hard poweroff is the only recovery.
>
> Pinging again to make sure the bug didn't disappear by chance :)
Unfortunately it hasn't gone away and is still present on today's sid and
kernel 2.6.26.3.
> Otherwise, we need to decide between:
> 1) reverting to 2.1.3-5 which worked fine except bug #481739 (which looks
> strange to me)
This version did/ does work fine for me.
> 2) reapplying 02_temporary_revert_pciaccess.diff and reverting 5 pciaccess
> commits (merging all of them in the 02_revert). Not sure how stable this
> one would be.
I've been running this[1] reliably since I first encountered the problem,
sorry I'm not too familiar with git and X.org's buildsystem generation, so
the quilt series looks a bit messy, but was nevertheless quite helpful for
me to pinpoint the problematic commits.
> Maybe we can start with (2) for now, and switch to (1) as the ultimate
> solution
> if needed before Lenny gets released?
Now I notice Tormod's suggestions and will try them tonight, hopefully I
get around to adding some ErrorF() breakpoints to the source tommorow, like
you suggested.
Regards
Stefan Lippers-Hollmann
[1]
http://sidux.com/debian/pool/fix.main/x/xserver-xorg-video-savage/xserver-xorg-video-savage_2.2.1-1+c0.sidux.1.dsc
signature.asc
Description: This is a digitally signed message part.
