mptcp: handle defer connect in mptcp_sendmsg

When TCP_FASTOPEN_CONNECT has been set on the socket before a connect,
the defer flag is set and must be handled when sendmsg is called.

This is similar to what is done in tcp_sendmsg_locked().

Acked-by: Paolo Abeni <[email protected]>
Reviewed-by: Matthieu Baerts <[email protected]>
Co-developed-by: Benjamin Hesmans <[email protected]>
Signed-off-by: Benjamin Hesmans <[email protected]>
Signed-off-by: Dmytro Shytyi <[email protected]>
Signed-off-by: Mat Martineau <[email protected]>
Signed-off-by: Jakub Kicinski <[email protected]>

Author: shytyi

net/mptcp/protocol.c

@@ -1677,6 +1677,7 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 {
 	struct mptcp_sock *msk = mptcp_sk(sk);
 	struct page_frag *pfrag;
+	struct socket *ssock;
 	size_t copied = 0;
 	int ret = 0;
 	long timeo;
@@ -1690,6 +1691,27 @@ static int mptcp_sendmsg(struct sock *sk, struct msghdr *msg, size_t len)
 
 	lock_sock(sk);
 
+	ssock = __mptcp_nmpc_socket(msk);
+	if (unlikely(ssock && inet_sk(ssock->sk)->defer_connect)) {
+		struct sock *ssk = ssock->sk;
+		int copied_syn = 0;
+
+		lock_sock(ssk);
+
+		ret = tcp_sendmsg_fastopen(ssk, msg, &copied_syn, len, NULL);
+		copied += copied_syn;
+		if (ret == -EINPROGRESS && copied_syn > 0) {
+			/* reflect the new state on the MPTCP socket */
+			inet_sk_state_store(sk, inet_sk_state_load(ssk));
+			release_sock(ssk);
+			goto out;
+		} else if (ret) {
+			release_sock(ssk);
+			goto out;
+		}
+		release_sock(ssk);
+	}
+
 	timeo = sock_sndtimeo(sk, msg->msg_flags & MSG_DONTWAIT);
 
 	if ((1 << sk->sk_state) & ~(TCPF_ESTABLISHED | TCPF_CLOSE_WAIT)) {