From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Thu, 21 Mar 2002 02:17:29 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Thu, 21 Mar 2002 02:17:18 -0500 Received: from poplar.cs.washington.edu ([128.95.2.24]:54534 "EHLO poplar.cs.washington.edu") by vger.kernel.org with ESMTP id ; Thu, 21 Mar 2002 02:17:14 -0500 Date: Wed, 20 Mar 2002 23:17:08 -0800 From: Neil Spring To: yasu@ysuzuki.net Cc: linux-kernel@vger.kernel.org, ysuzuki@bb.mbn.or.jp Subject: Re: PROBLEM: TCP CWR and ECE flags Message-ID: <20020321071658.GA13805@cs.washington.edu> In-Reply-To: <20020321024403.GB1190%ysuzuki@bb.mbn.or.jp> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.3.27i Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Thu, Mar 21, 2002 at 11:44:03AM +0900, SUZUKI Yasuhiro wrote: > A web page > http://www.icir.org/floyd/papers/ECN.Oct2000.txt > says that > * If Host A does not get any reply to its initial SYN (which had CWR > and ECE set) within the normal SYN retransmission timeout interval, > then Host A resends the SYN and any subsequent SYN retransmissions > with CWR and ECE cleared. > where Host A is a client and Host B is a server. It also says: This note outlines a procedure that ECN-Capable TCP implementations may wish to use in this case to ensure connectivity. We are not suggesting that TCP implementations should be required to use this procedure. If you don't like linux's behavior, su echo 0 > /proc/sys/net/ipv4/tcp_ecn echo "net/ipv4/tcp_ecn=0" >> /etc/sysctl.conf If you don't like nifty.com's ECN violating behavior, send them mail. See the faq, as mentioned at the footer of all Linux-kernel mailing lists, and specifically at http://www.tux.org/lkml/#s14-2 for additional information. -neil