From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id S270317AbTGSN0S (ORCPT ); Sat, 19 Jul 2003 09:26:18 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id S270333AbTGSN0S (ORCPT ); Sat, 19 Jul 2003 09:26:18 -0400 Received: from zork.zork.net ([64.81.246.102]:45963 "EHLO zork.zork.net") by vger.kernel.org with ESMTP id S270317AbTGSNZx (ORCPT ); Sat, 19 Jul 2003 09:25:53 -0400 To: netdev@oss.sgi.com Cc: linux-kernel@vger.kernel.org Subject: [2.6.0-test1-mm1] TCP connections over ipsec hang after a few seconds From: Sean Neakums Mail-Followup-To: netdev@oss.sgi.com, linux-kernel@vger.kernel.org Date: Sat, 19 Jul 2003 14:40:48 +0100 Message-ID: <6uk7aeab33.fsf@zork.zork.net> MIME-Version: 1.0 Content-Type: text/plain; charset=us-ascii Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org I've just set up transport mode IPsec between two machines with the aid of the LARTC IPsec docs. Both boxes are running 2.6.0-test1-mm1, with a wireless/wired bridge between, using ipsec-tools 0.2.2. Everything seems to be in order, except (big except) that TCP sessions hang after a small number of seconds of use, usually between ten and twenty. The problem seems unrelated to the amount of data transferred; I've tried both bulk rsync transfers and ssh sessions. I've also tested the same boxes over 100baseT; still happens. I'm not exactly sure what additional information to supply; my experience with IPsec is limited to this past few hours' experimentation.