btrfs: check if root is readonly while setting security xattr

commit b51111271b0352aa596c5ae8faf06939e91b3b68 upstream.

For a filesystem which has btrfs read-only property set to true, all
write operations including xattr should be denied. However, security
xattr can still be changed even if btrfs ro property is true.

This happens because xattr_permission() does not have any restrictions
on security.*, system.*  and in some cases trusted.* from VFS and
the decision is left to the underlying filesystem. See comments in
xattr_permission() for more details.

This patch checks if the root is read-only before performing the set
xattr operation.

Testcase:

  DEV=/dev/vdb
  MNT=/mnt

  mkfs.btrfs -f $DEV
  mount $DEV $MNT
  echo "file one" > $MNT/f1

  setfattr -n "security.one" -v 2 $MNT/f1
  btrfs property set /mnt ro true

  setfattr -n "security.one" -v 1 $MNT/f1

  umount $MNT

CC: stable@vger.kernel.org # 4.9+
Reviewed-by: Qu Wenruo <wqu@suse.com>
Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Goldwyn Rodrigues <rgoldwyn@suse.com>
Reviewed-by: David Sterba <dsterba@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

diff --git a/fs/btrfs/xattr.c b/fs/btrfs/xattr.c
index c5c5b97..43fe2c2 100644 (file)
--- a/fs/btrfs/xattr.c
+++ b/fs/btrfs/xattr.c
@@ -391,6 +391,9 @@ static int btrfs_xattr_handler_set(const struct xattr_handler *handler,
                                   const char *name, const void *buffer,
                                   size_t size, int flags)
 {
+       if (btrfs_root_readonly(BTRFS_I(inode)->root))
+               return -EROFS;
+
        name = xattr_full_name(handler, name);
        return btrfs_setxattr_trans(inode, name, buffer, size, flags);
 }