[openpgp] Re: session key length with SEIPDv2

[Daniel Kahn Gillmor <[e-mail]>]

Thanks Heiko for raising this, and thanks Daniel for commenting on it.

I agree it is a disappointing oversight that we didn't explicitly
specify the length of the session key for SEIPD v2.

I also think that no one implementing this is likely to just pick an
arbitrary size.

On Mon 2024-09-23 23:14:24 +0000, Daniel Huigens wrote:
> If others (e.g. the chairs) agree that we should file an erratum for this,
> I would tend to keep it simple and just say in Section 5.1 and 5.3 that:
>
>     The length of the session key MUST be the key size of the
>     symmetric algorithm used in the encryption container.

Sure, this covers the producer guidance neatly.

But I don't think the consumer guidance is trivially inferrable from the
above statement.  Should it fail to decrypt if it doesn't match?  What
if it's *larger* than the key size of the symmetric algorithm?  what if
it's half the size?

This might be resolvable with another single sentence (i'd love to hear
proposals), but leaving consumer guidance implicit seems like it is the
core of the original omission.  If we're going to fix this (and i think
we should), we should fix it fully.

I don't know whether such a normative fix is doable in errata, or
whether we need some additional machinery.

            --dkg