#include "aes.hpp"

#include <cstring>
#include <iostream>

#include "profiler.hpp"

const uint8_t AES::irreducible_polynomial = 0x1b;

const uint8_t AES::sbox[256] = {
	0x63, 0x7c, 0x77, 0x7b, 0xf2, 0x6b, 0x6f, 0xc5, 0x30, 0x01, 0x67, 0x2b, 0xfe, 0xd7, 0xab, 0x76,
	0xca, 0x82, 0xc9, 0x7d, 0xfa, 0x59, 0x47, 0xf0, 0xad, 0xd4, 0xa2, 0xaf, 0x9c, 0xa4, 0x72, 0xc0,
	0xb7, 0xfd, 0x93, 0x26, 0x36, 0x3f, 0xf7, 0xcc, 0x34, 0xa5, 0xe5, 0xf1, 0x71, 0xd8, 0x31, 0x15,
	0x04, 0xc7, 0x23, 0xc3, 0x18, 0x96, 0x05, 0x9a, 0x07, 0x12, 0x80, 0xe2, 0xeb, 0x27, 0xb2, 0x75,
	0x09, 0x83, 0x2c, 0x1a, 0x1b, 0x6e, 0x5a, 0xa0, 0x52, 0x3b, 0xd6, 0xb3, 0x29, 0xe3, 0x2f, 0x84,
	0x53, 0xd1, 0x00, 0xed, 0x20, 0xfc, 0xb1, 0x5b, 0x6a, 0xcb, 0xbe, 0x39, 0x4a, 0x4c, 0x58, 0xcf,
	0xd0, 0xef, 0xaa, 0xfb, 0x43, 0x4d, 0x33, 0x85, 0x45, 0xf9, 0x02, 0x7f, 0x50, 0x3c, 0x9f, 0xa8,
	0x51, 0xa3, 0x40, 0x8f, 0x92, 0x9d, 0x38, 0xf5, 0xbc, 0xb6, 0xda, 0x21, 0x10, 0xff, 0xf3, 0xd2,
	0xcd, 0x0c, 0x13, 0xec, 0x5f, 0x97, 0x44, 0x17, 0xc4, 0xa7, 0x7e, 0x3d, 0x64, 0x5d, 0x19, 0x73,
	0x60, 0x81, 0x4f, 0xdc, 0x22, 0x2a, 0x90, 0x88, 0x46, 0xee, 0xb8, 0x14, 0xde, 0x5e, 0x0b, 0xdb,
	0xe0, 0x32, 0x3a, 0x0a, 0x49, 0x06, 0x24, 0x5c, 0xc2, 0xd3, 0xac, 0x62, 0x91, 0x95, 0xe4, 0x79,
	0xe7, 0xc8, 0x37, 0x6d, 0x8d, 0xd5, 0x4e, 0xa9, 0x6c, 0x56, 0xf4, 0xea, 0x65, 0x7a, 0xae, 0x08,
	0xba, 0x78, 0x25, 0x2e, 0x1c, 0xa6, 0xb4, 0xc6, 0xe8, 0xdd, 0x74, 0x1f, 0x4b, 0xbd, 0x8b, 0x8a,
	0x70, 0x3e, 0xb5, 0x66, 0x48, 0x03, 0xf6, 0x0e, 0x61, 0x35, 0x57, 0xb9, 0x86, 0xc1, 0x1d, 0x9e,
	0xe1, 0xf8, 0x98, 0x11, 0x69, 0xd9, 0x8e, 0x94, 0x9b, 0x1e, 0x87, 0xe9, 0xce, 0x55, 0x28, 0xdf,
	0x8c, 0xa1, 0x89, 0x0d, 0xbf, 0xe6, 0x42, 0x68, 0x41, 0x99, 0x2d, 0x0f, 0xb0, 0x54, 0xbb, 0x16 };

