DHCP addressing (was: Broadcast traffic on vlans leaks into the parent interface on NetBSD-5.1)

[Edgar Fuß <ef%math.uni-bonn.de@localhost>]

> A dhcp client that cannot accept a broadcast reply is fundamentally broken.
> Sending the reply that way has always been one of the options for the
> server.
Hm, I would regard a server that violates an RFC SHOULD clause as somewhat
broken. RFC 2131, Section 4.1 reads:
> If the BROADCAST bit is cleared to 0, the message SHOULD be sent as an IP 
> unicast to the IP address specified in the 'yiaddr' field and the link-layer 
> address specified in the 'chaddr' field.
It's too long that I wrote a minimalistic DHCP server as part of a captive 
portal (no puns on that, please, I was forced to do that) and I think there 
were clients that would (as most clients do) clear the bradcast bit and then 
not receive server replies that had been broadcast.