Btrfs: use reserved space for creating a snapshot
authorLiu Bo <bo.li.liu@oracle.com>
Fri, 22 Feb 2013 04:33:36 +0000 (04:33 +0000)
committerJosef Bacik <jbacik@fusionio.com>
Tue, 26 Feb 2013 16:00:52 +0000 (11:00 -0500)
While inserting dir index and updating inode for a snapshot, we'd
add delayed items which consume trans->block_rsv, if we don't have
any space reserved in this trans handle, we either just return or
reserve space again.

But before creating pending snapshots during committing transaction,
we've done a release on this trans handle, so we don't have space reserved
in it at this stage.

What we're using is block_rsv of pending snapshots which has already
reserved well enough space for both inserting dir index and updating
inode, so we need to set trans handle to indicate that we have space
now.

Signed-off-by: Liu Bo <bo.li.liu@oracle.com>
Reviewed-by: Miao Xie <miaox@cn.fujitsu.com>
Signed-off-by: Josef Bacik <jbacik@fusionio.com>
fs/btrfs/transaction.c

index a83d486..4330433 100644 (file)
@@ -1123,6 +1123,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 
        rsv = trans->block_rsv;
        trans->block_rsv = &pending->block_rsv;
+       trans->bytes_reserved = trans->block_rsv->reserved;
 
        dentry = pending->dentry;
        parent = dget_parent(dentry);
@@ -1276,6 +1277,7 @@ static noinline int create_pending_snapshot(struct btrfs_trans_handle *trans,
 fail:
        dput(parent);
        trans->block_rsv = rsv;
+       trans->bytes_reserved = 0;
 no_free_objectid:
        kfree(new_root_item);
 root_item_alloc_fail: