From mboxrd@z Thu Jan 1 00:00:00 1970 From: Eric Dumazet Subject: Re: [PATCH net] ipv6: refresh rt6i_genid in ip6_pol_route() Date: Mon, 08 Sep 2014 03:28:12 -0700 Message-ID: <1410172092.11872.85.camel@edumazet-glaptop2.roam.corp.google.com> References: <1410149905.11872.73.camel@edumazet-glaptop2.roam.corp.google.com> <20140907.212742.893555292268057160.davem@davemloft.net> <1410151434.11872.82.camel@edumazet-glaptop2.roam.corp.google.com> <20140907.215943.213418445039957641.davem@davemloft.net> <1410152829.11872.84.camel@edumazet-glaptop2.roam.corp.google.com> <540D64C2.9050000@6wind.com> Mime-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: David Miller , therbert@google.com, alexander.h.duyck@intel.com, netdev@vger.kernel.org To: nicolas.dichtel@6wind.com Return-path: Received: from mail-pa0-f53.google.com ([209.85.220.53]:65319 "EHLO mail-pa0-f53.google.com" rhost-flags-OK-OK-OK-OK) by vger.kernel.org with ESMTP id S1752784AbaIHK2Q (ORCPT ); Mon, 8 Sep 2014 06:28:16 -0400 Received: by mail-pa0-f53.google.com with SMTP id rd3so2144182pab.40 for ; Mon, 08 Sep 2014 03:28:15 -0700 (PDT) In-Reply-To: <540D64C2.9050000@6wind.com> Sender: netdev-owner@vger.kernel.org List-ID: On Mon, 2014-09-08 at 10:11 +0200, Nicolas Dichtel wrote: > Le 08/09/2014 07:07, Eric Dumazet a =C3=A9crit : > > On Sun, 2014-09-07 at 21:59 -0700, David Miller wrote: > >> From: Eric Dumazet > >> Date: Sun, 07 Sep 2014 21:43:54 -0700 > >> > >>> On Sun, 2014-09-07 at 21:27 -0700, David Miller wrote: > >>>> From: Eric Dumazet > >>>> Date: Sun, 07 Sep 2014 21:18:25 -0700 > >>>> > >>>>> On Sun, 2014-09-07 at 15:54 -0700, David Miller wrote: > >>>>> > >>>>>> This might be broken. > >>>>>> > >>>>>> We are dealing here with persistent entries in the ipv6 routin= e trie. > >>>>>> > >>>>>> If you just bump the genid on the next person to look it up, o= ther > >>>>>> sockets and cached entities might not have validated the route= yet, > >>>>>> and now will falsely see the route as valid. We have to ensur= e that > >>>>>> they too drop this route and perform a relookup. > >>>>> > >>>>> I am confused, I thought it was the role of the cookie. > >>>>> > >>>>> (Ie socket has to store its own cookie to be able to validate a= route) > >>>>> > >>>>> Before 6f3118b571b8 patch, how was this done anyway ? > >>>>> > >>>>> If persistent routes cannot refresh the genid, then they are us= eless ? > >>>> > >>>> I just speak about the genid aspect. > >>>> > >>>> I understand that cookie (via node->fn_sernum) invalidates the p= ath > >>>> in the fib_trie, but the genid protects against other circumstan= ces > >>>> (matching IPSEC rule, f.e.) > >>>> > >>>> You have to make sure all other sockets did a full route lookup > >>>> (including IPSEC) before you can safely adjust the genid. > >>>> > >>>> I could be wrong, recheck my analysis please :-) > >>> > >>> So this whole genid protection can not work, unless we make sure = a > >>> socket cannot share a route with another socket. > >>> > >>> This means we have to clone all routes. > >> > >> I'm willing to revert the change in question if you think that's t= he > >> sanest way forward. > >> > >> The bug fix for more obscure use cases (IPSEC) if pointless if it > >> breaks more common things (TCP input route caching). > > > > Lets wait for Nicolas and/or Hannes input, they might have some ide= as... >=20 > The initial problem was in SCTP. Here is the thread after the v1 patc= h: > http://patchwork.ozlabs.org/patch/182235/ >=20 > Before the patch, SCTP stored the IPv6 route in its cache and if an I= Psec > rules was inserted after that operation, SCTP never invalidated the c= ached > route to use a new xfrm route. This thread mentions output route. The problem I currently have with IPv6 early demux is for input routes.