/**
 * 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_hash_sha1.cl"

#if   VECT_SIZE == 1
#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i)])
#elif VECT_SIZE == 2
#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1])
#elif VECT_SIZE == 4
#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3])
#elif VECT_SIZE == 8
#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7])
#elif VECT_SIZE == 16
#define uint_to_hex_lower8_le(i) (u32x) (l_bin2asc[(i).s0], l_bin2asc[(i).s1], l_bin2asc[(i).s2], l_bin2asc[(i).s3], l_bin2asc[(i).s4], l_bin2asc[(i).s5], l_bin2asc[(i).s6], l_bin2asc[(i).s7], l_bin2asc[(i).s8], l_bin2asc[(i).s9], l_bin2asc[(i).sa], l_bin2asc[(i).sb], l_bin2asc[(i).sc], l_bin2asc[(i).sd], l_bin2asc[(i).se], l_bin2asc[(i).sf])
#endif

__kernel void m13900_m04 (KERN_ATTR_BASIC ())
{
  /**
   * modifier
   */

  const u64 gid = get_global_id (0);
  const u64 lid = get_local_id (0);
  const u64 lsz = get_local_size (0);

  /**
   * shared
   */

  __local u32 l_bin2asc[256];

  for (u32 i = lid; i < 256; i += lsz)
  {
    const u32 i0 = (i >> 0) & 15;
    const u32 i1 = (i >> 4) & 15;

    l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'a' - 10 + i0) << 0
                 | ((i1 < 10) ? '0' + i1 : 'a' - 10 + i1) << 8;
  }

  barrier (CLK_LOCAL_MEM_FENCE);

  if (gid >= gid_max) return;

  /**
   * base
   */

  u32 pw_buf0[4];
  u32 pw_buf1[4];

  pw_buf0[0] = pws[gid].i[0];
  pw_buf0[1] = pws[gid].i[1];
  pw_buf0[2] = pws[gid].i[2];
  pw_buf0[3] = pws[gid].i[3];
  pw_buf1[0] = pws[gid].i[4];
  pw_buf1[1] = pws[gid].i[5];
  pw_buf1[2] = pws[gid].i[6];
  pw_buf1[3] = pws[gid].i[7];

  const u32 pw_l_len = pws[gid].pw_len & 63;

  /**
   * salt
   */

  u32 salt_buf0[3];

  salt_buf0[0] = swap32_S (salt_bufs[salt_pos].salt_buf[0]);
  salt_buf0[1] = swap32_S (salt_bufs[salt_pos].salt_buf[1]);
  salt_buf0[2] = swap32_S (salt_bufs[salt_pos].salt_buf[2]);

  const u32 salt_len = salt_bufs[salt_pos].salt_len;

  /**
   * loop
   */

  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
  {
    const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos) & 63;

    const u32x pw_len = (pw_l_len + pw_r_len) & 63;

    /**
     * concat password candidate
     */

    u32x wordl0[4] = { 0 };
    u32x wordl1[4] = { 0 };
    u32x wordl2[4] = { 0 };
    u32x wordl3[4] = { 0 };

    wordl0[0] = pw_buf0[0];
    wordl0[1] = pw_buf0[1];
    wordl0[2] = pw_buf0[2];
    wordl0[3] = pw_buf0[3];
    wordl1[0] = pw_buf1[0];
    wordl1[1] = pw_buf1[1];
    wordl1[2] = pw_buf1[2];
    wordl1[3] = pw_buf1[3];

    u32x wordr0[4] = { 0 };
    u32x wordr1[4] = { 0 };
    u32x wordr2[4] = { 0 };
    u32x wordr3[4] = { 0 };

    wordr0[0] = ix_create_combt (combs_buf, il_pos, 0);
    wordr0[1] = ix_create_combt (combs_buf, il_pos, 1);
    wordr0[2] = ix_create_combt (combs_buf, il_pos, 2);
    wordr0[3] = ix_create_combt (combs_buf, il_pos, 3);
    wordr1[0] = ix_create_combt (combs_buf, il_pos, 4);
    wordr1[1] = ix_create_combt (combs_buf, il_pos, 5);
    wordr1[2] = ix_create_combt (combs_buf, il_pos, 6);
    wordr1[3] = ix_create_combt (combs_buf, il_pos, 7);

