tcp: use RTT from SACK for RTO
If RTT is not available because Karn's check has failed or no
new packet is acked, use the RTT measured from SACK to estimate
the RTO. The sender can continue to estimate the RTO during loss
recovery or reordering event upon receiving non-partial ACKs.
This also changes when the RTO is re-armed. Previously it is
only re-armed when some data is cummulatively acknowledged (i.e.,
SND.UNA advances), but now it is re-armed whenever RTT estimator
is updated. This feature is particularly useful to reduce spurious
timeout for buffer bloat including cellular carriers [1], and
RTT estimation on reordering events.
[1] "An In-depth Study of LTE: Effect of Network Protocol and
Application Behavior on Performance", In Proc. of SIGCOMM 2013
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index b85bc7c..b61274b 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -2800,8 +2800,8 @@
tcp_xmit_retransmit_queue(sk);
}
-static inline void tcp_ack_update_rtt(struct sock *sk, const int flag,
- s32 seq_rtt)
+static inline bool tcp_ack_update_rtt(struct sock *sk, const int flag,
+ s32 seq_rtt, s32 sack_rtt)
{
const struct tcp_sock *tp = tcp_sk(sk);
@@ -2813,6 +2813,9 @@
if (flag & FLAG_RETRANS_DATA_ACKED)
seq_rtt = -1;
+ if (seq_rtt < 0)
+ seq_rtt = sack_rtt;
+
/* RTTM Rule: A TSecr value received in a segment is used to
* update the averaged RTT measurement only if the segment
* acknowledges some new data, i.e., only if it advances the
@@ -2823,13 +2826,14 @@
seq_rtt = tcp_time_stamp - tp->rx_opt.rcv_tsecr;
if (seq_rtt < 0)
- return;
+ return false;
tcp_rtt_estimator(sk, seq_rtt);
tcp_set_rto(sk);
/* RFC6298: only reset backoff on valid RTT measurement. */
inet_csk(sk)->icsk_backoff = 0;
+ return true;
}
/* Compute time elapsed between (last) SYNACK and the ACK completing 3WHS. */
@@ -2840,7 +2844,7 @@
if (tp->lsndtime && !tp->total_retrans)
seq_rtt = tcp_time_stamp - tp->lsndtime;
- tcp_ack_update_rtt(sk, FLAG_SYN_ACKED, seq_rtt);
+ tcp_ack_update_rtt(sk, FLAG_SYN_ACKED, seq_rtt, -1);
}
static void tcp_cong_avoid(struct sock *sk, u32 ack, u32 in_flight)
@@ -2929,7 +2933,7 @@
* arrived at the other end.
*/
static int tcp_clean_rtx_queue(struct sock *sk, int prior_fackets,
- u32 prior_snd_una)
+ u32 prior_snd_una, s32 sack_rtt)
{
struct tcp_sock *tp = tcp_sk(sk);
const struct inet_connection_sock *icsk = inet_csk(sk);
@@ -3019,6 +3023,10 @@
if (skb && (TCP_SKB_CB(skb)->sacked & TCPCB_SACKED_ACKED))
flag |= FLAG_SACK_RENEGING;
+ if (tcp_ack_update_rtt(sk, flag, seq_rtt, sack_rtt) ||
+ (flag & FLAG_ACKED))
+ tcp_rearm_rto(sk);
+
if (flag & FLAG_ACKED) {
const struct tcp_congestion_ops *ca_ops
= inet_csk(sk)->icsk_ca_ops;
@@ -3028,9 +3036,6 @@
tcp_mtup_probe_success(sk);
}
- tcp_ack_update_rtt(sk, flag, seq_rtt);
- tcp_rearm_rto(sk);
-
if (tcp_is_reno(tp)) {
tcp_remove_reno_sacks(sk, pkts_acked);
} else {
@@ -3339,7 +3344,7 @@
/* See if we can take anything off of the retransmit queue. */
acked = tp->packets_out;
- flag |= tcp_clean_rtx_queue(sk, prior_fackets, prior_snd_una);
+ flag |= tcp_clean_rtx_queue(sk, prior_fackets, prior_snd_una, sack_rtt);
acked -= tp->packets_out;
if (tcp_ack_is_dubious(sk, flag)) {
