IPv6 DNS config daemon

Justin Sherrill justin at shiningsilence.com
Thu Oct 22 20:05:15 PDT 2015


The question is: will you continue to maintain it, even if that's just
making sure it performs as advertised, for the foreseeable future?
Otherwise, it will bit-rot.

The one repeated reference to SLAAC that I've seen is that it's needed
to get Android clients on a IPv6 network; they won't talk DHCPv6.
Haven't tried it (yet) so I can't describe it better.

On Thu, Oct 22, 2015 at 10:57 PM, Charles Musser <cmusser at sonic.net> wrote:
> While experimenting with IPv6 SLAAC, I wrote a daemon that configures
> a host's DNS resolver using the scheme described in RFC 6106.
> FreeBSD's rtsold(1) and "radns" (in the ports collection) are two
> programs that already do this. However, this one might be nice to have
> in the base system. Features:
>
> - Small, simple design: around 750 LoC in a single source file.
>
> - Developed on DragonFly, so no portability overhead. Tested on
>   FreeBSD too, so probably works on other BSDs.
>
> - No configuration, other than command line flags for foreground
>   mode and debug output.
>
> - Data validation: Performs basic sanity checks on router
>   advertisement options; Checks validity of server addresses and search
>   domains; enforces system-defined size limits for lists of these
>   items.
>
> - Ran under Valgrind' memcheck (on FreeBSD) with no leaks found.
>
> - Not entangled with rtsold(1) which need not be running and might
>   not exist if the kernel took over soliciting router advertisements (as
>   is now done in OpenBSD).
>
> SLAAC is a somewhat dusty corner of IPv6 configuration, so maybe the
> existing programs are good enough. This limited use-case is both an
> argument for and against including a new implementation. Depending
> on your tolerance/enthusiasm level, it's either "why not?" or "why
> bother?"
>
> Is this something the project would want?
>
> Chuck



More information about the Users mailing list