///
/// Copyright (c) 2018 xdx. All rights reserved.
///
/// \file: block.go
///
/// \brief: sm4 block-function
///
/// \author: xdx
///

package sm4

import (
	"encoding/binary"
)

func rotl(x, n uint32) uint32 {
	return (x << n) | (x >> (32 - n))
}

/*
* Expanded SM4 S-boxes
* Sbox table: 8bits input convert to 8 bits output
 */

var sbox = [256]byte{
	0xd6, 0x90, 0xe9, 0xfe, 0xcc, 0xe1, 0x3d, 0xb7,
	0x16, 0xb6, 0x14, 0xc2, 0x28, 0xfb, 0x2c, 0x05,
	0x2b, 0x67, 0x9a, 0x76, 0x2a, 0xbe, 0x04, 0xc3,
	0xaa, 0x44, 0x13, 0x26, 0x49, 0x86, 0x06, 0x99,
	0x9c, 0x42, 0x50, 0xf4, 0x91, 0xef, 0x98, 0x7a,
	0x33, 0x54, 0x0b, 0x43, 0xed, 0xcf, 0xac, 0x62,
	0xe4, 0xb3, 0x1c, 0xa9, 0xc9, 0x08, 0xe8, 0x95,
	0x80, 0xdf, 0x94, 0xfa, 0x75, 0x8f, 0x3f, 0xa6,
	0x47, 0x07, 0xa7, 0xfc, 0xf3, 0x73, 0x17, 0xba,
	0x83, 0x59, 0x3c, 0x19, 0xe6, 0x85, 0x4f, 0xa8,
	0x68, 0x6b, 0x81, 0xb2, 0x71, 0x64, 0xda, 0x8b,
	0xf8, 0xeb, 0x0f, 0x4b, 0x70, 0x56, 0x9d, 0x35,
	0x1e, 0x24, 0x0e, 0x5e, 0x63, 0x58, 0xd1, 0xa2,
	0x25, 0x22, 0x7c, 0x3b, 0x01, 0x21, 0x78, 0x87,
	0xd4, 0x00, 0x46, 0x57, 0x9f, 0xd3, 0x27, 0x52,
	0x4c, 0x36, 0x02, 0xe7, 0xa0, 0xc4, 0xc8, 0x9e,
	0xea, 0xbf, 0x8a, 0xd2, 0x40, 0xc7, 0x38, 0xb5,
	0xa3, 0xf7, 0xf2, 0xce, 0xf9, 0x61, 0x15, 0xa1,
	0xe0, 0xae, 0x5d, 0xa4, 0x9b, 0x34, 0x1a, 0x55,
	0xad, 0x93, 0x32, 0x30, 0xf5, 0x8c, 0xb1, 0xe3,
	0x1d, 0xf6, 0xe2, 0x2e, 0x82, 0x66, 0xca, 0x60,
	0xc0, 0x29, 0x23, 0xab, 0x0d, 0x53, 0x4e, 0x6f,
	0xd5, 0xdb, 0x37, 0x45, 0xde, 0xfd, 0x8e, 0x2f,
	0x03, 0xff, 0x6a, 0x72, 0x6d, 0x6c, 0x5b, 0x51,
	0x8d, 0x1b, 0xaf, 0x92, 0xbb, 0xdd, 0xbc, 0x7f,
	0x11, 0xd9, 0x5c, 0x41, 0x1f, 0x10, 0x5a, 0xd8,
	0x0a, 0xc1, 0x31, 0x88, 0xa5, 0xcd, 0x7b, 0xbd,
	0x2d, 0x74, 0xd0, 0x12, 0xb8, 0xe5, 0xb4, 0xb0,
	0x89, 0x69, 0x97, 0x4a, 0x0c, 0x96, 0x77, 0x7e,
	0x65, 0xb9, 0xf1, 0x09, 0xc5, 0x6e, 0xc6, 0x84,
	0x18, 0xf0, 0x7d, 0xec, 0x3a, 0xdc, 0x4d, 0x20,
	0x79, 0xee, 0x5f, 0x3e, 0xd7, 0xcb, 0x39, 0x48,
}

func tau(x uint32) uint32 {
	return uint32(sbox[x&0xff]) |
		uint32(sbox[(x>>8)&0xff])<<8 |
		uint32(sbox[(x>>16)&0xff])<<16 |
		uint32(sbox[(x>>24)&0xff])<<24
}

func l(x uint32) uint32 {
	return x ^ rotl(x, 2) ^ rotl(x, 10) ^ rotl(x, 18) ^ rotl(x, 24)
}

// sbox8 分别rotl 8, 16, 24
var sbox8l8, sbox8l16, sbox8l24 [256]uint32

// sbox16 maps a uint16 to uint32
// sbox16(b0||b1) = L(t(b0)) ^ L(t(b1) << 8)
var sbox16 []uint32

