Re: [dhcwg] DHCPv4

On Sat, May 07, 2005 at 01:35:26PM +0200, Bud Millwood wrote:
> And for David: I seem to recall that the ISC DHCP server uses some type
> of raw 
> interface to the network layer; might the ethernet header size be included 
> for that reason?

Yes, this was probably done to ensure that the output packet would never
exceed 1500 bytes (due to some platform limitation in passing data in sizes
larger than that, even though through raw sockets, I am imagining).

But this is calculating the maximum size of the "options" space from the
client-supplied mms, and consequently the maximum size of the output packet
since that's the only variable, rather than directly calculating output
packet space.

It doesn't matter if the client (or server) is not ethernet, and it doesn't
matter if the server is using Berkeley sockets.  It's applied in every
case.

I think this is pretty clearly a bug on our part, but the fact that this
is in all released versions of the software of which I am aware (this
calculation of "udp overhead" is in RCS version 1.1 of the header, so it
appeared in all version-2 releases), I think muddies the 'what does MMS
mean' question that's being asked here.

-- 
David W. Hankins		"If you don't do it right the first time,
Software Engineer			you'll just have to do it again."
Internet Systems Consortium, Inc.		-- Jack T. Hankins

Re: [dhcwg] DHCPv4 - definition of maximum message size.