/**
 * Author......: See docs/credits.txt
 * License.....: MIT
 */

#define NEW_SIMD_CODE

#include "inc_vendor.cl"
#include "inc_hash_constants.h"
#include "inc_hash_functions.cl"
#include "inc_types.cl"
#include "inc_common.cl"
#include "inc_simd.cl"
#include "inc_cipher_aes.cl"

__constant u32a k_sha256[64] =
{
  SHA256C00, SHA256C01, SHA256C02, SHA256C03,
  SHA256C04, SHA256C05, SHA256C06, SHA256C07,
  SHA256C08, SHA256C09, SHA256C0a, SHA256C0b,
  SHA256C0c, SHA256C0d, SHA256C0e, SHA256C0f,
  SHA256C10, SHA256C11, SHA256C12, SHA256C13,
  SHA256C14, SHA256C15, SHA256C16, SHA256C17,
  SHA256C18, SHA256C19, SHA256C1a, SHA256C1b,
  SHA256C1c, SHA256C1d, SHA256C1e, SHA256C1f,
  SHA256C20, SHA256C21, SHA256C22, SHA256C23,
  SHA256C24, SHA256C25, SHA256C26, SHA256C27,
  SHA256C28, SHA256C29, SHA256C2a, SHA256C2b,
  SHA256C2c, SHA256C2d, SHA256C2e, SHA256C2f,
  SHA256C30, SHA256C31, SHA256C32, SHA256C33,
  SHA256C34, SHA256C35, SHA256C36, SHA256C37,
  SHA256C38, SHA256C39, SHA256C3a, SHA256C3b,
  SHA256C3c, SHA256C3d, SHA256C3e, SHA256C3f,
};

void sha256_transform_S (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[8])
{
  u32 a = digest[0];
  u32 b = digest[1];
  u32 c = digest[2];
  u32 d = digest[3];
  u32 e = digest[4];
  u32 f = digest[5];
  u32 g = digest[6];
  u32 h = digest[7];

  u32 w0_t = w0[0];
  u32 w1_t = w0[1];
  u32 w2_t = w0[2];
  u32 w3_t = w0[3];
  u32 w4_t = w1[0];
  u32 w5_t = w1[1];
  u32 w6_t = w1[2];
  u32 w7_t = w1[3];
  u32 w8_t = w2[0];
  u32 w9_t = w2[1];
  u32 wa_t = w2[2];
  u32 wb_t = w2[3];
  u32 wc_t = w3[0];
  u32 wd_t = w3[1];
  u32 we_t = w3[2];
  u32 wf_t = w3[3];

  #define ROUND_EXPAND_S()                            \
  {                                                   \
    w0_t = SHA256_EXPAND_S (we_t, w9_t, w1_t, w0_t);  \
    w1_t = SHA256_EXPAND_S (wf_t, wa_t, w2_t, w1_t);  \
    w2_t = SHA256_EXPAND_S (w0_t, wb_t, w3_t, w2_t);  \
    w3_t = SHA256_EXPAND_S (w1_t, wc_t, w4_t, w3_t);  \
    w4_t = SHA256_EXPAND_S (w2_t, wd_t, w5_t, w4_t);  \
    w5_t = SHA256_EXPAND_S (w3_t, we_t, w6_t, w5_t);  \
    w6_t = SHA256_EXPAND_S (w4_t, wf_t, w7_t, w6_t);  \
    w7_t = SHA256_EXPAND_S (w5_t, w0_t, w8_t, w7_t);  \
    w8_t = SHA256_EXPAND_S (w6_t, w1_t, w9_t, w8_t);  \
    w9_t = SHA256_EXPAND_S (w7_t, w2_t, wa_t, w9_t);  \
    wa_t = SHA256_EXPAND_S (w8_t, w3_t, wb_t, wa_t);  \
    wb_t = SHA256_EXPAND_S (w9_t, w4_t, wc_t, wb_t);  \
    wc_t = SHA256_EXPAND_S (wa_t, w5_t, wd_t, wc_t);  \
    wd_t = SHA256_EXPAND_S (wb_t, w6_t, we_t, wd_t);  \
    we_t = SHA256_EXPAND_S (wc_t, w7_t, wf_t, we_t);  \
    wf_t = SHA256_EXPAND_S (wd_t, w8_t, w0_t, wf_t);  \
  }

  #define ROUND_STEP_S(i)                                                                   \
  {                                                                                         \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha256[i +  0]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha256[i +  1]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha256[i +  2]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha256[i +  3]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha256[i +  4]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha256[i +  5]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha256[i +  6]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha256[i +  7]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha256[i +  8]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha256[i +  9]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha256[i + 10]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha256[i + 11]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha256[i + 12]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha256[i + 13]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, k_sha256[i + 14]); \
    SHA256_STEP_S (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha256[i + 15]); \
  }

  ROUND_STEP_S (0);

  #ifdef _unroll
  #pragma unroll
  #endif
  for (int i = 16; i < 64; i += 16)
  {
    ROUND_EXPAND_S (); ROUND_STEP_S (i);
  }

  digest[0] += a;
  digest[1] += b;
  digest[2] += c;
  digest[3] += d;
  digest[4] += e;
  digest[5] += f;
  digest[6] += g;
  digest[7] += h;
}

void hmac_sha256_pad_S (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[8], u32 opad[8])
{
  w0[0] = w0[0] ^ 0x36363636;
  w0[1] = w0[1] ^ 0x36363636;
  w0[2] = w0[2] ^ 0x36363636;
  w0[3] = w0[3] ^ 0x36363636;
  w1[0] = w1[0] ^ 0x36363636;
  w1[1] = w1[1] ^ 0x36363636;
  w1[2] = w1[2] ^ 0x36363636;
  w1[3] = w1[3] ^ 0x36363636;
  w2[0] = w2[0] ^ 0x36363636;
  w2[1] = w2[1] ^ 0x36363636;
  w2[2] = w2[2] ^ 0x36363636;
  w2[3] = w2[3] ^ 0x36363636;
  w3[0] = w3[0] ^ 0x36363636;
  w3[1] = w3[1] ^ 0x36363636;
  w3[2] = w3[2] ^ 0x36363636;
  w3[3] = w3[3] ^ 0x36363636;

  ipad[0] = SHA256M_A;
  ipad[1] = SHA256M_B;
  ipad[2] = SHA256M_C;
  ipad[3] = SHA256M_D;
  ipad[4] = SHA256M_E;
  ipad[5] = SHA256M_F;
  ipad[6] = SHA256M_G;
  ipad[7] = SHA256M_H;

  sha256_transform_S (w0, w1, w2, w3, ipad);

