Re: draft-ietf-6man-icmp-limits-05

On Mon, Sep 16, 2019 at 5:28 PM Ron Bonica <rbonica@juniper.net> wrote:
>
> Tom,
>
> When you send an ICMP Destination Unreachable with code equal to aggregate header limit exceeded, does the pointer always point beyond the end of the IPv6 header chain?

Not always. Consider this example: a device has a parsing buffer of
128 bytes. It starts parsing the packets, IP header, extension
headers, encapsulations, transport, headers, etc. For some packet it
finds that it needs parse beyond 128 bytes to get the information it
needs. If it drops the packet, then it sends Unreachable with point
equal to 128. In this case the pointer might be at an EH, an IP
header, transport header, encap header, ...

Note that this error has the lowest precedence so that Parameter
Problem is reported for any limits exceeded relating to EH.

Tom

>
>                                                                                             Ron
>
>
> Juniper Business Use Only
>
> -----Original Message-----
> From: Tom Herbert <tom@herbertland.com>
> Sent: Monday, September 16, 2019 6:00 PM
> To: Bob Hinden <bob.hinden@gmail.com>
> Cc: Ron Bonica <rbonica@juniper.net>; 6man <6man@ietf.org>
> Subject: Re: draft-ietf-6man-icmp-limits-05
>
> On Mon, Sep 16, 2019 at 2:10 PM Bob Hinden <bob.hinden@gmail.com> wrote:
> >
> >
> >
> > > On Sep 16, 2019, at 11:48 AM, Ron Bonica <rbonica=40juniper.net@dmarc.ietf.org> wrote:
> > >
> > > Tom,
> > >
> > > In draft-ietf-6man-icmp-limits-05, you propose sending and ICMP Parameter Problem message when the following errors occur:
> > >
> > >          - Extension header too big
> > >          - Extension header chain too long
> > >          - Too many options in extension header
> > >          - Option too big
> > >
> > > However, you send an ICMP Destination Unreachable message in response to aggregate header limits.
> > >
> > > Why not send an ICMP Parameter Problem message in cases?
> >
> > Good question.  It would make the draft simpler and avoid having to create RFC4884 extension headers.
> >
>
> Ron, Bob,
>
> There was already discussion of this on the list.
>
> Originally, Parameter Problem code was defined for this, however that was considered inconsistent with the definition of Parameter Problem
> (RFC4443):
>
> "If an IPv6 node processing a packet finds a problem with a field in the IPv6 header or extension headers such that it cannot complete processing the packet"
>
> Aggregate header limits can apply to headers other than just IP header and extension headers (e.g. UDP encapsulation headers). Hence, we went with Destination Unreachable and subsequently why we need to invoke the ICMP extended header format to get the error pointer.
>
> Tom
>
>
> > Bob
> >
> >
> > >
> > >
> > > Ron
> > >
> > >
> > > Juniper Business Use Only
> > > --------------------------------------------------------------------
> > > IETF IPv6 working group mailing list ipv6@ietf.org Administrative
> > > Requests:
> > > https://urldefense.com/v3/__https://www.ietf.org/mailman/listinfo/ip
> > > v6__;!8WoA6RjC81c!Tr9t_Imk6fHhld_RTnhilL4gppaKBxhvwff7Y6ZeeMo_eoEjPT
> > > TXdBfNvMophbc9$
> > > --------------------------------------------------------------------
> >