Re: Loopback interface terminology issue

[Mark Smith <[e-mail]>]

On 15 Oct. 2017 13:41, "Toerless Eckert" <[e-mail]> wrote:

On Sun, Oct 15, 2017 at 12:53:18PM +1100, Mark Andrews wrote:
> > IMHO: A loopback interface is an attachment to a node internal
> > "loopback link". If other addresses beside "loopback addresses" are
> > assigned to such a link, then this implies that the interface
> > is configured to forward non-self-destined packets.
>
> No.  Whether such addresses are forwarded by default or not depends on
> the weak/strong host model in use.

Not really. If the transport stack (a) logically attaches to the loopback
interface then the loopback source address is the source address
corresponding
to the interface. The difference between strong and weak model would only
come in if a transport socket (b) binds to a physical interface and tries
to send/receive packets for the loopback address.

I think the stacks i know logically do (a), althogh i am not sure how
cleanly it's implemented. If its 100% clean then the TTL of the sent packet
should be decremented by 1 before output on the physcial interface. Not
sure which stacks this is done in.

> Loopback interfaces are also used for test networks wholely constrained
> to the node.

Ok. How about:

If other addresses beside "loopback addresses" are assigned to such a link,
then the interface needs to be configured to forward non-self-destined
packets originated from the loopback interface (see rfc8200, section 2).



This draft might help.

In IPv4, there are different forwarding rules for hosts and routers over
the loopback interface for the 127/8 prefix.

This draft proposes a larger IPv6 prefix for loopback, and mirrors those
differences.

https://tools.ietf.org/html/draft-smith-v6ops-larger-ipv6-loopback-prefix-04




Cheers
    Toerless

> > On Sat, Oct 14, 2017 at 02:11:54PM +1300, Brian E Carpenter wrote:
> > > Hi,
> > >
> > > This come rather late in the day, but it results from some
wordsmithing
> > > difficulties in a draft in another WG:
> > >
> > > The IPv6 specs are very clear that an address is assigned to an
interface,
> > > not to a node, but they don't say anywhere that non-loopback
addresses may
> > > be assigned to the loopback interface. However, sometimes as a
practical
> > > matter, it's necessary to assign a routeable address (GUA or ULA) that
> > > is not created via a specific IPv6 interface. Conventionally, that is
> > > done by assigning it to the loopback interface.
> > >
> > > However, the only place I have found that defines 'loopback
interface' for
> > > IPv6 is in the addressing architecture: 'a virtual interface
(typically
> > > called the "loopback interface") to an imaginary link that goes
nowhere.'
> > > (RFC 4291, section 2.5.3). The text only mentions the loopback
address ::1.
> > >
> > > Do people agree with something like the following?
> > >
> > > 'Note that other forms of unicast IPv6 address may be assigned to the
> > > loopback interface, in cases where an address is assigned to the node
> > > without being assigned to the interface to a specific link.'
> > >
> > > This is after all common practice in various operating systems; we
> > > just never documented it, as far as I know.
> > >
> > > Regards
> > >    Brian Carpenter
> > >
> > >
> > > --------------------------------------------------------------------
> > > IETF IPv6 working group mailing list
> > > [e-mail]
> > > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> > > --------------------------------------------------------------------
> >
> > --
> > ---
> > [e-mail]
> >
> > --------------------------------------------------------------------
> > IETF IPv6 working group mailing list
> > [e-mail]
> > Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> > --------------------------------------------------------------------
> --
> Mark Andrews, ISC
> 1 Seymour St., Dundas Valley, NSW 2117, Australia
> PHONE: [phone]                 INTERNET: [e-mail]
>
> --------------------------------------------------------------------
> IETF IPv6 working group mailing list
> [e-mail]
> Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
> --------------------------------------------------------------------

--------------------------------------------------------------------
IETF IPv6 working group mailing list
[e-mail]
Administrative Requests: https://www.ietf.org/mailman/listinfo/ipv6
--------------------------------------------------------------------