nod/lib/aes.cpp

#include "nod/aes.hpp"
#include <cstdio>
#include <cstring>

#if _WIN32
#include <intrin.h>
#else
#include <cpuid.h>
#endif

#if __AES__ || (!defined(__clang__) && _MSC_VER >= 1800)
#define _AES_NI 1
#endif

namespace nod {

/* rotates x one bit to the left */

#define ROTL(x) (((x) >> 7) | ((x) << 1))

/* Rotates 32-bit word left by 1, 2 or 3 byte  */

#define ROTL8(x) (((x) << 8) | ((x) >> 24))
#define ROTL16(x) (((x) << 16) | ((x) >> 16))
#define ROTL24(x) (((x) << 24) | ((x) >> 8))

static const uint8_t InCo[4] = {0xB, 0xD, 0x9, 0xE}; /* Inverse Coefficients */

static inline uint32_t pack(const uint8_t* b) {
  /* pack bytes into a 32-bit Word */
  return ((uint32_t)b[3] << 24) | ((uint32_t)b[2] << 16) | ((uint32_t)b[1] << 8) | (uint32_t)b[0];
}

static inline void unpack(uint32_t a, uint8_t* b) {
  /* unpack bytes from a word */
  b[0] = (uint8_t)a;
  b[1] = (uint8_t)(a >> 8);
  b[2] = (uint8_t)(a >> 16);
  b[3] = (uint8_t)(a >> 24);
}

static inline uint8_t xtime(uint8_t a) { return ((a << 1) ^ (((a >> 7) & 1) * 0x11B)); }

static const struct SoftwareAESTables {
  uint8_t fbsub[256];
  uint8_t rbsub[256];
  uint8_t ptab[256], ltab[256];
  uint32_t ftable[256];
  uint32_t rtable[256];
  uint32_t rco[30];

  uint8_t bmul(uint8_t x, uint8_t y) const {
    /* x.y= AntiLog(Log(x) + Log(y)) */
    if (x && y)
      return ptab[(ltab[x] + ltab[y]) % 255];
    else
      return 0;
  }

  uint32_t SubByte(uint32_t a) const {
    uint8_t b[4];
    unpack(a, b);
    b[0] = fbsub[b[0]];
    b[1] = fbsub[b[1]];
    b[2] = fbsub[b[2]];
    b[3] = fbsub[b[3]];
    return pack(b);
  }

  uint8_t product(uint32_t x, uint32_t y) const {
    /* dot product of two 4-byte arrays */
    uint8_t xb[4], yb[4];
    unpack(x, xb);
    unpack(y, yb);
    return bmul(xb[0], yb[0]) ^ bmul(xb[1], yb[1]) ^ bmul(xb[2], yb[2]) ^ bmul(xb[3], yb[3]);
  }

  uint32_t InvMixCol(uint32_t x) const {
    /* matrix Multiplication */
    uint32_t y, m;
    uint8_t b[4];

    m = pack(InCo);
    b[3] = product(m, x);
    m = ROTL24(m);
    b[2] = product(m, x);
    m = ROTL24(m);
    b[1] = product(m, x);
    m = ROTL24(m);
    b[0] = product(m, x);
    y = pack(b);
    return y;
  }

  uint8_t ByteSub(uint8_t x) const {
    uint8_t y = ptab[255 - ltab[x]]; /* multiplicative inverse */
    x = y;
    x = ROTL(x);
    y ^= x;
    x = ROTL(x);
    y ^= x;
    x = ROTL(x);
    y ^= x;
    x = ROTL(x);
    y ^= x;
    y ^= 0x63;
    return y;
  }

  SoftwareAESTables() {
    /* generate tables */
    int i;
    uint8_t y, b[4];

    /* use 3 as primitive root to generate power and log tables */

    ltab[0] = 0;
    ptab[0] = 1;
    ltab[1] = 0;
    ptab[1] = 3;
    ltab[3] = 1;

    for (i = 2; i < 256; i++) {
      ptab[i] = ptab[i - 1] ^ xtime(ptab[i - 1]);
      ltab[ptab[i]] = i;
    }

    /* affine transformation:- each bit is xored with itself shifted one bit */

    fbsub[0] = 0x63;
    rbsub[0x63] = 0;

    for (i = 1; i < 256; i++) {
      y = ByteSub((uint8_t)i);
      fbsub[i] = y;
      rbsub[y] = i;
    }

    for (i = 0, y = 1; i < 30; i++) {
      rco[i] = y;
      y = xtime(y);
    }

    /* calculate forward and reverse tables */
    for (i = 0; i < 256; i++) {
      y = fbsub[i];
      b[3] = y ^ xtime(y);
      b[2] = y;
      b[1] = y;
      b[0] = xtime(y);
      ftable[i] = pack(b);

      y = rbsub[i];
      b[3] = bmul(InCo[0], y);
      b[2] = bmul(InCo[1], y);
      b[1] = bmul(InCo[2], y);
      b[0] = bmul(InCo[3], y);
      rtable[i] = pack(b);
    }
  }
} AEStb;

class SoftwareAES : public IAES {
protected:
  /* Parameter-dependent data */
  int Nk, Nb, Nr;
  uint8_t fi[24], ri[24];
  uint32_t fkey[120];
  uint32_t rkey[120];

  void gkey(int nb, int nk, const uint8_t* key);
  void _encrypt(uint8_t* buff);
  void _decrypt(uint8_t* buff);

public:
  void encrypt(const uint8_t* iv, const uint8_t* inbuf, uint8_t* outbuf, size_t len);
  void decrypt(const uint8_t* iv, const uint8_t* inbuf, uint8_t* outbuf, size_t len);
  void setKey(const uint8_t* key);
};

void SoftwareAES::gkey(int nb, int nk, const uint8_t* key) {
  /* blocksize=32*nb bits. Key=32*nk bits */
  /* currently nb,bk = 4, 6 or 8          */
  /* key comes as 4*Nk bytes              */
  /* Key Scheduler. Create expanded encryption key */
  int i, j, k, m, N;
  int C1, C2, C3;
  uint32_t CipherKey[8];

  Nb = nb;
  Nk = nk;

  /* Nr is number of rounds */
  if (Nb >= Nk)
    Nr = 6 + Nb;
  else
    Nr = 6 + Nk;

  C1 = 1;

  if (Nb < 8) {
    C2 = 2;
    C3 = 3;
  } else {
    C2 = 3;
    C3 = 4;
  }

  /* pre-calculate forward and reverse increments */
  for (m = j = 0; j < nb; j++, m += 3) {
    fi[m] = (j + C1) % nb;
    fi[m + 1] = (j + C2) % nb;
    fi[m + 2] = (j + C3) % nb;
    ri[m] = (nb + j - C1) % nb;
    ri[m + 1] = (nb + j - C2) % nb;
    ri[m + 2] = (nb + j - C3) % nb;
  }

  N = Nb * (Nr + 1);

  for (i = j = 0; i < Nk; i++, j += 4) {
    CipherKey[i] = pack(key + j);
  }

  for (i = 0; i < Nk; i++)
    fkey[i] = CipherKey[i];

  for (j = Nk, k = 0; j < N; j += Nk, k++) {
    fkey[j] = fkey[j - Nk] ^ AEStb.SubByte(ROTL24(fkey[j - 1])) ^ AEStb.rco[k];

    if (Nk <= 6) {
      for (i = 1; i < Nk && (i + j) < N; i++)
        fkey[i + j] = fkey[i + j - Nk] ^ fkey[i + j - 1];
    } else {
      for (i = 1; i < 4 && (i + j) < N; i++)
        fkey[i + j] = fkey[i + j - Nk] ^ fkey[i + j - 1];

      if ((j + 4) < N)
        fkey[j + 4] = fkey[j + 4 - Nk] ^ AEStb.SubByte(fkey[j + 3]);

      for (i = 5; i < Nk && (i + j) < N; i++)
        fkey[i + j] = fkey[i + j - Nk] ^ fkey[i + j - 1];
    }
  }

  /* now for the expanded decrypt key in reverse order */

  for (j = 0; j < Nb; j++)
    rkey[j + N - Nb] = fkey[j];

  for (i = Nb; i < N - Nb; i += Nb) {
    k = N - Nb - i;

    for (j = 0; j < Nb; j++)
      rkey[k + j] = AEStb.InvMixCol(fkey[i + j]);
  }

  for (j = N - Nb; j < N; j++)
    rkey[j - N + Nb] = fkey[j];
}

/* There is an obvious time/space trade-off possible here.     *
 * Instead of just one ftable[], I could have 4, the other     *
 * 3 pre-rotated to save the ROTL8, ROTL16 and ROTL24 overhead */

void SoftwareAES::_encrypt(uint8_t* buff) {
  int i, j, k, m;
  uint32_t a[8], b[8], *x, *y, *t;

  for (i = j = 0; i < Nb; i++, j += 4) {
    a[i] = pack(buff + j);
    a[i] ^= fkey[i];
  }

  k = Nb;
  x = a;
  y = b;

