Hi Alvaro -
Sorry for the slow reply on this one!
There was a bug in _pagebuf_page_io that would cause memory corruption
with RAID if you had a page size > 4k; almost certainly what this oops
is, and it should be fixed now in both CVS and the XFS 1.1 kernels.
-Eric
On 19 Mar 2002, Alvaro Figueroa wrote:
> > Today, I'd tried to use xfs on a linear raid, and it oops'ses.
>
> I've just tried it with a raid5, and the same thing happen.
>
> (BTW, I did forgot to mention (sorry) that the same raid formated with
> ext2, even with the same kernel, works perfectly)
>
> Here is the oops for the raid5 test.
<snip>
> >>I7; 004fcc50 <_pagebuf_page_io+230/2e0>
> Trace; 004fcc50 <_pagebuf_page_io+230/2e0>
|