[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: RFC 5006 status
On Thu, Mar 18, 2010 at 12:19:32PM +0100, Philip Homburg wrote:
> In your letter dated Thu, 18 Mar 2010 11:56:05 +0100 you wrote:
> >On Thu, Mar 18, 2010 at 10:38 AM, Philip Homburg
> ><pch-v6ops@u-1.phicoh.com> wrote:
> >> From an implementors point of view, the main thing I find annoying
> >> about RFC-5006 is that information about DNS servers ends up in the
> >> kernel instead of in user land where I need it. And I don't want
> >> the kernel to store and forward all kinds of user land data. Of
> >> course, it is always possible to send an extra router solicitation
> >> ICMP from user land...
> >>
> >
> >As for as I know, rdnssd and radns both process the RA information in
> >userspace. There's no handling of DNS in the kernel.
>
> A quick look at radns suggests that it doesn't send a router solicitation
> messages. So, if it misses the initial one that is requested by the kernel, it
> may take a while before you have your DNS addresses.
>
> I din't check the rdnssd code, but the manual says:
> "On Linux, since version 2.6.24, rdnssd takes advantage of a new
> "netlink interface, forwarding RDNSS options validated by the
> "kernel to user- land. Otherwise, it merely listens to all ICMPv6
> "traffic through a raw socket.
>
> Which suggests extra kernel code, or the same problem as radns.
Extra kernel code, yes. But it doesn't store or process anything related
to the DNS servers. In the same way it already forwards address, prefix,
lifetime, etc., it forwards the user options through netlink.
It is not much more than what was previously forwarded, and the code
is not specific to RDNSSD (it is a more generic "useropt" information).
See here for the code:
http://lxr.linux.no/linux+v2.6.33/net/ipv6/ndisc.c#L1373
regards,
Benoit
--
:wq