On Thu, Jun 14, 2018 at 8:35 PM, Dave Chinner [off-list ref] wrote:

On Thu, Jun 14, 2018 at 05:20:56PM -0500, Eric Sandeen wrote:

quoted

On 6/14/18 5:16 PM, Marco Benatto wrote:

quoted

Currently when root inode is in short-form and its parent ino
has an invalid value, process_sf_dir2() ends up not fixing it,
because if verify_inum() fails we never get to the next case which
would fix the root inode's parent pointer.

This behavior triggers the following assert on process_dir2():

   ASSERT((ino != mp->m_sb.sb_rootino && ino != *parent) ||
        (ino == mp->m_sb.sb_rootino &&
        (ino == *parent || need_root_dotdot == 1)));

This patch fixes this behavior by making sure we always properly
handle rootino parent pointer in process_sf_dir2()

Signed-off-by: Marco Benatto <redacted>

This looks correct to me, thanks.

FWIW the problem can be demonstrated by setting the root inode's
parent to 0 on a freshly made fs:

# mkfs.xfs -f -dfile,name=fsfile,size=1g
# xfs_db -x -c "sb 0" -c "addr rootino" -c "write u3.sfdir3.hdr.parent.i4 0" fsfile
# xfs_repair fsfile

xfstest please (also to check that xfs_scrub detects it)?

Sure, working on this. I'll send you once it's done.

Cheers,

Dave.
--
Dave Chinner
david@fromorbit.com

Cheers,

-- 
Marco Benatto