From cadf9df27e7cf40e390e060a1c71bb86ecde798b Mon Sep 17 00:00:00 2001
From: Stephen Mallon <stephen.mallon@sydney.edu.au>
Date: Tue, 20 Nov 2018 19:15:02 +1100
Subject: tcp: Fix SOF_TIMESTAMPING_RX_HARDWARE to use the latest timestamp
during TCP coalescing
During tcp coalescing ensure that the skb hardware timestamp refers to the
highest sequence number data.
Previously only the software timestamp was updated during coalescing.
Signed-off-by: Stephen Mallon <stephen.mallon@sydney.edu.au>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/ipv4/tcp_input.c | 1 +
1 file changed, 1 insertion(+)
(limited to 'net/ipv4/tcp_input.c')
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 2868ef28ce52..e695584bb33f 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -4363,6 +4363,7 @@ static bool tcp_try_coalesce(struct sock *sk,
if (TCP_SKB_CB(from)->has_rxtstamp) {
TCP_SKB_CB(to)->has_rxtstamp = true;
to->tstamp = from->tstamp;
+ skb_hwtstamps(to)->hwtstamp = skb_hwtstamps(from)->hwtstamp;
}
return true;
--
cgit
From 86de5921a3d5dd246df661e09bdd0a6131b39ae3 Mon Sep 17 00:00:00 2001
From: Eric Dumazet <edumazet@google.com>
Date: Tue, 20 Nov 2018 05:53:59 -0800
Subject: tcp: defer SACK compression after DupThresh
Jean-Louis reported a TCP regression and bisected to recent SACK
compression.
After a loss episode (receiver not able to keep up and dropping
packets because its backlog is full), linux TCP stack is sending
a single SACK (DUPACK).
Sender waits a full RTO timer before recovering losses.
While RFC 6675 says in section 5, "Algorithm Details",
(2) If DupAcks < DupThresh but IsLost (HighACK + 1) returns true --
indicating at least three segments have arrived above the current
cumulative acknowledgment point, which is taken to indicate loss
-- go to step (4).
...
(4) Invoke fast retransmit and enter loss recovery as follows:
there are old TCP stacks not implementing this strategy, and
still counting the dupacks before starting fast retransmit.
While these stacks probably perform poorly when receivers implement
LRO/GRO, we should be a little more gentle to them.
This patch makes sure we do not enable SACK compression unless
3 dupacks have been sent since last rcv_nxt update.
Ideally we should even rearm the timer to send one or two
more DUPACK if no more packets are coming, but that will
be work aiming for linux-4.21.
Many thanks to Jean-Louis for bisecting the issue, providing
packet captures and testing this patch.
Fixes: 5d9f4262b7ea ("tcp: add SACK compression")
Reported-by: Jean-Louis Dupond <jean-louis@dupond.be>
Tested-by: Jean-Louis Dupond <jean-louis@dupond.be>
Signed-off-by: Eric Dumazet <edumazet@google.com>
Acked-by: Neal Cardwell <ncardwell@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
include/linux/tcp.h | 1 +
net/ipv4/tcp_input.c | 14 ++++++++++++--
net/ipv4/tcp_output.c | 6 +++---
net/ipv4/tcp_timer.c | 2 +-
4 files changed, 17 insertions(+), 6 deletions(-)
(limited to 'net/ipv4/tcp_input.c')
diff --git a/include/linux/tcp.h b/include/linux/tcp.h
index 8ed77bb4ed86..a9b0280687d5 100644
--- a/include/linux/tcp.h
+++ b/include/linux/tcp.h
@@ -196,6 +196,7 @@ struct tcp_sock {
u32 rcv_tstamp; /* timestamp of last received ACK (for keepalives) */
u32 lsndtime; /* timestamp of last sent data packet (for restart window) */
u32 last_oow_ack_time; /* timestamp of last out-of-window ACK */
+ u32 compressed_ack_rcv_nxt;
u32 tsoffset; /* timestamp offset */
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index e695584bb33f..1e37c1388189 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -4268,7 +4268,7 @@ static void tcp_sack_new_ofo_skb(struct sock *sk, u32 seq, u32 end_seq)
* If the sack array is full, forget about the last one.