const uint8_t AES::r_sbox[256] = {
	0x52, 0x09, 0x6a, 0xd5, 0x30, 0x36, 0xa5, 0x38, 0xbf, 0x40, 0xa3, 0x9e, 0x81, 0xf3, 0xd7, 0xfb,
	0x7c, 0xe3, 0x39, 0x82, 0x9b, 0x2f, 0xff, 0x87, 0x34, 0x8e, 0x43, 0x44, 0xc4, 0xde, 0xe9, 0xcb,
	0x54, 0x7b, 0x94, 0x32, 0xa6, 0xc2, 0x23, 0x3d, 0xee, 0x4c, 0x95, 0x0b, 0x42, 0xfa, 0xc3, 0x4e,
	0x08, 0x2e, 0xa1, 0x66, 0x28, 0xd9, 0x24, 0xb2, 0x76, 0x5b, 0xa2, 0x49, 0x6d, 0x8b, 0xd1, 0x25,
	0x72, 0xf8, 0xf6, 0x64, 0x86, 0x68, 0x98, 0x16, 0xd4, 0xa4, 0x5c, 0xcc, 0x5d, 0x65, 0xb6, 0x92,
	0x6c, 0x70, 0x48, 0x50, 0xfd, 0xed, 0xb9, 0xda, 0x5e, 0x15, 0x46, 0x57, 0xa7, 0x8d, 0x9d, 0x84,
	0x90, 0xd8, 0xab, 0x00, 0x8c, 0xbc, 0xd3, 0x0a, 0xf7, 0xe4, 0x58, 0x05, 0xb8, 0xb3, 0x45, 0x06,
	0xd0, 0x2c, 0x1e, 0x8f, 0xca, 0x3f, 0x0f, 0x02, 0xc1, 0xaf, 0xbd, 0x03, 0x01, 0x13, 0x8a, 0x6b,
	0x3a, 0x91, 0x11, 0x41, 0x4f, 0x67, 0xdc, 0xea, 0x97, 0xf2, 0xcf, 0xce, 0xf0, 0xb4, 0xe6, 0x73,
	0x96, 0xac, 0x74, 0x22, 0xe7, 0xad, 0x35, 0x85, 0xe2, 0xf9, 0x37, 0xe8, 0x1c, 0x75, 0xdf, 0x6e,
	0x47, 0xf1, 0x1a, 0x71, 0x1d, 0x29, 0xc5, 0x89, 0x6f, 0xb7, 0x62, 0x0e, 0xaa, 0x18, 0xbe, 0x1b,
	0xfc, 0x56, 0x3e, 0x4b, 0xc6, 0xd2, 0x79, 0x20, 0x9a, 0xdb, 0xc0, 0xfe, 0x78, 0xcd, 0x5a, 0xf4,
	0x1f, 0xdd, 0xa8, 0x33, 0x88, 0x07, 0xc7, 0x31, 0xb1, 0x12, 0x10, 0x59, 0x27, 0x80, 0xec, 0x5f,
	0x60, 0x51, 0x7f, 0xa9, 0x19, 0xb5, 0x4a, 0x0d, 0x2d, 0xe5, 0x7a, 0x9f, 0x93, 0xc9, 0x9c, 0xef,
	0xa0, 0xe0, 0x3b, 0x4d, 0xae, 0x2a, 0xf5, 0xb0, 0xc8, 0xeb, 0xbb, 0x3c, 0x83, 0x53, 0x99, 0x61,
	0x17, 0x2b, 0x04, 0x7e, 0xba, 0x77, 0xd6, 0x26, 0xe1, 0x69, 0x14, 0x63, 0x55, 0x21, 0x0c, 0x7d };

const uint8_t AES::rc[10] = {
	0b00000001,
	0b00000010,
	0b00000100,
	0b00001000,
	0b00010000,
	0b00100000,
	0b01000000,
	0b10000000,
	0x1b,
	0x1b << 1
};

void AES::print_data(uint8_t data[16])
{
	PROFILER_RECORD;
	for (int i = 0;i < 16;i++)
	{
		std::cout << std::hex << (int) data[i] << " ";
	}
	std::cout << std::endl;
}

