tcp: RFC2988bis + taking RTT sample from 3WHS for the passive open side

This patch lowers the default initRTO from 3secs to 1sec per
RFC2988bis. It falls back to 3secs if the SYN or SYN-ACK packet
has been retransmitted, AND the TCP timestamp option is not on.

It also adds support to take RTT sample during 3WHS on the passive
open side, just like its active open counterpart, and uses it, if
valid, to seed the initRTO for the data transmission phase.

The patch also resets ssthresh to its initial default at the
beginning of the data transmission phase, and reduces cwnd to 1 if
there has been MORE THAN ONE retransmission during 3WHS per RFC5681.

Signed-off-by: H.K. Jerry Chu <hkchu@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

6 files changed, 45 insertions, 25 deletions

diff --git a/net/ipv4/syncookies.c b/net/ipv4/syncookies.c
index 26461492a847..92bb9434b338 100644
--- a/net/ipv4/syncookies.c
+++ b/net/ipv4/syncookies.c
@@ -316,6 +316,7 @@ struct sock *cookie_v4_check(struct sock *sk, struct sk_buff *skb,
         ireq->wscale_ok         = tcp_opt.wscale_ok;
         ireq->tstamp_ok         = tcp_opt.saw_tstamp;
         req->ts_recent          = tcp_opt.saw_tstamp ? tcp_opt.rcv_tsval : 0;
+        treq->snt_synack        = tcp_opt.saw_tstamp ? tcp_opt.rcv_tsecr : 0;
 
         /* We throwed the options of the initial SYN away, so we hope
          * the ACK carries the same options again (see RFC1122 4.2.3.8)
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index bef9f04c22ba..ea0d2183df4b 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -880,6 +880,11 @@ static void tcp_init_metrics(struct sock *sk)
                 tp->snd_ssthresh = dst_metric(dst, RTAX_SSTHRESH);
                 if (tp->snd_ssthresh > tp->snd_cwnd_clamp)
                         tp->snd_ssthresh = tp->snd_cwnd_clamp;
+        } else {
+                /* ssthresh may have been reduced unnecessarily during.
+                 * 3WHS. Restore it back to its initial default.
+                 */
+                tp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
         }
         if (dst_metric(dst, RTAX_REORDERING) &&
             tp->reordering != dst_metric(dst, RTAX_REORDERING)) {
@@ -887,10 +892,7 @@ static void tcp_init_metrics(struct sock *sk)
                 tp->reordering = dst_metric(dst, RTAX_REORDERING);
         }
 
-        if (dst_metric(dst, RTAX_RTT) == 0)
-                goto reset;
-
-        if (!tp->srtt && dst_metric_rtt(dst, RTAX_RTT) < (TCP_TIMEOUT_INIT << 3))
+        if (dst_metric(dst, RTAX_RTT) == 0 || tp->srtt == 0)
                 goto reset;
 
