Btrfs: Drop dirty roots created by log replay immediately when
authorYan Zheng <zheng.yan@oracle.com>
Thu, 20 Nov 2008 15:25:19 +0000 (10:25 -0500)
committerChris Mason <chris.mason@oracle.com>
Thu, 20 Nov 2008 15:25:19 +0000 (10:25 -0500)
The log replay produces dirty roots. These dirty roots
should be dropped immediately if the fs is mounted as
ro. Otherwise they can be added to the dirty root list
again when remounting the fs as rw. Thank you,

Signed-off-by: Yan Zheng <zheng.yan@oracle.com>
fs/btrfs/disk-io.c

index 9816522..fb04665 100644 (file)
@@ -1757,6 +1757,11 @@ struct btrfs_root *open_ctree(struct super_block *sb,
                                                      generation + 1);
                ret = btrfs_recover_log_trees(log_tree_root);
                BUG_ON(ret);
+
+               if (sb->s_flags & MS_RDONLY) {
+                       ret =  btrfs_commit_super(tree_root);
+                       BUG_ON(ret);
+               }
        }
 
        if (!(sb->s_flags & MS_RDONLY)) {