From owner-freebsd-performance@FreeBSD.ORG Wed Dec 5 22:28:21 2007 Return-Path: Delivered-To: freebsd-performance@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [IPv6:2001:4f8:fff6::34]) by hub.freebsd.org (Postfix) with ESMTP id 407EC16A418 for ; Wed, 5 Dec 2007 22:28:21 +0000 (UTC) (envelope-from mike@sentex.net) Received: from smarthost1.sentex.ca (smarthost1.sentex.ca [64.7.153.18]) by mx1.freebsd.org (Postfix) with ESMTP id 022B613C4CE for ; Wed, 5 Dec 2007 22:28:20 +0000 (UTC) (envelope-from mike@sentex.net) Received: from lava.sentex.ca (pyroxene.sentex.ca [199.212.134.18]) by smarthost1.sentex.ca (8.13.8/8.13.8) with ESMTP id lB5MSK9G014601; Wed, 5 Dec 2007 17:28:20 -0500 (EST) (envelope-from mike@sentex.net) Received: from mdt-xp.sentex.net (simeon.sentex.ca [192.168.43.27]) by lava.sentex.ca (8.13.8/8.13.3) with ESMTP id lB5MSJYA031170 (version=TLSv1/SSLv3 cipher=DHE-RSA-AES256-SHA bits=256 verify=NO); Wed, 5 Dec 2007 17:28:20 -0500 (EST) (envelope-from mike@sentex.net) Message-Id: <200712052228.lB5MSJYA031170@lava.sentex.ca> X-Mailer: QUALCOMM Windows Eudora Version 7.1.0.9 Date: Wed, 05 Dec 2007 17:30:22 -0500 To: Philipp Wuensche From: Mike Tancsa In-Reply-To: <475723FB.4020304@h3q.com> References: <4755ED57.6030603@h3q.com> <200712051616.lB5GGGvb029587@lava.sentex.ca> <4756DE7E.1000104@h3q.com> <200712051729.lB5HTAkn029896@lava.sentex.ca> <475723FB.4020304@h3q.com> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Cc: freebsd-performance@freebsd.org Subject: Re: tuning for high connection rates X-BeenThere: freebsd-performance@freebsd.org X-Mailman-Version: 2.1.5 Precedence: list List-Id: Performance/tuning List-Unsubscribe: , List-Archive: List-Post: List-Help: List-Subscribe: , X-List-Received-Date: Wed, 05 Dec 2007 22:28:21 -0000 At 05:19 PM 12/5/2007, Philipp Wuensche wrote: >After switching to net.isr.direct=0 and 346609775 good packets later, RX >overruns haven't increased by one! Thats nice. Still interrupt is using >up the CPU. I'm not quite sure if polling would help now!? Polling is helpful to prevent livelock. Not sure if thats happening to you. What firewall (if any) are you using ? pf used to be a lot slower than ipfw. The Yandex driver is at http://people.yandex-team.ru/~wawa/ but its against RELENG_6 only I think. Another thing to try is to turn back on Fast Interrupt handling. I think its currently disabled. In if_em.h, try adding #define EM_FAST_IRQ 1 and then recompile the kernel or just driver. >We will try disabling TSO to see if anything changes. If you have TCP in your app, it seems thats the thing to do according to the Intel developer. ---Mike