nfs: fix dio deadlock when O_DIRECT flag is flipped

We only support swap file calling nfs_direct_IO. However, application
might be able to get to nfs_direct_IO if it toggles O_DIRECT flag
during IO and it can deadlock because we grab inode->i_mutex in
nfs_file_direct_write(). So return 0 for such case. Then the generic
layer will fall back to buffer IO.

Signed-off-by: Peng Tao <tao.peng@primarydata.com>
Cc: stable@vger.kernel.org
Signed-off-by: Trond Myklebust <trond.myklebust@primarydata.com>

fs/nfs/direct.c

@@ -212,6 +212,12 @@ static int nfs_direct_cmp_commit_data_verf(struct nfs_direct_req *dreq,
  */
 ssize_t nfs_direct_IO(int rw, struct kiocb *iocb, struct iov_iter *iter, loff_t pos)
 {
+	struct inode *inode = iocb->ki_filp->f_mapping->host;
+
+	/* we only support swap file calling nfs_direct_IO */
+	if (!IS_SWAPFILE(inode))
+		return 0;
+
 #ifndef CONFIG_NFS_SWAP
 	dprintk("NFS: nfs_direct_IO (%pD) off/no(%Ld/%lu) EINVAL\n",
 			iocb->ki_filp, (long long) pos, iter->nr_segs);