From patchwork Fri Feb 13 08:16:44 2015 Content-Type: text/plain; charset="utf-8" MIME-Version: 1.0 Content-Transfer-Encoding: 7bit X-Patchwork-Submitter: Fan Du X-Patchwork-Id: 439433 X-Patchwork-Delegate: davem@davemloft.net Return-Path: X-Original-To: patchwork-incoming@ozlabs.org Delivered-To: patchwork-incoming@ozlabs.org Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by ozlabs.org (Postfix) with ESMTP id 76E9B1402D1 for ; Fri, 13 Feb 2015 19:21:34 +1100 (AEDT) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1752328AbbBMIVY (ORCPT ); Fri, 13 Feb 2015 03:21:24 -0500 Received: from mga03.intel.com ([134.134.136.65]:7254 "EHLO mga03.intel.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1751655AbbBMIVX (ORCPT ); Fri, 13 Feb 2015 03:21:23 -0500 Received: from orsmga001.jf.intel.com ([10.7.209.18]) by orsmga103.jf.intel.com with ESMTP; 13 Feb 2015 00:16:15 -0800 X-ExtLoop1: 1 X-IronPort-AV: E=Sophos;i="5.09,569,1418112000"; d="scan'208";a="651654078" Received: from dufan-optiplex-9010.bj.intel.com ([10.238.155.116]) by orsmga001.jf.intel.com with ESMTP; 13 Feb 2015 00:21:14 -0800 From: Fan Du To: davem@davemloft.net Cc: netdev@vger.kernel.org, fengyuleidian0615@gmail.com Subject: [PATCH net-next 2/3] ipv4: Use binary search to choose tcp PMTU probe_size Date: Fri, 13 Feb 2015 16:16:44 +0800 Message-Id: <1423815405-32644-3-git-send-email-fan.du@intel.com> X-Mailer: git-send-email 1.7.9.5 In-Reply-To: <1423815405-32644-1-git-send-email-fan.du@intel.com> References: <1423815405-32644-1-git-send-email-fan.du@intel.com> Sender: netdev-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org Current probe_size is chosen by doubling mss_cache, the initial mss base is 512 Bytes, as a result the converged probe_size will only be 1024 Bytes, there is still big gap between 1024 and common 1500 bytes of mtu. Use binary search to choose probe_size in a fine granularity manner, an optimal mss will be found to boost performance as its maxmium. Test env: Docker instance with vxlan encapuslation(82599EB) iperf -c 10.0.0.24 -t 60 before this patch: 1.26 Gbits/sec After this patch: increase 26% 1.59 Gbits/sec Signed-off-by: Fan Du --- include/net/inet_connection_sock.h | 3 +++ net/ipv4/tcp_input.c | 5 ++++- net/ipv4/tcp_output.c | 12 +++++++++--- net/ipv4/tcp_timer.c | 2 +- 4 files changed, 17 insertions(+), 5 deletions(-) diff --git a/include/net/inet_connection_sock.h b/include/net/inet_connection_sock.h index 5976bde..3d0932e 100644 --- a/include/net/inet_connection_sock.h +++ b/include/net/inet_connection_sock.h @@ -124,6 +124,9 @@ struct inet_connection_sock { int search_high; int search_low; + int search_high_sav; + int search_low_sav; + /* Information on the current probe. */ int probe_size; } icsk_mtup; diff --git a/net/ipv4/tcp_input.c b/net/ipv4/tcp_input.c index 8fdd27b..20b28e9 100644 --- a/net/ipv4/tcp_input.c +++ b/net/ipv4/tcp_input.c @@ -2613,7 +2613,10 @@ static void tcp_mtup_probe_success(struct sock *sk) tp->snd_cwnd_stamp = tcp_time_stamp; tp->snd_ssthresh = tcp_current_ssthresh(sk); - icsk->icsk_mtup.search_low = icsk->icsk_mtup.probe_size; + if (icsk->icsk_mtup.search_low == icsk->icsk_mtup.probe_size) + icsk->icsk_mtup.search_low = icsk->icsk_mtup.search_high; + else + icsk->icsk_mtup.search_low = icsk->icsk_mtup.probe_size; icsk->icsk_mtup.probe_size = 0; tcp_sync_mss(sk, icsk->icsk_pmtu_cookie); } diff --git a/net/ipv4/tcp_output.c b/net/ipv4/tcp_output.c index a2a796c..0a60deb 100644 --- a/net/ipv4/tcp_output.c +++ b/net/ipv4/tcp_output.c @@ -1349,10 +1349,13 @@ void tcp_mtup_init(struct sock *sk) struct inet_connection_sock *icsk = inet_csk(sk); struct net *net = sock_net(sk); - icsk->icsk_mtup.enabled = net->ipv4.sysctl_tcp_mtu_probing > 1; + icsk->icsk_mtup.enabled = net->ipv4.sysctl_tcp_mtu_probing; icsk->icsk_mtup.search_high = tp->rx_opt.mss_clamp + sizeof(struct tcphdr) + icsk->icsk_af_ops->net_header_len; icsk->icsk_mtup.search_low = tcp_mss_to_mtu(sk, net->ipv4.sysctl_tcp_base_mss); + + icsk->icsk_mtup.search_high_sav = icsk->icsk_mtup.search_high; + icsk->icsk_mtup.search_low_sav = icsk->icsk_mtup.search_low; icsk->icsk_mtup.probe_size = 0; } EXPORT_SYMBOL(tcp_mtup_init); @@ -1854,9 +1857,12 @@ static int tcp_mtu_probe(struct sock *sk) tp->rx_opt.num_sacks || tp->rx_opt.dsack) return -1; - /* Very simple search strategy: just double the MSS. */ + /* Use binary search for probe_size bewteen tcp_mss_base + * and current mss_clamp. + */ mss_now = tcp_current_mss(sk); - probe_size = 2 * tp->mss_cache; + + probe_size = (icsk->icsk_mtup.search_high + icsk->icsk_mtup.search_low) >> 1; size_needed = probe_size + (tp->reordering + 1) * tp->mss_cache; if (probe_size > tcp_mtu_to_mss(sk, icsk->icsk_mtup.search_high)) { /* TODO: set timer for probe_converge_event */ diff --git a/net/ipv4/tcp_timer.c b/net/ipv4/tcp_timer.c index 0732b78..9d1cfe0 100644 --- a/net/ipv4/tcp_timer.c +++ b/net/ipv4/tcp_timer.c @@ -113,7 +113,7 @@ static void tcp_mtu_probing(struct inet_connection_sock *icsk, struct sock *sk) struct tcp_sock *tp = tcp_sk(sk); int mss; - mss = tcp_mtu_to_mss(sk, icsk->icsk_mtup.search_low) >> 1; + mss = tcp_mtu_to_mss(sk, icsk->icsk_mtup.search_low); mss = min(net->ipv4.sysctl_tcp_base_mss, mss); mss = max(mss, 68 - tp->tcp_header_len); icsk->icsk_mtup.search_low = tcp_mss_to_mtu(sk, mss);