RFR: 8347938: Switch to latest ML-KEM private key encoding [v5]

[Ben Perez]

On Wed, 30 Jul 2025 15:45:50 GMT, Weijun Wang <weijun at openjdk.org> wrote:

>> The private key encoding formats of ML-KEM and ML-DSA are updated to match the latest IETF drafts at: https://datatracker.ietf.org/doc/html/draft-ietf-lamps-dilithium-certificates-11 and https://datatracker.ietf.org/doc/html/draft-ietf-lamps-kyber-certificates-10. New security/system properties are introduced to determine which CHOICE a private key is encoded when a new key pair is generated or when `KeyFactory::translateKey` is called.
>> 
>> By default, the choice is "seed".
>> 
>> Both the encoding and the expanded format are stored inside a `NamedPKCS8Key` now. When loading from a PKCS #8 key, the expanded format is calculated from the input if it's seed only.
>
> Weijun Wang has updated the pull request incrementally with one additional commit since the last revision:
> 
>   combine security properties description; remove one test

src/java.base/share/classes/sun/security/provider/ML_DSA.java line 593:

> 591:         var s1 = deepClone(sk.s1);
> 592:         mlDsaVectorNtt(s1); //s1 now in NTT domain
> 593:         int[][] As1 = new int[mlDsa_k][ML_DSA_N];

Multiarray allocations like this are quite slow - for better performance consider using the `integerMatrixAlloc` method. This also applies to the allocations on L598 and L599

-------------

PR Review Comment: https://git.openjdk.org/jdk/pull/24969#discussion_r2252637672