On Tue, 2012-08-14 at 19:35 +0100, Al Viro wrote:
> On Tue, Aug 14, 2012 at 06:04:08PM +0000, Myklebust, Trond wrote:
> > > Is this a known BUG ?
> > > Which other information or test can I do to contribute searching/resolving this issue?
> >
> > I'm not 100% certain, but it looks to me as if the call to dget_parent()
> > in nfs_lookup_revalidate is returning NULL.
> >
> > Could you please apply the following patch, and see if that triggers the
> > WARN_ON instead of the above Oops?
>
> Unlikely as hell. dget_parent() has only one return in it, and it's
> BUG_ON(!ret->d_count);
> ret->d_count++;
> spin_unlock(&ret->d_lock);
> return ret;
> so it's extremely unlikely that it would managed to return NULL. Not
> without oopsing...
Agreed, but the Oops is a NULL pointer dereference of offset 00038,
which looks suspiciously like a deference of (struct dentry)->d_inode.
There are only 2 of those in the vicinity of nfs_lookup_revalidate+0x2d:
one is a dereference of parent->d_inode, the other is dentry->d_inode...
--
Trond Myklebust
Linux NFS client maintainer
NetApp
Trond.Myklebust@xxxxxxxxxx
www.netapp.com
��.n��������+%������w��{.n�����{��w��� jg��� �����ݢj����G���� ���j:+v���w�m������w�������h�����٥
