From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Tue, 4 Sep 2001 17:07:53 -0400 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Tue, 4 Sep 2001 17:07:44 -0400 Received: from mailhost.nmt.edu ([129.138.4.52]:33553 "EHLO mailhost.nmt.edu") by vger.kernel.org with ESMTP id ; Tue, 4 Sep 2001 17:07:33 -0400 Date: Tue, 4 Sep 2001 15:07:50 -0600 From: Val Henson To: kuznet@ms2.inr.ac.ru Cc: davem@redhat.com, linux-kernel@vger.kernel.org Subject: Re: Lost TCP retransmission timer Message-ID: <20010904150750.D22301@boardwalk> In-Reply-To: <20010830161139.A18224@boardwalk> <200109011436.SAA18432@ms2.inr.ac.ru> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline User-Agent: Mutt/1.2.5i In-Reply-To: <200109011436.SAA18432@ms2.inr.ac.ru>; from kuznet@ms2.inr.ac.ru on Sat, Sep 01, 2001 at 06:36:35PM +0400 Favorite-Color: Polka dot Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Sat, Sep 01, 2001 at 06:36:35PM +0400, kuznet@ms2.inr.ac.ru wrote: > > Lots? I see only about 24K of data transmitted in both your samples. :) Okay, "lots" is relative. If you only telnet in and exit right away, you won't see it. > Actually, the problem is more or less clear from your /proc/net/tcp. > You use some funny device or netfilter plugin, which leak memory. > You can look into 7th column of /proc/net/tcp to estimate amount of leaked > buffers. When it reaches ~15, connection stalls. Seems, it raises > monotonically, so that it looks like all the buffers leak. > > What is output device? ncr885e, which I believe I am the de facto maintainer of... Dan Cox wrote it while he was working for Synergy Microsystems. When he quit, I was hired to replace him. Anyone else using this driver? I have a patch for it that fixes some other things (currently only in the LinuxPPC tree) but not this. Thanks for the help, I think this is my problem to fix now. -VAL