swsusp problems with 2.6.17-1.2139_FC5
swsusp has problems resumeing after upgrading my Acer Travelmate 4001WLMi from 2.6.16 to 2.6.17. Note that I'm running a Fedora 5 kernel, with the ATI proprietary video driver.
On this 512Mb system, I have typically 60% of memory in use when I go into hibernate. This is with X, wm, Emacs, some xterms, wireless.
With 2.6.16, when I initiate a hibernate with
echo shutdown > /sys/power/disk; echo disk > /sys/power/state
I get switched back to vt1(?), the "stopping tasks" messages, and so on, until the system shuts down in ordinary fashion.
With 2.6.17, when I initiate the hibernate, I get no apparent reaction (although the flashing disk led reveals what's going on) and the system shuts down.
Upon reboot, I get
Trying to resume from /dev/hda4
Resuming from /dev/hda4.
Attempting manual resume
and then the console does a half-hearted attemtp to restore X and freezes. The system is alive, I can login via the network. Restarting the X server (from another tty, the console remains stuck) seems to 'cure' the problem. However, when exiting X the console stays connected to vt7 and needs to be put back to vt1 manually.
It seems to be a problem with the video switching out/in X. When I change to a vt out of X, then suspend/resume seems to give no problem and after resume I can switch back to vt7 into X.
Any suggestions to cure this problem? Is it a know regression?
Thanks,
Johan
Hi,
Quoted 29 lines · Johan Vromans
swsusp has problems resumeing after upgrading my Acer Travelmate 4001WLMi from 2.6.16 to 2.6.17. Note that I'm running a Fedora 5 kernel, with the ATI proprietary video driver.
On this 512Mb system, I have typically 60% of memory in use when I go into hibernate. This is with X, wm, Emacs, some xterms, wireless.With 2.6.16, when I initiate a hibernate with
echo shutdown >
/sys/power/disk; echo disk >/sys/power/stateI get switched back to vt1(?), the "stopping tasks" messages, and so on, until the system shuts down in ordinary fashion.
With 2.6.17, when I initiate the hibernate, I get no apparent reaction (although the flashing disk led reveals what's going on) and the system shuts down.
Upon reboot, I get
Trying to resume from
/dev/hda4
Resuming from/dev/hda4.
Attempting manual resumeand then the console does a half-hearted attemtp to restore X and freezes. The system is alive, I can login via the network. Restarting the X server (from another tty, the console remains stuck) seems to 'cure' the problem. However, when exiting X the console stays connected to vt7 and needs to be put back to vt1 manually.
Quoted 5 lines
It seems to be a problem with the video switching out/in X. When I change to a vt out of X, then suspend/resume seems to give no problem and after resume I can switch back to vt7 into X.
Any suggestions to cure this problem? Is it a know regression?
It is not known, thanks for the report.
Pavel, is the recent Linus' patch likely to cause this?
Rafael
Hi!
Quoted 3 lines
swsusp has problems resumeing after upgrading my Acer Travelmate 4001WLMi from 2.6.16 to 2.6.17. Note that I'm running a Fedora 5 kernel, with the ATI proprietary video driver.
Stop right here. Can you reproduce the problem without ATI driver? Reproducing it on vanilla kernel (not -FC5) would be nice, too.
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html
Quoted 2 lines · Pavel Machek
Stop right here. Can you reproduce the problem without ATI driver? Reproducing it on vanilla kernel (not -FC5) would be nice, too.
A lot of suspend/reboot/resumes later...
The problem does not seem to be related to the ATI driver, but whether or not the pm-suspend program is used. With the Xorg driver I get the same problem when I suspend with
echo shutdown > /sys/power/disk; echo disk > /sys/power/state
When I use pm-hibernate suspend/resume seems works okay (with Xorg and ATI driver).
With 2.6.16, I did not have the need to use pm-hibernate. So something changed here.
As mentioned in my OP using pm-hibernate does not give any feedback what is going on (except for the disk led). I find this annoying. Another annoyance is that pm-hibernate locks this kernel for the next reboot, so it is not possible to boot something else and resume later.
Apart from that, suspend/resume is a life saver!
(Now it would be nice to get suspend to memory working. It seems to suspend okay, but I haven't found out how to resume...)
-- Johan
Hi!
Quoted 5 lines
Quoted 2 lines
Stop right here. Can you reproduce the problem without ATI driver? Reproducing it on vanilla kernel (not -FC5) would be nice, too.
A lot of suspend/reboot/resumes later...
Quoted 8 lines
The problem does not seem to be related to the ATI driver, but whether or not the pm-suspend program is used. With the Xorg driver I get the same problem when I suspend with
echo shutdown >
/sys/power/disk; echo disk >/sys/power/stateWhen I use pm-hibernate suspend/resume seems works okay (with Xorg and ATI driver).
What is pm-suspend and pm-hibernate, anyway?
Quoted 2 lines
With 2.6.16, I did not have the need to use pm-hibernate. So something changed here.
Okay, find out
Quoted 5 lines
As mentioned in my OP using pm-hibernate does not give any feedback what is going on (except for the disk led). I find this annoying. Another annoyance is that pm-hibernate locks this kernel for the next reboot, so it is not possible to boot something else and resume later.
grub lockup is a distro problem. Turn up console loglevel to see the messages.
Quoted 4 lines
Apart from that, suspend/resume is a life saver!
(Now it would be nice to get suspend to memory working. It seems to suspend okay, but I haven't found out how to resume...)
Resume is always harder :-). Is resume completely broken, or does it "only" break video?
Pavel
Signature · 2 lines
(english) http://www.livejournal.com/~pavelmachek (cesky, pictures) http://atrey.karlin.mff.cuni.cz/~pavel/picture/horses/blog.html