tcp: fix TFO SYN_RECV to not zero retrans_stamp with retransmits out

Fix tcp_rcv_synrecv_state_fastopen() to not zero retrans_stamp
if retransmits are outstanding.

tcp_fastopen_synack_timer() sets retrans_stamp, so typically we'll
need to zero retrans_stamp here to prevent spurious
retransmits_timed_out(). The logic to zero retrans_stamp is from this
2019 commit:

commit cd736d8 ("tcp: fix retrans timestamp on passive Fast Open")

However, in the corner case where the ACK of our TFO SYNACK carried
some SACK blocks that caused us to enter TCP_CA_Recovery then that
non-zero retrans_stamp corresponds to the active fast recovery, and we
need to leave retrans_stamp with its current non-zero value, for
correct ETIMEDOUT and undo behavior.

Fixes: cd736d8 ("tcp: fix retrans timestamp on passive Fast Open")
Signed-off-by: Neal Cardwell <[email protected]>
Signed-off-by: Yuchung Cheng <[email protected]>
Signed-off-by: Eric Dumazet <[email protected]>
Link: https://patch.msgid.link/[email protected]
Signed-off-by: Jakub Kicinski <[email protected]>

net/ipv4/tcp_input.c

@@ -6684,10 +6684,17 @@ static void tcp_rcv_synrecv_state_fastopen(struct sock *sk)
 	if (inet_csk(sk)->icsk_ca_state == TCP_CA_Loss && !tp->packets_out)
 		tcp_try_undo_recovery(sk);
 
-	/* Reset rtx states to prevent spurious retransmits_timed_out() */
 	tcp_update_rto_time(tp);
-	tp->retrans_stamp = 0;
 	inet_csk(sk)->icsk_retransmits = 0;
+	/* In tcp_fastopen_synack_timer() on the first SYNACK RTO we set
+	 * retrans_stamp but don't enter CA_Loss, so in case that happened we
+	 * need to zero retrans_stamp here to prevent spurious
+	 * retransmits_timed_out(). However, if the ACK of our SYNACK caused us
+	 * to enter CA_Recovery then we need to leave retrans_stamp as it was
+	 * set entering CA_Recovery, for correct retransmits_timed_out() and
+	 * undo behavior.
+	 */
+	tcp_retrans_stamp_cleanup(sk);
 
 	/* Once we leave TCP_SYN_RECV or TCP_FIN_WAIT_1,
 	 * we no longer need req so release it.