  /* State alternates between a and b */
  for (i = 1; i < Nr; i++) {
    /* Nr is number of rounds. May be odd. */

    /* if Nb is fixed - unroll this next
    loop and hard-code in the values of fi[]  */

    for (m = j = 0; j < Nb; j++, m += 3) {
      /* deal with each 32-bit element of the State */
      /* This is the time-critical bit */
      y[j] = fkey[k++] ^ AEStb.ftable[(uint8_t)x[j]] ^ ROTL8(AEStb.ftable[(uint8_t)(x[fi[m]] >> 8)]) ^
             ROTL16(AEStb.ftable[(uint8_t)(x[fi[m + 1]] >> 16)]) ^ ROTL24(AEStb.ftable[(uint8_t)(x[fi[m + 2]] >> 24)]);
    }

    t = x;
    x = y;
    y = t; /* swap pointers */
  }

  /* Last Round - unroll if possible */
  for (m = j = 0; j < Nb; j++, m += 3) {
    y[j] = fkey[k++] ^ (uint32_t)AEStb.fbsub[(uint8_t)x[j]] ^ ROTL8((uint32_t)AEStb.fbsub[(uint8_t)(x[fi[m]] >> 8)]) ^
           ROTL16((uint32_t)AEStb.fbsub[(uint8_t)(x[fi[m + 1]] >> 16)]) ^
           ROTL24((uint32_t)AEStb.fbsub[(uint8_t)(x[fi[m + 2]] >> 24)]);
  }

  for (i = j = 0; i < Nb; i++, j += 4) {
    unpack(y[i], (uint8_t*)&buff[j]);
    x[i] = y[i] = 0; /* clean up stack */
  }

  return;
}

void SoftwareAES::_decrypt(uint8_t* buff) {
  int i, j, k, m;
  uint32_t a[8], b[8], *x, *y, *t;

  for (i = j = 0; i < Nb; i++, j += 4) {
    a[i] = pack(buff + j);
    a[i] ^= rkey[i];
  }

  k = Nb;
  x = a;
  y = b;

  /* State alternates between a and b */
  for (i = 1; i < Nr; i++) {
    /* Nr is number of rounds. May be odd. */

    /* if Nb is fixed - unroll this next
    loop and hard-code in the values of ri[]  */

    for (m = j = 0; j < Nb; j++, m += 3) {
      /* This is the time-critical bit */
      y[j] = rkey[k++] ^ AEStb.rtable[(uint8_t)x[j]] ^ ROTL8(AEStb.rtable[(uint8_t)(x[ri[m]] >> 8)]) ^
             ROTL16(AEStb.rtable[(uint8_t)(x[ri[m + 1]] >> 16)]) ^ ROTL24(AEStb.rtable[(uint8_t)(x[ri[m + 2]] >> 24)]);
    }

    t = x;
    x = y;
    y = t; /* swap pointers */
  }

  /* Last Round - unroll if possible */
  for (m = j = 0; j < Nb; j++, m += 3) {
    y[j] = rkey[k++] ^ (uint32_t)AEStb.rbsub[(uint8_t)x[j]] ^ ROTL8((uint32_t)AEStb.rbsub[(uint8_t)(x[ri[m]] >> 8)]) ^
           ROTL16((uint32_t)AEStb.rbsub[(uint8_t)(x[ri[m + 1]] >> 16)]) ^
           ROTL24((uint32_t)AEStb.rbsub[(uint8_t)(x[ri[m + 2]] >> 24)]);
  }

  for (i = j = 0; i < Nb; i++, j += 4) {
    unpack(y[i], (uint8_t*)&buff[j]);
    x[i] = y[i] = 0; /* clean up stack */
  }

  return;
}

void SoftwareAES::setKey(const uint8_t* key) { gkey(4, 4, key); }

// CBC mode decryption
void SoftwareAES::decrypt(const uint8_t* iv, const uint8_t* inbuf, uint8_t* outbuf, size_t len) {
  uint8_t block[16];
  const uint8_t* ctext_ptr;
  unsigned int blockno = 0, i;

