From: Ulrich Pegelow (QPC98@DMSWWU1A.UNI-MUENSTER.DE)
Date: 12/31/69


Date: Monday, 1 Jun 1992 11:18:57 MES
From: Ulrich Pegelow <QPC98@DMSWWU1A.UNI-MUENSTER.DE>
Subject: fsck, 0.96a and bad blocks

Hi Linuxers,

I'm using Linux 0.96a on a 386sx equiped with two ST277R drives. With the
new kernel I always get harddisk I/O errors after a few seconds when
doing a file system check. Here's what the kernel says:

HD-controller reset
HD-controller reset
HD-controller reset
harddisk I/O error
dev 0342, block 6097
HD-controller reset
HD-controller reset
HD-controller reset
harddisk I/O error
dev 0342, block 6098
Kernel panic: next_buffer: request buffer list destroyed

In swapper task - not syncing

After this Linux hangs. Applying Linus' patch1 does not help except that
the kernel now doen't give up so fast. Blocks 6097 and 6098 *are* bad
blocks. However they were already detected when creating the file system.
Besides the file system appears to be consistent when I make a fsck using
a kernel other than 0.96a. Doing a

tar cvf /dev/null bin dev etc usr ....

also gives no I/O errors. Does this mean that the new HD driver won't work
with a harddisk that contains bad blocks? Or could there be some bug in my
file system older kernels did not detect?

Any help would be appreciated.

Ulrich

==============================================================================
Ulrich Pegelow Bitnet: QPC98@DMSWWU1A
Inst. f. Physical Chemistry Internet: pegelow@moorea.uni-muenster.de
University of Muenster (Germany)
==============================================================================