    if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
    {
      switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len);
    }
    else
    {
      switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len);
    }

    u32x w0[4];
    u32x w1[4];
    u32x w2[4];
    u32x w3[4];

    w0[0] = wordl0[0] | wordr0[0];
    w0[1] = wordl0[1] | wordr0[1];
    w0[2] = wordl0[2] | wordr0[2];
    w0[3] = wordl0[3] | wordr0[3];
    w1[0] = wordl1[0] | wordr1[0];
    w1[1] = wordl1[1] | wordr1[1];
    w1[2] = wordl1[2] | wordr1[2];
    w1[3] = wordl1[3] | wordr1[3];
    w2[0] = wordl2[0] | wordr2[0];
    w2[1] = wordl2[1] | wordr2[1];
    w2[2] = wordl2[2] | wordr2[2];
    w2[3] = wordl2[3] | wordr2[3];
    w3[0] = wordl3[0] | wordr3[0];
    w3[1] = wordl3[1] | wordr3[1];
    w3[2] = wordl3[2] | wordr3[2];
    w3[3] = wordl3[3] | wordr3[3];

    /**
     * SHA1($pass)
     */

    w0[0] = swap32 (w0[0]);
    w0[1] = swap32 (w0[1]);
    w0[2] = swap32 (w0[2]);
    w0[3] = swap32 (w0[3]);
    w1[0] = swap32 (w1[0]);
    w1[1] = swap32 (w1[1]);
    w1[2] = swap32 (w1[2]);
    w1[3] = swap32 (w1[3]);
    w2[0] = swap32 (w2[0]);
    w2[1] = swap32 (w2[1]);
    w2[2] = swap32 (w2[2]);
    w2[3] = swap32 (w2[3]);
    w3[0] = swap32 (w3[0]);
    w3[1] = swap32 (w3[1]);
    w3[2] = 0;
    w3[3] = pw_len * 8;

    u32x digest[5];

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

    sha1_transform_vector (w0, w1, w2, w3, digest);

    /**
     * SHA1($salt.SHA1($pass))
     */

    u32x a;
    u32x b;
    u32x c;
    u32x d;
    u32x e;

    a = digest[0];
    b = digest[1];
    c = digest[2];
    d = digest[3];
    e = digest[4];

    w0[0] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
    w0[1] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
    w0[2] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
    w0[3] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
    w1[0] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
    w1[1] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
    w1[2] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
    w1[3] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
    w2[0] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
    w2[1] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
    w2[2] = 0x80000000;
    w2[3] = 0;
    w3[0] = 0;
    w3[1] = 0;
    w3[2] = 0;
    w3[3] = 0;

    switch_buffer_by_offset_be (w0, w1, w2, w3, 9);

    w0[0] |= salt_buf0[0];
    w0[1] |= salt_buf0[1];
    w0[2] |= salt_buf0[2];
    w3[3] = 49 * 8;

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

    sha1_transform_vector (w0, w1, w2, w3, digest);

    /**
     * SHA1($salt.SHA1($salt.SHA1($pass)))
     */

    a = digest[0];
    b = digest[1];
    c = digest[2];
    d = digest[3];
    e = digest[4];

    w0[0] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
    w0[1] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
    w0[2] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
    w0[3] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
    w1[0] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
    w1[1] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
    w1[2] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
    w1[3] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
    w2[0] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
    w2[1] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
    w2[2] = 0x80000000;
    w2[3] = 0;
    w3[0] = 0;
    w3[1] = 0;
    w3[2] = 0;
    w3[3] = 0;

    switch_buffer_by_offset_be (w0, w1, w2, w3, 9);

    w0[0] |= salt_buf0[0];
    w0[1] |= salt_buf0[1];
    w0[2] |= salt_buf0[2];
    w3[3] = 49 * 8;

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

    sha1_transform_vector (w0, w1, w2, w3, digest);

    COMPARE_M_SIMD (digest[3], digest[4], digest[2], digest[1]);
  }
}

__kernel void m13900_m08 (KERN_ATTR_BASIC ())
{
}

__kernel void m13900_m16 (KERN_ATTR_BASIC ())
{
}

__kernel void m13900_s04 (KERN_ATTR_BASIC ())
{
  /**
   * modifier
   */

  const u64 gid = get_global_id (0);
  const u64 lid = get_local_id (0);
  const u64 lsz = get_local_size (0);