  // fprintf( stderr,"aes_decrypt(%p, %p, %p, %lld)\n", iv, inbuf, outbuf, len  );
  // printf("aes_decrypt(%p, %p, %p, %lld)\n", iv, inbuf, outbuf, len);

  for (blockno = 0; blockno <= (len / sizeof(block)); blockno++) {
    unsigned int fraction;

    if (blockno == (len / sizeof(block))) // last block
    {
      fraction = len % sizeof(block);

      if (fraction == 0)
        break;

      memset(block, 0, sizeof(block));
    } else
      fraction = 16;

    //    debug_printf("block %d: fraction = %d\n", blockno, fraction);
    memcpy(block, inbuf + blockno * sizeof(block), fraction);
    _decrypt(block);

    if (blockno == 0)
      ctext_ptr = iv;
    else
      ctext_ptr = (uint8_t*)(inbuf + (blockno - 1) * sizeof(block));

    for (i = 0; i < fraction; i++)
      outbuf[blockno * sizeof(block) + i] = ctext_ptr[i] ^ block[i];

    //    debug_printf("Block %d output: ", blockno);
    //    hexdump(outbuf + blockno*sizeof(block), 16);
  }
}

// CBC mode encryption
void SoftwareAES::encrypt(const uint8_t* iv, const uint8_t* inbuf, uint8_t* outbuf, size_t len) {
  uint8_t block[16];
  uint8_t feedback[16];
  memcpy(feedback, iv, 16);
  unsigned int blockno = 0, i;

  // printf("aes_decrypt(%p, %p, %p, %lld)\n", iv, inbuf, outbuf, len);
  // fprintf( stderr,"aes_encrypt(%p, %p, %p, %lld)\n", iv, inbuf, outbuf, len);

  for (blockno = 0; blockno <= (len / sizeof(block)); blockno++) {
    unsigned int fraction;

    if (blockno == (len / sizeof(block))) // last block
    {
      fraction = len % sizeof(block);

      if (fraction == 0)
        break;

      memset(block, 0, sizeof(block));
    } else
      fraction = 16;

    //    debug_printf("block %d: fraction = %d\n", blockno, fraction);
    memcpy(block, inbuf + blockno * sizeof(block), fraction);

    for (i = 0; i < fraction; i++)
      block[i] = inbuf[blockno * sizeof(block) + i] ^ feedback[i];

    _encrypt(block);
    memcpy(feedback, block, sizeof(block));
    memcpy(outbuf + blockno * sizeof(block), block, sizeof(block));
    //    debug_printf("Block %d output: ", blockno);
    //    hexdump(outbuf + blockno*sizeof(block), 16);
  }
}

#if _AES_NI

#include <wmmintrin.h>

class NiAES : public IAES {
  __m128i m_ekey[11];
  __m128i m_dkey[11];

public:
  void encrypt(const uint8_t* iv, const uint8_t* inbuf, uint8_t* outbuf, size_t len) {
    __m128i feedback, data;
    uint64_t i, j;
    if (len % 16)
      len = len / 16 + 1;
    else
      len /= 16;
    feedback = _mm_loadu_si128((__m128i*)iv);
    for (i = 0; i < len; i++) {
      data = _mm_loadu_si128(&((__m128i*)inbuf)[i]);
      feedback = _mm_xor_si128(data, feedback);
      feedback = _mm_xor_si128(feedback, m_ekey[0]);
      for (j = 1; j < 10; j++)
        feedback = _mm_aesenc_si128(feedback, m_ekey[j]);
      feedback = _mm_aesenclast_si128(feedback, m_ekey[j]);
      _mm_storeu_si128(&((__m128i*)outbuf)[i], feedback);
    }
  }
  void decrypt(const uint8_t* iv, const uint8_t* inbuf, uint8_t* outbuf, size_t len) {
    __m128i data, feedback, last_in;
    uint64_t i, j;
    if (len % 16)
      len = len / 16 + 1;
    else
      len /= 16;
    feedback = _mm_loadu_si128((__m128i*)iv);
    for (i = 0; i < len; i++) {
      last_in = _mm_loadu_si128(&((__m128i*)inbuf)[i]);
      data = _mm_xor_si128(last_in, m_dkey[0]);
      for (j = 1; j < 10; j++)
        data = _mm_aesdec_si128(data, m_dkey[j]);
      data = _mm_aesdeclast_si128(data, m_dkey[j]);
      data = _mm_xor_si128(data, feedback);
      _mm_storeu_si128(&((__m128i*)outbuf)[i], data);
      feedback = last_in;
    }
  }

