dm cache: Fix ABBA deadlock between shrink_slab and dm_cache_metadata_abort

Same ABBA deadlock pattern fixed in commit 4b60f452ec51 ("dm thin: Fix
ABBA deadlock between shrink_slab and dm_pool_abort_metadata") to
DM-cache's metadata.

Reported-by: Zhihao Cheng <chengzhihao1@huawei.com>
Cc: stable@vger.kernel.org
Fixes: 028ae9f76f29 ("dm cache: add fail io mode and needs_check flag")
Signed-off-by: Mike Snitzer <snitzer@kernel.org>

1 files changed, 47 insertions, 7 deletions

diff --git a/drivers/md/dm-cache-metadata.c b/drivers/md/dm-cache-metadata.c
index ab13b7380265..83a5975bcc72 100644
--- a/drivers/md/dm-cache-metadata.c
+++ b/drivers/md/dm-cache-metadata.c
@@ -551,11 +551,13 @@ static int __create_persistent_data_objects(struct dm_cache_metadata *cmd,
 	return r;
 }
 
-static void __destroy_persistent_data_objects(struct dm_cache_metadata *cmd)
+static void __destroy_persistent_data_objects(struct dm_cache_metadata *cmd,
+					      bool destroy_bm)
 {
 	dm_sm_destroy(cmd->metadata_sm);
 	dm_tm_destroy(cmd->tm);
-	dm_block_manager_destroy(cmd->bm);
+	if (destroy_bm)
+		dm_block_manager_destroy(cmd->bm);
 }
 
 typedef unsigned long (*flags_mutator)(unsigned long);
@@ -826,7 +828,7 @@ static struct dm_cache_metadata *lookup_or_open(struct block_device *bdev,
 		cmd2 = lookup(bdev);
 		if (cmd2) {
 			mutex_unlock(&table_lock);
-			__destroy_persistent_data_objects(cmd);
+			__destroy_persistent_data_objects(cmd, true);
 			kfree(cmd);
 			return cmd2;
 		}
@@ -874,7 +876,7 @@ void dm_cache_metadata_close(struct dm_cache_metadata *cmd)
 		mutex_unlock(&table_lock);
 
 		if (!cmd->fail_io)
-			__destroy_persistent_data_objects(cmd);
+			__destroy_persistent_data_objects(cmd, true);
 		kfree(cmd);
 	}
 }
@@ -1807,14 +1809,52 @@ int dm_cache_metadata_needs_check(struct dm_cache_metadata *cmd, bool *result)
 
 int dm_cache_metadata_abort(struct dm_cache_metadata *cmd)
 {
-	int r;
+	int r = -EINVAL;
+	struct dm_block_manager *old_bm = NULL, *new_bm = NULL;
+
+	/* fail_io is double-checked with cmd->root_lock held below */
+	if (unlikely(cmd->fail_io))
+		return r;
+
+	/*
+	 * Replacement block manager (new_bm) is created and old_bm destroyed outside of
+	 * cmd root_lock to avoid ABBA deadlock that would result (due to life-cycle of
+	 * shrinker associated with the block manager's bufio client vs cmd root_lock).
+	 * - must take shrinker_rwsem without holding cmd->root_lock
+	 */
+	new_bm = dm_block_manager_create(cmd->bdev, DM_CACHE_METADATA_BLOCK_SIZE << SECTOR_SHIFT,
+					 CACHE_MAX_CONCURRENT_LOCKS);
 
 	WRITE_LOCK(cmd);
-	__destroy_persistent_data_objects(cmd);
-	r = __create_persistent_data_objects(cmd, false);
+	if (cmd->fail_io) {
+		WRITE_UNLOCK(cmd);
+		goto out;
+	}
+
+	__destroy_persistent_data_objects(cmd, false);
+	old_bm = cmd->bm;
+	if (IS_ERR(new_bm)) {
+		DMERR("could not create block manager during abort");
+		cmd->bm = NULL;
+		r = PTR_ERR(new_bm);
+		goto out_unlock;
+	}
+
+	cmd->bm = new_bm;
+	r = __open_or_format_metadata(cmd, false);
+	if (r) {
+		cmd->bm = NULL;
+		goto out_unlock;
+	}
+	new_bm = NULL;
+out_unlock:
 	if (r)
 		cmd->fail_io = true;
 	WRITE_UNLOCK(cmd);
+	dm_block_manager_destroy(old_bm);
+out:
+	if (new_bm && !IS_ERR(new_bm))
+		dm_block_manager_destroy(new_bm);
 
 	return r;
 }