  w0[0] = w0[0] ^ 0x6a6a6a6a;
  w0[1] = w0[1] ^ 0x6a6a6a6a;
  w0[2] = w0[2] ^ 0x6a6a6a6a;
  w0[3] = w0[3] ^ 0x6a6a6a6a;
  w1[0] = w1[0] ^ 0x6a6a6a6a;
  w1[1] = w1[1] ^ 0x6a6a6a6a;
  w1[2] = w1[2] ^ 0x6a6a6a6a;
  w1[3] = w1[3] ^ 0x6a6a6a6a;
  w2[0] = w2[0] ^ 0x6a6a6a6a;
  w2[1] = w2[1] ^ 0x6a6a6a6a;
  w2[2] = w2[2] ^ 0x6a6a6a6a;
  w2[3] = w2[3] ^ 0x6a6a6a6a;
  w3[0] = w3[0] ^ 0x6a6a6a6a;
  w3[1] = w3[1] ^ 0x6a6a6a6a;
  w3[2] = w3[2] ^ 0x6a6a6a6a;
  w3[3] = w3[3] ^ 0x6a6a6a6a;

  opad[0] = SHA256M_A;
  opad[1] = SHA256M_B;
  opad[2] = SHA256M_C;
  opad[3] = SHA256M_D;
  opad[4] = SHA256M_E;
  opad[5] = SHA256M_F;
  opad[6] = SHA256M_G;
  opad[7] = SHA256M_H;

  sha256_transform_S (w0, w1, w2, w3, opad);
}

void hmac_sha256_run_S (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[8], u32 opad[8], u32 digest[8])
{
  digest[0] = ipad[0];
  digest[1] = ipad[1];
  digest[2] = ipad[2];
  digest[3] = ipad[3];
  digest[4] = ipad[4];
  digest[5] = ipad[5];
  digest[6] = ipad[6];
  digest[7] = ipad[7];

  sha256_transform_S (w0, w1, w2, w3, digest);

  w0[0] = digest[0];
  w0[1] = digest[1];
  w0[2] = digest[2];
  w0[3] = digest[3];
  w1[0] = digest[4];
  w1[1] = digest[5];
  w1[2] = digest[6];
  w1[3] = digest[7];
  w2[0] = 0x80000000;
  w2[1] = 0;
  w2[2] = 0;
  w2[3] = 0;
  w3[0] = 0;
  w3[1] = 0;
  w3[2] = 0;
  w3[3] = (64 + 32) * 8;

  digest[0] = opad[0];
  digest[1] = opad[1];
  digest[2] = opad[2];
  digest[3] = opad[3];
  digest[4] = opad[4];
  digest[5] = opad[5];
  digest[6] = opad[6];
  digest[7] = opad[7];

  sha256_transform_S (w0, w1, w2, w3, digest);
}

void sha256_transform_V (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[8])
{
  u32x a = digest[0];
  u32x b = digest[1];
  u32x c = digest[2];
  u32x d = digest[3];
  u32x e = digest[4];
  u32x f = digest[5];
  u32x g = digest[6];
  u32x h = digest[7];

  u32x w0_t = w0[0];
  u32x w1_t = w0[1];
  u32x w2_t = w0[2];
  u32x w3_t = w0[3];
  u32x w4_t = w1[0];
  u32x w5_t = w1[1];
  u32x w6_t = w1[2];
  u32x w7_t = w1[3];
  u32x w8_t = w2[0];
  u32x w9_t = w2[1];
  u32x wa_t = w2[2];
  u32x wb_t = w2[3];
  u32x wc_t = w3[0];
  u32x wd_t = w3[1];
  u32x we_t = w3[2];
  u32x wf_t = w3[3];

  #define ROUND_EXPAND()                            \
  {                                                 \
    w0_t = SHA256_EXPAND (we_t, w9_t, w1_t, w0_t);  \
    w1_t = SHA256_EXPAND (wf_t, wa_t, w2_t, w1_t);  \
    w2_t = SHA256_EXPAND (w0_t, wb_t, w3_t, w2_t);  \
    w3_t = SHA256_EXPAND (w1_t, wc_t, w4_t, w3_t);  \
    w4_t = SHA256_EXPAND (w2_t, wd_t, w5_t, w4_t);  \
    w5_t = SHA256_EXPAND (w3_t, we_t, w6_t, w5_t);  \
    w6_t = SHA256_EXPAND (w4_t, wf_t, w7_t, w6_t);  \
    w7_t = SHA256_EXPAND (w5_t, w0_t, w8_t, w7_t);  \
    w8_t = SHA256_EXPAND (w6_t, w1_t, w9_t, w8_t);  \
    w9_t = SHA256_EXPAND (w7_t, w2_t, wa_t, w9_t);  \
    wa_t = SHA256_EXPAND (w8_t, w3_t, wb_t, wa_t);  \
    wb_t = SHA256_EXPAND (w9_t, w4_t, wc_t, wb_t);  \
    wc_t = SHA256_EXPAND (wa_t, w5_t, wd_t, wc_t);  \
    wd_t = SHA256_EXPAND (wb_t, w6_t, we_t, wd_t);  \
    we_t = SHA256_EXPAND (wc_t, w7_t, wf_t, we_t);  \
    wf_t = SHA256_EXPAND (wd_t, w8_t, w0_t, wf_t);  \
  }

  #define ROUND_STEP(i)                                                                   \
  {                                                                                       \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha256[i +  0]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha256[i +  1]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha256[i +  2]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha256[i +  3]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha256[i +  4]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha256[i +  5]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha256[i +  6]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha256[i +  7]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha256[i +  8]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha256[i +  9]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha256[i + 10]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha256[i + 11]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha256[i + 12]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha256[i + 13]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, k_sha256[i + 14]); \
    SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha256[i + 15]); \
  }

  ROUND_STEP (0);

