From: Mike Snitzer snitzer@redhat.com
commit 54fa16ee532705985e6c946da455856f18f63ee1 upstream.
Check if in fail_io mode at start of dm_pool_metadata_set_needs_check(). Otherwise dm_pool_metadata_set_needs_check()'s superblock_lock() can crash in dm_bm_write_lock() while accessing the block manager object that was previously destroyed as part of a failed dm_pool_abort_metadata() that ultimately set fail_io to begin with.
Also, update DMERR() message to more accurately describe superblock_lock() failure.
Cc: stable@vger.kernel.org Reported-by: Zdenek Kabelac zkabelac@redhat.com Signed-off-by: Mike Snitzer snitzer@redhat.com Signed-off-by: Greg Kroah-Hartman gregkh@linuxfoundation.org
--- drivers/md/dm-thin-metadata.c | 7 +++++-- 1 file changed, 5 insertions(+), 2 deletions(-)
--- a/drivers/md/dm-thin-metadata.c +++ b/drivers/md/dm-thin-metadata.c @@ -2046,16 +2046,19 @@ int dm_pool_register_metadata_threshold(
int dm_pool_metadata_set_needs_check(struct dm_pool_metadata *pmd) { - int r; + int r = -EINVAL; struct dm_block *sblock; struct thin_disk_superblock *disk_super;
pmd_write_lock(pmd); + if (pmd->fail_io) + goto out; + pmd->flags |= THIN_METADATA_NEEDS_CHECK_FLAG;
r = superblock_lock(pmd, &sblock); if (r) { - DMERR("couldn't read superblock"); + DMERR("couldn't lock superblock"); goto out; }