Skip to contentSkip to conversation

tcp: fix double-counted receiver RTT when leaving receiver fast path

PATCH In review Neal Cardwell 3 participants 3 messages Mar 4, 2013

Subsystems (3)
lore
Share & tools
lore

Neal Cardwell

author maintainer
#0 lore raw

PATCH 1/1 tcp: fix double-counted receiver RTT when leaving receiver fast path

We should not update ts_recent and call tcp_rcv_rtt_measure_ts() both before and after going to step5. That wastes CPU and double-counts the receiver-side RTT sample.

---
 net/ipv4/tcp_input.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
net/ipv4/tcp_input.c +3 -3
index a759e19..0d9bdac 100644
@@ -5485,6 +5485,9 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb,
 				if (tcp_checksum_complete_user(sk, skb))
 					goto csum_error;
 
+				if ((int)skb->truesize > sk->sk_forward_alloc)
+					goto step5;
+
 				/* Predicted packet is in window by definition.
 				 * seq == rcv_nxt and rcv_wup <= rcv_nxt.
 				 * Hence, check seq<=rcv_wup reduces to:
@@ -5496,9 +5499,6 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb,
 
 				tcp_rcv_rtt_measure_ts(sk, skb);
 
-				if ((int)skb->truesize > sk->sk_forward_alloc)
-					goto step5;
-
 				NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPHPHITS);
 
 				/* Bulk data transfer: receiver */
Signature · 1 line
1.8.1.3
Acked-by on v1 1/1
#1 lore raw
Quoted 24 lines · Neal Cardwell

We should not update ts_recent and call tcp_rcv_rtt_measure_ts() both before and after going to step5. That wastes CPU and double-counts the receiver-side RTT sample.

---
 net/ipv4/tcp_input.c | 6 +++---
 1 file changed, 3 insertions(+), 3 deletions(-)
net/ipv4/tcp_input.c +3 -0
index a759e19..0d9bdac 100644
@@ -5485,6 +5485,9 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb,
 				if (tcp_checksum_complete_user(sk, skb))
 					goto csum_error;
 
+				if ((int)skb->truesize > sk->sk_forward_alloc)
+					goto step5;
+
 				/* Predicted packet is in window by definition.
 				 * seq == rcv_nxt and rcv_wup <= rcv_nxt.
 				 * Hence, check seq<=rcv_wup reduces to:

net/ipv4/tcp_input.c int tcp_rcv_established(struct sock *sk, struct sk_buff *skb, · line 5504

Quoted 9 lines
net/ipv4/tcp_input.c +0 -3
@@ -5496,9 +5499,6 @@ int tcp_rcv_established(struct sock *sk, struct sk_buff *skb,
 
 				tcp_rcv_rtt_measure_ts(sk, skb);
 
-				if ((int)skb->truesize > sk->sk_forward_alloc)
-					goto step5;
-
 				NET_INC_STATS_BH(sock_net(sk), LINUX_MIB_TCPHPHITS);
 
 				/* Bulk data transfer: receiver */