  #ifdef _unroll
  #pragma unroll
  #endif
  for (int i = 16; i < 64; i += 16)
  {
    ROUND_EXPAND (); ROUND_STEP (i);
  }

  digest[0] += a;
  digest[1] += b;
  digest[2] += c;
  digest[3] += d;
  digest[4] += e;
  digest[5] += f;
  digest[6] += g;
  digest[7] += h;
}

void hmac_sha256_run_V (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[8], u32x opad[8], u32x digest[8])
{
  digest[0] = ipad[0];
  digest[1] = ipad[1];
  digest[2] = ipad[2];
  digest[3] = ipad[3];
  digest[4] = ipad[4];
  digest[5] = ipad[5];
  digest[6] = ipad[6];
  digest[7] = ipad[7];

  sha256_transform_V (w0, w1, w2, w3, digest);

  w0[0] = digest[0];
  w0[1] = digest[1];
  w0[2] = digest[2];
  w0[3] = digest[3];
  w1[0] = digest[4];
  w1[1] = digest[5];
  w1[2] = digest[6];
  w1[3] = digest[7];
  w2[0] = 0x80000000;
  w2[1] = 0;
  w2[2] = 0;
  w2[3] = 0;
  w3[0] = 0;
  w3[1] = 0;
  w3[2] = 0;
  w3[3] = (64 + 32) * 8;

  digest[0] = opad[0];
  digest[1] = opad[1];
  digest[2] = opad[2];
  digest[3] = opad[3];
  digest[4] = opad[4];
  digest[5] = opad[5];
  digest[6] = opad[6];
  digest[7] = opad[7];

  sha256_transform_V (w0, w1, w2, w3, digest);
}

void sha1_transform_S (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[5])
{
  u32 A = digest[0];
  u32 B = digest[1];
  u32 C = digest[2];
  u32 D = digest[3];
  u32 E = digest[4];

  u32 w0_t = w0[0];
  u32 w1_t = w0[1];
  u32 w2_t = w0[2];
  u32 w3_t = w0[3];
  u32 w4_t = w1[0];
  u32 w5_t = w1[1];
  u32 w6_t = w1[2];
  u32 w7_t = w1[3];
  u32 w8_t = w2[0];
  u32 w9_t = w2[1];
  u32 wa_t = w2[2];
  u32 wb_t = w2[3];
  u32 wc_t = w3[0];
  u32 wd_t = w3[1];
  u32 we_t = w3[2];
  u32 wf_t = w3[3];

  #undef K
  #define K SHA1C00

  SHA1_STEP_S (SHA1_F0o, A, B, C, D, E, w0_t);
  SHA1_STEP_S (SHA1_F0o, E, A, B, C, D, w1_t);
  SHA1_STEP_S (SHA1_F0o, D, E, A, B, C, w2_t);
  SHA1_STEP_S (SHA1_F0o, C, D, E, A, B, w3_t);
  SHA1_STEP_S (SHA1_F0o, B, C, D, E, A, w4_t);
  SHA1_STEP_S (SHA1_F0o, A, B, C, D, E, w5_t);
  SHA1_STEP_S (SHA1_F0o, E, A, B, C, D, w6_t);
  SHA1_STEP_S (SHA1_F0o, D, E, A, B, C, w7_t);
  SHA1_STEP_S (SHA1_F0o, C, D, E, A, B, w8_t);
  SHA1_STEP_S (SHA1_F0o, B, C, D, E, A, w9_t);
  SHA1_STEP_S (SHA1_F0o, A, B, C, D, E, wa_t);
  SHA1_STEP_S (SHA1_F0o, E, A, B, C, D, wb_t);
  SHA1_STEP_S (SHA1_F0o, D, E, A, B, C, wc_t);
  SHA1_STEP_S (SHA1_F0o, C, D, E, A, B, wd_t);
  SHA1_STEP_S (SHA1_F0o, B, C, D, E, A, we_t);
  SHA1_STEP_S (SHA1_F0o, A, B, C, D, E, wf_t);
  w0_t = rotl32_S ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP_S (SHA1_F0o, E, A, B, C, D, w0_t);
  w1_t = rotl32_S ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP_S (SHA1_F0o, D, E, A, B, C, w1_t);
  w2_t = rotl32_S ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP_S (SHA1_F0o, C, D, E, A, B, w2_t);
  w3_t = rotl32_S ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP_S (SHA1_F0o, B, C, D, E, A, w3_t);

  #undef K
  #define K SHA1C01

  w4_t = rotl32_S ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, w4_t);
  w5_t = rotl32_S ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, w5_t);
  w6_t = rotl32_S ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, w6_t);
  w7_t = rotl32_S ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, w7_t);
  w8_t = rotl32_S ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, w8_t);
  w9_t = rotl32_S ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, w9_t);
  wa_t = rotl32_S ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, wa_t);
  wb_t = rotl32_S ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, wb_t);
  wc_t = rotl32_S ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, wc_t);
  wd_t = rotl32_S ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, wd_t);
  we_t = rotl32_S ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, we_t);
  wf_t = rotl32_S ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, wf_t);
  w0_t = rotl32_S ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, w0_t);
  w1_t = rotl32_S ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, w1_t);
  w2_t = rotl32_S ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, w2_t);
  w3_t = rotl32_S ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, w3_t);
  w4_t = rotl32_S ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, w4_t);
  w5_t = rotl32_S ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, w5_t);
  w6_t = rotl32_S ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, w6_t);
  w7_t = rotl32_S ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, w7_t);

  #undef K
  #define K SHA1C02

  w8_t = rotl32_S ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP_S (SHA1_F2o, A, B, C, D, E, w8_t);
  w9_t = rotl32_S ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP_S (SHA1_F2o, E, A, B, C, D, w9_t);
  wa_t = rotl32_S ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP_S (SHA1_F2o, D, E, A, B, C, wa_t);
  wb_t = rotl32_S ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP_S (SHA1_F2o, C, D, E, A, B, wb_t);
  wc_t = rotl32_S ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP_S (SHA1_F2o, B, C, D, E, A, wc_t);
  wd_t = rotl32_S ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP_S (SHA1_F2o, A, B, C, D, E, wd_t);
  we_t = rotl32_S ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP_S (SHA1_F2o, E, A, B, C, D, we_t);
  wf_t = rotl32_S ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP_S (SHA1_F2o, D, E, A, B, C, wf_t);
  w0_t = rotl32_S ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP_S (SHA1_F2o, C, D, E, A, B, w0_t);
  w1_t = rotl32_S ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP_S (SHA1_F2o, B, C, D, E, A, w1_t);
  w2_t = rotl32_S ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP_S (SHA1_F2o, A, B, C, D, E, w2_t);
  w3_t = rotl32_S ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP_S (SHA1_F2o, E, A, B, C, D, w3_t);
  w4_t = rotl32_S ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP_S (SHA1_F2o, D, E, A, B, C, w4_t);
  w5_t = rotl32_S ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP_S (SHA1_F2o, C, D, E, A, B, w5_t);
  w6_t = rotl32_S ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP_S (SHA1_F2o, B, C, D, E, A, w6_t);
  w7_t = rotl32_S ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP_S (SHA1_F2o, A, B, C, D, E, w7_t);
  w8_t = rotl32_S ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP_S (SHA1_F2o, E, A, B, C, D, w8_t);
  w9_t = rotl32_S ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP_S (SHA1_F2o, D, E, A, B, C, w9_t);
  wa_t = rotl32_S ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP_S (SHA1_F2o, C, D, E, A, B, wa_t);
  wb_t = rotl32_S ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP_S (SHA1_F2o, B, C, D, E, A, wb_t);

