Re: Debian IPv6 packages

[Brian May <bam@debian.org>]

>>>>> "Wichert" == Wichert Akkerman <wichert@cistron.nl> writes:

    Wichert> I don't think Debian should get to a point where we can't use a
    Wichert> stock Linus kernel without risking to break things.

    Wichert> I'm wondering though, how come that nobody is working in
    Wichert> getting USAGI merged in the main kernel tree, or fixing
    Wichert> its IPv6 support?

>>>>> "Arkadiusz" == Arkadiusz Miskiewicz <misiek@pld.ORG.PL> writes:

    Arkadiusz> Now the problem is that USAGI stuff isn't merged with
    Arkadiusz> main kernel source.  Just make USAGI people to merge
    Arkadiusz> it... Right now they are doing kind of ,,another''
    Arkadiusz> kernel && glibc as this is in *BSD family ;(

So everyone seems to agree that the only way to fix this is be
incorporating the USAGI stuff in the kernel?

Are there any downside to the USAGI stuff? ie. has it been tested to
the point where it is known to work, and not break anything else?

Still, I think the only way to get it incorporated by into the
upstream kernel source, is by demonstrating that lots of end users use
it (At least thats how I read Linus statement as why Reiserfs was
included in 2.4.1; his opinion seems to be that if it is only used for
experimentation then it is easy to apply the patch).

I think the only way we can do this is by making it USAGI a
requirement for IPv6 packages.

At which point, we start going around in circles again.
-- 
Brian May <bam@debian.org>