Re: mech-v2-05pre

[Radhakrishnan Suryanarayanan <rkrishnan.s@samsung.com>]

Hi pekka,
   I agree with your view that having those checks helps. I think we should
retain them.
The idea is only to indicate to the implementor of the need for the check to
filter out non-ipv6 packets while decapsulating like any other layer 2
protocol implementation does.
Rgds
Radhakrishnan
----- Original Message ----- 
From: "Pekka Savola" <pekkas@netcore.fi>
To: "Vladislav Yasevich" <vladislav.yasevich@hp.com>
Cc: "Fred Templin" <osprey67@yahoo.com>; "O.L.N.Rao" <olnrao@samsung.com>;
"Radhakrishnan Suryanarayanan" <rkrishnan.s@samsung.com>; "Erik Nordmark"
<Erik.Nordmark@sun.com>; <v6ops@ops.ietf.org>; "Alex Conta" <aconta@txc.com>
Sent: Thursday, August 26, 2004 10:13 AM
Subject: Re: mech-v2-05pre


> On Wed, 25 Aug 2004, Vladislav Yasevich wrote:
> > Wouldn't the version check be part of the standard IPv6 processing?
> > Why do we want to mandate it here as well?
> >
> > I can see the need to make sure that we have enough data, so the
> > decapsulator may want to check that it has at least 40 bytes of
> > payload, but the version check seems redundant to me.
>
> Well.. folks asked for the version check..
>
> I can see a justification in the sense that AFAIK it's not strictly
> necessary to check the IP version when processing the packets, because
> the lower layer (ethernet protocol type, protocol 41, etc.) already
> indicated the ip version of the payload.  When using a tunnel, the
> chance of someone sending packets which aren't really IPv6 but using
> proto-41 (or the like) seems significantly larger than e.g., using
> Ethernet.
>
> So, in that sense, both checks are redundant, just spelling out checks
> which should be done for IPv6 packets in any case, but I don't think
> it hurts to have them here.
>
> If other set of folks think it's bad idea to have them here, we could
> just remove the paragraph (and it would be as it was before), or add
> something like, "As with processing any other IPv6 packet received
> from link-layers, ...." to clarify that this particular link layer
> should not be special...
>
> Thoughts?
>
> Note: I already submitted -05 with the change.
>
> -- 
> Pekka Savola                 "You each name yourselves king, yet the
> Netcore Oy                    kingdom bleeds."
> Systems. Networks. Security. -- George R.R. Martin: A Clash of Kings
>
>
>