  #undef K
  #define K SHA1C03

  wc_t = rotl32_S ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, wc_t);
  wd_t = rotl32_S ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, wd_t);
  we_t = rotl32_S ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, we_t);
  wf_t = rotl32_S ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, wf_t);
  w0_t = rotl32_S ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, w0_t);
  w1_t = rotl32_S ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, w1_t);
  w2_t = rotl32_S ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, w2_t);
  w3_t = rotl32_S ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, w3_t);
  w4_t = rotl32_S ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, w4_t);
  w5_t = rotl32_S ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, w5_t);
  w6_t = rotl32_S ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, w6_t);
  w7_t = rotl32_S ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, w7_t);
  w8_t = rotl32_S ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, w8_t);
  w9_t = rotl32_S ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, w9_t);
  wa_t = rotl32_S ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, wa_t);
  wb_t = rotl32_S ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP_S (SHA1_F1, A, B, C, D, E, wb_t);
  wc_t = rotl32_S ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP_S (SHA1_F1, E, A, B, C, D, wc_t);
  wd_t = rotl32_S ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP_S (SHA1_F1, D, E, A, B, C, wd_t);
  we_t = rotl32_S ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP_S (SHA1_F1, C, D, E, A, B, we_t);
  wf_t = rotl32_S ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP_S (SHA1_F1, B, C, D, E, A, wf_t);

  digest[0] += A;
  digest[1] += B;
  digest[2] += C;
  digest[3] += D;
  digest[4] += E;
}

void hmac_sha1_pad_S (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 opad[5])
{
  w0[0] = w0[0] ^ 0x36363636;
  w0[1] = w0[1] ^ 0x36363636;
  w0[2] = w0[2] ^ 0x36363636;
  w0[3] = w0[3] ^ 0x36363636;
  w1[0] = w1[0] ^ 0x36363636;
  w1[1] = w1[1] ^ 0x36363636;
  w1[2] = w1[2] ^ 0x36363636;
  w1[3] = w1[3] ^ 0x36363636;
  w2[0] = w2[0] ^ 0x36363636;
  w2[1] = w2[1] ^ 0x36363636;
  w2[2] = w2[2] ^ 0x36363636;
  w2[3] = w2[3] ^ 0x36363636;
  w3[0] = w3[0] ^ 0x36363636;
  w3[1] = w3[1] ^ 0x36363636;
  w3[2] = w3[2] ^ 0x36363636;
  w3[3] = w3[3] ^ 0x36363636;

  ipad[0] = SHA1M_A;
  ipad[1] = SHA1M_B;
  ipad[2] = SHA1M_C;
  ipad[3] = SHA1M_D;
  ipad[4] = SHA1M_E;

  sha1_transform_S (w0, w1, w2, w3, ipad);

  w0[0] = w0[0] ^ 0x6a6a6a6a;
  w0[1] = w0[1] ^ 0x6a6a6a6a;
  w0[2] = w0[2] ^ 0x6a6a6a6a;
  w0[3] = w0[3] ^ 0x6a6a6a6a;
  w1[0] = w1[0] ^ 0x6a6a6a6a;
  w1[1] = w1[1] ^ 0x6a6a6a6a;
  w1[2] = w1[2] ^ 0x6a6a6a6a;
  w1[3] = w1[3] ^ 0x6a6a6a6a;
  w2[0] = w2[0] ^ 0x6a6a6a6a;
  w2[1] = w2[1] ^ 0x6a6a6a6a;
  w2[2] = w2[2] ^ 0x6a6a6a6a;
  w2[3] = w2[3] ^ 0x6a6a6a6a;
  w3[0] = w3[0] ^ 0x6a6a6a6a;
  w3[1] = w3[1] ^ 0x6a6a6a6a;
  w3[2] = w3[2] ^ 0x6a6a6a6a;
  w3[3] = w3[3] ^ 0x6a6a6a6a;

  opad[0] = SHA1M_A;
  opad[1] = SHA1M_B;
  opad[2] = SHA1M_C;
  opad[3] = SHA1M_D;
  opad[4] = SHA1M_E;

  sha1_transform_S (w0, w1, w2, w3, opad);
}

void hmac_sha1_run_S (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 opad[5], u32 digest[5])
{
  digest[0] = ipad[0];
  digest[1] = ipad[1];
  digest[2] = ipad[2];
  digest[3] = ipad[3];
  digest[4] = ipad[4];

  sha1_transform_S (w0, w1, w2, w3, digest);

  w0[0] = digest[0];
  w0[1] = digest[1];
  w0[2] = digest[2];
  w0[3] = digest[3];
  w1[0] = digest[4];
  w1[1] = 0x80000000;
  w1[2] = 0;
  w1[3] = 0;
  w2[0] = 0;
  w2[1] = 0;
  w2[2] = 0;
  w2[3] = 0;
  w3[0] = 0;
  w3[1] = 0;
  w3[2] = 0;
  w3[3] = (64 + 20) * 8;

  digest[0] = opad[0];
  digest[1] = opad[1];
  digest[2] = opad[2];
  digest[3] = opad[3];
  digest[4] = opad[4];

  sha1_transform_S (w0, w1, w2, w3, digest);
}

void sha1_transform_V (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[5])
{
  u32x A = digest[0];
  u32x B = digest[1];
  u32x C = digest[2];
  u32x D = digest[3];
  u32x E = digest[4];

  u32x w0_t = w0[0];
  u32x w1_t = w0[1];
  u32x w2_t = w0[2];
  u32x w3_t = w0[3];
  u32x w4_t = w1[0];
  u32x w5_t = w1[1];
  u32x w6_t = w1[2];
  u32x w7_t = w1[3];
  u32x w8_t = w2[0];
  u32x w9_t = w2[1];
  u32x wa_t = w2[2];
  u32x wb_t = w2[3];
  u32x wc_t = w3[0];
  u32x wd_t = w3[1];
  u32x we_t = w3[2];
  u32x wf_t = w3[3];

  #undef K
  #define K SHA1C00

  SHA1_STEP (SHA1_F0o, A, B, C, D, E, w0_t);
  SHA1_STEP (SHA1_F0o, E, A, B, C, D, w1_t);
  SHA1_STEP (SHA1_F0o, D, E, A, B, C, w2_t);
  SHA1_STEP (SHA1_F0o, C, D, E, A, B, w3_t);
  SHA1_STEP (SHA1_F0o, B, C, D, E, A, w4_t);
  SHA1_STEP (SHA1_F0o, A, B, C, D, E, w5_t);
  SHA1_STEP (SHA1_F0o, E, A, B, C, D, w6_t);
  SHA1_STEP (SHA1_F0o, D, E, A, B, C, w7_t);
  SHA1_STEP (SHA1_F0o, C, D, E, A, B, w8_t);
  SHA1_STEP (SHA1_F0o, B, C, D, E, A, w9_t);
  SHA1_STEP (SHA1_F0o, A, B, C, D, E, wa_t);
  SHA1_STEP (SHA1_F0o, E, A, B, C, D, wb_t);
  SHA1_STEP (SHA1_F0o, D, E, A, B, C, wc_t);
  SHA1_STEP (SHA1_F0o, C, D, E, A, B, wd_t);
  SHA1_STEP (SHA1_F0o, B, C, D, E, A, we_t);
  SHA1_STEP (SHA1_F0o, A, B, C, D, E, wf_t);
  w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F0o, E, A, B, C, D, w0_t);
  w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F0o, D, E, A, B, C, w1_t);
  w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F0o, C, D, E, A, B, w2_t);
  w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F0o, B, C, D, E, A, w3_t);

