From mboxrd@z Thu Jan 1 00:00:00 1970 From: David Miller Subject: Re: [PATCH] IPv6: fix rt_lookup in pmtu_discovery Date: Thu, 07 Jan 2010 01:27:01 -0800 (PST) Message-ID: <20100107.012701.257511338.davem@davemloft.net> References: <65634d661001062043s1b4eb204v63566149bb44f144@mail.gmail.com> Mime-Version: 1.0 Content-Type: Text/Plain; charset=us-ascii Content-Transfer-Encoding: 7bit Cc: netdev@vger.kernel.org, lorenzo@google.com To: therbert@google.com Return-path: Received: from 74-93-104-97-Washington.hfc.comcastbusiness.net ([74.93.104.97]:44760 "EHLO sunset.davemloft.net" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1750930Ab0AGJ0x (ORCPT ); Thu, 7 Jan 2010 04:26:53 -0500 In-Reply-To: <65634d661001062043s1b4eb204v63566149bb44f144@mail.gmail.com> Sender: netdev-owner@vger.kernel.org List-ID: From: Tom Herbert Date: Wed, 6 Jan 2010 20:43:48 -0800 > In rt6_pmtu_discovery a route lookup is done constrained to the > receiving interface of the ICMP message. In the case of an asymmetric > routing, the receive interface may be different than the sending > interface, so no route will be found in such cases and pmtu is never > correctly set in the route to the destination. This patch fixes that > by not using the receive interface in the lookup. > > Signed-off-by: Tom Herbert This needs to do what the IPV4 side does, iterate over specific then "any" device index, doing a lookup for each case until the route is found, therefore starting from more specific and going towards less specific routes.