vhost/vsock: fix reset orphans race with close timeout

If a local process has closed a connected socket and hasn't received a
RST packet yet, then the socket remains in the table until a timeout
expires.

When a vhost_vsock instance is released with the timeout still pending,
the socket is never freed because vhost_vsock has already set the
SOCK_DONE flag.

Check if the close timer is pending and let it close the socket.  This
prevents the race which can leak sockets.

Reported-by: Maximilian Riemensberger <riemensberger@cadami.net>
Cc: Graham Whaley <graham.whaley@gmail.com>
Signed-off-by: Stefan Hajnoczi <stefanha@redhat.com>
Signed-off-by: Michael S. Tsirkin <mst@redhat.com>

drivers/vhost/vsock.c

@@ -563,13 +563,21 @@ static void vhost_vsock_reset_orphans(struct sock *sk)
 	 * executing.
 	 */
 
-	if (!vhost_vsock_get(vsk->remote_addr.svm_cid)) {
-		sock_set_flag(sk, SOCK_DONE);
-		vsk->peer_shutdown = SHUTDOWN_MASK;
-		sk->sk_state = SS_UNCONNECTED;
-		sk->sk_err = ECONNRESET;
-		sk->sk_error_report(sk);
-	}
+	/* If the peer is still valid, no need to reset connection */
+	if (vhost_vsock_get(vsk->remote_addr.svm_cid))
+		return;
+
+	/* If the close timeout is pending, let it expire.  This avoids races
+	 * with the timeout callback.
+	 */
+	if (vsk->close_work_scheduled)
+		return;
+
+	sock_set_flag(sk, SOCK_DONE);
+	vsk->peer_shutdown = SHUTDOWN_MASK;
+	sk->sk_state = SS_UNCONNECTED;
+	sk->sk_err = ECONNRESET;
+	sk->sk_error_report(sk);
 }
 
 static int vhost_vsock_dev_release(struct inode *inode, struct file *file)