Skip to contentSkip to conversation

tcp: fix retrans timestamp on passive Fast Open

PATCH In review Yuchung Cheng 3 participants 3 messages May 13, 2019 – May 14, 2019

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

PATCH v2 1/1 tcp: fix retrans timestamp on passive Fast Open

Commit c7d13c8faa74 ("tcp: properly track retry time on passive Fast Open") sets the start of SYNACK retransmission time on passive Fast Open in "retrans_stamp". However the timestamp is not reset upon the handshake has completed. As a result, future data packet retransmission may not update it in tcp_retransmit_skb(). This may lead to socket aborting earlier unexpectedly by retransmits_timed_out() since retrans_stamp remains the SYNACK rtx time.

This bug only manifests on passive TFO sender that a) suffered SYNACK timeout and then b) stalls on very first loss recovery. Any successful loss recovery would reset the timestamp to avoid this issue.

Fixes c7d13c8faa74 ("tcp: properly track retry time on passive Fast Open")
---
 net/ipv4/tcp_input.c | 3 +++
 1 file changed, 3 insertions(+)
net/ipv4/tcp_input.c +3 -0
index 20f6fac5882e..cf69f50855ea 100644
@@ -6024,6 +6024,9 @@ static int tcp_rcv_synsent_state_process(struct sock *sk, struct sk_buff *skb,
 static void tcp_rcv_synrecv_state_fastopen(struct sock *sk)
 {
 	tcp_try_undo_loss(sk, false);
+
+	/* Reset rtx states to prevent spurious retransmits_timed_out() */
+	tcp_sk(sk)->retrans_stamp = 0;
 	inet_csk(sk)->icsk_retransmits = 0;
 
 	/* Once we leave TCP_SYN_RECV or TCP_FIN_WAIT_1,
Signature · 1 line
2.21.0.1020.gf2820cf01a-goog
#1 lore raw
Quoted 14 lines · Yuchung Cheng

Commit c7d13c8faa74 ("tcp: properly track retry time on passive Fast Open") sets the start of SYNACK retransmission time on passive Fast Open in "retrans_stamp". However the timestamp is not reset upon the handshake has completed. As a result, future data packet retransmission may not update it in tcp_retransmit_skb(). This may lead to socket aborting earlier unexpectedly by retransmits_timed_out() since retrans_stamp remains the SYNACK rtx time.

This bug only manifests on passive TFO sender that a) suffered SYNACK timeout and then b) stalls on very first loss recovery. Any successful loss recovery would reset the timestamp to avoid this issue.

Quoted 4 lines
Fixes c7d13c8faa74 ("tcp: properly track retry time on passive Fast Open")

---

maintainer
#2 lore raw
In reply to Yuchung Cheng

From: Yuchung Cheng <[email protected]>
Date: Mon, 13 May 2019 10:32:05 -0700

Quoted 9 lines

Commit c7d13c8faa74 ("tcp: properly track retry time on passive Fast Open") sets the start of SYNACK retransmission time on passive Fast Open in "retrans_stamp". However the timestamp is not reset upon the handshake has completed. As a result, future data packet retransmission may not update it in tcp_retransmit_skb(). This may lead to socket aborting earlier unexpectedly by retransmits_timed_out() since retrans_stamp remains the SYNACK rtx time.

Quoted 8 lines

This bug only manifests on passive TFO sender that a) suffered SYNACK timeout and then b) stalls on very first loss recovery. Any successful loss recovery would reset the timestamp to avoid this issue.

Fixes c7d13c8faa74 ("tcp: properly track retry time on passive Fast Open")

Applied and queued up for -stable, thanks.