3.18-stable review patch. If anyone has any objections, please let me know.
------------------
From: Eric Dumazet <edumazet@xxxxxxxxxx>
[ Upstream commit 9a9c9b51e54618861420093ae6e9b50a961914c5 ]
We want to add finer control of the number of ACK packets sent after
ECN events.
This patch is not changing current behavior, it only enables following
change.
Signed-off-by: Eric Dumazet <edumazet@xxxxxxxxxx>
Acked-by: Soheil Hassas Yeganeh <soheil@xxxxxxxxxx>
Acked-by: Neal Cardwell <ncardwell@xxxxxxxxxx>
Signed-off-by: David S. Miller <davem@xxxxxxxxxxxxx>
Signed-off-by: Greg Kroah-Hartman <gregkh@xxxxxxxxxxxxxxxxxxx>
---
include/net/tcp.h | 2 +-
net/ipv4/tcp_dctcp.c | 4 ++--
net/ipv4/tcp_input.c | 24 +++++++++++++-----------
3 files changed, 16 insertions(+), 14 deletions(-)
--- a/include/net/tcp.h
+++ b/include/net/tcp.h
@@ -372,7 +372,7 @@ ssize_t tcp_splice_read(struct socket *s
struct pipe_inode_info *pipe, size_t len,
unsigned int flags);
-void tcp_enter_quickack_mode(struct sock *sk);
+void tcp_enter_quickack_mode(struct sock *sk, unsigned int max_quickacks);
static inline void tcp_dec_quickack_mode(struct sock *sk,
const unsigned int pkts)
{
--- a/net/ipv4/tcp_dctcp.c
+++ b/net/ipv4/tcp_dctcp.c
@@ -135,7 +135,7 @@ static void dctcp_ce_state_0_to_1(struct
*/
if (inet_csk(sk)->icsk_ack.pending & ICSK_ACK_TIMER)
__tcp_send_ack(sk, ca->prior_rcv_nxt);
- tcp_enter_quickack_mode(sk);
+ tcp_enter_quickack_mode(sk, 1);
}
ca->prior_rcv_nxt = tp->rcv_nxt;
@@ -156,7 +156,7 @@ static void dctcp_ce_state_1_to_0(struct
*/
if (inet_csk(sk)->icsk_ack.pending & ICSK_ACK_TIMER)
__tcp_send_ack(sk, ca->prior_rcv_nxt);
- tcp_enter_quickack_mode(sk);
+ tcp_enter_quickack_mode(sk, 1);
}
ca->prior_rcv_nxt = tp->rcv_nxt;
--- a/net/ipv4/tcp_input.c
+++ b/net/ipv4/tcp_input.c
@@ -171,21 +171,23 @@ static void tcp_measure_rcv_mss(struct s
}
}
-static void tcp_incr_quickack(struct sock *sk)
+static void tcp_incr_quickack(struct sock *sk, unsigned int max_quickacks)
{
struct inet_connection_sock *icsk = inet_csk(sk);
unsigned int quickacks = tcp_sk(sk)->rcv_wnd / (2 * icsk->icsk_ack.rcv_mss);
if (quickacks == 0)
quickacks = 2;
+ quickacks = min(quickacks, max_quickacks);
if (quickacks > icsk->icsk_ack.quick)
- icsk->icsk_ack.quick = min(quickacks, TCP_MAX_QUICKACKS);
+ icsk->icsk_ack.quick = quickacks;
}
-void tcp_enter_quickack_mode(struct sock *sk)
+void tcp_enter_quickack_mode(struct sock *sk, unsigned int max_quickacks)
{
struct inet_connection_sock *icsk = inet_csk(sk);
- tcp_incr_quickack(sk);
+
+ tcp_incr_quickack(sk, max_quickacks);
icsk->icsk_ack.pingpong = 0;
icsk->icsk_ack.ato = TCP_ATO_MIN;
}
@@ -228,7 +230,7 @@ static void __tcp_ecn_check_ce(struct tc
* it is probably a retransmit.
*/
if (tp->ecn_flags & TCP_ECN_SEEN)
- tcp_enter_quickack_mode((struct sock *)tp);
+ tcp_enter_quickack_mode((struct sock *)tp, TCP_MAX_QUICKACKS);
break;
case INET_ECN_CE:
if (tcp_ca_needs_ecn((struct sock *)tp))
@@ -236,7 +238,7 @@ static void __tcp_ecn_check_ce(struct tc
if (!(tp->ecn_flags & TCP_ECN_DEMAND_CWR)) {
/* Better not delay acks, sender can have a very low cwnd */
- tcp_enter_quickack_mode((struct sock *)tp);
+ tcp_enter_quickack_mode((struct sock *)tp, TCP_MAX_QUICKACKS);
tp->ecn_flags |= TCP_ECN_DEMAND_CWR;
}
tp->ecn_flags |= TCP_ECN_SEEN;
@@ -644,7 +646,7 @@ static void tcp_event_data_recv(struct s
/* The _first_ data packet received, initialize
* delayed ACK engine.
*/
- tcp_incr_quickack(sk);
+ tcp_incr_quickack(sk, TCP_MAX_QUICKACKS);
icsk->icsk_ack.ato = TCP_ATO_MIN;
} else {
int m = now - icsk->icsk_ack.lrcvtime;
@@ -660,7 +662,7 @@ static void tcp_event_data_recv(struct s
/* Too long gap. Apparently sender failed to
* restart window, so that we send ACKs quickly.
*/
- tcp_incr_quickack(sk);
+ tcp_incr_quickack(sk, TCP_MAX_QUICKACKS);
sk_mem_reclaim(sk);
}
}
@@ -4001,7 +4003,7 @@ static void tcp_send_dupack(struct sock
if (TCP_SKB_CB(skb)->end_seq != TCP_SKB_CB(skb)->seq &&
before(TCP_SKB_CB(skb)->seq, tp->rcv_nxt)) {
NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_DELAYEDACKLOST);
- tcp_enter_quickack_mode(sk);
+ tcp_enter_quickack_mode(sk, TCP_MAX_QUICKACKS);
if (tcp_is_sack(tp) && sysctl_tcp_dsack) {
u32 end_seq = TCP_SKB_CB(skb)->end_seq;
@@ -4501,7 +4503,7 @@ queue_and_out:
tcp_dsack_set(sk, TCP_SKB_CB(skb)->seq, TCP_SKB_CB(skb)->end_seq);
out_of_window:
- tcp_enter_quickack_mode(sk);
+ tcp_enter_quickack_mode(sk, TCP_MAX_QUICKACKS);
inet_csk_schedule_ack(sk);
drop:
__kfree_skb(skb);
@@ -5521,7 +5523,7 @@ static int tcp_rcv_synsent_state_process
* to stand against the temptation 8) --ANK
*/
inet_csk_schedule_ack(sk);
- tcp_enter_quickack_mode(sk);
+ tcp_enter_quickack_mode(sk, TCP_MAX_QUICKACKS);
inet_csk_reset_xmit_timer(sk, ICSK_TIME_DACK,
TCP_DELACK_MAX, TCP_RTO_MAX);
