tcp: adjust tail loss probe timeout

This patch adjusts the timeout formula to schedule the TCP loss probe
(TLP). The previous formula uses 2*SRTT or 1.5*RTT + DelayACKMax if
only one packet is in flight. It keeps a lower bound of 10 msec which
is too large for short RTT connections (e.g. within a data-center).
The new formula = 2*RTT + (inflight == 1 ? 200ms : 2ticks) which
performs better for short and fast connections.

Signed-off-by: Yuchung Cheng <ycheng@google.com>
Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

diff --git a/include/net/tcp.h b/include/net/tcp.h
index 70483296157f87acdf5acd5e96eaa910119ba220..4f056ea79df2aa6d6f05a9d1894db1ed9d18b4e5 100644 (file)
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -139,6 +139,7 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 #endif
 #define TCP_RTO_MAX    ((unsigned)(120*HZ))
 #define TCP_RTO_MIN    ((unsigned)(HZ/5))
+#define TCP_TIMEOUT_MIN        (2U) /* Min timeout for TCP timers in jiffies */
 #define TCP_TIMEOUT_INIT ((unsigned)(1*HZ))    /* RFC6298 2.1 initial RTO value        */
 #define TCP_TIMEOUT_FALLBACK ((unsigned)(3*HZ))        /* RFC 1122 initial RTO value, now
                                                 * used as a fallback RTO for the
@@ -150,8 +151,6 @@ void tcp_time_wait(struct sock *sk, int state, int timeo);
 #define TCP_RESOURCE_PROBE_INTERVAL ((unsigned)(HZ/2U)) /* Maximal interval between probes
                                                         * for local resources.
                                                         */
-#define TCP_REO_TIMEOUT_MIN    (2000) /* Min RACK reordering timeout in usec */
-
 #define TCP_KEEPALIVE_TIME     (120*60*HZ)     /* two hours */
 #define TCP_KEEPALIVE_PROBES   9               /* Max of 9 keepalive probes    */
 #define TCP_KEEPALIVE_INTVL    (75*HZ)

diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 4e985dea1dd24fdecfbf9b47d51cff698e97cd2f..886d874775df3f4f6bebe17c93845a519e5b276b 100644 (file)
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -2377,7 +2377,6 @@ bool tcp_schedule_loss_probe(struct sock *sk)
        struct inet_connection_sock *icsk = inet_csk(sk);
        struct tcp_sock *tp = tcp_sk(sk);
        u32 timeout, tlp_time_stamp, rto_time_stamp;
-       u32 rtt = usecs_to_jiffies(tp->srtt_us >> 3);
 
        /* No consecutive loss probes. */
        if (WARN_ON(icsk->icsk_pending == ICSK_TIME_LOSS_PROBE)) {
@@ -2406,15 +2405,19 @@ bool tcp_schedule_loss_probe(struct sock *sk)
             tcp_send_head(sk))
                return false;
 
-       /* Probe timeout is at least 1.5*rtt + TCP_DELACK_MAX to account
+       /* Probe timeout is 2*rtt. Add minimum RTO to account
         * for delayed ack when there's one outstanding packet. If no RTT
         * sample is available then probe after TCP_TIMEOUT_INIT.
         */
-       timeout = rtt << 1 ? : TCP_TIMEOUT_INIT;
-       if (tp->packets_out == 1)
-               timeout = max_t(u32, timeout,
-                               (rtt + (rtt >> 1) + TCP_DELACK_MAX));
-       timeout = max_t(u32, timeout, msecs_to_jiffies(10));
+       if (tp->srtt_us) {
+               timeout = usecs_to_jiffies(tp->srtt_us >> 2);
+               if (tp->packets_out == 1)
+                       timeout += TCP_RTO_MIN;
+               else
+                       timeout += TCP_TIMEOUT_MIN;
+       } else {
+               timeout = TCP_TIMEOUT_INIT;
+       }
 
        /* If RTO is shorter, just schedule TLP in its place. */
        tlp_time_stamp = tcp_jiffies32 + timeout;

diff --git a/net/ipv4/tcp_recovery.c b/net/ipv4/tcp_recovery.c
index fe9a493d02082d3830f37854d5f169f769844ffb..449cd914d58e1df8352fd9180fbdb5b0feef683c 100644 (file)
--- a/net/ipv4/tcp_recovery.c
+++ b/net/ipv4/tcp_recovery.c
@@ -113,7 +113,7 @@ void tcp_rack_mark_lost(struct sock *sk)
        tp->rack.advanced = 0;
        tcp_rack_detect_loss(sk, &timeout);
        if (timeout) {
-               timeout = usecs_to_jiffies(timeout + TCP_REO_TIMEOUT_MIN);
+               timeout = usecs_to_jiffies(timeout) + TCP_TIMEOUT_MIN;
                inet_csk_reset_xmit_timer(sk, ICSK_TIME_REO_TIMEOUT,
                                          timeout, inet_csk(sk)->icsk_rto);
        }