// sbox16Rotl = sbox16[i] << 16
var sbox16Rotl []uint32

const useSbox65536 = true

func init() {
	if useSbox65536 {
		sbox16 = make([]uint32, 65536)
		sbox16Rotl = make([]uint32, 65536)
		for i := 0; i < 65536; i++ {
			x := l(uint32(sbox[i&0xff])) ^ l(uint32(sbox[(i>>8)&0xff])<<8)
			sbox16[i] = x
			sbox16Rotl[i] = (x >> 16) | (x << 16)
		}
	}

	for i := 0; i < 256; i++ {
		sbox8l8[i] = rotl(ltTable8[i], 8)
		sbox8l16[i] = rotl(ltTable8[i], 16)
		sbox8l24[i] = rotl(ltTable8[i], 24)
	}

}

// lt return L(Tau(x))
func lt(x uint32) uint32 {
	if useSbox65536 {
		// FIXME: 为什么会更慢？（C下更快）
		// sbox16 and sbox16Rotl taken 512KB memorys. too big and cache may always missing.
		// thus is slower than 1.
		return sbox16Rotl[x>>16] ^ sbox16[x&0xffff]
		// return ltauTable16[x&0xffff] ^ rotl(ltauTable16[x>>16], 16)
	} else {
		// 1. 4次右移，4次&， 4次查表
		return sbox8l24[(x>>24)&0xff] ^ sbox8l16[(x>>16)&0xff] ^ sbox8l8[(x>>8)&0xff] ^ ltTable8[x&0xff]

		// 2. 4次右移，4次&， 4次rotl 比1略慢
		// return rotl(sbox8[(x>>24)&0xff], 24) ^ rotl(sbox8[(x>>16)&0xff], 16) ^ rotl(sbox8[(x>>8)&0xff], 8) ^ sbox8[x&0xff]

		// 3. 至少慢一半
		// return l(tau(x))
	}
}

/*
Encrypt one block
Can asm code be much more faster?
*/
func encryptBlockGo(key []uint32, output, input []byte) {
	// var a, b, c, d uint32
	a := binary.BigEndian.Uint32(input[:4])
	b := binary.BigEndian.Uint32(input[4:8])
	c := binary.BigEndian.Uint32(input[8:12])
	d := binary.BigEndian.Uint32(input[12:16])

	for i := 0; i < 8; i++ {
		e := c ^ d
		a ^= lt(e ^ b ^ key[4*i+0])
		b ^= lt(e ^ a ^ key[4*i+1])
		e = a ^ b
		c ^= lt(e ^ d ^ key[4*i+2])
		d ^= lt(e ^ c ^ key[4*i+3])
	}
	binary.BigEndian.PutUint32(output[:4], d)
	binary.BigEndian.PutUint32(output[4:8], c)
	binary.BigEndian.PutUint32(output[8:12], b)
	binary.BigEndian.PutUint32(output[12:16], a)
}

/*
Decrypt one block, same as encryption except the key are reversed
*/
func decryptBlockGo(key []uint32, output, input []byte) {
	encryptBlockGo(key, output, input)
}

////////////////////////////////////////////////////////////////////
//
// Key schedule
//
////////////////////////////////////////////////////////////////////

/* fixed parameter */
var ck = [32]uint32{
	0x00070e15, 0x1c232a31, 0x383f464d, 0x545b6269,
	0x70777e85, 0x8c939aa1, 0xa8afb6bd, 0xc4cbd2d9,
	0xe0e7eef5, 0xfc030a11, 0x181f262d, 0x343b4249,
	0x50575e65, 0x6c737a81, 0x888f969d, 0xa4abb2b9,
	0xc0c7ced5, 0xdce3eaf1, 0xf8ff060d, 0x141b2229,
	0x30373e45, 0x4c535a61, 0x686f767d, 0x848b9299,
	0xa0a7aeb5, 0xbcc3cad1, 0xd8dfe6ed, 0xf4fb0209,
	0x10171e25, 0x2c333a41, 0x484f565d, 0x646b7279,
}
var fk = [4]uint32{0xa3b1bac6, 0x56aa3350, 0x677d9197, 0xb27022dc}

// Key expansion algorithm. See FIPS-197, Figure 11.
// Their rcon[i] is our powx[i-1] << 24.
func expandKeyGo(key []byte, enc, dec []uint32) {
	a := binary.BigEndian.Uint32(key[:4]) ^ fk[0]
	b := binary.BigEndian.Uint32(key[4:8]) ^ fk[1]
	c := binary.BigEndian.Uint32(key[8:12]) ^ fk[2]
	d := binary.BigEndian.Uint32(key[12:16]) ^ fk[3]
	var t uint32
	for i := 0; i < 32; {
		t = tau(b ^ c ^ d ^ ck[i])
		a ^= t ^ rotl(t, 13) ^ rotl(t, 23)
		enc[i] = a
		dec[31-i] = a
		i++

		t = tau(a ^ c ^ d ^ ck[i])
		b ^= t ^ rotl(t, 13) ^ rotl(t, 23)
		enc[i] = b
		dec[31-i] = b
		i++

		t = tau(a ^ b ^ d ^ ck[i])
		c ^= t ^ rotl(t, 13) ^ rotl(t, 23)
		enc[i] = c
		dec[31-i] = c
		i++

		t = tau(a ^ c ^ b ^ ck[i])
		d ^= t ^ rotl(t, 13) ^ rotl(t, 23)
		enc[i] = d
		dec[31-i] = d
		i++
	}
}