  #undef K
  #define K SHA1C01

  w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w4_t);
  w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w5_t);
  w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w6_t);
  w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w7_t);
  w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w8_t);
  w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w9_t);
  wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wa_t);
  wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wb_t);
  wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wc_t);
  wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wd_t);
  we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, we_t);
  wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wf_t);
  w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w0_t);
  w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w1_t);
  w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w2_t);
  w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w3_t);
  w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w4_t);
  w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w5_t);
  w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w6_t);
  w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w7_t);

  #undef K
  #define K SHA1C02

  w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w8_t);
  w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w9_t);
  wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wa_t);
  wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wb_t);
  wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wc_t);
  wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, wd_t);
  we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, we_t);
  wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wf_t);
  w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w0_t);
  w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w1_t);
  w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w2_t);
  w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w3_t);
  w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w4_t);
  w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w5_t);
  w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w6_t);
  w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w7_t);
  w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w8_t);
  w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w9_t);
  wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wa_t);
  wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wb_t);

  #undef K
  #define K SHA1C03

  wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wc_t);
  wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wd_t);
  we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, we_t);
  wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wf_t);
  w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w0_t);
  w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w1_t);
  w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w2_t);
  w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w3_t);
  w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w4_t);
  w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w5_t);
  w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w6_t);
  w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w7_t);
  w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w8_t);
  w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w9_t);
  wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wa_t);
  wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wb_t);
  wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wc_t);
  wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wd_t);
  we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, we_t);
  wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wf_t);

  digest[0] += A;
  digest[1] += B;
  digest[2] += C;
  digest[3] += D;
  digest[4] += E;
}

void hmac_sha1_pad_V (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[5], u32x opad[5])
{
  w0[0] = w0[0] ^ 0x36363636;
  w0[1] = w0[1] ^ 0x36363636;
  w0[2] = w0[2] ^ 0x36363636;
  w0[3] = w0[3] ^ 0x36363636;
  w1[0] = w1[0] ^ 0x36363636;
  w1[1] = w1[1] ^ 0x36363636;
  w1[2] = w1[2] ^ 0x36363636;
  w1[3] = w1[3] ^ 0x36363636;
  w2[0] = w2[0] ^ 0x36363636;
  w2[1] = w2[1] ^ 0x36363636;
  w2[2] = w2[2] ^ 0x36363636;
  w2[3] = w2[3] ^ 0x36363636;
  w3[0] = w3[0] ^ 0x36363636;
  w3[1] = w3[1] ^ 0x36363636;
  w3[2] = w3[2] ^ 0x36363636;
  w3[3] = w3[3] ^ 0x36363636;

  ipad[0] = SHA1M_A;
  ipad[1] = SHA1M_B;
  ipad[2] = SHA1M_C;
  ipad[3] = SHA1M_D;
  ipad[4] = SHA1M_E;

  sha1_transform_V (w0, w1, w2, w3, ipad);

  w0[0] = w0[0] ^ 0x6a6a6a6a;
  w0[1] = w0[1] ^ 0x6a6a6a6a;
  w0[2] = w0[2] ^ 0x6a6a6a6a;
  w0[3] = w0[3] ^ 0x6a6a6a6a;
  w1[0] = w1[0] ^ 0x6a6a6a6a;
  w1[1] = w1[1] ^ 0x6a6a6a6a;
  w1[2] = w1[2] ^ 0x6a6a6a6a;
  w1[3] = w1[3] ^ 0x6a6a6a6a;
  w2[0] = w2[0] ^ 0x6a6a6a6a;
  w2[1] = w2[1] ^ 0x6a6a6a6a;
  w2[2] = w2[2] ^ 0x6a6a6a6a;
  w2[3] = w2[3] ^ 0x6a6a6a6a;
  w3[0] = w3[0] ^ 0x6a6a6a6a;
  w3[1] = w3[1] ^ 0x6a6a6a6a;
  w3[2] = w3[2] ^ 0x6a6a6a6a;
  w3[3] = w3[3] ^ 0x6a6a6a6a;

  opad[0] = SHA1M_A;
  opad[1] = SHA1M_B;
  opad[2] = SHA1M_C;
  opad[3] = SHA1M_D;
  opad[4] = SHA1M_E;

  sha1_transform_V (w0, w1, w2, w3, opad);
}

void hmac_sha1_run_V (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[5], u32x opad[5], u32x digest[5])
{
  digest[0] = ipad[0];
  digest[1] = ipad[1];
  digest[2] = ipad[2];
  digest[3] = ipad[3];
  digest[4] = ipad[4];

  sha1_transform_V (w0, w1, w2, w3, digest);

  w0[0] = digest[0];
  w0[1] = digest[1];
  w0[2] = digest[2];
  w0[3] = digest[3];
  w1[0] = digest[4];
  w1[1] = 0x80000000;
  w1[2] = 0;
  w1[3] = 0;
  w2[0] = 0;
  w2[1] = 0;
  w2[2] = 0;
  w2[3] = 0;
  w3[0] = 0;
  w3[1] = 0;
  w3[2] = 0;
  w3[3] = (64 + 20) * 8;

  digest[0] = opad[0];
  digest[1] = opad[1];
  digest[2] = opad[2];
  digest[3] = opad[3];
  digest[4] = opad[4];

  sha1_transform_V (w0, w1, w2, w3, digest);
}

__kernel void m14800_init (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global pbkdf2_sha256_tmp_t *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global itunes_backup_t *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
  /**
   * base
   */

  const u32 gid = get_global_id (0);

  if (gid >= gid_max) return;

  u32 w0[4];

  w0[0] = swap32_S (pws[gid].i[ 0]);
  w0[1] = swap32_S (pws[gid].i[ 1]);
  w0[2] = swap32_S (pws[gid].i[ 2]);
  w0[3] = swap32_S (pws[gid].i[ 3]);

  u32 w1[4];

  w1[0] = swap32_S (pws[gid].i[ 4]);
  w1[1] = swap32_S (pws[gid].i[ 5]);
  w1[2] = swap32_S (pws[gid].i[ 6]);
  w1[3] = swap32_S (pws[gid].i[ 7]);

  u32 w2[4];

  w2[0] = swap32_S (pws[gid].i[ 8]);
  w2[1] = swap32_S (pws[gid].i[ 9]);
  w2[2] = swap32_S (pws[gid].i[10]);
  w2[3] = swap32_S (pws[gid].i[11]);

  u32 w3[4];

  w3[0] = swap32_S (pws[gid].i[12]);
  w3[1] = swap32_S (pws[gid].i[13]);
  w3[2] = swap32_S (pws[gid].i[14]);
  w3[3] = swap32_S (pws[gid].i[15]);

