Cryptodoc Update for ML-KEM #239
Conversation
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
This comment was marked as resolved.
This comment was marked as resolved.
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
This comment was marked as resolved.
This comment was marked as resolved.
Sorry, something went wrong.
This comment was marked as resolved.
This comment was marked as resolved.
reneme
force-pushed
the
cryptodoc/ml-kem
branch
from
e5131ed to
6cd646c
Compare
This comment was marked as outdated.
This comment was marked as outdated.
This comment was marked as resolved.
This comment was marked as resolved.
reneme
force-pushed
the
cryptodoc/ml-kem
branch
from
616c7f5 to
19684d4
Compare
FAlbertDev
force-pushed
the
cryptodoc/ml-kem
branch
from
e3f313a to
254a4b2
Compare
docs/cryptodoc/src/05_09_ml_kem.rst
Outdated
| ``KyberSymmetricPrimitives`` class | ||
| (:srcref:`[src/lib/pubkey/kyber/kyber_common]/kyber_symmetric_primitives.h:30|Kyber_Symmetric_Primitives`). | ||
| This class provides an interface for the primitives, which are defined as | ||
| :math:`PRF`, :math:`H`, :math:`J`, :math:`G`, and :math:`XOF` in Section 4.1 of |
There was a problem hiding this comment.
In the algorithms below, double backticks are used mostly, it seems
There was a problem hiding this comment.
I agree, double backticks are more consistent 👍
| **Steps:** | ||
|
|
||
| 1. Generate the random seed ``seed.d`` and the implicit rejection value ``seed.z`` at random using ``rng`` | ||
| 2. ``(rho, sigma) = G(seed.d || k)`` |
There was a problem hiding this comment.
ML-DSA uses more math notation for rho and sigma etc (which I like) but this has to be a bit more harmonized. Also, see above, G is written in math mode in the "Symmetric primitives" section but here part of a text in double backticks.
There was a problem hiding this comment.
I think I would personally prefer something like
:math:(\rho, \sigma) = G(seed.d || k)
but I don't have very strong feelings and it applies to quite a few places.
There was a problem hiding this comment.
For the sake of consistency, I suggest that we stay with double backticks. This is more consistent with all other algorithms (besides ML-DSA). We will align ML-DSA with this syntax as well.
With the published ML-KEM specification (FIPS 203) and Botan's changes, our cryptodoc must be adapted accordingly.
The implementation and structure of ML-KEM have changed significantly due to extensive refactoring and changes from the specification. Therefore, I rewrote most parts (almost everything) and restructured the format so it matches with the other new algorithms.