From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: X-Spam-Checker-Version: SpamAssassin 3.4.0 (2014-02-07) on aws-us-west-2-korg-lkml-1.web.codeaurora.org X-Spam-Level: X-Spam-Status: No, score=-6.0 required=3.0 tests=DKIMWL_WL_HIGH,DKIM_SIGNED, DKIM_VALID,HEADER_FROM_DIFFERENT_DOMAINS,MAILING_LIST_MULTI,SIGNED_OFF_BY, SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT autolearn=unavailable autolearn_force=no version=3.4.0 Received: from mail.kernel.org (mail.kernel.org [198.145.29.99]) by smtp.lore.kernel.org (Postfix) with ESMTP id C63DDC10F0E for ; Mon, 15 Apr 2019 18:50:06 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 913C12073F for ; Mon, 15 Apr 2019 18:50:06 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1555354206; bh=1/SyIu0OXXX2OtItZtMTZDyFUPZSjpx3GPUXkohNUIQ=; h=From:To:Cc:Subject:Date:In-Reply-To:References:List-ID:From; b=1cs83ESf4s+lZhBd2o7MNUg7DWTAFo0dMFdnELbqCgcfSQxE2YbrurlASw0/i8fef 3CrD0bRuAByZ+vpCs+vTfhl77Pu9iO6fN2JOal41CHFLEoenWxEcafrCQWbZImys25 Alm0lt6fHrPCQsItlvz07s5jdb3zYHDlIwSNy60I= Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1728772AbfDOSq6 (ORCPT ); Mon, 15 Apr 2019 14:46:58 -0400 Received: from mail.kernel.org ([198.145.29.99]:49862 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1728764AbfDOSq5 (ORCPT ); Mon, 15 Apr 2019 14:46:57 -0400 Received: from localhost (83-86-89-107.cable.dynamic.v4.ziggo.nl [83.86.89.107]) (using TLSv1.2 with cipher ECDHE-RSA-AES256-GCM-SHA384 (256/256 bits)) (No client certificate requested) by mail.kernel.org (Postfix) with ESMTPSA id F303A218A1; Mon, 15 Apr 2019 18:46:55 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=kernel.org; s=default; t=1555354016; bh=1/SyIu0OXXX2OtItZtMTZDyFUPZSjpx3GPUXkohNUIQ=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=xV5TU11AX/9VmDylbmwToVuRpW14+M6CLYqsrxLGx5eXuUhyww6asJ4homosdYLBv 8AenXc4PECnNlMuUhvEvf97Rc3Op5GEeKNqAAzRAz+ghnkLypUbKHCQmJwOLhYiyqy o9dT2vr0c6tHx2QYlLBW9lA/xrVe3xPCRrDdfHsk= From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Koen De Schepper , Olivier Tilmans , Bob Briscoe , Lawrence Brakmo , Florian Westphal , Daniel Borkmann , Yuchung Cheng , Neal Cardwell , Eric Dumazet , Andrew Shewmaker , Glenn Judd , Daniel Borkmann , "David S. Miller" Subject: [PATCH 4.9 55/76] tcp: Ensure DCTCP reacts to losses Date: Mon, 15 Apr 2019 20:44:19 +0200 Message-Id: <20190415183723.537531204@linuxfoundation.org> X-Mailer: git-send-email 2.21.0 In-Reply-To: <20190415183707.712011689@linuxfoundation.org> References: <20190415183707.712011689@linuxfoundation.org> User-Agent: quilt/0.66 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Sender: stable-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: stable@vger.kernel.org From: Koen De Schepper [ Upstream commit aecfde23108b8e637d9f5c5e523b24fb97035dc3 ] RFC8257 ยง3.5 explicitly states that "A DCTCP sender MUST react to loss episodes in the same way as conventional TCP". Currently, Linux DCTCP performs no cwnd reduction when losses are encountered. Optionally, the dctcp_clamp_alpha_on_loss resets alpha to its maximal value if a RTO happens. This behavior is sub-optimal for at least two reasons: i) it ignores losses triggering fast retransmissions; and ii) it causes unnecessary large cwnd reduction in the future if the loss was isolated as it resets the historical term of DCTCP's alpha EWMA to its maximal value (i.e., denoting a total congestion). The second reason has an especially noticeable effect when using DCTCP in high BDP environments, where alpha normally stays at low values. This patch replace the clamping of alpha by setting ssthresh to half of cwnd for both fast retransmissions and RTOs, at most once per RTT. Consequently, the