[Cryptech Tech] Cuve25519 HW implementation now available

[Randy Bush]

i would have approached lange first.  she is radically open source and
tim was her student.

> I've looked through the code. As far as I can see, there is no top level
> tests that actuallt checks that the core perform a correct Curve25519
> operation. There are testvectors for addition and multiplication but no
> expected result at the top.

make a long tape of rng, cleartext, etc.  play it against verilog
version and softwre version.

randy