Btrfs: fix race when listing an inode's xattrs

commit f1cd1f0b upstream.

When listing a inode's xattrs we have a time window where we race against
a concurrent operation for adding a new hard link for our inode that makes
us not return any xattr to user space. In order for this to happen, the
first xattr of our inode needs to be at slot 0 of a leaf and the previous
leaf must still have room for an inode ref (or extref) item, and this can
happen because an inode's listxattrs callback does not lock the inode's
i_mutex (nor does the VFS does it for us), but adding a hard link to an
inode makes the VFS lock the inode's i_mutex before calling the inode's
link callback.

If we have the following leafs:

               Leaf X (has N items)                    Leaf Y

 [ ... (257 INODE_ITEM 0) (257 INODE_REF 256) ]  [ (257 XATTR_ITEM 12345), ... ]
           slot N - 2         slot N - 1              slot 0

The race illustrated by the following sequence diagram is possible:

       CPU 1                                               CPU 2

  btrfs_listxattr()

    searches for key (257 XATTR_ITEM 0)

    gets path with path->nodes[0] == leaf X
    and path->slots[0] == N

    because path->slots[0] is >=
    btrfs_header_nritems(leaf X), it calls
    btrfs_next_leaf()

    btrfs_next_leaf()
      releases the path

                                                   adds key (257 INODE_REF 666)
                                                   to the end of leaf X (slot N),
                                                   and leaf X now has N + 1 items

      searches for the key (257 INODE_REF 256),
      with path->keep_locks == 1, because that
      is the last key it saw in leaf X before
      releasing the path

      ends up at leaf X again and it verifies
      that the key (257 INODE_REF 256) is no
      longer the last key in leaf X, so it
      returns with path->nodes[0] == leaf X
      and path->slots[0] == N, pointing to
      the new item with key (257 INODE_REF 666)

    btrfs_listxattr's loop iteration sees that
    the type of the key pointed by the path is
    different from the type BTRFS_XATTR_ITEM_KEY
    and so it breaks the loop and stops looking
    for more xattr items
      --> the application doesn't get any xattr
          listed for our inode

So fix this by breaking the loop only if the key's type is greater than
BTRFS_XATTR_ITEM_KEY and skip the current key if its type is smaller.
Signed-off-by: Filipe Manana <fdmanana@suse.com>
[ luis: backported to 3.16:
  - drop btrfs_key_type(), which was dropped upstream by
    962a298f ("btrfs: kill the key type accessor helpers") ]
Signed-off-by: Luis Henriques <luis.henriques@canonical.com>

fs/btrfs/xattr.c

@@ -309,8 +309,10 @@ ssize_t btrfs_listxattr(struct dentry *dentry, char *buffer, size_t size)
 		/* check to make sure this item is what we want */
 		if (found_key.objectid != key.objectid)
 			break;
-		if (btrfs_key_type(&found_key) != BTRFS_XATTR_ITEM_KEY)
+		if (found_key.type > BTRFS_XATTR_ITEM_KEY)
 			break;
+		if (found_key.type < BTRFS_XATTR_ITEM_KEY)
+			goto next;
 
 		di = btrfs_item_ptr(leaf, slot, struct btrfs_dir_item);
 		if (verify_dir_item(root, leaf, di))