On Wed, Dec 06, 2023 at 10:03:36PM -0800, Christoph Hellwig wrote:
> > +/*
> > + * Decide if this directory has been zapped to satisfy the inode and ifork
> > + * verifiers. Checking and repairing should be postponed until the directory
> > + * is fixed.
> > + */
> > +bool
> > +xchk_dir_looks_zapped(
> > + struct xfs_inode *dp)
> > +{
> > + /* Repair zapped this dir's data fork a short time ago */
> > + if (xfs_ifork_zapped(dp, XFS_DATA_FORK))
> > + return true;
> > +
> > + /*
> > + * If the dinode repair found a bad data fork, it will reset the fork
> > + * to extents format with zero records and wait for the bmapbtd
> > + * scrubber to reconstruct the block mappings. Directories always
> > + * contain some content, so this is a clear sign of a zapped directory.
> > + */
> > + return dp->i_df.if_format == XFS_DINODE_FMT_EXTENTS &&
> > + dp->i_df.if_nextents == 0;
>
> Correct me if I'm wrong: in general the xfs_ifork_zapped should be
> all that's needed here now, and the check below just finds another
> obvious case if we crashed / unmounted and lost the zapped flag?
> If so maybe update the comment a bit.
Correct. The comment now reads:
/*
* If the dinode repair found a bad data fork, it will reset the fork
* to extents format with zero records and wait for the bmapbtd
* scrubber to reconstruct the block mappings. Directories always
* contain some content, so this is a clear sign of a zapped directory.
* The state checked by xfs_ifork_zapped is not persisted, so this is
* our backup strategy if repairs are interrupted by a crash or an
* unmount.
*/
> Otherwise:
>
> Reviewed-by: Christoph Hellwig <hch@xxxxxx>
Thanks!
--D
