(usagi-users 00520) Re: USAGI integration policy [was: Problem receiving the IPv6 headers]

[Pekka Savola <pekkas@xxxxxxxxxx>]

On Fri, 11 May 2001 itojun@xxxxxxxxxx wrote:
> >For example, KAME would be nothing (compared to what it is now, anyway) if
> >it wasn't integrated and maintained (in an excellent fashion, btw.) in
> >different BSD's; as a separate patch, it'd only be used by people who are
> >very seriously into IPv6.
>
> 	FYI, KAME has been distributed as a separate patch for 2+ years prior
> 	to the merge, and the situations that days were just like above
> 	("only be used by serious IPv6-ers").
>
> 	and prior to the distribution, we have been working on the codebase
> 	for 2+ years.  it takes time.

Yes, I'm aware of that :-).

However, I think the scenario is a little different here.  As far as I
know, there was no IPv6 implementation in BSD's _at all_ before KAME.
Only "competing" IPv6 patch by INRIA I think.

The only way to get IPv6 gracefully was a complete merge when it had
proved stable and when people started getting interested of IPv6 once
again.

Here, there already exists, to varying extent, an IPv6 implementation that
kinda works (to varying degree).  The patching is currently the only way
to go here.  The maturing period does not need to be so long when
singular issues, not the whole package, is considered.

I feel that the Linux kernel implementation will have to rewritten, at
least some parts of it, at some point.  Reading certain parts of it,
e.g. ND and verifying it against RFC2461 almost made me feel sick.

It isn't time for that yet, though; a lot more operational experience and
development effort is required first, in order to be able to get the full
benefits from that effort if it's undertaken.

-- 
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