tcp: switch snt_synack back to measuring transmit time of first SYNACK
authorNeal Cardwell <ncardwell@google.com>
Mon, 30 Jun 2014 19:09:49 +0000 (15:09 -0400)
committerDavid S. Miller <davem@davemloft.net>
Tue, 8 Jul 2014 02:26:37 +0000 (19:26 -0700)
Always store in snt_synack the time at which the server received the
first client SYN and attempted to send the first SYNACK.

Recent commit aa27fc501 ("tcp: tcp_v[46]_conn_request: fix snt_synack
initialization") resolved an inconsistency between IPv4 and IPv6 in
the initialization of snt_synack. This commit brings back the idea
from 843f4a55e (tcp: use tcp_v4_send_synack on first SYN-ACK), which
was going for the original behavior of snt_synack from the commit
where it was added in 9ad7c049f0f79 ("tcp: RFC2988bis + taking RTT
sample from 3WHS for the passive open side") in v3.1.

In addition to being simpler (and probably a tiny bit faster),
unconditionally storing the time of the first SYNACK attempt has been
useful because it allows calculating a performance metric quantifying
how long it took to establish a passive TCP connection.

Signed-off-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: Yuchung Cheng <ycheng@google.com>
Cc: Octavian Purdila <octavian.purdila@intel.com>
Cc: Jerry Chu <hkchu@google.com>
Acked-by: Octavian Purdila <octavian.purdila@intel.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
include/net/tcp.h
net/ipv4/tcp_ipv4.c
net/ipv6/tcp_ipv6.c

index 0d5389aecf182fecf4c43054f7548dd275be8411..c9a75dbba0c7fc01a0971ebcbc3fb477a53da461 100644 (file)
@@ -1093,7 +1093,7 @@ static inline void tcp_openreq_init(struct request_sock *req,
        req->cookie_ts = 0;
        tcp_rsk(req)->rcv_isn = TCP_SKB_CB(skb)->seq;
        tcp_rsk(req)->rcv_nxt = TCP_SKB_CB(skb)->seq + 1;
-       tcp_rsk(req)->snt_synack = 0;
+       tcp_rsk(req)->snt_synack = tcp_time_stamp;
        req->mss = rx_opt->mss_clamp;
        req->ts_recent = rx_opt->saw_tstamp ? rx_opt->rcv_tsval : 0;
        ireq->tstamp_ok = rx_opt->tstamp_ok;
index 5dfebd2f2e382269f5327cef7f53ea1bc68a95fc..52d0f6a1ec2c9b1b87310037c817f94a98238c82 100644 (file)
@@ -838,8 +838,6 @@ static int tcp_v4_send_synack(struct sock *sk, struct dst_entry *dst,
                                            ireq->ir_rmt_addr,
                                            ireq->opt);
                err = net_xmit_eval(err);
-               if (!tcp_rsk(req)->snt_synack && !err)
-                       tcp_rsk(req)->snt_synack = tcp_time_stamp;
        }
 
        return err;
index bc24ee21339a4a6c10e2f314e1a95ce78e26351c..a97c95585da8debd5960c4e8b15b9c923329513c 100644 (file)
@@ -498,8 +498,6 @@ static int tcp_v6_send_synack(struct sock *sk, struct dst_entry *dst,
                skb_set_queue_mapping(skb, queue_mapping);
                err = ip6_xmit(sk, skb, fl6, np->opt, np->tclass);
                err = net_xmit_eval(err);
-               if (!tcp_rsk(req)->snt_synack && !err)
-                       tcp_rsk(req)->snt_synack = tcp_time_stamp;
        }
 
 done: