RISC-V extension selection is not completely accurate

[EAlexJ]

This might be a bit knit picky, but the the --enable-riscv=<extensions> is not quite correct.

Namely, the Zkn extension is a superset of others, currently supported and used are Zbkb, Zbkc, and Zkned.
When selecting Zkn it gets treated as if only Zkned is selected.
This means when working with a core that fully implements the Zkn or even the Zk extension, the correct string at this point needs to be --enable-riscv=zbkc,zbkb,zkn, while just specifying Zkn should be sufficient.

This does not apply to the vector set of extensions, as there simply is no Zvkn option (and I do not know if there could be one, it would depend on Zvknhb being used in the library).
I'd argue having no superset for vector extensions is less grave, the naming for the vector cryptography extensions becomes quite involved anyways.

I tried patching this myself, but I have no experience working with an autogen/configure flow.

Let me know if I left anything unclear.

-Alex

[anhu]

I think you mean --enable-riscv-asm, not --enable-riscv ; please correct me if I'm wrong.

[SparkiDev]

Hi @EAlexJ,

I've removed zkn as an option in PR: #10530
Is this what you were expecting?

Thanks,
Sean

[EAlexJ]

Yes, I meant enable-riscv-asm.

Removing zkn works just fine, as it does not suggest that all the extensions are actually used or required