RFR: 8298127: HSS/LMS Signature Verification [v5]

[Ferenc Rakoczi]

On Mon, 8 May 2023 16:38:01 GMT, Weijun Wang <weijun at openjdk.org> wrote:

>> src/java.base/share/classes/sun/security/provider/HSS.java line 528:
>> 
>>> 526:         // update()-digest() sequence) which is parametrized so that the digest output is copied back into this buffer.
>>> 527:         // This way, we avoid memory allocations and some computations that would have to be done otherwise.
>>> 528:         final byte[] hashBuf;
>> 
>> I'm a little worried about the mutability of `hashBuf` and whether it's suitable to be put inside `LMOTSParams`.  By using `of` to return an `LMOTSParams` object we have the chance to return cached objects in the future. There should always be one `hashBuf` for each LM-OTS verification, and this is not clear from the current code.
>
> How will the performance change if we make `hashbufSha256_24` and `hashbufSha256_32` static and each time we want to verify an LM-OTS signature we clone one of them?

Changed. There should not be noticeable performance difference.

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

PR Review Comment: https://git.openjdk.org/jdk/pull/13691#discussion_r1188534788