Re: Question about <draft-ietf-6man-deprecate-atomfrag-generation-03>

[Bob Hinden <[e-mail]>]

Tore,

> I just realised that this would also remove an important piece of
> information that is would still be relevant to implementers:
> 
> «[An] IPv6 node may receive an ICMP Packet Too Big message reporting a
> Next-Hop MTU less than 1280. In that case, the IPv6 node is not
> required to reduce the size of subsequent packets to less than 1280[.]»

You may be reading this out of context.  The start of this section starts with:

   IPv6 requires that every link in the internet have an MTU of 1280
   octets or greater.

RFC443 ICMPv6 that defines the ICMP Packet too Big message references RFC1981 Path MTU Discovery (also referenced in this section) says:

   A node MUST NOT reduce its estimate of the Path MTU below the IPv6
   minimum link MTU.

I think this makes it clear what to do if an ICMP Packet Too Big message is received with an MTU less than 1280.

However, this then goes on to say:

      Note: A node may receive a Packet Too Big message reporting a
      next-hop MTU that is less than the IPv6 minimum link MTU.  In that
      case, the node is not required to reduce the size of subsequent
      packets sent on the path to less than the IPv6 minimun link MTU,
      but rather must include a Fragment header in those packets [IPv6-
      SPEC].

That would also need to be updated (that is, remove this note), or file an errata once the rfc2460bis is done.  The latter would be easier.

Bob






> 
> That is, it becomes somewhat ambigiuous what an IPv6 node should
> actually do upon receiving an ICMPv6 PTB with an MTU of, say, 1000. I
> can imagine three possible ways of understanding the new text:
> 
> A) the PTB may be completely ignored
> B) subsequent packets must be reduced to 1280
> C) subsequent packets must be reduced to 1000
> 
> Of these, option A would be damaging because it could create PMTUD
> blackholes even with translators conforming to 6145bis. Both B and C
> would work, though. So I think there needs to be replacement text that
> mandates at least option B while at the same time does not prohibit
> option C. Something like this, maybe? (Of course, do feel free to come
> up with something better.)
> 
> «An IPv6 node may receive an ICMP Packet Too Big message reporting a
> Next-Hop MTU less than 1280. In that case, the IPv6 node must limit the
> size of subsequent packets to 1280 or less.»
> 
> (This also applies to draft-ietf-6man-deprecate-atomfrag-generation-03.)
> 
> Tore