Hi Ard, On Tue, Jan 4, 2022 at 6:03 PM Ard Biesheuvel <ardb@xxxxxxxxxx> wrote: > The only downside here is that the ARM/x86 accelerated shashes and the > generic shash now use the same core transform, right? Given that the > generic blake2s shash is never used for anything in the kernel, the > only reason for its existence was to be able to use the randomized > crypto testing infrastructure to test the arch code. > > Ergo, there is no point in retaining the blake2s shashes and we can > simply remove all of them. (Note that blake2b is used as an shash via > the crypto API by btrfs, but blake2s is only used via the library API) That makes sense and is fine with me. Let's do that in a separate commit later. I've got a bunch of things I'd like to fix up in the general lib/crypto vs crypto split that are kind of interrelated. Jason
