Package: linux-image-2.6.32-5-686
Version: 2.6.32-30

I can hibernate my laptop successfully, but some proportion of the
time (maybe one time in six?) it crashes during the resume.

The last message printed by the kernel is "Suspending consoles" which
I'm afraid isn't in any log file, so is reproduced from memory.  After
that the system is unresponsive other than to holding down the power
button to force a poweroff.  The message mentions "no_console_suspend"
which I have now added to my boot options.

I did get one stack trace from "no_console_suspend" during resume, but
I was in a hurry trying to get something else done so didn't have time
to write down the message.  Also, unfortunately, the top of the stack
trace fell off the top of my 80x25 text console.

Normally I hibernate my laptop using a "hibernate" script which came
from an ancient swsusp2 package.  I am going to try using
  echo disk >/sys/power/state
(after switching from X to a text VC) instead.

I have disabled KMS (see my other bug report) by putting
"options radeon modeset=0" in /etc/modprobe.d/radeon-kms.conf.

Attached is /var/log/dmesg.

The machine is an IBM Thinkpad R50p.  Hibernation used to work
perfectly with my own kernel some time ago.  In lenny, it would work
nearly all of the time but would occasionally come back in a state
where the disk subsystem was wedged.  In squeeze it fails a lot.

My "resume=" argument refers to an LVM LV, where the VG is backed by a
devmapper dm-crypt device set up by cryptsetup in my initramfs; the
dm-crypt device is backed by a partition on the ide disk.  The same VG
contains the root filesystem, but not /boot.

I appreciate that this may not be very helpful or informative.  If
you'd like to suggest ways to collect more data, please let me know.

Ian.



-- 
To UNSUBSCRIBE, email to debian-kernel-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org
Archive: 
http://lists.debian.org/19833.12046.451431.178...@chiark.greenend.org.uk

Reply via email to