Re: SCSI I/O error generating a kernel (parport? reiserfs?) bug (2.4.21-99)

[Jeff Mahoney <jeffm@suse.com>]

-----BEGIN PGP SIGNED MESSAGE-----
Hash: SHA1

Antonio Larrosa Jiménez wrote:
> Hello,
> 
> Last night I saw an I/O error in a RAID device on a SuSE 9.0 system (with the 
> stock 2.4.21-99 kernel, not tainted). I don't know if it's useful for anyone 
> given that the kernel has changed much since then, but I report it just in 
> case the problem is still in there.
> 
> I find strange that the backtrace (below) talks about parport being a SCSI I/O 
> error, so maybe it's not related to the SCSI problem, but since at the end it 
> mentions reiserfs, it makes me wonder.
> 
> Btw, the RAID is on a cciss controller.
> 
> Greetings,
> 
> These are the contents of the syslog:
> Mar 31 04:34:01 baja1 kernel: journal-601, buffer write failed
> Mar 31 04:34:01 baja1 kernel: kernel BUG at prints.c:334!

Although it's unclear by the kernel output, this is a reiserfs panic due
to an i/o error in the journal, not a BUG. Kernels prior to 2.6.10,
unless specifically patched, could not handle i/o errors in the journal.
This functionality has not been backported to 2.4 kernels.

I suspect that perhaps klogd was matched against the wrong kernel symbol
table, and that's why you're seeing odd output in your syslog. The
entire call chain, after sync_supers, should be in reiserfs if you're
getting that panic.

- -Jeff

- --
Jeff Mahoney
SuSE Labs
-----BEGIN PGP SIGNATURE-----
Version: GnuPG v1.4.0 (GNU/Linux)

iD8DBQFCnzwFLPWxlyuTD7IRAkIMAJwJ6c99kqoenFH+joLytWmUa7vS7ACeNKAc
z2REA+4/g6QB0cAA4XiHIAE=
=Jk7H
-----END PGP SIGNATURE-----