  /**
   * dpsl (salt for PBKDF2-HMAC-SHA256)
   */

  u32 dpsl0[4];
  u32 dpsl1[4];

  dpsl0[0] = esalt_bufs[digests_offset].dpsl[0];
  dpsl0[1] = esalt_bufs[digests_offset].dpsl[1];
  dpsl0[2] = esalt_bufs[digests_offset].dpsl[2];
  dpsl0[3] = esalt_bufs[digests_offset].dpsl[3];
  dpsl1[0] = esalt_bufs[digests_offset].dpsl[4];
  dpsl1[1] = 0;
  dpsl1[2] = 0;
  dpsl1[3] = 0;

  /**
   * pads
   */

  u32 ipad[8];
  u32 opad[8];

  hmac_sha256_pad_S (w0, w1, w2, w3, ipad, opad);

  tmps[gid].ipad[0] = ipad[0];
  tmps[gid].ipad[1] = ipad[1];
  tmps[gid].ipad[2] = ipad[2];
  tmps[gid].ipad[3] = ipad[3];
  tmps[gid].ipad[4] = ipad[4];
  tmps[gid].ipad[5] = ipad[5];
  tmps[gid].ipad[6] = ipad[6];
  tmps[gid].ipad[7] = ipad[7];

  tmps[gid].opad[0] = opad[0];
  tmps[gid].opad[1] = opad[1];
  tmps[gid].opad[2] = opad[2];
  tmps[gid].opad[3] = opad[3];
  tmps[gid].opad[4] = opad[4];
  tmps[gid].opad[5] = opad[5];
  tmps[gid].opad[6] = opad[6];
  tmps[gid].opad[7] = opad[7];

  w0[0] = dpsl0[0];
  w0[1] = dpsl0[1];
  w0[2] = dpsl0[2];
  w0[3] = dpsl0[3];
  w1[0] = dpsl1[0];
  w1[1] = 1;
  w1[2] = 0x80000000;
  w1[3] = 0;
  w2[0] = 0;
  w2[1] = 0;
  w2[2] = 0;
  w2[3] = 0;
  w3[0] = 0;
  w3[1] = 0;
  w3[2] = 0;
  w3[3] = (64 + 20 + 4) * 8;

  u32 dgst[8];

  hmac_sha256_run_S (w0, w1, w2, w3, ipad, opad, dgst);

  tmps[gid].dgst[0] = dgst[0];
  tmps[gid].dgst[1] = dgst[1];
  tmps[gid].dgst[2] = dgst[2];
  tmps[gid].dgst[3] = dgst[3];
  tmps[gid].dgst[4] = dgst[4];
  tmps[gid].dgst[5] = dgst[5];
  tmps[gid].dgst[6] = dgst[6];
  tmps[gid].dgst[7] = dgst[7];

  tmps[gid].out[0] = dgst[0];
  tmps[gid].out[1] = dgst[1];
  tmps[gid].out[2] = dgst[2];
  tmps[gid].out[3] = dgst[3];
  tmps[gid].out[4] = dgst[4];
  tmps[gid].out[5] = dgst[5];
  tmps[gid].out[6] = dgst[6];
  tmps[gid].out[7] = dgst[7];
}

__kernel void m14800_loop (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global pbkdf2_sha256_tmp_t *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global itunes_backup_t *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
  const u32 gid = get_global_id (0);

  if ((gid * VECT_SIZE) >= gid_max) return;

  u32x ipad[8];
  u32x opad[8];

  ipad[0] = packv (tmps, ipad, gid, 0);
  ipad[1] = packv (tmps, ipad, gid, 1);
  ipad[2] = packv (tmps, ipad, gid, 2);
  ipad[3] = packv (tmps, ipad, gid, 3);
  ipad[4] = packv (tmps, ipad, gid, 4);
  ipad[5] = packv (tmps, ipad, gid, 5);
  ipad[6] = packv (tmps, ipad, gid, 6);
  ipad[7] = packv (tmps, ipad, gid, 7);

  opad[0] = packv (tmps, opad, gid, 0);
  opad[1] = packv (tmps, opad, gid, 1);
  opad[2] = packv (tmps, opad, gid, 2);
  opad[3] = packv (tmps, opad, gid, 3);
  opad[4] = packv (tmps, opad, gid, 4);
  opad[5] = packv (tmps, opad, gid, 5);
  opad[6] = packv (tmps, opad, gid, 6);
  opad[7] = packv (tmps, opad, gid, 7);

  u32x dgst[8];
  u32x out[8];

  dgst[0] = packv (tmps, dgst, gid, 0);
  dgst[1] = packv (tmps, dgst, gid, 1);
  dgst[2] = packv (tmps, dgst, gid, 2);
  dgst[3] = packv (tmps, dgst, gid, 3);
  dgst[4] = packv (tmps, dgst, gid, 4);
  dgst[5] = packv (tmps, dgst, gid, 5);
  dgst[6] = packv (tmps, dgst, gid, 6);
  dgst[7] = packv (tmps, dgst, gid, 7);

  out[0] = packv (tmps, out, gid, 0);
  out[1] = packv (tmps, out, gid, 1);
  out[2] = packv (tmps, out, gid, 2);
  out[3] = packv (tmps, out, gid, 3);
  out[4] = packv (tmps, out, gid, 4);
  out[5] = packv (tmps, out, gid, 5);
  out[6] = packv (tmps, out, gid, 6);
  out[7] = packv (tmps, out, gid, 7);

  for (u32 j = 0; j < loop_cnt; j++)
  {
    u32x w0[4];
    u32x w1[4];
    u32x w2[4];
    u32x w3[4];

    w0[0] = dgst[0];
    w0[1] = dgst[1];
    w0[2] = dgst[2];
    w0[3] = dgst[3];
    w1[0] = dgst[4];
    w1[1] = dgst[5];
    w1[2] = dgst[6];
    w1[3] = dgst[7];
    w2[0] = 0x80000000;
    w2[1] = 0;
    w2[2] = 0;
    w2[3] = 0;
    w3[0] = 0;
    w3[1] = 0;
    w3[2] = 0;
    w3[3] = (64 + 32) * 8;

