Re: [TLS] Update spec to match current practices for certificate chain order

[[e-mail] (Martin Rex)]

Kemp, David P. wrote:
> Martin Rex wrote:
>> Clients do not have to accomodate server that violate MUSTS.
> 
> Absolutely correct.   Accommodating a compliant server is the
> only mandatory to implement requirement on clients.
> 
>> In most cases, it is detrimental and often dangerous to work around
>> such protocol violations by the peer.
> 
> How ???

Detrimental because it hides a server configuration error that can
only be fixed on the server.  The resulting behaviour of the end-user
becomes seemingly non-deterministic.  It may work with some clients
and not work with others, and it may work after having been to other
servers, but not directly after reboot.  Just crazy.


> 
>> Either the certification path presented by the server crosses through
>> the trust anchor(s) of the client or ends there -- then you get interop,
>> or it doesn't then you get interop failure.
> 
> Clients that provide value-added (optional) path construction
> functionality do not fail, either operationally or security-wise.

I've seen epic failures here before.  Such as chasing AIA for missing
intermediate certs (not having a trusted cert store that contains it).
That's a real security goof, and it was implemented by some clients.


>
> Perhaps no such clients exist, now or in the future.  But if they do,
> satisfying more than the MUSTS does not make them non-compliant.

For a number implementors, convenience is more important than security,
and they will skip anything that is difficult.  Such as skipping
certificate and certicate path validation for the peer completely.


-Martin