tcp: delete redundant calls of tcp_mtup_init() and

I find that both tcp_v4_syn_recv_sock() and tcp_v6_syn_recv_sock() call them for new sock, so delete redundant calls of them.

Weiping Pan (2):
tcp: delete redundant calls of tcp_mtup_init()
tcp: delete redundant call of tcp_initialize_rcv_mss()

 net/ipv4/tcp_input.c |    2 --
 net/ipv4/tcp_ipv4.c  |    1 -
 2 files changed, 0 insertions(+), 3 deletions(-)

1.7.4

As both tcp_v4_syn_recv_sock() and tcp_v6_syn_recv_sock() already call tcp_mtup_init() for new sock, the calls of it in tcp_rcv_state_process() and tcp_v4_conn_req_fastopen() are redundant.

---
 net/ipv4/tcp_input.c |    1 -
 net/ipv4/tcp_ipv4.c  |    1 -
 2 files changed, 0 insertions(+), 2 deletions(-)

index c53b7f3..8add6bc 100644
@@ -5731,7 +5731,6 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
 			icsk->icsk_af_ops->rebuild_header(sk);
 			tcp_init_congestion_control(sk);
 
-			tcp_mtup_init(sk);
 			tp->copied_seq = tp->rcv_nxt;
 			tcp_init_buffer_space(sk);
 		}

index 0672139..6578047 100644
@@ -1410,7 +1410,6 @@ static int tcp_v4_conn_req_fastopen(struct sock *sk,
 	/* Now finish processing the fastopen child socket. */
 	inet_csk(child)->icsk_af_ops->rebuild_header(child);
 	tcp_init_congestion_control(child);
-	tcp_mtup_init(child);
 	tcp_init_metrics(child);
 	tcp_init_buffer_space(child);
 

1.7.4

As both tcp_v4_syn_recv_sock() and tcp_v6_syn_recv_sock() already call tcp_initialize_rcv_mss() for new sock, the call of it in tcp_rcv_state_process() is redundant.

---
 net/ipv4/tcp_input.c |    1 -
 1 files changed, 0 insertions(+), 1 deletions(-)

index 8add6bc..8a2b5c7 100644
@@ -5771,7 +5771,6 @@ int tcp_rcv_state_process(struct sock *sk, struct sk_buff *skb,
 		/* Prevent spurious tcp_cwnd_restart() on first data packet */
 		tp->lsndtime = tcp_time_stamp;
 
-		tcp_initialize_rcv_mss(sk);
 		tcp_fast_path_on(tp);
 		break;
 

1.7.4

From: Weiping Pan <[email protected]>
Date: Sun, 12 Jan 2014 15:54:29 +0800

I do not agree with the approach of these two patches.

It is better to have tcp_rcv_state_process() (one location) make these calls rather than each and every inet sock operations instance.

Therefore you should remove the calls from tcp_v{4,6}_recv_sock() and keep the one in tcp_rcv_state_process().

Thanks.

As tcp_rcv_state_process() has already calls tcp_mtup_init() for non-fastopen sock, we can delete the redundant calls of tcp_mtup_init() in tcp_{v4,v6}_syn_recv_sock().

---
 net/ipv4/tcp_ipv4.c |    1 -
 net/ipv6/tcp_ipv6.c |    1 -
 2 files changed, 0 insertions(+), 2 deletions(-)

index 0672139..4176606 100644
@@ -1668,7 +1668,6 @@ struct sock *tcp_v4_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
 	}
 	sk_setup_caps(newsk, dst);
 
-	tcp_mtup_init(newsk);
 	tcp_sync_mss(newsk, dst_mtu(dst));
 	newtp->advmss = dst_metric_advmss(dst);
 	if (tcp_sk(sk)->rx_opt.user_mss &&

index f67033b..c5b0e1f 100644
@@ -1230,7 +1230,6 @@ static struct sock * tcp_v6_syn_recv_sock(struct sock *sk, struct sk_buff *skb,
 		inet_csk(newsk)->icsk_ext_hdr_len = (newnp->opt->opt_nflen +
 						     newnp->opt->opt_flen);
 
-	tcp_mtup_init(newsk);
 	tcp_sync_mss(newsk, dst_mtu(dst));
 	newtp->advmss = dst_metric_advmss(dst);
 	if (tcp_sk(sk)->rx_opt.user_mss &&

1.7.4.4

From: Weiping Pan <[email protected]>
Date: Sun, 19 Jan 2014 20:44:46 +0800

Looks good, applied.