Re: 2.0.36pre10 oops

Stephen C. Tweedie (sct@redhat.com)
Mon, 28 Sep 1998 15:01:16 +0100


Hi,

On Sat, 26 Sep 1998 22:07:01 -0400 (EDT), "Andrew J. Anderson"
<andrew@db.erau.edu> said:

> Unfortunately, yes, I am still able to trigger an oops doing a bru backup
> of my system. It does not oops while doing a tar, so it is most odd. I
> don't think it's hardware related, because I have seen no other symptoms
> of bad ram or flaky equipment. So far I am only able to trigger this
> using bru from BRU2000-15.0P-2.

It looks hardware-related to me. tar is pretty primitive: it just
throws data from one device to another with very little other processing
happening. BRU does much more: in particular, becase BRU performs
extensive CRC operations on all of its backups, it stresses memory and
the CPU during those intensive IO operations much more than tar does.
(Gzip-compressed tar will also do this.)

>>> EIP: 122e29 <__free_page+39/100>

>>> EIP: 11eba7 <shrink_mmap+157/210>

__free_page() oopses are 99%-reliable indicators of hardware problems in
the 2.0 kernels. However, if you can definitely pin this problem down
to a particular kernel version, we might be able to get somewhere with
it. As it stands, I don't think we can start poking the finger at
Linux.

--Stephen.

-
To unsubscribe from this list: send the line "unsubscribe linux-kernel" in
the body of a message to majordomo@vger.rutgers.edu
Please read the FAQ at http://www.tux.org/lkml/