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=-16.6 required=3.0 tests=DKIMWL_WL_MED,DKIM_SIGNED, DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS,INCLUDES_PATCH, MAILING_LIST_MULTI,SIGNED_OFF_BY,SPF_PASS,URIBL_BLOCKED,USER_AGENT_GIT, USER_IN_DEF_DKIM_WL autolearn=ham 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 A3761C10F13 for ; Thu, 11 Apr 2019 12:55:32 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [209.132.180.67]) by mail.kernel.org (Postfix) with ESMTP id 6940C217F4 for ; Thu, 11 Apr 2019 12:55:32 +0000 (UTC) Authentication-Results: mail.kernel.org; dkim=pass (2048-bit key) header.d=google.com header.i=@google.com header.b="JgyHmQTu" Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S1726603AbfDKMzb (ORCPT ); Thu, 11 Apr 2019 08:55:31 -0400 Received: from mail-ot1-f74.google.com ([209.85.210.74]:44869 "EHLO mail-ot1-f74.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1726014AbfDKMzb (ORCPT ); Thu, 11 Apr 2019 08:55:31 -0400 Received: by mail-ot1-f74.google.com with SMTP id w3so2830141otg.11 for ; Thu, 11 Apr 2019 05:55:30 -0700 (PDT) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20161025; h=date:message-id:mime-version:subject:from:to:cc; bh=B8s8kiCB+oeDeA+/KtEWG5OSeWBD+EBomltNKl3HdAQ=; b=JgyHmQTuBLT+H0M319tnTE6teyqRKw0WnRxSqQNzp7d7NoNTVsgONFU7lrgR9DO8BT Ke+M6xk/LQztvmVT+aJhnAY0p89W31eYLaZSrJycfSjgJ9WVcrV1pbz7j/3GNWgMPOex imYLf5o2hj/cr32UQRlQ0cFB1gM13toOjAKrZQH9JG22ZAe8v2pN0KN/7FbMlz2mVgm1 u6bL2ZRVnGhT/A0jziH9bwp68h+yIGLxKkN5RzgPFOs8zjpv6ms+lKOVkc65EZKzshBj X5oyBNVhrTKqYXMkWfe68j9FqmSE4bgfZBPM5JYE0PcSiV3RV7XMzQFCd3LgezCzmIP2 7YEg== X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=1e100.net; s=20161025; h=x-gm-message-state:date:message-id:mime-version:subject:from:to:cc; bh=B8s8kiCB+oeDeA+/KtEWG5OSeWBD+EBomltNKl3HdAQ=; b=YRpHGCVOey4FL1gbse2hIxd0hT/IG0/Hh1AMPQrGegyeQGxHhGMjRsRbemwPPofIfN xXzEQwkGG2QDHWA4t64glCPIB7MDPUqgZBVzDHjjPjYzRag80uQQWK2R00TTtHJJArzy tj3iK63uViMX/mKzE/524AkPGq+D/du1/Z0jqg71WrzGmm6rJKykn3vHLdxWH3DCFVEX XD+9pC3kbcVUU1e7zEQ3puuUKPH40CQmxgvObPpJ450WhgeXqqJ1KxInMADZ5KBenuzR MWNiAdRfv9QimA/Rh/+HlLdyMwymnbSmFE5pCYwOaKkBjTYx4WTsuGYjCv29jn+QZ/1L z8zg== X-Gm-Message-State: APjAAAUf9/he+ZC0HypXl9w+78OaxBLEwAXv9ATTOQ3quMqa0GFpU3GY YVSC5ifiaDbaKeo+YhOUSv0ppQNcafCxFg== X-Google-Smtp-Source: APXvYqytzX1lMTJ0D1DZ8J+3UfD+tMJAL44TXijXIX58kyULY/y8LGLg8rqhizj99Dx0IM9z8ByHaKtg2ke0yg== X-Received: by 2002:aca:cc8c:: with SMTP id c134mr4154217oig.3.1554987329890; Thu, 11 Apr 2019 05:55:29 -0700 (PDT) Date: Thu, 11 Apr 2019 05:55:23 -0700 Message-Id: <20190411125523.1613-1-edumazet@google.com> Mime-Version: 1.0 X-Mailer: git-send-email 2.21.0.392.gf8f6787159e-goog Subject: [PATCH net] dctcp: more accurate tracking of packets delivery From: Eric Dumazet To: "David S . Miller" Cc: netdev , Eric Dumazet , Eric Dumazet , Yuchung Cheng , Neal Cardwell , Soheil Hassas Yeganeh , Florian Westphal , Daniel Borkmann , Lawrence Brakmo , Abdul Kabbani Content-Type: text/plain; charset="UTF-8" Sender: netdev-owner@vger.kernel.org Precedence: bulk List-ID: X-Mailing-List: netdev@vger.kernel.org After commit e21db6f69a95 ("tcp: track total bytes delivered with ECN CE marks") core TCP stack does a very good job tracking ECN signals. The "sender's best estimate of CE information" Yuchung mentioned in his patch is indeed the best we can do. DCTCP can use tp->delivered_ce and tp->delivered to not duplicate the logic, and use the existing best estimate. This solves some problems, since current DCTCP logic does not deal with losses and/or GRO or ack aggregation very well. This also removes a dubious use of inet_csk(sk)->icsk_ack.rcv_mss (this should have been tp->mss_cache), and a 64 bit divide. Finally, we can see that