void AES::encrypt_ecb(uint8_t data[16], const uint8_t key[16])
{
	PROFILER_RECORD;
	uint8_t round_keys[11][16];
	construct_round_keys(key, round_keys);
	add_round_key(data, round_keys[0]);
	for (int r = 1;r < 10;r++)
	{
		sub_bytes(data);
		shift_rows(data);
		mix_columns(data);
		add_round_key(data, round_keys[r]);
	}
	sub_bytes(data);
	shift_rows(data);
	add_round_key(data, round_keys[10]);
}

void AES::decrypt_ecb(uint8_t data[16], const uint8_t key[16])
{
	PROFILER_RECORD;
	uint8_t round_keys[11][16];
	construct_round_keys(key, round_keys);
	add_round_key(data, round_keys[10]);
	r_shift_rows(data);
	r_sub_bytes(data);
	for (int r = 9;r > 0;r--)
	{
		add_round_key(data, round_keys[r]);
		r_mix_columns(data);
		r_shift_rows(data);
		r_sub_bytes(data);
	}
	add_round_key(data, round_keys[0]);
}

void AES::add_round_key(uint8_t data_[16], const uint8_t key_[16])
{
	PROFILER_RECORD;
	const uint32_t (&key)[4] = *reinterpret_cast<const uint32_t (*)[4]>(key_);
	uint32_t (&data)[4] = *reinterpret_cast<uint32_t (*)[4]>(data_);
	for (int i = 0;i < 4;i++)
	{
		data[i] ^= key[i];
	}
}

void AES::sub_bytes(uint8_t data[16])
{
	PROFILER_RECORD;
	for (int i = 0;i < 16;i++)
	{
		data[i] = sbox[data[i]];
	}
}

void AES::r_sub_bytes(uint8_t data[16])
{
	PROFILER_RECORD;
	for (int i = 0;i < 16;i++)
	{
		data[i] = r_sbox[data[i]];
	}
}

#define SWAP(x, y, tmp) tmp = x; x = y; y = tmp

void AES::shift_rows(uint8_t data[16])
{
	PROFILER_RECORD;
	uint8_t tmp = data[1];
	data[1] = data[5];
	data[5] = data[9];
	data[9] = data[13];
	data[13] = tmp;

	SWAP(data[2], data[10], tmp);
	SWAP(data[6], data[14], tmp);

	tmp = data[15];
	data[15] = data[11];
	data[11] = data[7];
	data[7] = data[3];
	data[3] = tmp;
}

void AES::r_shift_rows(uint8_t data[16])
{
	PROFILER_RECORD;
	uint8_t tmp = data[13];
	data[13] = data[9];
	data[9] = data[5];
	data[5] = data[1];
	data[1] = tmp;

	SWAP(data[2], data[10], tmp);
	SWAP(data[6], data[14], tmp);

	tmp = data[3];
	data[3] = data[7];
	data[7] = data[11];
	data[11] = data[15];
	data[15] = tmp;
}

#undef SWAP

void AES::mix_columns(uint8_t data[16])
{
	PROFILER_RECORD;
	uint8_t data_new[16];
	for (int i = 0;i < 4;i++)
	{
		data_new[i * 4 + 0] = gf_mult2(data[i * 4 + 0]) ^ gf_mult3(data[i * 4 + 1]) ^ data[i * 4 + 2] ^ data[i * 4 + 3];
		data_new[i * 4 + 1] = data[i * 4 + 0] ^ gf_mult2(data[i * 4 + 1]) ^ gf_mult3(data[i * 4 + 2]) ^ data[i * 4 + 3];
		data_new[i * 4 + 2] = data[i * 4 + 0] ^ data[i * 4 + 1] ^ gf_mult2(data[i * 4 + 2]) ^ gf_mult3(data[i * 4 + 3]);
		data_new[i * 4 + 3] = gf_mult3(data[i * 4 + 0]) ^ data[i * 4 + 1] ^ data[i * 4 + 2] ^ gf_mult2(data[i * 4 + 3]);
	}
	memcpy(data, data_new, 16);
}

