From mboxrd@z Thu Jan 1 00:00:00 1970 Return-Path: Received: (majordomo@vger.kernel.org) by vger.kernel.org via listexpand id ; Wed, 2 Apr 2003 01:50:14 -0500 Received: (majordomo@vger.kernel.org) by vger.kernel.org id ; Wed, 2 Apr 2003 01:50:14 -0500 Received: from mail4.bluewin.ch ([195.186.4.74]:1433 "EHLO mail4.bluewin.ch") by vger.kernel.org with ESMTP id ; Wed, 2 Apr 2003 01:50:13 -0500 Date: Wed, 2 Apr 2003 09:01:15 +0200 From: Roger Luethi To: Brad Campbell Cc: linux-kernel@vger.kernel.org Subject: Re: via-rhine problem on EPIAV-1Ghz 2.4.21-pre6 Message-ID: <20030402070115.GA2667@k3.hellgate.ch> Mail-Followup-To: Brad Campbell , linux-kernel@vger.kernel.org References: <3E88FA24.7040406@seme.com.au> <20030401042734.GA21273@gtf.org> <3E89171A.8010506@seme.com.au> <20030401185258.GC3736@arthur.home> <3E8A6298.8060600@seme.com.au> Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Disposition: inline In-Reply-To: <3E8A6298.8060600@seme.com.au> User-Agent: Mutt/1.3.27i X-Operating-System: Linux 2.5.65 on i686 X-GPG-Fingerprint: 92 F4 DC 20 57 46 7B 95 24 4E 9E E7 5A 54 DC 1B X-GPG: 1024/80E744BD wwwkeys.ch.pgp.net Sender: linux-kernel-owner@vger.kernel.org X-Mailing-List: linux-kernel@vger.kernel.org On Wed, 02 Apr 2003 12:10:00 +0800, Brad Campbell wrote: > It solved the timeout propblems, I do get these however. > > Not that they cause any hiccups in throughput. > > eth0: Tx descriptor write-back race. > eth0: Tx descriptor write-back race. > eth0: Tx descriptor write-back race. > eth0: Transmit error, Tx status 00008800. > eth0: Transmitter underrun, Tx threshold now 40. > eth0: Tx descriptor write-back race. > eth0: Transmit error, Tx status 00008800. > eth0: Transmitter underrun, Tx threshold now 60. > eth0: Tx descriptor write-back race. These are only informational messages on errors handled by the new code. You won't see them at the default debug level. I made a few observations leading me to believe that the number of those errors can be further reduced, but that will take quite a bit of work. And there is still a Tx pause of up to a couple of seconds (net watchdog kicking in) on every GB or so, especially under high load. Roger