p384: v0.11 release tracking issue

[tarcieri]

This is a tracking ticket for work items it would be nice to have before cutting a final v0.11 release which includes an initial arithmetic implementation.

NOTE: not all of these need to be completed prior to a release and they can be added after-the-fact.

• p384: initial 32-bit support #573
• Test vectors
  • p384: initial test vectors for group operations #567
  • p384: scalar multiplication test vectors #570
  • p384: Initial inclusion of FIPS 186-4 CAVP test vectors for ECDSA #569
  • p384: add Wycheproof test vectors #574

cc @brycx @jedisct1

[tarcieri]

@daviddrysdale would you be interested in helping add Wycheproof test vectors?

[tarcieri]

@jedisct1 did you want to take a look at adding 32-bit support? If not, I probably can in the next few days

[tarcieri]

#573 implemented (currently untested) 32-bit support except for field and scalar inversions, which presently use a 64-bit only implementation.

Once that has been addressed, we can add cross tests for 32-bit targets, and after that I'd say it's ready to release.

[daviddrysdale]

@daviddrysdale would you be interested in helping add Wycheproof test vectors?

Happy to put in something analogous to the P256 test vectors, but it doesn't seem like the various traits (e.g. SigningKey, VerifyingKey) needed to use new_wycheproof_test! are available yet?

[tarcieri]

@daviddrysdale I think I got it taken care of in #574

[brycx]

@tarcieri Is there anything left blocking the 0.11.0 release? I see you've changed the inversion to work with the 32-bit backend as well 🚀

[tarcieri]

I was curious to try this transformation on the field arithmetic and what impact it would have on performance: RustCrypto/utils#768

I wouldn't consider it a release blocker, but it would be nice to have at least compile-time transformations to/from Montgomery form, if only so field element constants could be given in canonical form.

It would also potentially enable computing basepoint tables at compile time.

[tarcieri]

#589 adds const fn support to field elements.

After that it would be nice to expose const fn wrappers for the field arithmetic functions, and then rewrite the various constants back into canonical form, transforming them into Montgomery form at compile time.

And after that, I'm good to ship it!

[tarcieri]

Released in #593