[openpgp] Re: Persistent Symmetric Keys: new algorithms or new packets?

Andrew Gallagher <[email protected]> Wed, 23 Jul 2025 15:37:35 +0100

Newsgroups	gmane.ietf.openpgp
Message-ID	<[email protected]>

On 23 Jul 2025, at 14:58, Daniel Huigens <[email protected]> wrote:
> 
> Or we could set it to 0, which is reserved?

0 has a consistent interpretation (if allocated!) in the algorithm registries as “None”. So the “natural" meaning of public-key algorithm 0 in a PKESK or signature packet would be “no asymmetric algorithm” - which is more or less the meaning we wish to convey; so I think the use of 0 as an exceptional code point here is justifiable.

The only concern I would have is that at some point we may want to specify something other than AEAD/HMAC, and there is only one exceptional code point, so we would need to prefix a “version” or “mode" octet to the algorithm-specific data to allow future flexibility. But on the other hand the use of an exceptional 0 allows implementations to easily detect the use of persistent symmetric modes, without reserving blocks in the public-key registry.

A

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

signature.asc (application/pgp-signature, 833 B)