From owner-freebsd-net@FreeBSD.ORG Fri Jun 6 20:18:45 2008 Return-Path: Delivered-To: freebsd-net@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id B1C921065678 for ; Fri, 6 Jun 2008 20:18:45 +0000 (UTC) (envelope-from mainland@eecs.harvard.edu) Received: from mail.eecs.harvard.edu (bowser.eecs.harvard.edu [140.247.60.24]) by mx1.freebsd.org (Postfix) with ESMTP id A16D58FC1B for ; Fri, 6 Jun 2008 20:18:45 +0000 (UTC) (envelope-from mainland@eecs.harvard.edu) Received: from minipax.eecs.harvard.edu (minipax.eecs.harvard.edu [140.247.60.67]) (using TLSv1 with cipher DHE-RSA-AES256-SHA (256/256 bits)) (No client certificate requested) by mail.eecs.harvard.edu (Postfix) with ESMTP id 6BC061A3C87; Fri, 6 Jun 2008 16:02:00 -0400 (EDT) Message-ID: <484997B1.5020708@eecs.harvard.edu> Date: Fri, 06 Jun 2008 16:01:53 -0400 From: Geoffrey Mainland User-Agent: Thunderbird 2.0.0.14 (X11/20080503) MIME-Version: 1.0 To: freebsd-net@freebsd.org Content-Type: text/plain; charset=ISO-8859-1; format=flowed Content-Transfer-Encoding: 7bit Subject: Major Atheros issues with Ubiquiti SR9/XR9 X-BeenThere: freebsd-net@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Networking and TCP/IP with FreeBSD List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Fri, 06 Jun 2008 20:18:45 -0000 Hi, I'm running 7-CURRENT on several ALIX 3c2 boards with SR9 radios and having major performance problems: throughput on TCP streams generated by iperf often falls to zero, and, depending on the HAL I use, I see ping latencies during one of these iperf transfers of up to more than a *minute*. I've tried both the 0.9.30.13 HAL and the new 0.10.5.6 HAL. I see this bad behavior with both, but the new HAL seems even worse. I've documented my configuration and the test results at http://www.eecs.harvard.edu/~mainland/freebsd/sr9/. This includes the kernel configuration, dmesg output, pciconf output, rc.conf, athstats output, appropriate sysctl values, iperf/ping output generated by my driver script and some plots showing the problem. We are trying to deploy a bunch of these nodes outdoors, using the 900MHz radios as a backhaul, so the drop-outs I see with SR9s are a major problem for us. I've also run the same tests on Soekris nodes, used different SR9s, and also tried the new XR9s on the same ALIX boards---all with similar results. We also have a bunch of Wistrom CM9s which seem to work just fine. Any idea what could be going on? Thanks! Geoff