From mboxrd@z Thu Jan 1 00:00:00 1970 From: Hannes Frederic Sowa Subject: Re: [PATCH] ipv6: default route for link local address is not added while assigning a address Date: Mon, 3 Feb 2014 17:08:38 +0100 Message-ID: <20140203160838.GA17999@order.stressinduktion.org> References: <52E8A2AA.3090507@linux.vnet.ibm.com> <52E8DA37.7010208@6wind.com> <20140130232909.GH25336@order.stressinduktion.org> <52EF42FF.60907@linux.vnet.ibm.com> <52EFB454.1040908@6wind.com> Mime-Version: 1.0 Content-Type: text/plain; charset=utf-8 Content-Transfer-Encoding: QUOTED-PRINTABLE Cc: Sohny Thomas , netdev , linux-kernel@vger.kernel.org, yoshfuji@linux-ipv6.org, davem@davemloft.net, kumuda To: Nicolas Dichtel Return-path: Content-Disposition: inline In-Reply-To: <52EFB454.1040908@6wind.com> Sender: linux-kernel-owner@vger.kernel.org List-Id: netdev.vger.kernel.org Hello! On Mon, Feb 03, 2014 at 04:23:00PM +0100, Nicolas Dichtel wrote: > Le 03/02/2014 08:19, Sohny Thomas a =C3=A9crit : > > > >>Actually I am not so sure, there is no defined semantic of flush. I= would > >>be ok with all three solutions: leave it as is, always add link-loc= al > >>address (it does not matter if we don't have a link-local address o= n > It matters. This address is required. > RFC 4291 > Section 2.1: > All interfaces are required to have at least one Link-Local unicas= t > address (see Section 2.8 for additional required addresses). > Section 2.8: > o Its required Link-Local address for each interface. Yes, sure, it is required. But you also can manually delete the LL addr= ess and we don't guard against that. > >>that interface, as a global scoped one is just fine enough) or make= flush=20 > >>not > >>remove the link-local address (but this seems a bit too special cas= ed for=20 > >>me). > > > >1) In case if we leave it as it is, there is rfc 6724 rule 2 to be=20 > >considered ( > >previously rfc 3484) > > > >Rule 2: Prefer appropriate scope. > > If Scope(SA) < Scope(SB): If Scope(SA) < Scope(D), then prefer S= B and > > otherwise prefer SA. Similarly, if Scope(SB) < Scope(SA): If > > Scope(SB) < Scope(D), then prefer SA and otherwise prefer SB. > > > >Test: > > > > Destination: fe80::2(LS) > > Candidate Source Addresses: 3ffe::1(GS) or fec0::1(SS) or LLA(L= S) > > Result: LLA(LS) > > Scope(LLA) < Scope(fec0::1): If Scope(LLA) < Scope(fe80::2), n= o,=20 > > prefer LLA > > Scope(LLA) < Scope(3ffe::1): If Scope(LLA) < Scope(fe80::2), n= o,=20 > > prefer LLA > > > > > >Now the above test fails since the route itself is not present, and = the=20 > >test > >assumes that the route gets added since the LLA is not removed durin= g the=20 > >test > In your scenario, the link local route has been removed manually, not= by the > kernel. What is your network manager? The test scenario is outlined here: Basically, the command in question is this one: [root@localhost ~]# ip -6 -statistics -statistics route flush dev eth0 which removes the fe80::/64 route. > >2) having a LLA always helps in NDP i think > A link-local Address yes, it's a MUST. But having only the link local= route=20 > will > not help. Agreed, the LL address should be available, too. I currently don't know what will break if LL address is not available. I guess MLD won't work properly and thus even basic connectivity won't work with some switches= =2E > >3) making flush not remove link-local address will be chnaging=20 > >functionality of > >ip flush command > You can flush by specifying the prototype: > ip -6 route flush proto static So we have four possiblities now: 1) leave it as is seems still acceptable to me 2) add fe80::/64 route unconditionally if any address gets added Sohny's patch already looks good in doing so at first look. 3) add fe80::/64 route in case LL address gets added via inet6_rtm_newa= ddr would be ok, too. I tend towards this solution somehow by now. 4) make flush not remove the fe80::/64 address Least favourable to me. I guess this also woud need iproute change and seems most difficult to do. Any opionions? Greetings, Hannes