         /* Initial rtt is determined from SYN,SYN-ACK.
@@ -916,19 +918,26 @@ static void tcp_init_metrics(struct sock *sk)
                 tp->mdev_max = tp->rttvar = max(tp->mdev, tcp_rto_min(sk));
         }
         tcp_set_rto(sk);
-        if (inet_csk(sk)->icsk_rto < TCP_TIMEOUT_INIT && !tp->rx_opt.saw_tstamp) {
 reset:
-                /* Play conservative. If timestamps are not
-                 * supported, TCP will fail to recalculate correct
-                 * rtt, if initial rto is too small. FORGET ALL AND RESET!
+        if (tp->srtt == 0) {
+                /* RFC2988bis: We've failed to get a valid RTT sample from
+                 * 3WHS. This is most likely due to retransmission,
+                 * including spurious one. Reset the RTO back to 3secs
+                 * from the more aggressive 1sec to avoid more spurious
+                 * retransmission.
                  */
-                if (!tp->rx_opt.saw_tstamp && tp->srtt) {
-                        tp->srtt = 0;
-                        tp->mdev = tp->mdev_max = tp->rttvar = TCP_TIMEOUT_INIT;
-                        inet_csk(sk)->icsk_rto = TCP_TIMEOUT_INIT;
-                }
+                tp->mdev = tp->mdev_max = tp->rttvar = TCP_TIMEOUT_FALLBACK;
+                inet_csk(sk)->icsk_rto = TCP_TIMEOUT_FALLBACK;
         }
-        tp->snd_cwnd = tcp_init_cwnd(tp, dst);
+        /* Cut cwnd down to 1 per RFC5681 if SYN or SYN-ACK has been
+         * retransmitted. In light of RFC2988bis' more aggressive 1sec
+         * initRTO, we only reset cwnd when more than 1 SYN/SYN-ACK
+         * retransmission has occurred.
+         */
+        if (tp->total_retrans > 1)
+                tp->snd_cwnd = 1;
+        else
+                tp->snd_cwnd = tcp_init_cwnd(tp, dst);
         tp->snd_cwnd_stamp = tcp_time_stamp;
 }
 
@@ -3112,12 +3121,13 @@ static void tcp_fastretrans_alert(struct sock *sk, int pkts_acked, int flag)
         tcp_xmit_retransmit_queue(sk);
 }
 
-static void tcp_valid_rtt_meas(struct sock *sk, u32 seq_rtt)
+void tcp_valid_rtt_meas(struct sock *sk, u32 seq_rtt)
 {
         tcp_rtt_estimator(sk, seq_rtt);
         tcp_set_rto(sk);
         inet_csk(sk)->icsk_backoff = 0;
 }
+EXPORT_SYMBOL(tcp_valid_rtt_meas);
 
 /* Read draft-ietf-tcplw-high-performance before mucking
  * with this code. (Supersedes RFC1323)
@@ -5806,12 +5816,6 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
                                               tp->rx_opt.snd_wscale;
                                 tcp_init_wl(tp, TCP_SKB_CB(skb)->seq);
 
-                                /* tcp_ack considers this ACK as duplicate
-                                 * and does not calculate rtt.
-                                 * Force it here.
-                                 */
-                                tcp_ack_update_rtt(sk, 0, 0);
-
                                 if (tp->rx_opt.tstamp_ok)
                                         tp->advmss -= TCPOLEN_TSTAMP_ALIGNED;
 
diff --git a/net/ipv4/tcp_ipv4.c b/net/ipv4/tcp_ipv4.c
index a7d6671e33b8..617dee3ccfb1 100644
--- a/net/ipv4/tcp_ipv4.c
+++ b/net/ipv4/tcp_ipv4.c
@@ -429,8 +429,8 @@ void tcp_v4_err(struct sk_buff *icmp_skb, u32 info)
                         break;
 
                 icsk->icsk_backoff--;
-                inet_csk(sk)->icsk_rto = __tcp_set_rto(tp) <<
-                                         icsk->icsk_backoff;
+                inet_csk(sk)->icsk_rto = (tp->srtt ? __tcp_set_rto(tp) :
+                        TCP_TIMEOUT_INIT) << icsk->icsk_backoff;
                 tcp_bound_rto(sk);
 
                 skb = tcp_write_queue_head(sk);
@@ -1384,6 +1384,7 @@ int tcp_v4_conn_request(struct sock *sk, struct sk_buff *skb)
                 isn = tcp_v4_init_sequence(skb);
         }
         tcp_rsk(req)->snt_isn = isn;
+        tcp_rsk(req)->snt_synack = tcp_time_stamp;
 
         if (tcp_v4_send_synack(sk, dst, req,
                                (struct request_values *)&tmp_ext) ||
@@ -1458,6 +1459,10 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
                 newtp->advmss = tcp_sk(sk)->rx_opt.user_mss;
 
         tcp_initialize_rcv_mss(newsk);
+        if (tcp_rsk(req)->snt_synack)
+                tcp_valid_rtt_meas(newsk,
+                    tcp_time_stamp - tcp_rsk(req)->snt_synack);
+        newtp->total_retrans = req->retrans;
 
