Re: Oops on reiserfs corruption

[Edward Shishkin <edward.shishkin@gmail.com>]

On 11/10/2011 11:13 PM, Pierre Ynard wrote:
> Hello,

Hello.

>
> A reiserfs filesystem on my system got randomly corrupted, and then
> would trigger a 100% reproducible oops:


The whole sequence (making corruption - trigger an oops)
is reproducible?

If yes, please describe steps how to reproduce.
If no, then please extract file system metadata,
with debugreiserfs -p and let us to download this
before fsck-ing, if possible.

Thanks,
Edward.


>
>
> [  178.016969] PAX: suspicious general protection fault: 0000 [#1]
> [  178.017141] last sysfs file: /sys/module/x_tables/initstate
> [  178.017234] Modules linked in: af_packet ts_bm xt_tcpudp xt_string tun dummy sit tunnel4 ipt_MASQUERADE iptable_nat nf_nat nf_conntrack_ipv4 nf_conntrack nf_defrag_ipv4 ip_tables x_tables dm_mod pata_ali 3c59x mii ipv6 unix
> [  178.018237]
> [  178.018312] Pid: 1816, comm: rm Not tainted 2.6.39.3-grsec #1
> [  178.018481] EIP: 0060:[<000e29b2>] EFLAGS: 00010286 CPU: 0
> [  178.018577] EAX: 00000000 EBX: c204fb4b ECX: 00000000 EDX: c03c3e00
> [  178.018682] ESI: ffffffff EDI: c204f73e EBP: c03c3e00 ESP: c03c3d70
> [  178.018784]  DS: 0068 ES: 0068 FS: 0000 GS: 007b SS: 0068
> [  178.018885] Process rm (pid: 1816, ti=c7b585a8 task=c7b58380 task.ti=c7b585a8)
> [  178.018996] Stack:
> [  178.019063]  c03a3380 c204fb3e c03c3e28 c7b12200 c01ffab8 000e8fda 00000001 00000000
> [  178.019452]  00000000 00000000 00000000 c76f3a48 00000001 c76f38c0 00000088 c76f3888
> [  178.019840]  00000074 000f360b c7b12200 c8cd7000 000f0369 00000000 c7b12200 c1868d7b
> [  178.020230] Call Trace:
> [  178.020335]  [<000e8fda>] ? search_for_position_by_key+0x1a/0x2b0
> [  178.020463]  [<000f360b>] ? reiserfs_write_lock+0x1b/0x30
> [  178.020572]  [<000f0369>] ? do_journal_end.clone.38+0x209/0xc60
> [  178.020703]  [<000e2c7c>] ? __reiserfs_error+0x1c/0xb0
> [  178.020807]  [<000eac5e>] ? reiserfs_do_truncate+0x47e/0x580
> [  178.020920]  [<000f31bf>] ? reiserfs_for_each_xattr+0x23f/0x3c0
> [  178.021047]  [<00076dfe>] ? try_to_unmap_file+0xfe/0x470
> [  178.021154]  [<00076dff>] ? try_to_unmap_file+0xff/0x470
> [  178.021271]  [<006b113d>] ? 0x6b113c
> [  178.021365]  [<000ead85>] ? reiserfs_delete_object+0x25/0x60
> [  178.021478]  [<000d77fb>] ? reiserfs_evict_inode+0xab/0xe0
> [  178.021584]  [<006b113d>] ? 0x6b113c
> [  178.021688]  [<0009dde6>] ? evict+0x46/0x110
> [  178.021784]  [<00094d38>] ? do_unlinkat+0xe8/0x1f0
> [  178.021888]  [<00076dff>] ? try_to_unmap_file+0xff/0x470
> [  178.021915]  [<00207c60>] ? syscall_call+0x7/0xb
> [  178.021915] Code: ff 00 00 52 8b 48 08 51 8b 50 04 52 8b 00 50 68 28 52 86 c1 57 e8 2f 4f 04 00 83 c4 1c e9 e7 fd ff ff 8d 6e 04 8b 36 85 f6 74 75<8a>  46 10 ba 5b f6 86 c1 84 c0 74 1b 3c 03 74 79 ba 2c 88 86 c1
> [  178.021915] EIP: [<000e29b2>] prepare_error_buf+0x3b2/0x4b0 SS:ESP 0068:c03c3d70
> [  178.024828] ---[ end trace a7de6d386ed5b2d8 ]---
> [  178.024925] ------------[ cut here ]------------
>
>
> I tried and ran into it on both 3.0.8-grsec and 2.6.39.3-grsec kernels.
> Sorry for the non-vanilla kernel, I just got rid of the problem with a
> fsck as quickly as possible. I thought I'd report it anyway.
>
> Regards,
>