Subject: Re: gre(4) conformance
To: None <email@example.com>
From: Hitoshi Asaeda <Hitoshi.Asaeda@sophia.inria.fr>
Date: 06/11/2002 11:38:15
> gre(4) basically talks about RFC1701/1702 GRE (status - Informational),
> and many of the header fields are defined (but not used).
> now there's a standard track document RFC2784 GRE, which simplifies
> many of the protocol features. which one do we want to conform to?
> are there needs for RFC1701/1702 support?
In fact, at the design stage of new GRE (rfc2784) after rfc1701/1702
was published, we needed to define it as a PS, since several protocol,
like MSDP, UDLR, mobileIP etc. were about to use it as a regular
tunnel link. To be simplify it, we decided only the basic components
should be defined as a PS, and rfc2784 was published.
But after that, mobileIP team insisted on the lack of components in
rfc2784. So, one addendum, rfc2890, was issued. It is a PS.
Also, PPTP (rfc2637) uses GRE with specifying version field. It is
I don't have strong opinion to eliminate "greconfig" and keep current
gre(4) implementation only to support rfc2784, but we need to
understand if these protocol will be raised to be supported in a
pseudo device, we may resume obsoleted greconfig again.
(Anyway, only supporting rfc2784 means no support IPv6.;)
But, I voted "keep greconfig" to retain the possibility to enhance
rfc2784 for the future use, if we don't have time to rewrite gre(4)
and greconfig now. Just my comments.