Re: [PATCH] xfs: don't fail when converting shortform attr to long form during ATTR_REPLACE

[Dave Chinner <david@fromorbit.com>]

On Tue, Apr 17, 2018 at 05:51:07PM -0700, Darrick J. Wong wrote:
> On Wed, Apr 18, 2018 at 10:45:21AM +1000, Dave Chinner wrote:
> > On Tue, Apr 17, 2018 at 05:31:30PM -0700, Darrick J. Wong wrote:
> > > From: Darrick J. Wong <darrick.wong@oracle.com>
> > > 
> > > Kanda Motohiro reported that expanding a tiny xattr into a large xattr
> > > fails on XFS because we remove the tiny xattr from a shortform fork and
> > > then try to re-add it after converting the fork to extents format having
> > > not removed the ATTR_REPLACE flag.  This fails because the attr is no
> > > longer present, causing a fs shutdown.
> > > 
> > > This is derived from the patch in his bug report, but we really
> > > shouldn't ignore a nonzero retval from the remove call.  Something's
> > > broken and we should bail out and shut down.
> > > 
> > > Bugzilla: https://bugzilla.kernel.org/show_bug.cgi?id=199119
> > > Reported-by: kanda.motohiro@gmail.com
> > > Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
> > > ---
> > >  fs/xfs/libxfs/xfs_attr.c |    9 ++++++++-
> > >  1 file changed, 8 insertions(+), 1 deletion(-)
> > > 
> > > diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c
> > > index ce4a34a..2588c73 100644
> > > --- a/fs/xfs/libxfs/xfs_attr.c
> > > +++ b/fs/xfs/libxfs/xfs_attr.c
> > > @@ -511,7 +511,14 @@ xfs_attr_shortform_addname(xfs_da_args_t *args)
> > >  		if (args->flags & ATTR_CREATE)
> > >  			return retval;
> > >  		retval = xfs_attr_shortform_remove(args);
> > > -		ASSERT(retval == 0);
> > > +		if (retval)
> > > +			return retval;
> > > +		/*
> > > +		 * Since we have removed the old attr here,
> > > +		 * further lookup will fail with ENOATTR.
> > > +		 * Ignore this was a replace and go on creating new attr.
> > 
> > The last line of that comment doesn't make sense to me. not sure
> > what it's meant to say...
> 
> That was what was in the patch attached to the bugzilla. :)
> 
> How about:
> 
> /*
>  * Since we have removed the old attr, clear ATTR_REPLACE so that the
>  * leaf format add routine won't trip over the attr not being around.
>  */
> 
> instead?

Yup, that makes sense now.

Reviewed-by: Dave Chinner <dchinner@redhat.com>

-Dave.
-- 
Dave Chinner
david@fromorbit.com