*/
if (this_sack >= TCP_NUM_SACKS) {
- if (tp->compressed_ack)
+ if (tp->compressed_ack > TCP_FASTRETRANS_THRESH)
tcp_send_ack(sk);
this_sack--;
tp->rx_opt.num_sacks--;
@@ -5189,7 +5189,17 @@ send_now:
if (!tcp_is_sack(tp) ||
tp->compressed_ack >= sock_net(sk)->ipv4.sysctl_tcp_comp_sack_nr)
goto send_now;
- tp->compressed_ack++;
+
+ if (tp->compressed_ack_rcv_nxt != tp->rcv_nxt) {
+ tp->compressed_ack_rcv_nxt = tp->rcv_nxt;
+ if (tp->compressed_ack > TCP_FASTRETRANS_THRESH)
+ NET_ADD_STATS(sock_net(sk), LINUX_MIB_TCPACKCOMPRESSED,
+ tp->compressed_ack - TCP_FASTRETRANS_THRESH);
+ tp->compressed_ack = 0;
+ }
+
+ if (++tp->compressed_ack <= TCP_FASTRETRANS_THRESH)
+ goto send_now;
if (hrtimer_is_queued(&tp->compressed_ack_timer))
return;
diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c
index 9c34b97d365d..3f510cad0b3e 100644
--- a/net/ipv4/tcp_output.c
+++ b/net/ipv4/tcp_output.c
@@ -180,10 +180,10 @@ static inline void tcp_event_ack_sent(struct sock *sk, unsigned int pkts,
{
struct tcp_sock *tp = tcp_sk(sk);
- if (unlikely(tp->compressed_ack)) {
+ if (unlikely(tp->compressed_ack > TCP_FASTRETRANS_THRESH)) {
NET_ADD_STATS(sock_net(sk), LINUX_MIB_TCPACKCOMPRESSED,
- tp->compressed_ack);
- tp->compressed_ack = 0;
+ tp->compressed_ack - TCP_FASTRETRANS_THRESH);
+ tp->compressed_ack = TCP_FASTRETRANS_THRESH;
if (hrtimer_try_to_cancel(&tp->compressed_ack_timer) == 1)
__sock_put(sk);
}
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 676020663ce8..5f8b6d3cd855 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -740,7 +740,7 @@ static enum hrtimer_restart tcp_compressed_ack_kick(struct hrtimer *timer)
bh_lock_sock(sk);
if (!sock_owned_by_user(sk)) {
- if (tp->compressed_ack)
+ if (tp->compressed_ack > TCP_FASTRETRANS_THRESH)
tcp_send_ack(sk);
} else {
if (!test_and_set_bit(TCP_DELACK_TIMER_DEFERRED,
--
cgit
From 9efdda4e3abed13f0903b7b6e4d4c2102019440a Mon Sep 17 00:00:00 2001
From: Eric Dumazet <edumazet@google.com>
Date: Sat, 24 Nov 2018 09:12:24 -0800
Subject: tcp: address problems caused by EDT misshaps
When a qdisc setup including pacing FQ is dismantled and recreated,
some TCP packets are sent earlier than instructed by TCP stack.
TCP can be fooled when ACK comes back, because the following
operation can return a negative value.
tcp_time_stamp(tp) - tp->rx_opt.rcv_tsecr;
Some paths in TCP stack were not dealing properly with this,
this patch addresses four of them.
Fixes: ab408b6dc744 ("tcp: switch tcp and sch_fq to new earliest departure time model")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Signed-off-by: David S. Miller <davem@davemloft.net>
---
net/ipv4/tcp_input.c | 16 ++++++++++------
net/ipv4/tcp_timer.c | 10 ++++++----
2 files changed, 16 insertions(+), 10 deletions(-)
(limited to 'net/ipv4/tcp_input.c')
diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c
index 1e37c1388189..a9d9555a973f 100644
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -579,10 +579,12 @@ static inline void tcp_rcv_rtt_measure_ts(struct sock *sk,
u32 delta = tcp_time_stamp(tp) - tp->rx_opt.rcv_tsecr;
u32 delta_us;
- if (!delta)
- delta = 1;
- delta_us = delta * (USEC_PER_SEC / TCP_TS_HZ);
- tcp_rcv_rtt_update(tp, delta_us, 0);
+ if (likely(delta < INT_MAX / (USEC_PER_SEC / TCP_TS_HZ))) {
+ if (!delta)
+ delta = 1;
+ delta_us = delta * (USEC_PER_SEC / TCP_TS_HZ);
+ tcp_rcv_rtt_update(tp, delta_us, 0);
+ }
}
}
@@ -2910,9 +2912,11 @@ static bool tcp_ack_update_rtt(struct sock *sk, const int flag,
if (seq_rtt_us < 0 && tp->rx_opt.saw_tstamp && tp->rx_opt.rcv_tsecr &&
flag & FLAG_ACKED) {
u32 delta = tcp_time_stamp(tp) - tp->rx_opt.rcv_tsecr;
- u32 delta_us = delta * (USEC_PER_SEC / TCP_TS_HZ);
- seq_rtt_us = ca_rtt_us = delta_us;
+ if (likely(delta < INT_MAX / (USEC_PER_SEC / TCP_TS_HZ))) {
+ seq_rtt_us = delta * (USEC_PER_SEC / TCP_TS_HZ);
+ ca_rtt_us = seq_rtt_us;
+ }
}
rs->rtt_us = ca_rtt_us; /* RTT of last (S)ACKed packet (or -1) */
if (seq_rtt_us < 0)
diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c
index 5f8b6d3cd855..091c53925e4d 100644
--- a/net/ipv4/tcp_timer.c
+++ b/net/ipv4/tcp_timer.c
@@ -40,15 +40,17 @@ static u32 tcp_clamp_rto_to_user_timeout(const struct sock *sk)
{
struct inet_connection_sock *icsk = inet_csk(sk);
u32 elapsed, start_ts;
+ s32 remaining;
start_ts = tcp_retransmit_stamp(sk);
if (!icsk->icsk_user_timeout || !start_ts)
return icsk->icsk_rto;
elapsed = tcp_time_stamp(tcp_sk(sk)) - start_ts;
- if (elapsed >= icsk->icsk_user_timeout)
+ remaining = icsk->icsk_user_timeout - elapsed;
+ if (remaining <= 0)
return 1; /* user timeout has passed; fire ASAP */
- else
- return min_t(u32, icsk->icsk_rto, msecs_to_jiffies(icsk->icsk_user_timeout - elapsed));
+
+ return min_t(u32, icsk->icsk_rto, msecs_to_jiffies(remaining));
}
/**
@@ -209,7 +211,7 @@ static bool retransmits_timed_out(struct sock *sk,
(boundary - linear_backoff_thresh) * TCP_RTO_MAX;
timeout = jiffies_to_msecs(timeout);
}
- return (tcp_time_stamp(tcp_sk(sk)) - start_ts) >= timeout;
+ return (s32)(tcp_time_stamp(tcp_sk(sk)) - start_ts - timeout) >= 0;
}
/* A write timeout has occurred. Process the after effects. */
--
cgit