    hmac_sha256_run_V (w0, w1, w2, w3, ipad, opad, dgst);

    out[0] ^= dgst[0];
    out[1] ^= dgst[1];
    out[2] ^= dgst[2];
    out[3] ^= dgst[3];
    out[4] ^= dgst[4];
    out[5] ^= dgst[5];
    out[6] ^= dgst[6];
    out[7] ^= dgst[7];
  }

  unpackv (tmps, dgst, gid, 0, dgst[0]);
  unpackv (tmps, dgst, gid, 1, dgst[1]);
  unpackv (tmps, dgst, gid, 2, dgst[2]);
  unpackv (tmps, dgst, gid, 3, dgst[3]);
  unpackv (tmps, dgst, gid, 4, dgst[4]);
  unpackv (tmps, dgst, gid, 5, dgst[5]);
  unpackv (tmps, dgst, gid, 6, dgst[6]);
  unpackv (tmps, dgst, gid, 7, dgst[7]);

  unpackv (tmps, out, gid, 0, out[0]);
  unpackv (tmps, out, gid, 1, out[1]);
  unpackv (tmps, out, gid, 2, out[2]);
  unpackv (tmps, out, gid, 3, out[3]);
  unpackv (tmps, out, gid, 4, out[4]);
  unpackv (tmps, out, gid, 5, out[5]);
  unpackv (tmps, out, gid, 6, out[6]);
  unpackv (tmps, out, gid, 7, out[7]);
}

__kernel void m14800_init2 (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global pbkdf2_sha256_tmp_t *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global itunes_backup_t *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
  /**
   * base
   */

  const u32 gid = get_global_id (0);

  if (gid >= gid_max) return;

  u32 w0[4];
  u32 w1[4];
  u32 w2[4];
  u32 w3[4];

  w0[0] = tmps[gid].out[0];
  w0[1] = tmps[gid].out[1];
  w0[2] = tmps[gid].out[2];
  w0[3] = tmps[gid].out[3];
  w1[0] = tmps[gid].out[4];
  w1[1] = tmps[gid].out[5];
  w1[2] = tmps[gid].out[6];
  w1[3] = tmps[gid].out[7];
  w2[0] = 0;
  w2[1] = 0;
  w2[2] = 0;
  w2[3] = 0;
  w3[0] = 0;
  w3[1] = 0;
  w3[2] = 0;
  w3[3] = 0;

  /**
   * salt
   */

  const u32 salt_len = salt_bufs[salt_pos].salt_len;

  u32 ipad[5];
  u32 opad[5];

  hmac_sha1_pad_S (w0, w1, w2, w3, ipad, opad);

  tmps[gid].ipad[0] = ipad[0];
  tmps[gid].ipad[1] = ipad[1];
  tmps[gid].ipad[2] = ipad[2];
  tmps[gid].ipad[3] = ipad[3];
  tmps[gid].ipad[4] = ipad[4];

  tmps[gid].opad[0] = opad[0];
  tmps[gid].opad[1] = opad[1];
  tmps[gid].opad[2] = opad[2];
  tmps[gid].opad[3] = opad[3];
  tmps[gid].opad[4] = opad[4];

  for (u32 i = 0, j = 1; i < 8; i += 5, j += 1)
  {
    u32 w0[4];
    u32 w1[4];
    u32 w2[4];
    u32 w3[4];

    w0[0] = salt_bufs[salt_pos].salt_buf[0];
    w0[1] = salt_bufs[salt_pos].salt_buf[1];
    w0[2] = salt_bufs[salt_pos].salt_buf[2];
    w0[3] = salt_bufs[salt_pos].salt_buf[3];
    w1[0] = salt_bufs[salt_pos].salt_buf[4];
    w1[1] = j;
    w1[2] = 0x80000000;
    w1[3] = 0;
    w2[0] = 0;
    w2[1] = 0;
    w2[2] = 0;
    w2[3] = 0;
    w3[0] = 0;
    w3[1] = 0;
    w3[2] = 0;
    w3[3] = (64 + salt_len + 4) * 8;

    u32 dgst[5];

    hmac_sha1_run_S (w0, w1, w2, w3, ipad, opad, dgst);

    tmps[gid].dgst[i + 0] = dgst[0];
    tmps[gid].dgst[i + 1] = dgst[1];
    tmps[gid].dgst[i + 2] = dgst[2];
    tmps[gid].dgst[i + 3] = dgst[3];
    tmps[gid].dgst[i + 4] = dgst[4];

    tmps[gid].out[i + 0] = dgst[0];
    tmps[gid].out[i + 1] = dgst[1];
    tmps[gid].out[i + 2] = dgst[2];
    tmps[gid].out[i + 3] = dgst[3];
    tmps[gid].out[i + 4] = dgst[4];
  }
}

__kernel void m14800_loop2 (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global pbkdf2_sha256_tmp_t *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global itunes_backup_t *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
  const u32 gid = get_global_id (0);

  if ((gid * VECT_SIZE) >= gid_max) return;

  u32x ipad[5];
  u32x opad[5];

  ipad[0] = packv (tmps, ipad, gid, 0);
  ipad[1] = packv (tmps, ipad, gid, 1);
  ipad[2] = packv (tmps, ipad, gid, 2);
  ipad[3] = packv (tmps, ipad, gid, 3);
  ipad[4] = packv (tmps, ipad, gid, 4);

  opad[0] = packv (tmps, opad, gid, 0);
  opad[1] = packv (tmps, opad, gid, 1);
  opad[2] = packv (tmps, opad, gid, 2);
  opad[3] = packv (tmps, opad, gid, 3);
  opad[4] = packv (tmps, opad, gid, 4);

  for (u32 i = 0; i < 8; i += 5)
  {
    u32x dgst[5];
    u32x out[5];

    dgst[0] = packv (tmps, dgst, gid, i + 0);
    dgst[1] = packv (tmps, dgst, gid, i + 1);
    dgst[2] = packv (tmps, dgst, gid, i + 2);
    dgst[3] = packv (tmps, dgst, gid, i + 3);
    dgst[4] = packv (tmps, dgst, gid, i + 4);

    out[0] = packv (tmps, out, gid, i + 0);
    out[1] = packv (tmps, out, gid, i + 1);
    out[2] = packv (tmps, out, gid, i + 2);
    out[3] = packv (tmps, out, gid, i + 3);
    out[4] = packv (tmps, out, gid, i + 4);

    for (u32 j = 0; j < loop_cnt; j++)
    {
      u32x w0[4];
      u32x w1[4];
      u32x w2[4];
      u32x w3[4];

