Btrfs: hash the btree inode during fill_super
authorYan Zheng <zheng.yan@oracle.com>
Thu, 24 Sep 2009 13:24:43 +0000 (09:24 -0400)
committerChris Mason <chris.mason@oracle.com>
Thu, 24 Sep 2009 13:24:43 +0000 (09:24 -0400)
The snapshot deletion  patches dropped this line, but the inode
needs to be hashed.

Signed-off-by: Chris Mason <chris.mason@oracle.com>
fs/btrfs/disk-io.c

index a0d41e713f3caff2fbfe4ddfad4acd907bb1d7ad..f1e905f7e7018e6f8ab689cd6e5aeb6137785f1c 100644 (file)
@@ -1664,6 +1664,7 @@ struct btrfs_root *open_ctree(struct super_block *sb,
        memset(&BTRFS_I(fs_info->btree_inode)->location, 0,
               sizeof(struct btrfs_key));
        BTRFS_I(fs_info->btree_inode)->dummy_inode = 1;
+       insert_inode_hash(fs_info->btree_inode);
 
        spin_lock_init(&fs_info->block_group_cache_lock);
        fs_info->block_group_cache_tree.rb_node = NULL;