---

yaml
---
r: 156509
b: refs/heads/master
c: 013f1b1
h: refs/heads/master
i:
  156507: 6e4d146
v: v3

[refs]

@@ -1,2 +1,2 @@
 ---
-refs/heads/master: 6606bb97e146a387932efee263745b7240a11193
+refs/heads/master: 013f1b12f4fc479f697acae2f31bad220162cd03

trunk/fs/btrfs/extent-tree.c

@@ -265,10 +265,6 @@ static int caching_kthread(void *data)
 
 	atomic_inc(&block_group->space_info->caching_threads);
 	last = max_t(u64, block_group->key.objectid, BTRFS_SUPER_INFO_OFFSET);
-again:
-	/* need to make sure the commit_root doesn't disappear */
-	down_read(&fs_info->extent_commit_sem);
-
 	/*
 	 * We don't want to deadlock with somebody trying to allocate a new
 	 * extent for the extent root while also trying to search the extent
@@ -282,6 +278,10 @@ static int caching_kthread(void *data)
 	key.objectid = last;
 	key.offset = 0;
 	btrfs_set_key_type(&key, BTRFS_EXTENT_ITEM_KEY);
+again:
+	/* need to make sure the commit_root doesn't disappear */
+	down_read(&fs_info->extent_commit_sem);
+
 	ret = btrfs_search_slot(NULL, fs_info->extent_root, &key, path, 0, 0);
 	if (ret < 0)
 		goto err;
@@ -304,6 +304,19 @@ static int caching_kthread(void *data)
 
 			if (need_resched() ||
 			    btrfs_transaction_in_commit(fs_info)) {
+				leaf = path->nodes[0];
+
+				/* this shouldn't happen, but if the
+				 * leaf is empty just move on.
+				 */
+				if (btrfs_header_nritems(leaf) == 0)
+					break;
+				/*
+				 * we need to copy the key out so that
+				 * we are sure the next search advances
+				 * us forward in the btree.
+				 */
+				btrfs_item_key_to_cpu(leaf, &key, 0);
 				btrfs_release_path(fs_info->extent_root, path);
 				up_read(&fs_info->extent_commit_sem);
 				schedule_timeout(1);