On Thu, Nov 20, 2025 at 09:35:24PM +0000, Sipos, Brian J. wrote:
> WG,
> 
> In reviewing existing symmetric keyed algorithms defined in RFC 9053, it
> seems like there is an implicit constraint on input key length which is not
> actually required anywhere. I'm not sure that this constraint actually
> matters in practice, because I expect that using an alternate input key
> value would fail verification/decryption but it seems like something that
> would be a useful check for an implementation to make (rather than wasting
> the time to verify / decrypt and then discover the failure).
> 
>  
> 
> The algorithms are in Section 3.1, 3.2, 4.1, 4.2, 4.3, 6.1.1, 6.1.2, and
> 6.2.1 all with the constraint:
> 
> The "kty" field MUST be present, and it MUST be "Symmetric".
> 
>  
> 
> The definition in Sections 3.1 actually does add a constraint that the
> others appear to not have:
> 
> Implementations creating and validating MAC values MUST validate that the
> key type, key length, and algorithm are correct and appropriate for the
> entities involved.
>
> Or maybe the general requirement in Section 3.1 actually is meant to apply
> to all MAC algorithms and not just the ones in that subsection..?

Well, this gets even odder:

- Section 3.1 (HMAC) can accept key of any length.
- As can Section 6.1.2 (Direct Key with KDF).
- The rest can only take key of fixed length.
- The only other MAC algorithm is section 3.2 (which takes key of fixed
  length).




-Ilari

_______________________________________________
COSE mailing list -- [email protected]
To unsubscribe send an email to [email protected]

Reply via email to