Machine freezing with 'kernel BUG at dcache.c:350'

Submitted by Anonymous
on March 8, 2004 - 2:25am

Hi

The following entries in /var/log/messages are connected to 'freezing' of the machine.
Occurred 3 times since last Nov. with the postfix/trivial-rewrite process as well as the nscd.

Feb 16 02:11:42 SRV0018 kernel: kernel BUG at dcache.c:350!
Feb 16 02:11:42 SRV0018 kernel: invalid operand: 0000 2.4.20-4GB #1 Mon Mar 17 17:54:44 UTC 2003
Feb 16 02:11:42 SRV0018 kernel: CPU: 0
Feb 16 02:11:42 SRV0018 kernel: EIP: 0010:[prune_dcache+114/304] Not tainted
Feb 16 02:11:42 SRV0018 kernel: EIP: 0010:[c0157432] Not tainted
Feb 16 02:11:42 SRV0018 kernel: EFLAGS: 00010282
Feb 16 02:11:42 SRV0018 kernel: eax: ffffffdf ebx: c3b4ecf8 ecx: c1d80860 edx: c3b4ed78
Feb 16 02:11:42 SRV0018 kernel: esi: c3b4ece0 edi: 00000716 ebp: c5cd2800 esp: c3fcdf14
Feb 16 02:11:42 SRV0018 kernel: ds: 0018 es: 0018 ss: 0018
Feb 16 02:11:42 SRV0018 kernel: Process trivial-rewrite (pid: 24108, stackpage=c3fcd000)
Feb 16 02:11:42 SRV0018 kernel: Stack: c112b920 c5d08444 c112b920 c015774d 00000733 c5d08400 c014a91f c112b920
Feb 16 02:11:42 SRV0018 kernel: c5cd2850 c3cc5ce0 c1126360 c2721ce0 c2721ce0 c0145703 c5d08400 c3cc5ce0
Feb 16 02:11:42 SRV0018 kernel: c2ec5c68 00000000 c5f2e1e0 c01445de c3cc5ce0 c5f2e1e0 c3cc5ce0 c5f2e1e0
Feb 16 02:11:43 SRV0018 kernel: Call Trace:[shrink_dcache_parent+13/32] [kill_super+127/304] [ieee1394:ieee1394_procfs_entry+48605476/40422361] [fput+147/224] [filp_close+94/176]
Feb 16 02:11:43 SRV0018 kernel: Call Trace:
[c015774d] [c014a91f] [c5cd2850] [c0145703] [c01445de]
Feb 16 02:11:43 SRV0018 kernel:[put_files_struct+78/176] [do_exit+166/688] [sys_exit+15/16] [system_call+51/64]
Feb 16 02:11:43 SRV0018 kernel:
[c012116e] [c0121796] [c01219cf] [c0108c33]
Feb 16 02:11:43 SRV0018 kernel: Modules: [(ext3:c5cc0060:c5cd2de8)]
Feb 16 02:11:43 SRV0018 kernel: Code: 0f 0b 5e 01 fd ec 29 c0 8d 43 f8 8b 53 f8 8b 48 04 89 4a 04
(Please note that I had to remove various < and > for proper displaying)

I'm not primarily interested in a detailed analysis of all this Hexcode.
But in first place I would like to know how certain this can be related to a defective hardware.

Some more info on the system:
HP Netserver e30: PII with 166MHz and 92MB RAM, IDE HD
Doing only nonanonymous ftp with very low amount of transferred data.

Minimal system without graphics; based on SuSE 8.2 with:
- 2.4.20-4GB kernel
- pure-ftpd 1.0.14-31
- postfix-2.0.6-8
- some security stuff like aide, sudo, harden scripts, ...
- secumod 1.6c: security kernel module by Mark Vogelsberger

If necessary more info can be provided.
Many thanks for any comments and hints
Georg