In may opinion ( already stated about 2 months ago) there are really no
problems with Routing Headers.

I am really thinking about this in terms of forwarding.  If while
processing
the routing header, I have to forward the packet off the node, then all
the
"router" rules apply (i.e I have to have forwarding enabled).  If the
packet
is forwarded to the same node (looped back), then I am not really
forwading this,
and the node consumes the packet.

We don't needlessly drop packets, MIPv6 is happy, and there are no holes
that
I can see (I may be blind though:)

-vlad

Pekka Savola wrote:
> 
> On Wed, 12 Dec 2001, Robert Elz wrote:
> >     Date:        Tue, 11 Dec 2001 19:05:37 +0100 (CET)
> >     From:        Erik Nordmark <[EMAIL PROTECTED]>
> >     Message-ID:  <[EMAIL PROTECTED]>
> >
> >   | It isn't obvious to me from reading RFC 2460 whether a host can drop packets
> >   | with routing headers instead of resubmitting them for transmissing to the
> >   | next hop.
> >
> [snip]
> > There is though a different, related, question - one where what 2460 says
> > is applicable ... that is whether a host implementation should simply drop
> > packets containing source routes, rather than forwarding them.
> >
> > Here we get the questions of the conformance of the implementation to the
> > spec.
> >
> > And there I think the answer is a clear no - implementations are required
> > to be at least capable of processing routing headers (even in hosts).
> 
> The last sentence IMO doesn't answer the real question.
> 
> I think it's quite clear that all nodes must support routing headers, that
> is, recognize the header type, be able to parse it, not send ICMP
> parameter problem packets back to the source, etc.
> 
> The real question how one processes a datagram with a routing header when
> the forwarding of routing headers has been disabled.  This issue is not
> mentioned in the specification (one could argue it should not be).
> 
> The options basically are:
> 
>  1) ignore the routing header, that is, act like segments left would have
> been = 0
>  2) drop the packet silently
>  3) drop the packet and send back an ICMP message
>     - with some already defined type/code
>     - with some new type/code
> 
> --
> Pekka Savola                 "Tell me of difficulties surmounted,
> Netcore Oy                   not those you stumble over and fall"
> Systems. Networks. Security.  -- Robert Jordan: A Crown of Swords
> 
> --------------------------------------------------------------------
> IETF IPng Working Group Mailing List
> IPng Home Page:                      http://playground.sun.com/ipng
> FTP archive:                      ftp://playground.sun.com/pub/ipng
> Direct all administrative requests to [EMAIL PROTECTED]
> --------------------------------------------------------------------
--------------------------------------------------------------------
IETF IPng Working Group Mailing List
IPng Home Page:                      http://playground.sun.com/ipng
FTP archive:                      ftp://playground.sun.com/pub/ipng
Direct all administrative requests to [EMAIL PROTECTED]
--------------------------------------------------------------------

Reply via email to