mladder.S, fe25519_{mul, square, nsquare}.S: Kaushik Nath
Some optimization ideas have been taken from the s2n-bignum library.
other code: see amd64-64