Bug 68641 - kernel panic at shutdown or reboot
Summary: kernel panic at shutdown or reboot
Status: NEW
Alias: None
Product: Other
Classification: Unclassified
Component: Other (show other bugs)
Hardware: x86-64 Linux
: P1 normal
Assignee: other_other
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2014-01-13 14:00 UTC by ribiere
Modified: 2014-01-17 11:42 UTC (History)
1 user (show)

See Also:
Kernel Version: 3.10.17
Subsystem:
Regression: No
Bisected commit-id:


Attachments
crash data with program crash and lspci, lsusb, lscpu (718 bytes, text/plain)
2014-01-13 14:00 UTC, ribiere
Details
crash (1.99 KB, text/plain)
2014-01-14 09:33 UTC, ribiere
Details
bt (1.57 KB, text/plain)
2014-01-14 09:33 UTC, ribiere
Details
log (60.00 KB, text/plain)
2014-01-14 09:33 UTC, ribiere
Details
lscpu (719 bytes, text/plain)
2014-01-14 09:33 UTC, ribiere
Details
lspci (2.36 KB, text/plain)
2014-01-14 09:33 UTC, ribiere
Details

Description ribiere 2014-01-13 14:00:35 UTC
Created attachment 121811 [details]
crash data with program crash and lspci, lsusb, lscpu

kernel panic but only at shutdown or reboot.

Sometimes I have a kernel panic but only at shutdown or reboot.
That occurs about one time on 7 to 10 shutdowns.

Previously I sent a report on the same problem, but at this time
the kernel was tainted by the Nvidia proprietary driver.
I redirected the report to Nvidia and Nvidia did not found anything.

Context :
- slackware-14.1 (x86_64)
- kernel-3.10.17 compiled from sources
- motherboard gigabyte GA-P67A-UD4-B3
- 8gb memory
- 3 hard disks
- 1 external disk connected by E-SATA
- nvidia-gtx460
- driver nouveau (but not operational)

After searches on internet I have seen some similar problems but
without solution.

I suspect E-SATA because I have installed a similar PC for a friend
with a more recent Gigabyte motherboard and nvidia-gtx580. This PC
has no E-SATA connector and no external disk and my friend has no kernel
panic.

I join informations obtained with the program crash analyzing a vmcore,
lspci, lsusb, lscpu

Let me know if you need more information (kernel config, hardware...)

Thanks for your assistance
 
note : I am not sure if the attachements are joined
Comment 1 Alan 2014-01-13 15:32:40 UTC
Only lsusb - can you attach the actual messages from the crash

Thanks
Alan
Comment 2 ribiere 2014-01-14 09:33:17 UTC
Created attachment 121961 [details]
crash

Le 13/01/2014 16:32, bugzilla-daemon@bugzilla.kernel.org a écrit :
> https://bugzilla.kernel.org/show_bug.cgi?id=68641
>
> --- Comment #1 from Alan <alan@lxorguk.ukuu.org.uk> ---
> Only lsusb - can you attach the actual messages from the crash
>
> Thanks
> Alan
>
Thanks,

I am sorry  but how to join an attachment is clear for me.

So I join :
- crash
- crash-bt
- crash-log
- lspci
- lscpu

As I have a vmcore, if you need some information...

Note : le cpu is a Intel I72600K


Best regards
Comment 3 ribiere 2014-01-14 09:33:18 UTC
Created attachment 121971 [details]
bt
Comment 4 ribiere 2014-01-14 09:33:18 UTC
Created attachment 121981 [details]
log
Comment 5 ribiere 2014-01-14 09:33:18 UTC
Created attachment 121991 [details]
lscpu
Comment 6 ribiere 2014-01-14 09:33:18 UTC
Created attachment 122001 [details]
lspci
Comment 7 Alan 2014-01-16 22:57:53 UTC
[ 2612.779607] BUG: unable to handle kernel paging request at ffffc9000873f250
[ 2612.779824] IP: [<ffffffff8104ee35>] get_next_timer_interrupt+0xa5/0x260
[ 2612.779991] PGD 21680f067 PUD 216830067 PMD 21289a067 PTE 0
[ 2612.780254] Oops: 0000 [#1] SMP 
[ 2612.780441] Modules linked in: it87 hwmon_vid nouveau cfbfillrect cfbimgblt video cfbcopyarea i2c_algo_bit ttm drm_kms_helper fbcon bitblit softcursor fb fbdev font drm agpgart snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq snd_seq_device snd_pcm_oss snd_mixer_oss xt_LOG ipt_ULOG ipt_REJECT xt_mark xt_iprange iptable_raw iptable_mangle iptable_filter ip_tables lp fuse snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_intel snd_hda_codec snd_hwdep snd_pcm snd_page_alloc snd_timer snd rtc_cmos parport_pc i2c_i801 i2c_core coretemp r8169 mxm_wmi crc32c_intel ghash_clmulni_intel wmi button processor


So it looks like a driver leaving a timer when unloading.

