Re: [dhcwg] [v6ops] Re: Question to DHCPv6 Relay Implementors regarding draft-ietf-dhc-dhcpv6-pd-relay-requirements

Ian,

> For multi-access links, when the packet's
> ingress and egress interface match, and the source MAC and next-hop MAC addresses
> match.

As I said, this gets very tricky if the client has multiple MACs. If Client A has MAC addresses
a1, a2, a3, a4, etc. it becomes very difficult for the relay to know that a packet received
from one of the MAC addresses (e.g., a1) must not be sent back to another of the MAC
addresses (e.g., a3). I think another failure case is if there is a proxy between the client
and relay. In that case, the relay will see the MAC address of the proxy and not the
MAC address of client A. And, if there were multiple additional clients B, C, D, etc.
sharing the same proxy then the proposed check could block legitimate traffic.

As I said before, I think the better fix is to instrument the client. If the client receives
a packet on its relay-facing interface, and the routing system determines that the
packet should be forwarded out the same interface via a default route, the client
must drop the packet. That way, the relay never sees a looped packet, and there
is no extraneous traffic on the client/relay interface.

Fred

> -----Original Message-----
> From: dhcwg [mailto:dhcwg-bounces@ietf.org] On Behalf Of ianfarrer@gmx.com
> Sent: Tuesday, October 13, 2020 6:16 AM
> To: Michael Richardson <mcr+ietf@sandelman.ca>; Jen Linkova <furry13@gmail.com>
> Cc: dhcwg <dhcwg@ietf.org>; v6ops list <v6ops@ietf.org>; Bernie Volz (volz) <volz=40cisco.com@dmarc.ietf.org>; 6man
> <ipv6@ietf.org>
> Subject: Re: [dhcwg] [v6ops] [EXTERNAL] Re: Question to DHCPv6 Relay Implementors regarding draft-ietf-dhc-dhcpv6-pd-relay-
> requirements
> 
> Hi,
> 
> Thanks for all of the discussion on this. We’ve reworked the requirement as follows:
> 
> R-4
> To prevent routing loops, the relay SHOULD implement a configurable policy to
> drop client traffic as follows:  For point-to-point links, when the packet's
> ingress and egress interfaces match. For multi-access links, when the packet's
> ingress and egress interface match, and the source MAC and next-hop MAC addresses
> match. An ICMPv6 Type 1, Code 6 (Destination Unreachable, reject route to
> destination) error message MAY be sent back to the client.  The ICMP policy
> SHOULD be configurable.
> 
> Thanks,
> Ian
> 
> > On 9. Oct 2020, at 17:41, Michael Richardson <mcr+ietf@sandelman.ca> wrote:
> >
> >
> > Jen Linkova <furry13@gmail.com> wrote:
> >> I think there is confusion re: the scenario we are talking about. I've
> >> attached the diagram for the case which concerns me.
> >> So:
> >> - The Relay R has an interface eth0 connected to a switch S.
> >> - Devices A and B are connected to the same switch and using R as a
> >> default gateway.
> >> - The prefix 2001:db8::/56 was delegated to a client A via the relay R.
> >
> > a friendly amendment to your example to aid in human comprehension:
> >     } - The prefix 2001:db8:0000:0123:/64 was delegated to a client A via the relay R.
> >     }  - R installs a route for 2001:db8:0000:0123:/64 towards A via eth0.
> >
> >> - The device B (which has an address NOT from the delegated prefix,
> >> but from another /64 assigned to that common link, let's sat
> >> 2001:db8:cafe::/64) sends a packet to an address from the delegated
> >
> > now, my brain can more clearly see that 2001:db8:cafe::/64 is not
> > within 2001:db8:0000:0123:/64, while I had to use a few extra brain cells to
> > see that it wasn't in that ::/56 :-)
> >
> >> What I'd expect to happen (with DHCP-PD or without - e.g. if R has a
> >> static route towards A, not a dynamic route produced by PD):
> >> - the packet is sent to A. Well, if A does not have a route to
> >> 2001:db8::42 then indeed a routing loop might happen. But if A does
> >> have a route, the packet will be delivered.
> >
> >> What seems to be required by R4:
> >> - R detects that the packet is received via eth0 and needs to be sent
> >> back to eth0. R4 seems to require such packets to be dropped.
> >> So if B would never be able to communicate to any address in the
> >> delegated prefix, right?
> >
> >> Am I missing anything?
> >
> > I think that you got it right.
> >
> >>> Perhaps the missing piece of the rule is don’t send it back to where it came from, based on link layer addresses (or link if point-to-
> point).
> >
> >> Yes. If R4 was saying 'drop the packet if it comes from the same
> >> link-layer address you are going to send it back' - it would make
> >> total sense. But I don't think routers do *that*.
> >
> > Yes, if we made the check on L2 address, then it would work.
> > And I agree that routers are exactly doing that.
> >
> > I think that it also works if B is a router with additional interfaces
> > downstream, unless there are multiple paths.
> >
> > --
> > Michael Richardson <mcr+IETF@sandelman.ca>   . o O ( IPv6 IøT consulting )
> >           Sandelman Software Works Inc, Ottawa and Worldwide
> >
> > --------------------------------------------------------------------
> > IETF IPv6 working group mailing list
> > ipv6@ietf.org
> > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> > --------------------------------------------------------------------
> 
> _______________________________________________
> dhcwg mailing list
> dhcwg@ietf.org
> https://www.ietf.org/mailman/listinfo/dhcwg