package sm234

import (
	"crypto/cipher"
	"encoding/binary"
	"fmt"
	"math/bits"
)

type sm4ctx struct {
	sk [32]uint32
}

const (
	blockSize = 16
)

var (
	sboxTable = [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,
	}

	fk = [4]uint32{
		0xa3b1bac6, 0x56aa3350, 0x677d9197, 0xb27022dc,
	}

	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,
	}
)

func sm4SBox(idat uint32) uint32 {
	var a, b, c, d byte

	a = sboxTable[idat&0xff]
	b = sboxTable[(idat>>8)&0xff]
	c = sboxTable[(idat>>16)&0xff]
	d = sboxTable[(idat>>24)&0xff]

	odat := uint32(d)
	odat = (odat << 8) | uint32(c)
	odat = (odat << 8) | uint32(b)
	odat = (odat << 8) | uint32(a)

	return odat
}

// key 长度必须是 16 bytes = 128 bits
func NewSM4Block(key []byte) (cipher.Block, error) {

	var nctx sm4ctx
	var kkk [36]uint32
	var temp uint32

	if len(key) < 16 {
		return nil, fmt.Errorf("key size error, only 16")
	}

	// round key.
	kkk[0] = binary.BigEndian.Uint32(key[0:]) ^ fk[0]
	kkk[1] = binary.BigEndian.Uint32(key[4:]) ^ fk[1]
	kkk[2] = binary.BigEndian.Uint32(key[8:]) ^ fk[2]
	kkk[3] = binary.BigEndian.Uint32(key[12:]) ^ fk[3]

	for i := 0; i < 32; i++ {
		temp = kkk[i+1] ^ kkk[i+2] ^ kkk[i+3] ^ ck[i]
		temp = sm4SBox(temp)
		temp = temp ^ bits.RotateLeft32(temp, 13) ^ bits.RotateLeft32(temp, 23)
		kkk[i+4] = kkk[i] ^ temp
		nctx.sk[i] = kkk[i+4]
	}

	return &nctx, nil
}

// BlockSize returns the cipher's block size.
func (ctx *sm4ctx) BlockSize() int {
	return blockSize
}

// Encrypt encrypts the first block in src into dst.
// Dst and src must overlap entirely or not at all.
func (ctx *sm4ctx) Encrypt(dst, src []byte) {

	var temp uint32
	var ulbuf [36]uint32

	if len(src) < blockSize {
		panic("sm4: input not full block")
	}

	if len(dst) < blockSize {
		panic("sm4: output not full block")
	}

	ulbuf[0] = binary.BigEndian.Uint32(src[0:])
	ulbuf[1] = binary.BigEndian.Uint32(src[4:])
	ulbuf[2] = binary.BigEndian.Uint32(src[8:])
	ulbuf[3] = binary.BigEndian.Uint32(src[12:])

	for i := 0; i < 32; i++ {
		temp = ulbuf[i+1] ^ ulbuf[i+2] ^ ulbuf[i+3] ^ ctx.sk[i]
		temp = sm4SBox(temp)
		uvv := temp ^ bits.RotateLeft32(temp, 2) ^ bits.RotateLeft32(temp, 10)
		uvv = uvv ^ bits.RotateLeft32(temp, 18) ^ bits.RotateLeft32(temp, 24)
		ulbuf[i+4] = ulbuf[i] ^ uvv
	}

	binary.BigEndian.PutUint32(dst[0:], ulbuf[35])
	binary.BigEndian.PutUint32(dst[4:], ulbuf[34])
	binary.BigEndian.PutUint32(dst[8:], ulbuf[33])
	binary.BigEndian.PutUint32(dst[12:], ulbuf[32])
}

// Decrypt decrypts the first block in src into dst.
// Dst and src must overlap entirely or not at all.
func (ctx *sm4ctx) Decrypt(dst, src []byte) {

	var temp uint32
	var ulbuf [36]uint32

	if len(src) < blockSize {
		panic("sm4: input not full block")
	}

	if len(dst) < blockSize {
		panic("sm4: output not full block")
	}

	ulbuf[0] = binary.BigEndian.Uint32(src[0:])
	ulbuf[1] = binary.BigEndian.Uint32(src[4:])
	ulbuf[2] = binary.BigEndian.Uint32(src[8:])
	ulbuf[3] = binary.BigEndian.Uint32(src[12:])

	for i := 0; i < 32; i++ {
		temp = ulbuf[i+1] ^ ulbuf[i+2] ^ ulbuf[i+3] ^ ctx.sk[31-i]
		temp = sm4SBox(temp)
		uvv := temp ^ bits.RotateLeft32(temp, 2) ^ bits.RotateLeft32(temp, 10)
		uvv = uvv ^ bits.RotateLeft32(temp, 18) ^ bits.RotateLeft32(temp, 24)
		ulbuf[i+4] = ulbuf[i] ^ uvv
	}

	binary.BigEndian.PutUint32(dst[0:], ulbuf[35])
	binary.BigEndian.PutUint32(dst[4:], ulbuf[34])
	binary.BigEndian.PutUint32(dst[8:], ulbuf[33])
	binary.BigEndian.PutUint32(dst[12:], ulbuf[32])
}