Bug#774526:

2015-01-14 Thread Jan Michael Greiner
I put this bug on kernel.org, not knowing that it was already solved by benh
https://bugzilla.kernel.org/show_bug.cgi?id=91321



Bug#774526: linux-image-3.2.0-4-amd64: system crash with sudden rebo

2015-01-12 Thread John Kennerson
The kernel from benh [1] works for me.

I am happy; thank you!
Julius

[1] 
https://people.debian.org/~benh/packages/linux-image-3.2.0-4-amd64_3.2.65-1+deb7u1~test_amd64.deb


-- 
To UNSUBSCRIBE, email to debian-bugs-dist-requ...@lists.debian.org
with a subject of unsubscribe. Trouble? Contact listmas...@lists.debian.org



Bug#774526: linux-image-3.2.0-4-amd64: system crash with sudden reboot when hotplugging a CPU - suspend functionality broken

2015-01-03 Thread Julius
Package: src:linux
Version: 3.2.65-1
Severity: normal

Dear Maintainer,

please note that not all information that are automatically included
here represent the state of the system at the observation of the bug.
Details at the end of my explanation. I am sorry for that.

   * What led up to the situation?
==

Via the proposed updates I upgraded the package
linux-image-3.2.0-4-amd64 and headers to 3.2.65-1. After that, my
system did not wake up from suspend as usual. Instead, it booted as
if it had crashed (short filesystem check and raid resync).

   * What exactly did you do (or not do) that was effective (or
 ineffective)?
==

After experiencing the broken suspend functionality multiple times,
here is what I tried to investigate the problem.

I did some hibernation tests using /sys/power/pm_test as described in
https://www.kernel.org/doc/Documentation/power/basic-pm-debugging.txt
and found that the processor test (enabling/disabling cores) is the
first that is failing, i.e., leading to a sudden reboot. To confirm
the CPU issue I manually tested that individually:

# cat /proc/cpuinfo | grep processor
processor   : 0
processor   : 1
# echo 0  /sys/devices/system/cpu/cpu1/online
# cat /proc/cpuinfo | grep processor
processor   : 0
# echo 1  /sys/devices/system/cpu/cpu1/online

   * What was the outcome of this action?
==

What happens then is a boot from BIOS as if I switched-on the
computer after sudden shutdown.

   * What outcome did you expect instead?
==

I expect that the last echo re-enables the CPU so that it is listed
in /proc/cpuinfo again. I confirmed that this is the behavior when
using the version 3.2.63-2+deb7u2 that also succeeds to resume after
suspending.

Something else that I tried
==

Adding maxcpus=1 as boot argument made the system boot with only one
processor in /proc/cpuinfo. With this, I made the observation that
the pm tests for hibernation work. However, resuming after a real
suspend again leads to a reboot. Hence I removed this boot argument
again.

What else I (unfortunately) did after experiencing the bug
==

I downgraded my kernel to the previous version I found by apt-cache.

# aptitude install linux-image-3.2.0-4-amd64=3.2.63-2+deb7u2 \
linux-headers-3.2.0-4-amd64=3.2.63-2+deb7u2 \
linux-headers-3.2.0-4-common=3.2.63-2+deb7u2
# update-grub
# reboot

After reboot, everything worked as expected. I thus removed the
proposed updates from /etc/apt/sources.list.

Note again that due to these changes, some of the automatically
generated information here are wrong. Now I am back to version 
3.2.63-2+deb7u2 which works. The version that fails for me is
3.2.65-1. The version part in the pseudo header of this message has
been modified by me. All other information represent the current
(working) configuration of my system. It should still be helpful
because the hardware is the same.

Finally some proof (?) that I had another version once. After typing

$ zless /var/log/apt/history.log.1.gz

there is one paragraph (besides many others) which reads

Start-Date: 2014-12-31  11:37:04
Commandline: apt-get upgrade
Upgrade: wireless-regdb:amd64 (2014.06.13-1~deb7u1, 2014.10.07-1~deb7u1), 
linux-image-3.2.0-4-amd64:amd64 (3.2.63-2+deb7u2, 3.2.65-1), 
linux-headers-3.2.0-4-amd64:amd64 (3.2.63-2+deb7u2, 3.2.65-1), 
linux-headers-3.2.0-4-common:amd64 (3.2.63-2+deb7u2, 3.2.65-1), 
linux-libc-dev:amd64 (3.2.63-2+deb7u2, 3.2.65-1)
End-Date: 2014-12-31  11:38:00

I hope that helps. Anyway thanks for all your dedication!

-- Package-specific info:
** Version:
Linux version 3.2.0-4-amd64 (debian-ker...@lists.debian.org) (gcc version 4.6.3 
(Debian 4.6.3-14) ) #1 SMP Debian 3.2.63-2+deb7u2

** Command line:
BOOT_IMAGE=/boot/vmlinuz-3.2.0-4-amd64 
root=UUID=fd570e99-d11c-4a60-b919-4f8ad6beb1bd ro quiet

** Tainted: O (4096)
 * Out-of-tree module has been loaded.

** Kernel log:
[  735.674096] Calibrating delay loop (skipped) already calibrated this CPU
[  750.627738] NMI watchdog enabled, takes one hw-pmu counter.
[  750.628318] coretemp coretemp.0: Using relative temperature scale!
[ 3778.728164] CPU 1 is now offline
[ 3778.728170] SMP alternatives: switching to UP code
[ 3791.810370] SMP alternatives: switching to SMP code
[ 3791.819085] Booting Node 0 Processor 1 APIC 0x1
[ 3791.819089] smpboot cpu 1: start_ip = 9a000
[ 3778.729199] Calibrating delay loop (skipped) already calibrated this CPU
[ 3791.867545] NMI watchdog enabled, takes one hw-pmu counter.
[ 3791.868752] coretemp coretemp.0: Using relative temperature scale!
[ 3874.716032] CPU 1 is now offline
[ 3874.716038] SMP alternatives: switching to UP code
[ 3879.894306] SMP alternatives: switching to SMP code
[ 3879.897941] Booting Node 0 Processor 1 APIC 0x1
[ 3879.897943] smpboot cpu 1: start_ip = 9a000
[ 3874.616699] Calibrating delay loop (skipped)