You should be able to shut the box down to single user mode rather than rebooting, and then you can use rmmod to remove modules one at a time and see if you can find which one blows up.

"telinit S" is the usual root command for that, but some distributions vary
Comment 8 ribiere 2014-01-17 11:42:10 UTC
Le 16/01/2014 23:57, bugzilla-daemon@bugzilla.kernel.org a écrit :
> https://bugzilla.kernel.org/show_bug.cgi?id=68641
>
> --- Comment #7 from Alan <alan@lxorguk.ukuu.org.uk> ---
> [ 2612.779607] BUG: unable to handle kernel paging request at
> ffffc9000873f250
> [ 2612.779824] IP: [<ffffffff8104ee35>] get_next_timer_interrupt+0xa5/0x260
> [ 2612.779991] PGD 21680f067 PUD 216830067 PMD 21289a067 PTE 0
> [ 2612.780254] Oops: 0000 [#1] SMP
> [ 2612.780441] Modules linked in: it87 hwmon_vid nouveau cfbfillrect
> cfbimgblt
> video cfbcopyarea i2c_algo_bit ttm drm_kms_helper fbcon bitblit softcursor fb
> fbdev font drm agpgart snd_seq_dummy snd_seq_oss snd_seq_midi_event snd_seq
> snd_seq_device snd_pcm_oss snd_mixer_oss xt_LOG ipt_ULOG ipt_REJECT xt_mark
> xt_iprange iptable_raw iptable_mangle iptable_filter ip_tables lp fuse
> snd_hda_codec_hdmi snd_hda_codec_realtek snd_hda_intel snd_hda_codec
> snd_hwdep
> snd_pcm snd_page_alloc snd_timer snd rtc_cmos parport_pc i2c_i801 i2c_core
> coretemp r8169 mxm_wmi crc32c_intel ghash_clmulni_intel wmi button processor
>
>
> So it looks like a driver leaving a timer when unloading.
>
> You should be able to shut the box down to single user mode rather than
> rebooting, and then you can use rmmod to remove modules one at a time and see
> if you can find which one blows up.
>
> "telinit S" is the usual root command for that, but some distributions vary
>

Thanks for your answer

I Did as you said : telinit S in two cases :
- first with the kernel not tainted
- second with the kernel tainted in an other partition

In both cases I can remove all modules not in use without any error
message.

At this point lsmod for the first case :
Module                  Size  Used by
xt_LOG                  7616  1
ipt_ULOG                4303  3
ipt_REJECT              2185  4
xt_mark                 1085  1
xt_iprange              1448  19
iptable_mangle          1456  1
iptable_filter          1328  1
ip_tables              15459  2 iptable_filter,iptable_mangle
fuse                   70856  1
nouveau               937208  1
cfbfillrect             3714  1 nouveau
cfbimgblt               2007  1 nouveau
video                  10816  1 nouveau
cfbcopyarea             3318  1 nouveau
i2c_algo_bit            5135  1 nouveau
ttm                    60559  1 nouveau
fbcon                  37235  71
bitblit                 4961  1 fbcon
softcursor              1237  1 bitblit
font                    7292  1 fbcon
drm_kms_helper         27608  1 nouveau
drm                   218588  3 ttm,drm_kms_helper,nouveau
agpgart                26212  2 drm,ttm
fb                     50582  5 
fbcon,drm_kms_helper,nouveau,softcursor,bitblit
i2c_core               18325  4 drm,drm_kms_helper,i2c_algo_bit,nouveau
fbdev                    734  2 fb,fbcon
mxm_wmi                 1307  1 nouveau
wmi                     7867  2 mxm_wmi,nouveau
button                  4242  1 nouveau


lsmod for the second case :
Module                  Size  Used by
xt_LOG                  7616  1
ipt_ULOG                4303  3
ipt_REJECT              2185  4
xt_mark                 1085  1
xt_iprange              1448  19
iptable_mangle          1456  1
iptable_filter          1328  1
ip_tables              15459  2 iptable_filter,iptable_mangle
fuse                   70856  1

Then I did a shutdown -h 0 and obtain a kernel panic
with the message in the first case :
"unable to handle kernel paging request..."
In the second case I could not seen the exact message because
the console is in normal vga mode and the output is too fast.

But in these cases I cannot obtain a vmcore.

What can I do to continue on this problem ?

Best regards.

Note You need to log in before you can comment on or make changes to this bug.