Received:  by oss.sgi.com id <S553699AbRAFDfh>;
	Fri, 5 Jan 2001 19:35:37 -0800
Received: from shell.cyberus.ca ([209.195.95.7]:6564 "EHLO shell.cyberus.ca")
	by oss.sgi.com with ESMTP id <S553688AbRAFDfY>;
	Fri, 5 Jan 2001 19:35:24 -0800
Received: from localhost (hadi@localhost)
	by shell.cyberus.ca (8.9.3/666/Cyberus Online Inc.) with ESMTP id WAA19172;
	Fri, 5 Jan 2001 22:34:37 -0500 (EST)
X-Authentication-Warning: shell.cyberus.ca: hadi owned process doing -bs
Date:   Fri, 5 Jan 2001 22:34:37 -0500 (EST)
From:   jamal <hadi@cyberus.ca>
To:     Jordan Mendelson <jordy@napster.com>
cc:     <netdev@oss.sgi.com>
Subject: Re: Mediocre TCP Performance 2.4.0 <-> Win98SE PPP
In-Reply-To: <3A5665CB.30AB2158@napster.com>
Message-ID: <Pine.GSO.4.30.0101052220500.18916-100000@shell.cyberus.ca>
MIME-Version: 1.0
Content-Type: TEXT/PLAIN; charset=US-ASCII
Sender: owner-netdev@oss.sgi.com
Precedence: bulk
Return-Path: <owner-netdev@oss.sgi.com>
X-Orcpt: rfc822;netdev-outgoing
Content-Length: 920
Lines: 28



I am not on lk so i missed your initial post; i apologize if
i am repeating things you might have tried out.
It does not make sense that it works in 2.2.16 but not 2.4.0.

How reproducible is this?
The only visible difference between the two appears to be the mss. It'll
be a shocker that would cause problems.
They both seem to be happily SACKing.
>From the traces it appears 64.124.41.136 is Linux, no? And your traces
must be from the Linux side, right? Why are those  packets being dropped
only when using 2.4.0?
Some suggestions:
Force the MTU on linux to be 536+40.
If that doesnt fix it turn off SACK.
If that doesnt fix it i would start suspecting the NIC. What NIC
do you have on Linux? Might be alignment screw ups.
If that doesnt fix it, i would start suspecting windows.
Capture traces on both sides (windows and Linux side) for both kernels
preferably using -w option and post the URL.

cheers,
jamal