Btrfs: fix unexpected balance crash due to BUG_ON
authorLiu Bo <bo.li.liu@oracle.com>
Tue, 12 Jul 2016 18:24:21 +0000 (11:24 -0700)
committerDavid Sterba <dsterba@suse.com>
Tue, 26 Jul 2016 11:52:25 +0000 (13:52 +0200)
Mounting a btrfs can resume previous balance operations asynchronously.
An user got a crash when one drive has some corrupt sectors.

Since balance can cancel itself in case of any error, we can gracefully
return errors to upper layers and let balance do the cancel job.

Reported-by: sash <master.b.at.raven@chefmail.de>
Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
fs/btrfs/volumes.c

index 51bd3ee8e64f43c9271245d179e287a841ebbc8c..05aec96e4c96abab7f0a6818157a173c95ce18d0 100644 (file)
@@ -3460,7 +3460,7 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info)
        u64 size_to_free;
        u64 chunk_type;
        struct btrfs_chunk *chunk;
-       struct btrfs_path *path;
+       struct btrfs_path *path = NULL;
        struct btrfs_key key;
        struct btrfs_key found_key;
        struct btrfs_trans_handle *trans;
@@ -3494,13 +3494,33 @@ static int __btrfs_balance(struct btrfs_fs_info *fs_info)
                ret = btrfs_shrink_device(device, old_size - size_to_free);
                if (ret == -ENOSPC)
                        break;
-               BUG_ON(ret);
+               if (ret) {
+                       /* btrfs_shrink_device never returns ret > 0 */
+                       WARN_ON(ret > 0);
+                       goto error;
+               }
 
                trans = btrfs_start_transaction(dev_root, 0);
-               BUG_ON(IS_ERR(trans));
+               if (IS_ERR(trans)) {
+                       ret = PTR_ERR(trans);
+                       btrfs_info_in_rcu(fs_info,
+                "resize: unable to start transaction after shrinking device %s (error %d), old size %llu, new size %llu",
+                                         rcu_str_deref(device->name), ret,
+                                         old_size, old_size - size_to_free);
+                       goto error;
+               }
 
                ret = btrfs_grow_device(trans, device, old_size);
-               BUG_ON(ret);
+               if (ret) {
+                       btrfs_end_transaction(trans, dev_root);
+                       /* btrfs_grow_device never returns ret > 0 */
+                       WARN_ON(ret > 0);
+                       btrfs_info_in_rcu(fs_info,
+                "resize: unable to grow device after shrinking device %s (error %d), old size %llu, new size %llu",
+                                         rcu_str_deref(device->name), ret,
+                                         old_size, old_size - size_to_free);
+                       goto error;
+               }
 
                btrfs_end_transaction(trans, dev_root);
        }