      w0[0] = dgst[0];
      w0[1] = dgst[1];
      w0[2] = dgst[2];
      w0[3] = dgst[3];
      w1[0] = dgst[4];
      w1[1] = 0x80000000;
      w1[2] = 0;
      w1[3] = 0;
      w2[0] = 0;
      w2[1] = 0;
      w2[2] = 0;
      w2[3] = 0;
      w3[0] = 0;
      w3[1] = 0;
      w3[2] = 0;
      w3[3] = (64 + 20) * 8;

      hmac_sha1_run_V (w0, w1, w2, w3, ipad, opad, dgst);

      out[0] ^= dgst[0];
      out[1] ^= dgst[1];
      out[2] ^= dgst[2];
      out[3] ^= dgst[3];
      out[4] ^= dgst[4];
    }

    unpackv (tmps, dgst, gid, i + 0, dgst[0]);
    unpackv (tmps, dgst, gid, i + 1, dgst[1]);
    unpackv (tmps, dgst, gid, i + 2, dgst[2]);
    unpackv (tmps, dgst, gid, i + 3, dgst[3]);
    unpackv (tmps, dgst, gid, i + 4, dgst[4]);

    unpackv (tmps, out, gid, i + 0, out[0]);
    unpackv (tmps, out, gid, i + 1, out[1]);
    unpackv (tmps, out, gid, i + 2, out[2]);
    unpackv (tmps, out, gid, i + 3, out[3]);
    unpackv (tmps, out, gid, i + 4, out[4]);
  }
}

__kernel void m14800_comp (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global pbkdf2_sha256_tmp_t *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global itunes_backup_t *esalt_bufs, __global u32 *d_return_buf, __global u32 *d_scryptV0_buf, __global u32 *d_scryptV1_buf, __global u32 *d_scryptV2_buf, __global u32 *d_scryptV3_buf, const u32 bitmap_mask, const u32 bitmap_shift1, const u32 bitmap_shift2, const u32 salt_pos, const u32 loop_pos, const u32 loop_cnt, const u32 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
  const u32 gid = get_global_id (0);
  const u32 lid = get_local_id (0);
  const u32 lsz = get_local_size (0);

  /**
   * aes shared
   */

  #ifdef REAL_SHM

  __local u32 s_td0[256];
  __local u32 s_td1[256];
  __local u32 s_td2[256];
  __local u32 s_td3[256];
  __local u32 s_td4[256];

  __local u32 s_te0[256];
  __local u32 s_te1[256];
  __local u32 s_te2[256];
  __local u32 s_te3[256];
  __local u32 s_te4[256];

  for (u32 i = lid; i < 256; i += lsz)
  {
    s_td0[i] = td0[i];
    s_td1[i] = td1[i];
    s_td2[i] = td2[i];
    s_td3[i] = td3[i];
    s_td4[i] = td4[i];

    s_te0[i] = te0[i];
    s_te1[i] = te1[i];
    s_te2[i] = te2[i];
    s_te3[i] = te3[i];
    s_te4[i] = te4[i];
  }

  barrier (CLK_LOCAL_MEM_FENCE);

  #else

  __constant u32a *s_td0 = td0;
  __constant u32a *s_td1 = td1;
  __constant u32a *s_td2 = td2;
  __constant u32a *s_td3 = td3;
  __constant u32a *s_td4 = td4;

  __constant u32a *s_te0 = te0;
  __constant u32a *s_te1 = te1;
  __constant u32a *s_te2 = te2;
  __constant u32a *s_te3 = te3;
  __constant u32a *s_te4 = te4;

  #endif

  if (gid >= gid_max) return;

  /**
   * aes
   */

  u32 ukey[8];

  ukey[0] = tmps[gid].out[0];
  ukey[1] = tmps[gid].out[1];
  ukey[2] = tmps[gid].out[2];
  ukey[3] = tmps[gid].out[3];
  ukey[4] = tmps[gid].out[4];
  ukey[5] = tmps[gid].out[5];
  ukey[6] = tmps[gid].out[6];
  ukey[7] = tmps[gid].out[7];

  #define KEYLEN 60

  u32 ks[KEYLEN];

  AES256_set_decrypt_key (ks, ukey, s_te0, s_te1, s_te2, s_te3, s_te4, s_td0, s_td1, s_td2, s_td3, s_td4);

  u32 cipher[4];

  cipher[0] = esalt_bufs[digests_offset].wpky[0];
  cipher[1] = esalt_bufs[digests_offset].wpky[1];
  cipher[2] = 0;
  cipher[3] = 0;

  u32 lsb[8];

  lsb[0] = esalt_bufs[digests_offset].wpky[8];
  lsb[1] = esalt_bufs[digests_offset].wpky[9];
  lsb[2] = esalt_bufs[digests_offset].wpky[6];
  lsb[3] = esalt_bufs[digests_offset].wpky[7];
  lsb[4] = esalt_bufs[digests_offset].wpky[4];
  lsb[5] = esalt_bufs[digests_offset].wpky[5];
  lsb[6] = esalt_bufs[digests_offset].wpky[2];
  lsb[7] = esalt_bufs[digests_offset].wpky[3];

  for (int j = 5; j >= 0; j--)
  {
    // 1st

    cipher[1] ^= (4 * j + 4);

    cipher[2] = lsb[0];
    cipher[3] = lsb[1];

    AES256_decrypt (ks, cipher, cipher, s_td0, s_td1, s_td2, s_td3, s_td4);

    lsb[0] = cipher[2];
    lsb[1] = cipher[3];

    // 2nd

    cipher[1] ^= (4 * j + 3);

    cipher[2] = lsb[2];
    cipher[3] = lsb[3];

    AES256_decrypt (ks, cipher, cipher, s_td0, s_td1, s_td2, s_td3, s_td4);

    lsb[2] = cipher[2];
    lsb[3] = cipher[3];

    // 3rd

    cipher[1] ^= (4 * j + 2);

    cipher[2] = lsb[4];
    cipher[3] = lsb[5];

    AES256_decrypt (ks, cipher, cipher, s_td0, s_td1, s_td2, s_td3, s_td4);

    lsb[4] = cipher[2];
    lsb[5] = cipher[3];

    // 4th

    cipher[1] ^= (4 * j + 1);

    cipher[2] = lsb[6];
    cipher[3] = lsb[7];

    AES256_decrypt (ks, cipher, cipher, s_td0, s_td1, s_td2, s_td3, s_td4);

    lsb[6] = cipher[2];
    lsb[7] = cipher[3];
  }

  if ((cipher[0] == 0xa6a6a6a6) && (cipher[1] == 0xa6a6a6a6))
  {
    if (atomic_inc (&hashes_shown[digests_offset]) == 0)
    {
      mark_hash (plains_buf, d_return_buf, salt_pos, digests_cnt, 0, digests_offset + 0, gid, 0);
    }

    return;
  }
}