  static inline __m128i AES_128_ASSIST(__m128i temp1, __m128i temp2) {
    __m128i temp3;
    temp2 = _mm_shuffle_epi32(temp2, 0xff);
    temp3 = _mm_slli_si128(temp1, 0x4);
    temp1 = _mm_xor_si128(temp1, temp3);
    temp3 = _mm_slli_si128(temp3, 0x4);
    temp1 = _mm_xor_si128(temp1, temp3);
    temp3 = _mm_slli_si128(temp3, 0x4);
    temp1 = _mm_xor_si128(temp1, temp3);
    temp1 = _mm_xor_si128(temp1, temp2);
    return temp1;
  }

  void setKey(const uint8_t* key) {
    __m128i temp1, temp2;

    temp1 = _mm_loadu_si128((__m128i*)key);
    m_ekey[0] = temp1;
    m_dkey[10] = temp1;
    temp2 = _mm_aeskeygenassist_si128(temp1, 0x1);
    temp1 = AES_128_ASSIST(temp1, temp2);
    m_ekey[1] = temp1;
    m_dkey[9] = _mm_aesimc_si128(temp1);
    temp2 = _mm_aeskeygenassist_si128(temp1, 0x2);
    temp1 = AES_128_ASSIST(temp1, temp2);
    m_ekey[2] = temp1;
    m_dkey[8] = _mm_aesimc_si128(temp1);
    temp2 = _mm_aeskeygenassist_si128(temp1, 0x4);
    temp1 = AES_128_ASSIST(temp1, temp2);
    m_ekey[3] = temp1;
    m_dkey[7] = _mm_aesimc_si128(temp1);
    temp2 = _mm_aeskeygenassist_si128(temp1, 0x8);
    temp1 = AES_128_ASSIST(temp1, temp2);
    m_ekey[4] = temp1;
    m_dkey[6] = _mm_aesimc_si128(temp1);
    temp2 = _mm_aeskeygenassist_si128(temp1, 0x10);
    temp1 = AES_128_ASSIST(temp1, temp2);
    m_ekey[5] = temp1;
    m_dkey[5] = _mm_aesimc_si128(temp1);
    temp2 = _mm_aeskeygenassist_si128(temp1, 0x20);
    temp1 = AES_128_ASSIST(temp1, temp2);
    m_ekey[6] = temp1;
    m_dkey[4] = _mm_aesimc_si128(temp1);
    temp2 = _mm_aeskeygenassist_si128(temp1, 0x40);
    temp1 = AES_128_ASSIST(temp1, temp2);
    m_ekey[7] = temp1;
    m_dkey[3] = _mm_aesimc_si128(temp1);
    temp2 = _mm_aeskeygenassist_si128(temp1, 0x80);
    temp1 = AES_128_ASSIST(temp1, temp2);
    m_ekey[8] = temp1;
    m_dkey[2] = _mm_aesimc_si128(temp1);
    temp2 = _mm_aeskeygenassist_si128(temp1, 0x1b);
    temp1 = AES_128_ASSIST(temp1, temp2);
    m_ekey[9] = temp1;
    m_dkey[1] = _mm_aesimc_si128(temp1);
    temp2 = _mm_aeskeygenassist_si128(temp1, 0x36);
    temp1 = AES_128_ASSIST(temp1, temp2);
    m_ekey[10] = temp1;
    m_dkey[0] = temp1;
  }
};

static int HAS_AES_NI = -1;

#endif

std::unique_ptr<IAES> NewAES() {
#if _AES_NI
  if (HAS_AES_NI == -1) {
#if _MSC_VER
    int info[4];
    __cpuid(info, 1);
    HAS_AES_NI = ((info[2] & 0x2000000) != 0);
#else
    unsigned int a, b, c, d;
    __cpuid(1, a, b, c, d);
    HAS_AES_NI = ((c & 0x2000000) != 0);
#endif
  }
  if (HAS_AES_NI)
    return std::unique_ptr<IAES>(new NiAES);
  else
    return std::unique_ptr<IAES>(new SoftwareAES);
#else
  return std::unique_ptr<IAES>(new SoftwareAES);
#endif
}

} // namespace nod