void AES::r_mix_columns(uint8_t data[16])
{
	PROFILER_RECORD;
	uint8_t data_new[16];
	for (int i = 0;i < 4;i++)
	{
		data_new[i * 4 + 0] = gf_multE(data[i * 4 + 0]) ^ gf_multB(data[i * 4 + 1]) ^ gf_multD(data[i * 4 + 2]) ^ gf_mult9(data[i * 4 + 3]);
		data_new[i * 4 + 1] = gf_mult9(data[i * 4 + 0]) ^ gf_multE(data[i * 4 + 1]) ^ gf_multB(data[i * 4 + 2]) ^ gf_multD(data[i * 4 + 3]);
		data_new[i * 4 + 2] = gf_multD(data[i * 4 + 0]) ^ gf_mult9(data[i * 4 + 1]) ^ gf_multE(data[i * 4 + 2]) ^ gf_multB(data[i * 4 + 3]);
		data_new[i * 4 + 3] = gf_multB(data[i * 4 + 0]) ^ gf_multD(data[i * 4 + 1]) ^ gf_mult9(data[i * 4 + 2]) ^ gf_multE(data[i * 4 + 3]);
	}
	mempcpy(data, data_new, 16);
}

void AES::construct_round_keys(const uint8_t initial_key_[16], uint8_t round_keys_[11][16])
{
	PROFILER_RECORD;
	const uint32_t (&initial_key)[4] = *reinterpret_cast<const uint32_t (*)[4]>(initial_key_);
	uint32_t (&round_keys)[11][4] = *reinterpret_cast<uint32_t (*)[11][4]>(round_keys_);
	for (int k = 0;k < 4;k++)
	{
		round_keys[0][k] = initial_key[k];
	}
	for (int r = 1;r < 11;r++)
	{
		round_keys[r][0] = round_keys[r - 1][0] ^ roundkey_g(round_keys[r - 1][3], r - 1);
		round_keys[r][1] = round_keys[r][0] ^ round_keys[r - 1][1];
		round_keys[r][2] = round_keys[r][1] ^ round_keys[r - 1][2];
		round_keys[r][3] = round_keys[r][2] ^ round_keys[r - 1][3];
	}
}

uint32_t AES::roundkey_g(uint32_t last_roundkey, uint8_t round)
{
	PROFILER_RECORD;
	uint32_t result;
	uint8_t *last = (uint8_t*) &last_roundkey;
	uint8_t *next = (uint8_t*) &result;
	next[3] = sbox[last[0]];
	next[0] = sbox[last[1]] ^ rc[round]; // TODO Check round
	next[1] = sbox[last[2]];
	next[2] = sbox[last[3]];
	return result;
}

uint8_t AES::gf_reduce(uint8_t value)
{
	PROFILER_RECORD;
	return value ^ irreducible_polynomial;
}

uint8_t AES::gf_mult2(uint8_t value)
{
	PROFILER_RECORD;
	uint16_t result = ((uint16_t) value) << 1;
	if ((result & 0x100) == 0)
	{
		return (uint8_t) result;
	}
	else
	{
		return gf_reduce(result);
	}
}

uint8_t AES::gf_mult3(uint8_t value)
{
	PROFILER_RECORD;
	return gf_mult2(value) ^ value;
}

uint8_t AES::gf_mult4(uint8_t value)
{
	PROFILER_RECORD;
	return gf_mult2(gf_mult2(value));
}

uint8_t AES::gf_mult8(uint8_t value)
{
	PROFILER_RECORD;
	return gf_mult2(gf_mult4(value));
}

uint8_t AES::gf_mult9(uint8_t value)
{
	PROFILER_RECORD;
	return gf_mult8(value) ^ value;
}

uint8_t AES::gf_multB(uint8_t value)
{
	PROFILER_RECORD;
	return gf_mult8(value) ^ gf_mult3(value);
}

uint8_t AES::gf_multD(uint8_t value)
{
	PROFILER_RECORD;
	uint8_t gf4 = gf_mult4(value);
	return gf_mult2(gf4) ^ gf4 ^ value;
}

uint8_t AES::gf_multE(uint8_t value)
{
	PROFILER_RECORD;
	uint8_t gf2 = gf_mult2(value);
	uint8_t gf4 = gf_mult2(gf2);
	return gf_mult2(gf4) ^ gf4 ^ gf2;
}