> On Jan 18, 2023, at 12:55 PM, Ilari Liusvaara <[email protected]> 
> wrote:
> 
> On Wed, Jan 18, 2023 at 01:53:10PM -0500, Russ Housley wrote:
>> Laurence:
>> 
>> Obviously it is too late to add anything to RFC 9052.
>> 
>> I propose the following for draft-ietf-cose-aes-ctr-and-cbc:
>> 
>> X.  Implementation Considerations
>> 
>>   COSE libraries that support either AES-CTR or AES-CBC and accept
>>   Additional Authenticated Data (AAD) as input should return an error
>>   if one of these non-AEAD content encryption algorithm is selected.
>>   This would ensure that a caller does not expect the AAD to be
>>   protected when the cryptographic algorithm is unable to do so.
> 
> RFC 9052 already requires that if using AE algorithm:
> 
> - Attempt to encrypt or decrypt with non-empty external aad MUST fail.
> - Attempt to encrypt with any protected header parameters MUST fail.
> - Atttept to decrypt message with any protected header parameters MUST
>  fail.
> 
> It seems to me that AES-CTR and AES-CBC are analogous to these, but
> with even more serious limitations, so all these limitations should be
> inherited.

After reading section 5 in 9052 more carefully, I’d like to suggest that 
draft-ietf-cose-aes-ctr-and-cbc give normative instructions that say how to use 
a fully non-authenticated encryption algorithm. Technically speaking, neither 
5.2 (AEAD) or 5.3 (AE) apply to AES-CTR or AES-CBC since they have no 
authentication. Those instructions amount to following 5.3 so it’s not a big 
deal, but from a standards completeness view it seems necessary.

We double-warn about the danger of AES-CTR in draft-ietf-cose-aes-ctr-and-cbc 
(in the main body and in security considerations) plus its dangers are also 
well described externally. Thus seems about equivalent to also warn about AAD 
in draft-ietf-cose-aes-ctr-and-cbc.


(For me, there is also still the issue of warning that AEAD doesn’t really 
provide the same level of authentication that public key signing does, 
especially with one-time random CEKs, but I will get to that in another message 
once I understand better how SUIT does sign-then-encrypt).

LL
_______________________________________________
COSE mailing list
[email protected]
https://www.ietf.org/mailman/listinfo/cose

Reply via email to