Added new 64-bit implementations of Curve25519 and P-256.
Fixed endianness in Curve25519 implementation (no consequence on security). Also added new Curve25519 code for 64-bit platforms.
Changed speed benchmark for i31 to a 521-bit modulus.
Added SHAKE implementation.
Added POWER8 implementation for AES/CTR+CBC-MAC (for CCM and EAX modes).
Added AESCTR_DRBG implementation (beta).
Added RSA key generation code (i15, i31, i62).
Fixed test code (removed static reference to aes_x86ni code).
Added generic EAX and CCM implementations.
Added ChaCha20 implementation with SSE2 opcodes.
New "i62" code for big integers with 64x64->128 opcodes; also improved "i31" modular exponentiation.
Added "ctmulq" implementation of Poly1305 (using 64->128 multiplications when available).
New AES and GHASH implementations using POWER8 crypto opcodes.
Added AES+GHASH implementation using AES-NI opcodes; also ARM-Thumb assembly for faster Montgomery multiplication on Cortex-M0+. Added selection functions for "default" implementations.
Some cleanups (removed unused files, split i15 code into per-function files).
Added optimised implementation of P-256 that uses 32->64 multiplications (MUL31).
Two new Curve25519 implementations (generic "i31" code, and optimised code with MUL31).
New implementation of Curve25519 (using multiplications of words of 15 bits or so, should be much faster on Cortex M0).
New basic implementation of Curve25519 (generic i15 code, experimental).
More optimisations for EC P-256 "i15" (specialised squaring function, mixed coordinates addition with a 4-bit window when the base point is the conventional generator).