From owner-freebsd-net@FreeBSD.ORG Fri Sep 21 12:23:47 2007 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 14ED216A417 for ; Fri, 21 Sep 2007 12:23:47 +0000 (UTC) (envelope-from iaccounts@ibctech.ca) Received: from pearl.ibctech.ca (pearl.ibctech.ca [208.70.104.210]) by mx1.freebsd.org (Postfix) with ESMTP id 9B8D013C46A for ; Fri, 21 Sep 2007 12:23:46 +0000 (UTC) (envelope-from iaccounts@ibctech.ca) Received: (qmail 70238 invoked by uid 1002); 21 Sep 2007 12:23:45 -0000 Received: from iaccounts@ibctech.ca by pearl.ibctech.ca by uid 89 with qmail-scanner-1.22 (spamassassin: 2.64. Clear:RC:1(208.70.104.100):. Processed in 11.3624 secs); 21 Sep 2007 12:23:45 -0000 Received: from unknown (HELO ?192.168.30.110?) (steve@ibctech.ca@208.70.104.100) by pearl.ibctech.ca with (DHE-RSA-AES256-SHA encrypted) SMTP; 21 Sep 2007 12:23:33 -0000 Message-ID: <46F3B7C9.7050605@ibctech.ca> Date: Fri, 21 Sep 2007 08:23:37 -0400 From: Steve Bertrand User-Agent: Thunderbird 2.0.0.6 (Windows/20070728) MIME-Version: 1.0 To: Sten Daniel Soersdal References: <46F1AC0B.9040109@ibctech.ca> <46F1BDE1.8090102@gmail.com> <46F1F136.3010203@ibctech.ca> <46F23D74.9000701@gmail.com> In-Reply-To: <46F23D74.9000701@gmail.com> X-Enigmail-Version: 0.95.3 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit Cc: mattr@eagle.ca, freebsd-net@freebsd.org Subject: Re: Quagga as border router 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, 21 Sep 2007 12:23:47 -0000 > I'm not saying you should use polling. I'm saying that not using polling > makes for more context switches. 64bit registers are twice as large as > 32bit registers. There will be a bigger penalty on stack/memory usage > and therefore slower transitions from one context to another (read: > handling a packet). > This might be mitigated by having a very large cpu cache. > > It may or may not make much of a difference considering stacks are > aligned, i was just theorizing. Ok, I get what you are saying now. > Just curious: Is there a reason you can't advertise your entire > allocated block and receive two full feeds? Well to be honest, there is no reason why I couldn't receive full routes with the new box. My current router couldn't handle the full route table, but any new router will be able to. Regards, Steve