tcp: TLP loss detection.
This is the second of the TLP patch series; it augments the basic TLP
algorithm with a loss detection scheme.
This patch implements a mechanism for loss detection when a Tail
loss probe retransmission plugs a hole thereby masking packet loss
from the sender. The loss detection algorithm relies on counting
TLP dupacks as outlined in Sec. 3 of:
http://tools.ietf.org/html/draft-dukkipati-tcpm-tcp-loss-probe-01
The basic idea is: Sender keeps track of TLP "episode" upon
retransmission of a TLP packet. An episode ends when the sender receives
an ACK above the SND.NXT (tracked by tlp_high_seq) at the time of the
episode. We want to make sure that before the episode ends the sender
receives a "TLP dupack", indicating that the TLP retransmission was
unnecessary, so there was no loss/hole that needed plugging. If the
sender gets no TLP dupack before the end of the episode, then it reduces
ssthresh and the congestion window, because the TLP packet arriving at
the receiver probably plugged a hole.
Signed-off-by: Nandita Dukkipati <nanditad@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
diff --git a/net/ipv4/proc.c b/net/ipv4/proc.c
index 4c35911..b6f2ea1 100644
--- a/net/ipv4/proc.c
+++ b/net/ipv4/proc.c
@@ -225,6 +225,7 @@
SNMP_MIB_ITEM("TCPSlowStartRetrans", LINUX_MIB_TCPSLOWSTARTRETRANS),
SNMP_MIB_ITEM("TCPTimeouts", LINUX_MIB_TCPTIMEOUTS),
SNMP_MIB_ITEM("TCPLossProbes", LINUX_MIB_TCPLOSSPROBES),
+ SNMP_MIB_ITEM("TCPLossProbeRecovery", LINUX_MIB_TCPLOSSPROBERECOVERY),
SNMP_MIB_ITEM("TCPRenoRecoveryFail", LINUX_MIB_TCPRENORECOVERYFAIL),
SNMP_MIB_ITEM("TCPSackRecoveryFail", LINUX_MIB_TCPSACKRECOVERYFAIL),
SNMP_MIB_ITEM("TCPSchedulerFailed", LINUX_MIB_TCPSCHEDULERFAILED),
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index b794f89..836d74d 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -2682,6 +2682,7 @@
struct tcp_sock *tp = tcp_sk(sk);
tp->high_seq = tp->snd_nxt;
+ tp->tlp_high_seq = 0;
tp->snd_cwnd_cnt = 0;
tp->prior_cwnd = tp->snd_cwnd;
tp->prr_delivered = 0;
@@ -3569,6 +3570,38 @@
}
}
+/* This routine deals with acks during a TLP episode.
+ * Ref: loss detection algorithm in draft-dukkipati-tcpm-tcp-loss-probe.
+ */
+static void tcp_process_tlp_ack(struct sock *sk, u32 ack, int flag)
+{
+ struct tcp_sock *tp = tcp_sk(sk);
+ bool is_tlp_dupack = (ack == tp->tlp_high_seq) &&
+ !(flag & (FLAG_SND_UNA_ADVANCED |
+ FLAG_NOT_DUP | FLAG_DATA_SACKED));
+
+ /* Mark the end of TLP episode on receiving TLP dupack or when
+ * ack is after tlp_high_seq.
+ */
+ if (is_tlp_dupack) {
+ tp->tlp_high_seq = 0;
+ return;
+ }
+
+ if (after(ack, tp->tlp_high_seq)) {
+ tp->tlp_high_seq = 0;
+ /* Don't reduce cwnd if DSACK arrives for TLP retrans. */
+ if (!(flag & FLAG_DSACKING_ACK)) {
+ tcp_init_cwnd_reduction(sk, true);
+ tcp_set_ca_state(sk, TCP_CA_CWR);
+ tcp_end_cwnd_reduction(sk);
+ tcp_set_ca_state(sk, TCP_CA_Open);
+ NET_INC_STATS_BH(sock_net(sk),
+ LINUX_MIB_TCPLOSSPROBERECOVERY);
+ }
+ }
+}
+
/* This routine deals with incoming acks, but not outgoing ones. */
static int tcp_ack(struct sock *sk, const struct sk_buff *skb, int flag)
{
@@ -3676,6 +3709,9 @@
tcp_cong_avoid(sk, ack, prior_in_flight);
}
+ if (tp->tlp_high_seq)
+ tcp_process_tlp_ack(sk, ack, flag);
+
if ((flag & FLAG_FORWARD_PROGRESS) || !(flag & FLAG_NOT_DUP)) {
struct dst_entry *dst = __sk_dst_get(sk);
if (dst)
@@ -3697,6 +3733,9 @@
*/
if (tcp_send_head(sk))
tcp_ack_probe(sk);
+
+ if (tp->tlp_high_seq)
+ tcp_process_tlp_ack(sk, ack, flag);
return 1;
invalid_ack:
diff --git a/net/ipv4/tcp_minisocks.c b/net/ipv4/tcp_minisocks.c
index b83a49c..4bdb09f 100644
--- a/net/ipv4/tcp_minisocks.c
+++ b/net/ipv4/tcp_minisocks.c
@@ -440,6 +440,7 @@
newtp->fackets_out = 0;
newtp->snd_ssthresh = TCP_INFINITE_SSTHRESH;
tcp_enable_early_retrans(newtp);
+ newtp->tlp_high_seq = 0;
/* So many TCP implementations out there (incorrectly) count the
* initial SYN frame in their delayed-ACK and congestion control
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index beb63db..8e7742f 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -2132,6 +2132,7 @@
*/
void tcp_send_loss_probe(struct sock *sk)
{
+ struct tcp_sock *tp = tcp_sk(sk);
struct sk_buff *skb;
int pcount;
int mss = tcp_current_mss(sk);
@@ -2142,6 +2143,10 @@
goto rearm_timer;
}
+ /* At most one outstanding TLP retransmission. */
+ if (tp->tlp_high_seq)
+ goto rearm_timer;
+
/* Retransmit last segment. */
skb = tcp_write_queue_tail(sk);
if (WARN_ON(!skb))
@@ -2164,6 +2169,10 @@
if (skb->len > 0)
err = __tcp_retransmit_skb(sk, skb);
+ /* Record snd_nxt for loss detection. */
+ if (likely(!err))
+ tp->tlp_high_seq = tp->snd_nxt;
+
rearm_timer:
inet_csk_reset_xmit_timer(sk, ICSK_TIME_RETRANS,
inet_csk(sk)->icsk_rto,
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index ecd61d5..eeccf79 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -356,6 +356,8 @@
WARN_ON(tcp_write_queue_empty(sk));
+ tp->tlp_high_seq = 0;
+
if (!tp->snd_wnd && !sock_flag(sk, SOCK_DEAD) &&
!((1 << sk->sk_state) & (TCPF_SYN_SENT | TCPF_SYN_RECV))) {
/* Receiver dastardly shrinks window. Our retransmits