Re: [PATCH 09/14] xfs/348: dir->symlink corruption must not be allowed

[Eryu Guan <eguan@redhat.com>]

On Thu, Nov 02, 2017 at 09:37:54AM -0700, Darrick J. Wong wrote:
> On Thu, Nov 02, 2017 at 08:42:14PM +0800, Eryu Guan wrote:
> > On Wed, Nov 01, 2017 at 02:47:23PM -0700, Darrick J. Wong wrote:
> > > From: Darrick J. Wong <darrick.wong@oracle.com>
> > > 
> > > A directory corrupted into a symlink will be caught by the upcoming
> > > local format ifork verifiers.
> > > 
> > > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > > ---
> > >  tests/xfs/348.out |    2 +-
> > >  1 file changed, 1 insertion(+), 1 deletion(-)
> > > 
> > > 
> > > diff --git a/tests/xfs/348.out b/tests/xfs/348.out
> > > index f4a7a71..17d9be2 100644
> > > --- a/tests/xfs/348.out
> > > +++ b/tests/xfs/348.out
> > > @@ -239,7 +239,7 @@ would have junked entry "DATA" in directory PARENT_INO
> > >  would have junked entry "DIR" in directory PARENT_INO
> > >  would have junked entry "EMPTY" in directory PARENT_INO
> > >  would have junked entry "FIFO" in directory PARENT_INO
> > > -stat: 'SCRATCH_MNT/test/DIR' is a symbolic link
> > > +stat: cannot stat 'SCRATCH_MNT/test/DIR': Structure needs cleaning
> > 
> > But this breaks tests on old kernels. Or with the new ifork verifiers,
> > old kernels can be considered as buggy?
> 
> Yes, they're buggy since we shouldn't be interpreting directory entries
> as a link target given that the zero bytes in the "link target" will
> screw things up.

Then I'm fine with taking it, and probably will edit the commit log a
bit to reflect that we were missing a error case previously. Thanks!

Eryu