 #ifdef CONFIG_TCP_MD5SIG
         /* Copy over the MD5 key from the original socket */
@@ -1854,7 +1859,7 @@ static int tcp_v4_init_sock(struct sock *sk)
          * algorithms that we must have the following bandaid to talk
          * efficiently to them.  -DaveM
          */
-        tp->snd_cwnd = 2;
+        tp->snd_cwnd = TCP_INIT_CWND;
 
         /* See draft-stevens-tcpca-spec-01 for discussion of the
          * initialization of these values.
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index 80b1f80759ab..d2fe4e06b472 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -486,7 +486,7 @@ struct sock *tcp_create_openreq_child(struct sock *sk, struct request_sock *req,
                  * algorithms that we must have the following bandaid to talk
                  * efficiently to them.  -DaveM
                  */
-                newtp->snd_cwnd = 2;
+                newtp->snd_cwnd = TCP_INIT_CWND;
                 newtp->snd_cwnd_cnt = 0;
                 newtp->bytes_acked = 0;
 
@@ -720,6 +720,10 @@ struct sock *tcp_check_req(struct sock *sk, struct sk_buff *skb,
                 NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPDEFERACCEPTDROP);
                 return NULL;
         }
+        if (tmp_opt.saw_tstamp && tmp_opt.rcv_tsecr)
+                tcp_rsk(req)->snt_synack = tmp_opt.rcv_tsecr;
+        else if (req->retrans) /* don't take RTT sample if retrans && ~TS */
+                tcp_rsk(req)->snt_synack = 0;
 
         /* OK, ACK is valid, create big socket and
          * feed this segment to it. It will repeat all
diff --git a/net/ipv6/syncookies.c b/net/ipv6/syncookies.c
index 8b9644a8b697..89d5bf806222 100644
--- a/net/ipv6/syncookies.c
+++ b/net/ipv6/syncookies.c
@@ -223,6 +223,7 @@ struct sock *cookie_v6_check(struct sock *sk, struct sk_buff *skb)
         ireq->wscale_ok         = tcp_opt.wscale_ok;
         ireq->tstamp_ok         = tcp_opt.saw_tstamp;
         req->ts_recent          = tcp_opt.saw_tstamp ? tcp_opt.rcv_tsval : 0;
+        treq->snt_synack        = tcp_opt.saw_tstamp ? tcp_opt.rcv_tsecr : 0;
         treq->rcv_isn = ntohl(th->seq) - 1;
         treq->snt_isn = cookie;
 
diff --git a/net/ipv6/tcp_ipv6.c b/net/ipv6/tcp_ipv6.c
index d1fd28711ba5..a1ef61a889c3 100644
--- a/net/ipv6/tcp_ipv6.c
+++ b/net/ipv6/tcp_ipv6.c
@@ -1341,6 +1341,7 @@ static int tcp_v6_conn_request(struct sock *sk, struct sk_buff *skb)
         }
 have_isn:
         tcp_rsk(req)->snt_isn = isn;
+        tcp_rsk(req)->snt_synack = tcp_time_stamp;
 
         security_inet_conn_request(sk, skb, req);
 
@@ -1509,6 +1510,10 @@ static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
         tcp_sync_mss(newsk, dst_mtu(dst));
         newtp->advmss = dst_metric_advmss(dst);
         tcp_initialize_rcv_mss(newsk);
+        if (tcp_rsk(req)->snt_synack)
+                tcp_valid_rtt_meas(newsk,
+                    tcp_time_stamp - tcp_rsk(req)->snt_synack);
+        newtp->total_retrans = req->retrans;
 
         newinet->inet_daddr = newinet->inet_saddr = LOOPBACK4_IPV6;
         newinet->inet_rcv_saddr = LOOPBACK4_IPV6;