Hi;
This does look suspicious, but it is reproducible for 3 systems
with completely different boards, memory, and CPUs so I don't
think it is memory in this case although it might be related
to how memory is used in general. I should mention a little
more about the hardware though:
IDE disks
raidtools-0.90-6
I will go put in some debug statements to see if I can flush
out more things so to speak.
I attached one of the other kernel oops but they are consistent
and reproduceable.
thanks for the suggestion though,
Scott
-----Original Message-----
From: Jens Axboe
To: Scott Smyth
Cc: 'linux-xfs@xxxxxxxxxxx'
Sent: 1/12/01 6:25 PM
Subject: Re: XFS 2.4.0 tree w/ LVM_0_9-patches: buffer_flush problem w/
raid5
On Fri, Jan 12 2001, Scott Smyth wrote:
> Unable to handle kernel NULL pointer dereference at virtual address
00000020
Sync buffers oopsing there most likely means that a bad bit flip has
corrupted the buffer list. I would suspect hw problems in this case,
you should run some memory tests as this is most likely where the
problem
is.
--
* Jens Axboe <axboe@xxxxxxx>
* SuSE Labs
raid5_xfs_umount_ksymoops.txt
Description: Binary data
|