the DCTCP logic, calling dctcp_update_alpha() for every ACK could be done differently, calling it only once per RTT. Signed-off-by: Eric Dumazet Cc: Yuchung Cheng Cc: Neal Cardwell Cc: Soheil Hassas Yeganeh Cc: Florian Westphal Cc: Daniel Borkmann Cc: Lawrence Brakmo Cc: Abdul Kabbani --- net/ipv4/tcp_dctcp.c | 45 +++++++++++++++++--------------------------- 1 file changed, 17 insertions(+), 28 deletions(-) diff --git a/net/ipv4/tcp_dctcp.c b/net/ipv4/tcp_dctcp.c index 359da68d7c0628360d5f1b727c878f678e28d39a..477cb4aa456c11c70185a982cbadafba857d3619 100644 --- a/net/ipv4/tcp_dctcp.c +++ b/net/ipv4/tcp_dctcp.c @@ -49,9 +49,8 @@ #define DCTCP_MAX_ALPHA 1024U struct dctcp { - u32 acked_bytes_ecn; - u32 acked_bytes_total; - u32 prior_snd_una; + u32 old_delivered; + u32 old_delivered_ce; u32 prior_rcv_nxt; u32 dctcp_alpha; u32 next_seq; @@ -73,8 +72,8 @@ static void dctcp_reset(const struct tcp_sock *tp, struct dctcp *ca) { ca->next_seq = tp->snd_nxt; - ca->acked_bytes_ecn = 0; - ca->acked_bytes_total = 0; + ca->old_delivered = tp->delivered; + ca->old_delivered_ce = tp->delivered_ce; } static void dctcp_init(struct sock *sk) @@ -86,7 +85,6 @@ static void dctcp_init(struct sock *sk) sk->sk_state == TCP_CLOSE)) { struct dctcp *ca = inet_csk_ca(sk); - ca->prior_snd_una = tp->snd_una; ca->prior_rcv_nxt = tp->rcv_nxt; ca->dctcp_alpha = min(dctcp_alpha_on_init, DCTCP_MAX_ALPHA); @@ -118,37 +116,25 @@ static void dctcp_update_alpha(struct sock *sk, u32 flags) { const struct tcp_sock *tp = tcp_sk(sk); struct dctcp *ca = inet_csk_ca(sk); - u32 acked_bytes = tp->snd_una - ca->prior_snd_una; - - /* If ack did not advance snd_una, count dupack as MSS size. - * If ack did update window, do not count it at all. - */ - if (acked_bytes == 0 && !(flags & CA_ACK_WIN_UPDATE)) - acked_bytes = inet_csk(sk)->icsk_ack.rcv_mss; - if (acked_bytes) { - ca->acked_bytes_total += acked_bytes; - ca->prior_snd_una = tp->snd_una; - - if (flags & CA_ACK_ECE) - ca->acked_bytes_ecn += acked_bytes; - } /* Expired RTT */ if (!before(tp->snd_una, ca->next_seq)) { - u64 bytes_ecn = ca->acked_bytes_ecn; + u32 delivered_ce = tp->delivered_ce - ca->old_delivered_ce; u32 alpha = ca->dctcp_alpha; /* alpha = (1 - g) * alpha + g * F */ alpha -= min_not_zero(alpha, alpha >> dctcp_shift_g); - if (bytes_ecn) { + if (delivered_ce) { + u32 delivered = tp->delivered - ca->old_delivered; + /* If dctcp_shift_g == 1, a 32bit value would overflow - * after 8 Mbytes. + * after 8 M packets. */ - bytes_ecn <<= (10 - dctcp_shift_g); - do_div(bytes_ecn, max(1U, ca->acked_bytes_total)); + delivered_ce <<= (10 - dctcp_shift_g); + delivered_ce /= max(1U, delivered); - alpha = min(alpha + (u32)bytes_ecn, DCTCP_MAX_ALPHA); + alpha = min(alpha + delivered_ce, DCTCP_MAX_ALPHA); } /* dctcp_alpha can be read from dctcp_get_info() without * synchro, so we ask compiler to not use dctcp_alpha @@ -200,6 +186,7 @@ static size_t dctcp_get_info(struct sock *sk, u32 ext, int *attr, union tcp_cc_info *info) { const struct dctcp *ca = inet_csk_ca(sk); + const struct tcp_sock *tp = tcp_sk(sk); /* Fill it also in case of VEGASINFO due to req struct limits. * We can still correctly retrieve it later. @@ -211,8 +198,10 @@ static size_t dctcp_get_info(struct sock *sk, u32 ext, int *attr, info->dctcp.dctcp_enabled = 1; info->dctcp.dctcp_ce_state = (u16) ca->ce_state; info->dctcp.dctcp_alpha = ca->dctcp_alpha; - info->dctcp.dctcp_ab_ecn = ca->acked_bytes_ecn; - info->dctcp.dctcp_ab_tot = ca->acked_bytes_total; + info->dctcp.dctcp_ab_ecn = tp->mss_cache * + (tp->delivered_ce - ca->old_delivered_ce); + info->dctcp.dctcp_ab_tot = tp->mss_cache * + (tp->delivered - ca->old_delivered); } *attr = INET_DIAG_DCTCPINFO; -- 2.21.0.392.gf8f6787159e-goog