Andi Kleen wrote:
> On Tue, Apr 11, 2000 at 02:28:06PM -0500, Steve Lord wrote:
> > I think the bug here is going to be a long way away from this code - there
> > always should be an inode passed into pagebuf_get - the comment is wrong.
> > The inode here is coming from the vnode passed in, where it should also
> > always be set - I bet NFS is using a way into the filesystem to reference
> > an inode which XFS is not setting things up properly, that is where to look.
> fs/nfsfh/nfsfh.c does a lot of dirty things with the VFS, especially
> with inodes. Especially iget()/read_inode() must handle the case of
> doing iget for an already deleted file.
Figures, I never liked NFS anyways.. Maybe it will work better with coda.
I'll recreate the traceback here shortly and send it to the list.