On Fri, Apr 22, 2005 at 04:47:11PM -0700, Bill Studenmund wrote:
> On Fri, Apr 22, 2005 at 03:18:56PM -0700, Jonathan Stone wrote:
> > 
> > I have no clue what the motivation is here.  But I have a good working
> > knowledge of IPv4, and with that knowledge (but no idea what you are
> > trying to achieve), this entire proposal looks wrong, unacceptably
> > wrong -- to the point where I would back out such a patch, with
> > preduice, as being clearly wrong.
> > 
> > In particular, trying to force IPv6-style "link local" semantics on to
> > RFC-1918 or zeroconf addresses *is* incorrect, and will break valid
> > uses of IPv4.
> 
> How so?
> 
> As I understand what's being suggested, it's a form of policy to be
> applied when determining the source IP of packets in the case where: 1)
> the sending application has not bound the source of the packets, and 2) we
> have multiple IP addresses on the outgoing NIC.

That is correct.

> We thus at present have an
> ambiguity as to which IP to choose. So how is adding policy to pick one of
> them going to break valid uses of IPv4? Those uses have no assurances at 
> present, so how will they get something they might not have gotten now?

If anybody can think of a case, please let me know!

Dave

-- 
David Young             OJC Technologies
dyoung@ojctech.com      Urbana, IL * (217) 278-3933