Commit 6bcfd7f8 authored Oct 08, 2015 by Eric Dumazet Committed by David S. Miller Oct 11, 2015

tcp: fix RFS vs lockless listeners



Before recent TCP listener patches, we were updating listener
sk->sk_rxhash before the cloning of master socket.

children sk_rxhash was therefore correct after the normal 3WHS.

But with lockless listener, we no longer dirty/change listener sk_rxhash
as it would be racy.

We need to correctly update the child sk_rxhash, otherwise first data
packet wont hit correct cpu if RFS is used.

Fixes: 079096f1 ("tcp/dccp: install syn_recv requests into ehash table")
Signed-off-by: Eric Dumazet <edumazet@google.com>
Reported-by: Willem de Bruijn <willemb@google.com>
Cc: Tom Herbert <tom@herbertland.com>
Acked-by: Tom Herbert <tom@herbertland.com>
Signed-off-by: David S. Miller <davem@davemloft.net>

parent 5010ea59

net/ipv4/syncookies.c

+1 −0

Original line number	Diff line number	Diff line
		@@ -225,6 +225,7 @@ struct sock tcp_get_cookie_sock(struct sock sk, struct sk_buff *skb,
		child = icsk->icsk_af_ops->syn_recv_sock(sk, skb, req, dst);
		if (child) {
		atomic_set(&req->rsk_refcnt, 1);
		sock_rps_save_rxhash(child, skb);
		inet_csk_reqsk_queue_add(sk, req, child);
		} else {
		reqsk_free(req);

net/ipv4/tcp_minisocks.c

+1 −0

Original line number	Diff line number	Diff line
		@@ -768,6 +768,7 @@ struct sock tcp_check_req(struct sock sk, struct sk_buff *skb,
		if (!child)
		goto listen_overflow;

		sock_rps_save_rxhash(child, skb);
		tcp_synack_rtt_meas(child, req);
		inet_csk_reqsk_queue_drop(sk, req);
		inet_csk_reqsk_queue_add(sk, req, child);