net/ulp: prevent ULP without clone op from entering the LISTEN status

commit 2c02d41 upstream.

When an ULP-enabled socket enters the LISTEN status, the listener ULP data
pointer is copied inside the child/accepted sockets by sk_clone_lock().

The relevant ULP can take care of de-duplicating the context pointer via
the clone() operation, but only MPTCP and SMC implement such op.

Other ULPs may end-up with a double-free at socket disposal time.

We can't simply clear the ULP data at clone time, as TLS replaces the
socket ops with custom ones assuming a valid TLS ULP context is
available.

Instead completely prevent clone-less ULP sockets from entering the
LISTEN status.

Fixes: 734942c ("tcp: ULP infrastructure")
Reported-by: slipper <slipper.alive@gmail.com>
Signed-off-by: Paolo Abeni <pabeni@redhat.com>
Link: https://lore.kernel.org/r/4b80c3d1dbe3d0ab072f80450c202d9bc88b4b03.1672740602.git.pabeni@redhat.com
Signed-off-by: Jakub Kicinski <kuba@kernel.org>
Signed-off-by: Greg Kroah-Hartman <gregkh@linuxfoundation.org>

net/ipv4/inet_connection_sock.c

@@ -916,11 +916,25 @@ void inet_csk_prepare_forced_close(struct sock *sk)
 }
 EXPORT_SYMBOL(inet_csk_prepare_forced_close);
 
+static int inet_ulp_can_listen(const struct sock *sk)
+{
+	const struct inet_connection_sock *icsk = inet_csk(sk);
+
+	if (icsk->icsk_ulp_ops && !icsk->icsk_ulp_ops->clone)
+		return -EINVAL;
+
+	return 0;
+}
+
 int inet_csk_listen_start(struct sock *sk, int backlog)
 {
 	struct inet_connection_sock *icsk = inet_csk(sk);
 	struct inet_sock *inet = inet_sk(sk);
-	int err = -EADDRINUSE;
+	int err;
+
+	err = inet_ulp_can_listen(sk);
+	if (unlikely(err))
+		return err;
 
 	reqsk_queue_alloc(&icsk->icsk_accept_queue);
 

net/ipv4/tcp_ulp.c

@@ -136,6 +136,10 @@ static int __tcp_set_ulp(struct sock *sk, const struct tcp_ulp_ops *ulp_ops)
 	if (icsk->icsk_ulp_ops)
 		goto out_err;
 
+	err = -EINVAL;
+	if (!ulp_ops->clone && sk->sk_state == TCP_LISTEN)
+		goto out_err;
+
 	err = ulp_ops->init(sk);
 	if (err)
 		goto out_err;