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=-18.8 required=3.0 tests=BAYES_00,DKIMWL_WL_HIGH, DKIM_SIGNED,DKIM_VALID,DKIM_VALID_AU,HEADER_FROM_DIFFERENT_DOMAINS, INCLUDES_CR_TRAILER,INCLUDES_PATCH,MAILING_LIST_MULTI,SPF_HELO_NONE,SPF_PASS, USER_AGENT_GIT 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 AC154C4332B for ; Mon, 15 Feb 2021 15:35:20 +0000 (UTC) Received: from vger.kernel.org (vger.kernel.org [23.128.96.18]) by mail.kernel.org (Postfix) with ESMTP id 9224F64EFD for ; Mon, 15 Feb 2021 15:35:20 +0000 (UTC) Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S231131AbhBOPfL (ORCPT ); Mon, 15 Feb 2021 10:35:11 -0500 Received: from mail.kernel.org ([198.145.29.99]:46916 "EHLO mail.kernel.org" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S231224AbhBOPdE (ORCPT ); Mon, 15 Feb 2021 10:33:04 -0500 Received: by mail.kernel.org (Postfix) with ESMTPSA id 09E6564EB1; Mon, 15 Feb 2021 15:30:29 +0000 (UTC) DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/simple; d=linuxfoundation.org; s=korg; t=1613403030; bh=j1nNR3ryxJs6LxaHt8kb1VfQ/3ogQuet5q+d/qNmY1E=; h=From:To:Cc:Subject:Date:In-Reply-To:References:From; b=tXZD6qNkLILziRPL/LtRMwZIWb/4wu/aY8slnuAn3eDz5Ntk+t3XTIn54TekwqkhD fscvi0NwliysyUbu+NPZDDwHZ8YyNt+4La3PAIvzSJTM/h/WQQX+BE+Oj3CEfKt1fG lI49z1aSrOtIBdDsnJnqEUUNMtcsRErEoLWkHFsg= From: Greg Kroah-Hartman To: linux-kernel@vger.kernel.org Cc: Greg Kroah-Hartman , stable@vger.kernel.org, Eric Dumazet , Jian Yang , Maxim Mikityanskiy , Saeed Mahameed , Edward Cree , Alexander Lobakin , Jakub Kicinski , John Sperbeck Subject: [PATCH 5.4 52/60] net: gro: do not keep too many GRO packets in napi->rx_list Date: Mon, 15 Feb 2021 16:27:40 +0100 Message-Id: <20210215152717.045112340@linuxfoundation.org> X-Mailer: git-send-email 2.30.1 In-Reply-To: <20210215152715.401453874@linuxfoundation.org> References: <20210215152715.401453874@linuxfoundation.org> User-Agent: quilt/0.66 MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Precedence: bulk List-ID: X-Mailing-List: stable@vger.kernel.org From: Eric Dumazet commit 8dc1c444df193701910f5e80b5d4caaf705a8fb0 upstream. Commit c80794323e82 ("net: Fix packet reordering caused by GRO and listified RX cooperation") had the unfortunate effect of adding latencies in common workloads. Before the patch, GRO packets were immediately passed to upper stacks. After the patch, we can accumulate quite a lot of GRO packets (depdending on NAPI budget). My fix is counting in napi->rx_count number of segments instead of number of logical packets. Fixes: c80794323e82 ("net: Fix packet reordering caused by GRO and listified RX cooperation") Signed-off-by: Eric Dumazet Bisected-by: John Sperbeck Tested-by: Jian Yang Cc: Maxim Mikityanskiy Reviewed-by: Saeed Mahameed Reviewed-by: Edward Cree Reviewed-by: Alexander Lobakin Link: https://lore.kernel.org/r/20210204213146.4192368-1-eric.dumazet@gmail.com Signed-off-by: Jakub Kicinski Signed-off-by: Greg Kroah-Hartman --- net/core/dev.c | 11 ++++++----- 1 file changed, 6 insertions(+), 5 deletions(-) --- a/net/core/dev.c +++ b/net/core/dev.c @@ -5275,10 +5275,11 @@ static void gro_normal_list(struct napi_ /* Queue one GRO_NORMAL SKB up for list processing. If batch size exceeded, * pass the whole batch up to the stack. */ -static void gro_normal_one(struct napi_struct *napi, struct sk_buff *skb) +static void gro_normal_one(struct napi_struct *napi, struct sk_buff *skb, int segs) { list_add_tail(&skb->list, &napi->rx_list); - if (++napi->rx_count >= gro_normal_batch) + napi->rx_count += segs; + if (napi->rx_count >= gro_normal_batch) gro_normal_list(napi); } @@ -5317,7 +5318,7 @@ static int napi_gro_complete(struct napi } out: - gro_normal_one(napi, skb); + gro_normal_one(napi, skb, NAPI_GRO_CB(skb)->count); return NET_RX_SUCCESS; } @@ -5608,7 +5609,7 @@ static gro_result_t napi_skb_finish(stru { switch (ret) { case GRO_NORMAL: - gro_normal_one(napi, skb); + gro_normal_one(napi, skb, 1); break; case GRO_DROP: @@ -5696,7 +5697,7 @@ static gro_result_t napi_frags_finish(st __skb_push(skb, ETH_HLEN); skb->protocol = eth_type_trans(skb, skb->dev); if (ret == GRO_NORMAL) - gro_normal_one(napi, skb); + gro_normal_one(napi, skb, 1); break; case GRO_DROP: