2004-11-22 18:42:21

by Jan Engelhardt

[permalink] [raw]
Subject: [reiser4 bug] Whoops on module unload

Hi,


I just picked this up from /dev/vcsa10 (and /var/log/{kernel,messages})...
seems to have happened upon `rmmod reiser4`. Did not luckily seem to make the
system unusable.

Kernel-version: SUSE Kernel of the day 20041029
Reiser4 fs were not mounted at the time of the rmmod (that's logic), but were
mounted before.

The backtrace does not give much info (and ksymoops neither), so the best I
could suggest is checking whether these codepaths allow unfreed objects. Since
the KOTD I'm using is also a month old by now, don't invest too much time into
this, if at all.

Nov 22 19:29:13 otto kernel: slab error in kmem_cache_destroy(): cache
`plugin_set': Can't free all objects
Nov 22 19:29:13 otto kernel: [kmem_cache_destroy+216/304]
kmem_cache_destroy+0xd8/0x130
Nov 22 19:29:13 otto kernel: [<c0137b88>] kmem_cache_destroy+0xd8/0x130
Nov 22 19:29:13 otto kernel: [pg0+276180016/1069949952]
plugin_set_done+0x10/0x40 [reiser4]
Nov 22 19:29:13 otto kernel: [<d0afec30>] plugin_set_done+0x10/0x40 [reiser4]
Nov 22 19:29:13 otto kernel: [pg0+276150806/1069949952]
shutdown_reiser4+0x126/0x270 [reiser4]
Nov 22 19:29:13 otto kernel: [<d0af7a16>] shutdown_reiser4+0x126/0x270
[reiser4]
Nov 22 19:29:13 otto kernel: [sys_delete_module+415/432]
sys_delete_module+0x19f/0x1b0
Nov 22 19:29:13 otto kernel: [<c012ba5f>] sys_delete_module+0x19f/0x1b0
Nov 22 19:29:13 otto kernel: [do_munmap+289/368] do_munmap+0x121/0x170
Nov 22 19:29:13 otto kernel: [<c0141351>] do_munmap+0x121/0x170
Nov 22 19:29:13 otto kernel: [syscall_call+7/11] syscall_call+0x7/0xb
Nov 22 19:29:13 otto kernel: [<c010414f>] syscall_call+0x7/0xb


Jan Engelhardt
--
Gesellschaft für Wissenschaftliche Datenverarbeitung
Am Fassberg, 37077 Göttingen, http://www.gwdg.de


2004-11-22 19:00:46

by Hans Reiser

[permalink] [raw]
Subject: Re: [reiser4 bug] Whoops on module unload

Jan Engelhardt wrote:

>Hi,
>
>
>I just picked this up from /dev/vcsa10 (and /var/log/{kernel,messages})...
>seems to have happened upon `rmmod reiser4`. Did not luckily seem to make the
>system unusable.
>
>Kernel-version: SUSE Kernel of the day 20041029
>Reiser4 fs were not mounted at the time of the rmmod (that's logic), but were
>mounted before.
>
>The backtrace does not give much info (and ksymoops neither), so the best I
>could suggest is checking whether these codepaths allow unfreed objects. Since
>the KOTD I'm using is also a month old by now, don't invest too much time into
>this, if at all.
>
>Nov 22 19:29:13 otto kernel: slab error in kmem_cache_destroy(): cache
>`plugin_set': Can't free all objects
>Nov 22 19:29:13 otto kernel: [kmem_cache_destroy+216/304]
>kmem_cache_destroy+0xd8/0x130
>Nov 22 19:29:13 otto kernel: [<c0137b88>] kmem_cache_destroy+0xd8/0x130
>Nov 22 19:29:13 otto kernel: [pg0+276180016/1069949952]
>plugin_set_done+0x10/0x40 [reiser4]
>Nov 22 19:29:13 otto kernel: [<d0afec30>] plugin_set_done+0x10/0x40 [reiser4]
>Nov 22 19:29:13 otto kernel: [pg0+276150806/1069949952]
>shutdown_reiser4+0x126/0x270 [reiser4]
>Nov 22 19:29:13 otto kernel: [<d0af7a16>] shutdown_reiser4+0x126/0x270
>[reiser4]
>Nov 22 19:29:13 otto kernel: [sys_delete_module+415/432]
>sys_delete_module+0x19f/0x1b0
>Nov 22 19:29:13 otto kernel: [<c012ba5f>] sys_delete_module+0x19f/0x1b0
>Nov 22 19:29:13 otto kernel: [do_munmap+289/368] do_munmap+0x121/0x170
>Nov 22 19:29:13 otto kernel: [<c0141351>] do_munmap+0x121/0x170
>Nov 22 19:29:13 otto kernel: [syscall_call+7/11] syscall_call+0x7/0xb
>Nov 22 19:29:13 otto kernel: [<c010414f>] syscall_call+0x7/0xb
>
>
>Jan Engelhardt
>
>
vs, please look into this, Jan, it will probably be several days before
vs gets to this, as he is a bit swamped this week.

Hans