dctcp_clamp_alpha_on_loss module parameter has been removed. The table below shows experimental results where we measured the drop probability of a PIE AQM (not applying ECN marks) at a bottleneck in the presence of a single TCP flow with either the alpha-clamping option enabled or the cwnd halving proposed by this patch. Results using reno or cubic are given for comparison. | Link | RTT | Drop TCP CC | speed | base+AQM | probability ==================|=========|==========|============ CUBIC | 40Mbps | 7+20ms | 0.21% RENO | | | 0.19% DCTCP-CLAMP-ALPHA | | | 25.80% DCTCP-HALVE-CWND | | | 0.22% ------------------|---------|----------|------------ CUBIC | 100Mbps | 7+20ms | 0.03% RENO | | | 0.02% DCTCP-CLAMP-ALPHA | | | 23.30% DCTCP-HALVE-CWND | | | 0.04% ------------------|---------|----------|------------ CUBIC | 800Mbps | 1+1ms | 0.04% RENO | | | 0.05% DCTCP-CLAMP-ALPHA | | | 18.70% DCTCP-HALVE-CWND | | | 0.06% We see that, without halving its cwnd for all source of losses, DCTCP drives the AQM to large drop probabilities in order to keep the queue length under control (i.e., it repeatedly faces RTOs). Instead, if DCTCP reacts to all source of losses, it can then be controlled by the AQM using similar drop levels than cubic or reno. Signed-off-by: Koen De Schepper Signed-off-by: Olivier Tilmans Cc: Bob Briscoe Cc: Lawrence Brakmo Cc: Florian Westphal Cc: Daniel Borkmann Cc: Yuchung Cheng Cc: Neal Cardwell Cc: Eric Dumazet Cc: Andrew Shewmaker Cc: Glenn Judd Acked-by: Florian Westphal Acked-by: Neal Cardwell Acked-by: Daniel Borkmann Signed-off-by: David S. Miller Signed-off-by: Greg Kroah-Hartman --- net/ipv4/tcp_dctcp.c | 36 ++++++++++++++++++------------------ 1 file changed, 18 insertions(+), 18 deletions(-) --- a/net/ipv4/tcp_dctcp.c +++ b/net/ipv4/tcp_dctcp.c @@ -66,11 +66,6 @@ static unsigned int dctcp_alpha_on_init module_param(dctcp_alpha_on_init, uint, 0644); MODULE_PARM_DESC(dctcp_alpha_on_init, "parameter for initial alpha value"); -static unsigned int dctcp_clamp_alpha_on_loss __read_mostly; -module_param(dctcp_clamp_alpha_on_loss, uint, 0644); -MODULE_PARM_DESC(dctcp_clamp_alpha_on_loss, - "parameter for clamping alpha on loss"); - static struct tcp_congestion_ops dctcp_reno; static void dctcp_reset(const struct tcp_sock *tp, struct dctcp *ca) @@ -211,21 +206,23 @@ static void dctcp_update_alpha(struct so } } -static void dctcp_state(struct sock *sk, u8 new_state) +static void dctcp_react_to_loss(struct sock *sk) { - if (dctcp_clamp_alpha_on_loss && new_state == TCP_CA_Loss) { - struct dctcp *ca = inet_csk_ca(sk); + struct dctcp *ca = inet_csk_ca(sk); + struct tcp_sock *tp = tcp_sk(sk); - /* If this extension is enabled, we clamp dctcp_alpha to - * max on packet loss; the motivation is that dctcp_alpha - * is an indicator to the extend of congestion and packet - * loss is an indicator of extreme congestion; setting - * this in practice turned out to be beneficial, and - * effectively assumes total congestion which reduces the - * window by half. - */ - ca->dctcp_alpha = DCTCP_MAX_ALPHA; - } + ca->loss_cwnd = tp->snd_cwnd; + tp->snd_ssthresh = max(tp->snd_cwnd >> 1U, 2U); +} + +static void dctcp_state(struct sock *sk, u8 new_state) +{ + if (new_state == TCP_CA_Recovery && + new_state != inet_csk(sk)->icsk_ca_state) + dctcp_react_to_loss(sk); + /* We handle RTO in dctcp_cwnd_event to ensure that we perform only + * one loss-adjustment per RTT. + */ } static void dctcp_cwnd_event(struct sock *sk, enum tcp_ca_event ev) @@ -237,6 +234,9 @@ static void dctcp_cwnd_event(struct sock case CA_EVENT_ECN_NO_CE: dctcp_ce_state_1_to_0(sk); break; + case CA_EVENT_LOSS: + dctcp_react_to_loss(sk); + break; default: /* Don't care for the rest. */ break;