Watson Ladd wrote on 27.05.2014 15:47:
> On Mon, May 26, 2014 at 2:18 AM, Johannes Merkle
> <[email protected]> wrote:
>> I had suggested two additions, for which there has been no decision yet. I 
>> would like to see more discussion on these.
>>
>> 1. Usage of Extension SignatureAlgorithms by client to signal minimum 
>> required hash functions. Ralph's feedback on that
>> suggestion was that we need more data on the extend of SHA-2 support by 
>> current and future implementations. But I think
>> it's not just about migrating to SHA-2 it could also be useful to prevent 
>> usage of MD5.
> 
> MD5 in TLS 1.2 should not be used anymore. There are some real clever
> tricks that can be played. We need a secure hash every time. However,

I absolutely agree.

> the ciphersuite, not the SignatureAlgorithms, determines which hash to
> use. If we are talking about certificates, then I think this is the
> correct thing, but I don't know.
> 
the Signature Algorithm extension is neither about HMAC nor about certificates, 
its about signatures on key exchange
messages, in case of PFS the signature of the ephemeral DH keys.


> (I don't understand why MD5 was still included as an option in any RFC
> after 2004)
> 

+1



>>
>> 2. When re-using keys for ECDHE (which is the default behavior in some 
>> implementations, e.g. OpenSSL) or when using
>> non-ephemeral ECDH, the validity of the received public DH-key should be 
>> checked to avoid non-group attacks. That is, it
>> should be checked that the received point P is on the curve (unless point 
>> compression was used). Small subgroup checks
>> could even be recommended for classical DH. Something in the spirit of RFC 
>> 6989.
> 
> This is a problem for ephemeral DH as well due to Triple Handshake. We
> might as well throw this in: it doesn't hurt. However, if you aren't
> doing it already, odds are you aren't capable of implementing TLS
> correctly, because you don't understand the issues associated with
> implementing cryptography.

True, but I'm not sure if we can assume that this is not the case.

-- 
Johannes

_______________________________________________
Uta mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/uta

Reply via email to