dkm

Lwt approach might be a little heavy, but my inclination goes in that general direction: instad of making sure all code paths are the same length, sometimes at increasing costs...

@hannesm Exactly. Threat model. In my mind, the threat model for the first cut of the code is strictly a network attacker. Cache-eviction, power measurements and other local side channel...

@vouillon Are you sure `Cstruct.sub` is not constant time? Are you thinking about something subtle in the ordering of its three checks (cstruct.ml: 73)? Because the core operation is still...

In essence, I believe it is just sound engineering to go step by step not to get lost along the way. And even the core logic is a mouthful. (Cf....

My (rather vague) idea was that this is the `_unix` module, so I didn't make any effort to make it Win compatible. Can you even compile the library on Windows...

@toolslive What is your scenario here? Where do you deploy to, and what capabilities do CPUs there have? How unacceptable is it for you to `opam pin nocrypto` to the...

You mean, a little like [this](https://github.com/mirleft/ocaml-nocrypto/blob/master/src/native/cpudetect_static.c)? It's quite clear how to go about detecting features, the problem is that the dispatch phase complicates code paths and, as I said, sufficiently...

I'm still trying to gauge how to proceed here. While it is possible to influence the build with configure flags, I'm not sure that the current state is good. I'm...

Around [five](https://github.com/mirleft/ocaml-nocrypto/blob/master/src/native/aes/aesni.c#L324) just for AES. I am working on an accelerated GCM which adds several more varying entry-points, and actually the OCaml code has to change depending on presence of...

That was my idea, yeah. I wanted to have a few distinct bits of code that have different and feature-dependent code paths, then figure out how to neatly swap out...