[PATCH net v4] mptcp: do not use Fast Open on MP_JOIN subflows

From: Yilin Zhang <redacted>

tcp_fastopen_create_child() hands the SYN packet itself to
subflow_syn_recv_sock(). For an MP_JOIN request this takes the
fatal fallback: the cloned child is destroyed and handed back with
drop_req flagged, but tcp_fastopen_create_child() does not check
the flag and queues it, so accept() can expose the freed child.

MP_JOIN cannot use Fast Open: data on a subflow requires the
completed HMAC exchange (RFC 8684, sec. 3.2). Refuse it (only the
TFO path passes a SYN skb here) and let tcp_conn_request() fall
back to the regular MP_JOIN handshake; also strip the Fast Open
cookie from MP_JOIN SYN/ACKs.

Fixes: 90bf45134d55 ("mptcp: add new sock flag to deal with join subflows")
Cc: <redacted>
Reported-by: Kimi Security Team <redacted>
Suggested-by: Jiayuan Chen <jiayuan.chen@linux.dev>
Suggested-by: Paolo Abeni <pabeni@redhat.com>
Signed-off-by: Yilin Zhang <redacted>
[ Moving code to avoid clearing TS and consolidating checks. ]
Signed-off-by: Matthieu Baerts (NGI0) <matttbe@kernel.org>
---
Note: I'm taking over this series because the v3 got lost a bit, the v4
      only contains minor modifications, no need to delay this fix any
      longer.
v4:
- Disable TFO without clearing tstamp_ok.
- Moving SYN check inside the existing MPJ section.
- Check for SYN only. (Paolo)
v3:
- https://lore.kernel.org/20260920061904.3575780-1-yilinzhang@moonshot.ai (local)
v2:
- https://lore.kernel.org/20260903094010.4066892-1-yilinzhang@moonshot.ai (local)
v1:
- https://lore.kernel.org/20260902121247.3248539-1-yilinzhang@moonshot.ai (local)
---
 net/mptcp/subflow.c | 11 +++++++++++
 1 file changed, 11 insertions(+)

diff --git a/net/mptcp/subflow.c b/net/mptcp/subflow.c
index 78e886b20258..00677fb3efcd 100644
--- a/net/mptcp/subflow.c
+++ b/net/mptcp/subflow.c

@@ -379,6 +379,10 @@ static void subflow_prep_synack(const struct sock *sk, struct request_sock *req,
 	struct mptcp_subflow_context *subflow = mptcp_subflow_ctx(sk);
 	struct inet_request_sock *ireq = inet_rsk(req);
 
+	/* MP_JOIN cannot use TFO, do not send a cookie in the SYN/ACK */
+	if (foc && mptcp_subflow_rsk(req)->mp_join)
+		foc->len = -1;
+
 	/* clear tstamp_ok, as needed depending on cookie */
 	if (foc && foc->len > -1)
 		ireq->tstamp_ok = 0;

@@ -887,6 +891,13 @@ static struct sock *subflow_syn_recv_sock(const struct sock *sk,
 			fallback = true;
 
 	} else if (subflow_req->mp_join) {
+		/* a SYN skb here comes from TFO, which MP_JOIN cannot use: just
+		 * fall back to the regular path.
+		 */
+		if ((TCP_SKB_CB(skb)->tcp_flags & (TCPHDR_SYN | TCPHDR_ACK)) ==
+		    TCPHDR_SYN)
+			return NULL;
+
 		mptcp_get_options(skb, &mp_opt);
 		if (!(mp_opt.suboptions & OPTION_MPTCP_MPJ_ACK))
 			fallback = true;

---
base-commit: d8a4f4c59da52b7c31ae337243b8473a19a3f819
change-id: 20261030-mptcp-tfo-mpj-55fedc3189fc

Best regards,
--  
Matthieu Baerts (NGI0) [off-list ref]