tcp: sk_add_backlog() is too agressive for TCP

While investigating TCP performance problems on 10Gb+ links, we found a
tcp sender was dropping lot of incoming ACKS because of sk_rcvbuf limit
in sk_add_backlog(), especially if receiver doesnt use GRO/LRO and sends
one ACK every two MSS segments.

A sender usually tweaks sk_sndbuf, but sk_rcvbuf stays at its default
value (87380), allowing a too small backlog.

A TCP ACK, even being small, can consume nearly same truesize space than
outgoing packets. Using sk_rcvbuf + sk_sndbuf as a limit makes sense and
is fast to compute.

Performance results on netperf, single flow, receiver with disabled
GRO/LRO : 7500 Mbits instead of 6050 Mbits, no more TCPBacklogDrop
increments at sender.

Signed-off-by: Eric Dumazet <edumazet@google.com>
Cc: Neal Cardwell <ncardwell@google.com>
Cc: Tom Herbert <therbert@google.com>
Cc: Maciej Żenczykowski <maze@google.com>
Cc: Yuchung Cheng <ycheng@google.com>
Cc: Ilpo Järvinen <ilpo.jarvinen@helsinki.fi>
Cc: Rick Jones <rick.jones2@hp.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

net/ipv4/tcp_ipv4.c

@@ -1752,7 +1752,8 @@ int tcp_v4_rcv(struct sk_buff *skb)
 			if (!tcp_prequeue(sk, skb))
 				ret = tcp_v4_do_rcv(sk, skb);
 		}
-	} else if (unlikely(sk_add_backlog(sk, skb, sk->sk_rcvbuf))) {
+	} else if (unlikely(sk_add_backlog(sk, skb,
+					   sk->sk_rcvbuf + sk->sk_sndbuf))) {
 		bh_unlock_sock(sk);
 		NET_INC_STATS_BH(net, LINUX_MIB_TCPBACKLOGDROP);
 		goto discard_and_relse;

net/ipv6/tcp_ipv6.c

@@ -1654,7 +1654,8 @@ static int tcp_v6_rcv(struct sk_buff *skb)
 			if (!tcp_prequeue(sk, skb))
 				ret = tcp_v6_do_rcv(sk, skb);
 		}
-	} else if (unlikely(sk_add_backlog(sk, skb, sk->sk_rcvbuf))) {
+	} else if (unlikely(sk_add_backlog(sk, skb,
+					   sk->sk_rcvbuf + sk->sk_sndbuf))) {
 		bh_unlock_sock(sk);
 		NET_INC_STATS_BH(net, LINUX_MIB_TCPBACKLOGDROP);
 		goto discard_and_relse;