Kernel Oops

I have tested SGI's linux-2.4-xfs in it's 2.4.7 incarnation. Compilled with gcc-2.96-81 on a Dell Latitude CPt400 with an Adaptec 1460A PCMCIA SCSI adapter.

I can do whatever I want until I mount an LVM volume. After some time or immediately it will panic with a dump like this:

Aug  1 18:57:26 localhost kernel: Unable to handle kernel paging 
request at virtual address c489c4f0
Aug  1 18:57:26 localhost kernel:  printing eip:
Aug  1 18:57:26 localhost kernel: c489c4f0
Aug  1 18:57:26 localhost kernel: *pde = 03bd1067
Aug  1 18:57:26 localhost kernel: *pte = 00000000
Aug  1 18:57:26 localhost kernel: Oops: 0000
Aug  1 18:57:26 localhost kernel: CPU:    0
Aug  1 18:57:26 localhost kernel: EIP:    
0010:[fat:__insmod_fat_S.bss_L2264+450160/60315488]
Aug  1 18:57:26 localhost kernel: EIP:    0010:[<c489c4f0>]
Aug  1 18:57:26 localhost kernel: EFLAGS: 00010086
Aug  1 18:57:26 localhost kernel: eax: c0331c20   ebx: 00000202   ecx: 
00000801   edx: c489c4f0
Aug  1 18:57:26 localhost kernel: esi: c250f620   edi: 00000002   ebp: 
00073000   esp: c1913c98
Aug  1 18:57:26 localhost kernel: ds: 0018   es: 0018   ss: 0018
Aug  1 18:57:26 localhost kernel: Process mount (pid: 19930, 
stackpage=c1913000)
Aug  1 18:57:26 localhost kernel: Stack: c01e6d2a 00000801 00000001 
c01e7ada 00000801 c10e6c84 00000000 c02c5cb4
Aug  1 18:57:26 localhost kernel:        c02c5e20 c250f620 00000001 
00000000 00000000 c01e7b94 00000000 c250f620
Aug  1 18:57:26 localhost kernel:        c250f620 c250f620 00000000 
c01e7df7 00000000 c250f620 c0133ae3 c1912000
Aug  1 18:57:27 localhost kernel: Call Trace: [blk_get_queue+42/48] 
[generic_make_request+186/288] [submit_bh+84/112] [ll_rw_block+583/672] 
[getblk+211/256] [block_read+722/1184] [do_ide_request+15/32]
Aug  1 18:57:27 localhost kernel: Call Trace: [<c01e6d2a>] [<c01e7ada>] 
[<c01e7b94>] [<c01e7df7>] [<c0133ae3>] [<c0138aa2>] [<c01f498f>]
Aug  1 18:57:27 localhost kernel:        [__alloc_pages+116/624] 
[do_anonymous_page+58/272] [do_no_page+48/192] 
[filemap_nopage+185/1264] [filemap_nopage+185/1264] [do_no_page+77/192] 
[handle_mm_fault+97/208] [__alloc_pages+116/624]
Aug  1 18:57:27 localhost kernel:        [<c012bb84>] [<c012189a>] 
[<c01219a0>] [<c0124c49>] [<c0124c49>] [<c01219bd>] [<c0121a91>] 
[<c012bb84>]
Aug  1 18:57:27 localhost kernel:        [do_anonymous_page+58/272] 
[cached_lookup+16/80] [path_walk+1788/2016] [file_read_actor+115/240] 
[cached_lookup+16/80] [blkdev_open+71/128] [dentry_open+195/320] 
[sys_read+150/208]
Aug  1 18:57:27 localhost kernel:        [<c012189a>] [<c013bf60>] 
[<c013c86c>] [<c0124643>] [<c013bf60>] [<c01392a7>] [<c01316a3>] 
[<c0131cc6>]
Aug  1 18:57:27 localhost kernel:        [sys_llseek+201/224] 
[system_call+51/56]
Aug  1 18:57:27 localhost kernel:        [<c0131c19>] [<c0106ecb>]
Aug  1 18:57:27 localhost kernel:
Aug  1 18:57:27 localhost kernel: Code:  Bad EIP value.

This also happens with 2.4.8-pre3-xfs. I'm not sure if this is a LVM issue, a general kernel issue or an SGi/XFS issue, but I thought I'd bring it up here before I talked to the XFS guys as this is directly related to LVM (#define LVM_RELEASE_NAME "0.9.1_beta6" - perhaps a known problem?)

Regards,
-Frode

^ Frode Tennebø | email: [email protected] | Frode@IRC  ^
| with Standard.Disclaimer; use Standard.Disclaimer; |

It appeard that a filesystem error caused the panic when mounting one of the LVs. A fsck on a newly booted system cured that problem.

[snap]

Furhtermore the oops happening after doing some work (eg. compiling) while the LV in question was mounted, but working on a different filesystem, also seems to have dissappeard. However, doing some work on the filesystem of the LV in question (from simple ls, to compilations) brings back the oops. I'm unable to umount the filesystem and eventually everything hangs.

What could be wrong?

Regards,
-Frode

^ Frode Tennebø | email: [email protected] | Frode@IRC  ^
| with Standard.Disclaimer; use Standard.Disclaimer; |