From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S268534AbUIXHiq (ORCPT ); Fri, 24 Sep 2004 03:38:46 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S268526AbUIXHiq (ORCPT ); Fri, 24 Sep 2004 03:38:46 -0400 Received: from scanner2.mail.elte.hu ([157.181.151.9]:55787 "EHLO mx2.elte.hu") by vger.kernel.org with ESMTP id S268534AbUIXHio (ORCPT ); Fri, 24 Sep 2004 03:38:44 -0400 Date: Fri, 24 Sep 2004 09:40:26 +0200 From: Ingo Molnar To: "K.R. Foley" Cc: linux-kernel@vger.kernel.org, Lee Revell , Mark_H_Johnson@Raytheon.com, Rui Nuno Capela Subject: Re: [patch] voluntary-preempt-2.6.9-rc2-mm3-S5 Message-ID: <20040924074026.GB17368@elte.hu> References: <20040909061729.GH1362@elte.hu> <20040919122618.GA24982@elte.hu> <414F8CFB.3030901@cybsft.com> <20040921071854.GA7604@elte.hu> <20040921074426.GA10477@elte.hu> <20040922103340.GA9683@elte.hu> <20040923122838.GA9252@elte.hu> <20040923211206.GA2366@elte.hu> <415384E1.2080907@cybsft.com> <415394EE.50106@cybsft.com> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <415394EE.50106@cybsft.com> User-Agent: Mutt/1.4.1i X-ELTE-SpamVersion: MailScanner 4.31.6-itk1 (ELTE 1.2) SpamAssassin 2.63 ClamAV 0.73 X-ELTE-VirusStatus: clean X-ELTE-SpamCheck: no X-ELTE-SpamCheck-Details: score=-4.9, required 5.9, autolearn=not spam, BAYES_00 -4.90 X-ELTE-SpamLevel: X-ELTE-SpamScore: -4 Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org * K.R. Foley wrote: > The following, on top of Ingo's patch above, fixes the problem with > dropping new connections and doesn't have any adverse affects that > I've seen: > > --- linux-2.6.9-rc2-pre-mm3/net/ipv4/tcp_output.c.orig 2004-09-23 > 22:16:42.249435870 -0500 > +++ linux-2.6.9-rc2-pre-mm3/net/ipv4/tcp_output.c 2004-09-23 > 22:12:03.911811945 -0500 > @@ -699,11 +699,6 @@ > > tcp_minshall_update(tp, mss_now, skb); > sent_pkts = 1; > - /* > - * Break out early - we'll continue later: > - */ > - if (softirq_need_resched()) > - break; hm, ok, i'll revert this in my tree. I suspect we'll see some latencies resurfacing under high network load again, but correctness goes first obviously. If then we'll have to find some other method to break that critical path. Ingo