xfs: fix an incore inode UAF in xfs_bui_recover
authorDarrick J. Wong <darrick.wong@oracle.com>
Sat, 26 Sep 2020 00:39:51 +0000 (17:39 -0700)
committerDarrick J. Wong <darrick.wong@oracle.com>
Wed, 7 Oct 2020 15:40:28 +0000 (08:40 -0700)
commitff4ab5e02a0447dd1e290883eb6cd7d94848e590
tree59451bf62ddec9ce297594b6500ae56b2586d46d
parent64a3f3315bc60f710a0a25c1798ac0ea58c6fa1f
xfs: fix an incore inode UAF in xfs_bui_recover

In xfs_bui_item_recover, there exists a use-after-free bug with regards
to the inode that is involved in the bmap replay operation.  If the
mapping operation does not complete, we call xfs_bmap_unmap_extent to
create a deferred op to finish the unmapping work, and we retain a
pointer to the incore inode.

Unfortunately, the very next thing we do is commit the transaction and
drop the inode.  If reclaim tears down the inode before we try to finish
the defer ops, we dereference garbage and blow up.  Therefore, create a
way to join inodes to the defer ops freezer so that we can maintain the
xfs_inode reference until we're done with the inode.

Note: This imposes the requirement that there be enough memory to keep
every incore inode in memory throughout recovery.

Signed-off-by: Darrick J. Wong <darrick.wong@oracle.com>
Reviewed-by: Brian Foster <bfoster@redhat.com>
Reviewed-by: Christoph Hellwig <hch@lst.de>
fs/xfs/libxfs/xfs_defer.c
fs/xfs/libxfs/xfs_defer.h
fs/xfs/xfs_bmap_item.c
fs/xfs/xfs_extfree_item.c
fs/xfs/xfs_log_recover.c
fs/xfs/xfs_refcount_item.c
fs/xfs/xfs_rmap_item.c