evp_test (ML_KEM): IKME instead of entropy name, bug fixes & a small refactor#26082
evp_test (ML_KEM): IKME instead of entropy name, bug fixes & a small refactor#26082paulidale wants to merge 2 commits intoopenssl:feature/ml-kemfrom
Conversation
vdukhovni
left a comment
There was a problem hiding this comment.
Generally fine, but please rebase.
|
|
||
| if ((p = OSSL_PARAM_locate_const(params, OSSL_KEM_PARAM_MLKEM_ENC_ENTROPY)) != NULL | ||
| && (p->data_size != MLKEM_ENCAP_ENTROPY | ||
| if ((p = OSSL_PARAM_locate_const(params, OSSL_KEM_PARAM_IKME)) != NULL | ||
| && (p->data_size < MLKEM_ENCAP_ENTROPY | ||
| || (ctx->entropy = OPENSSL_memdup(p->data, MLKEM_ENCAP_ENTROPY)) == NULL)) | ||
| return 0; |
There was a problem hiding this comment.
This entire file goes away in the upcoming refactor. So it likely best if you rebase on the ML-KEM refactor PR.
There was a problem hiding this comment.
Okay. CI failures are relevant but don't reproduce locally.
There was a problem hiding this comment.
Might it be easier for you to pull the commits here into the refactor branch?
Otherwise, we're going to encounter an ordering conflict which serialises PRs.
There was a problem hiding this comment.
Which CI failures? As for the conflict, I am happy to pull your changes in, but this PR will become obsolete and should be closed. Works for you?
The recent additon of EVP KEM tests for ML-KEM used a new parameter to pass the input entropy for the key. The ECX tests already had a different parameter for this purpose, so the ML-KEM code is updated.
0659143 to
c7b3bd6
Compare
|
Integrated as fixup into the ML-KEM PR. Closing. |
|
Thanks. |
Use the existing IKME param instead of a new one for the key generation input.
Avoid using param builder.
Fix some mistakes in the return code.