From owner-freebsd-net@FreeBSD.ORG Tue Mar 20 20:05:29 2007 Return-Path: X-Original-To: freebsd-net@freebsd.org Delivered-To: freebsd-net@freebsd.org Received: from mx1.freebsd.org (mx1.freebsd.org [69.147.83.52]) by hub.freebsd.org (Postfix) with ESMTP id 25B8216A408 for ; Tue, 20 Mar 2007 20:05:29 +0000 (UTC) (envelope-from kml@patheticgeek.net) Received: from patheticgeek.net (dsl092-035-004.lax1.dsl.speakeasy.net [66.92.35.4]) by mx1.freebsd.org (Postfix) with SMTP id C6F7613C4C3 for ; Tue, 20 Mar 2007 20:05:28 +0000 (UTC) (envelope-from kml@patheticgeek.net) Received: (qmail 11740 invoked from network); 20 Mar 2007 20:04:29 -0000 Received: from dsl092-035-004.lax1.dsl.speakeasy.net (HELO yakko.patheticgeek.net) (66.92.35.4) by dsl092-035-004.lax1.dsl.speakeasy.net with SMTP; 20 Mar 2007 20:04:29 -0000 Date: Tue, 20 Mar 2007 13:05:25 -0700 From: Kevin Lahey To: "Bruce M. Simpson" Message-ID: <20070320130525.01caad3d@yakko.patheticgeek.net> In-Reply-To: <45FF3D2F.3040000@FreeBSD.org> References: <994cd1cf0703050842r5e54daa6y5fe6af3083e15cd@mail.gmail.com> <45EC6E88.3080101@tomjudge.com> <20070305115615.W38684@orthanc.ca> <994cd1cf0703052105y375679a4t482f4e35988f9daf@mail.gmail.com> <20070319145422.39bfddcd@yakko.patheticgeek.net> <20070319172156.68cba0a9@yakko.patheticgeek.net> <45FF3D2F.3040000@FreeBSD.org> X-Mailer: Sylpheed-Claws 2.5.5 (GTK+ 2.10.6; i386--netbsdelf) Mime-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Cc: freebsd-net@freebsd.org Subject: Re: PMTU Discovery support 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: Tue, 20 Mar 2007 20:05:29 -0000 On Tue, 20 Mar 2007 01:47:27 +0000 "Bruce M. Simpson" wrote: > I believe PMTU measurements may only be relied upon for active TCP > connections, but it's been a while since I read this code. I think, according to RFC1191, that the probing should happen on any given connection after 10 minutes. Otherwise, a single ICMP message could permanently reduce the MTU of a connection for good. I believe that FreeBSD and others did it that way 10 years ago, when I did a survey of available PMTUD implementations. I wonder if all applicable TCP connections ought to be notified when the hostcache MTU is updated and when a hostcache entry with an MTU set is deleted? (For that matter, can an entry ever be deleted while outstanding connections are going to that host?) Kevin kml@patheticgeek.net