Btrfs: make sure to record the transid in new inodes
authorChris Mason <chris.mason@oracle.com>
Fri, 24 Jun 2011 17:13:29 +0000 (13:13 -0400)
committerChris Mason <chris.mason@oracle.com>
Fri, 24 Jun 2011 17:13:29 +0000 (13:13 -0400)
When we create a new inode, we aren't filling in the
field that records the transaction that last changed this
inode.

If we then go to fsync that inode, it will be skipped because the field
isn't filled in.

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

index 5813dec5101cc1856762d33b5f1370069c8ecef5..87f1e0cf26f8c4c7e8737348164edf56c7bce7ea 100644 (file)
@@ -4520,6 +4520,7 @@ static struct inode *btrfs_new_inode(struct btrfs_trans_handle *trans,
        inode_tree_add(inode);
 
        trace_btrfs_inode_new(inode);
+       btrfs_set_inode_last_trans(trans, inode);
 
        return inode;
 fail: