DHCP over PPP - or how to use sockets instead of BPF?
Ted.Lemon at nominum.com
Sat Apr 17 00:04:01 UTC 2010
Client identifier trumps chaddr, so it's no problem. Just don't set the broadcast bit. This is how DHCP-over-infiniband works, IIRC.
On Apr 16, 2010, at 12:16 PM, Patrik Lahti <plahti at qnx.com> wrote:
> On 16/04/10 02:04 PM, sthaug at nethelp.no wrote:
>>> But it would be nice to allow DHCPv4 to "work" somehow. Otherwise, how
>>> do people automatically get DNS servers on PPP links?
>> Basic PPP negotiation can give you an IP address. And then DHCPINFORM
>> can give you DNS servers and other interesting parameters. DHCPINFORM
>> requires a working IP address.
>> Steinar Haug, Nethelp consulting, sthaug at nethelp.no
> That's a good idea. I haven't really dived into DHCPINFORM, I'm assuming
> it's a bit like stateless DHCPv6 (Information-Request)?
> It's just that in order to form a proper DHCPv4 message one would need
> to fill the CHADDR field. And from a quick look through RFC 2131 it
> seems to require CHADDR for DHCPINFORM too. I realize that in most cases
> the CHADDR field wouldn't matter for DHCPINFORM. But it would be
> important to be interoperable, in case servers/relays expect something
> useful there (e.g. for matching the CHADDR to certain config parameters).
> Are you suggesting faking this hardware address or do one of the options
> in in RFC 3456?
> I noted:
> https://lists.isc.org/pipermail/dhcp-users/2009-June/009072.html which
> suggests at least one client implementation fakes it.
> dhcp-users mailing list
> dhcp-users at lists.isc.org
More information about the dhcp-users