On Thu, 21 Aug 1997, Greg A. Woods wrote:
> > RFC 821 specifically forbids refusal of mail on the basis of what the
> > sending host sends in the HELO/EHLO command. However, you can get Exim
> > to check it. Just set the helo_verify_hosts or helo_verify_nets option.
>
> Actually it's RFC 1123 that says this, though the folks working on the
> "son-of-821" draft have merged most of the 1123 recommendations.
Oops. Sorry. I confess not to having checked precisely where it was
said. Too many RFCs and proto-RFCs around these days!
> In a corporate world justifying non-verification and early rejection by
> citing the "robustness principle" is probably enough to get you fired as
> a security officer.
Exim can be configured to do the reverse lookup, and if what HELO said
was "wrong", it will correct the data in what it logs. However, it can't
be configured to reject the message, and it also ploughs on if the
lookup fails. Obviously it would be fairly straightforward to add
options to make it fail if it can't look up the name.
> I think
> by now we can depend on experts to correctly configure their software.
Maybe, but you can't depend on non-experts to configure their PCs!
Personally, I don't like the idea of wasting bandwidth doing all those
lookups when 99% of the time you get the "right" answer and in any case
the IP address gets logged. I've configured our mailers here just do do
the check for machines on the local net. That way I correct the
information in the Received: fields for messages from local PCs that are
misconfigured.
> there are a few so-called experts
> who will use the robustness principle to justify their own intentional
> non-compliance of other rules.
Agreed.
--
Philip Hazel University Computing Service,
ph10@??? New Museums Site, Cambridge CB2 3QG,
P.Hazel@??? England. Phone: +44 1223 334714