  /**
   * shared
   */

  __local u32 l_bin2asc[256];

  for (u32 i = lid; i < 256; i += lsz)
  {
    const u32 i0 = (i >> 0) & 15;
    const u32 i1 = (i >> 4) & 15;

    l_bin2asc[i] = ((i0 < 10) ? '0' + i0 : 'a' - 10 + i0) << 0
                 | ((i1 < 10) ? '0' + i1 : 'a' - 10 + i1) << 8;
  }

  barrier (CLK_LOCAL_MEM_FENCE);

  if (gid >= gid_max) return;

  /**
   * base
   */

  u32 pw_buf0[4];
  u32 pw_buf1[4];

  pw_buf0[0] = pws[gid].i[0];
  pw_buf0[1] = pws[gid].i[1];
  pw_buf0[2] = pws[gid].i[2];
  pw_buf0[3] = pws[gid].i[3];
  pw_buf1[0] = pws[gid].i[4];
  pw_buf1[1] = pws[gid].i[5];
  pw_buf1[2] = pws[gid].i[6];
  pw_buf1[3] = pws[gid].i[7];

  const u32 pw_l_len = pws[gid].pw_len & 63;

  /**
   * salt
   */

  u32 salt_buf0[3];

  salt_buf0[0] = swap32_S (salt_bufs[salt_pos].salt_buf[0]);
  salt_buf0[1] = swap32_S (salt_bufs[salt_pos].salt_buf[1]);
  salt_buf0[2] = swap32_S (salt_bufs[salt_pos].salt_buf[2]);

  const u32 salt_len = salt_bufs[salt_pos].salt_len;

  /**
   * digest
   */

  const u32 search[4] =
  {
    digests_buf[digests_offset].digest_buf[DGST_R0],
    digests_buf[digests_offset].digest_buf[DGST_R1],
    digests_buf[digests_offset].digest_buf[DGST_R2],
    digests_buf[digests_offset].digest_buf[DGST_R3]
  };

  /**
   * loop
   */

  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
  {
    const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos) & 63;

    const u32x pw_len = (pw_l_len + pw_r_len) & 63;

    /**
     * concat password candidate
     */

    u32x wordl0[4] = { 0 };
    u32x wordl1[4] = { 0 };
    u32x wordl2[4] = { 0 };
    u32x wordl3[4] = { 0 };

    wordl0[0] = pw_buf0[0];
    wordl0[1] = pw_buf0[1];
    wordl0[2] = pw_buf0[2];
    wordl0[3] = pw_buf0[3];
    wordl1[0] = pw_buf1[0];
    wordl1[1] = pw_buf1[1];
    wordl1[2] = pw_buf1[2];
    wordl1[3] = pw_buf1[3];

    u32x wordr0[4] = { 0 };
    u32x wordr1[4] = { 0 };
    u32x wordr2[4] = { 0 };
    u32x wordr3[4] = { 0 };

    wordr0[0] = ix_create_combt (combs_buf, il_pos, 0);
    wordr0[1] = ix_create_combt (combs_buf, il_pos, 1);
    wordr0[2] = ix_create_combt (combs_buf, il_pos, 2);
    wordr0[3] = ix_create_combt (combs_buf, il_pos, 3);
    wordr1[0] = ix_create_combt (combs_buf, il_pos, 4);
    wordr1[1] = ix_create_combt (combs_buf, il_pos, 5);
    wordr1[2] = ix_create_combt (combs_buf, il_pos, 6);
    wordr1[3] = ix_create_combt (combs_buf, il_pos, 7);

    if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
    {
      switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len);
    }
    else
    {
      switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len);
    }

    u32x w0[4];
    u32x w1[4];
    u32x w2[4];
    u32x w3[4];

    w0[0] = wordl0[0] | wordr0[0];
    w0[1] = wordl0[1] | wordr0[1];
    w0[2] = wordl0[2] | wordr0[2];
    w0[3] = wordl0[3] | wordr0[3];
    w1[0] = wordl1[0] | wordr1[0];
    w1[1] = wordl1[1] | wordr1[1];
    w1[2] = wordl1[2] | wordr1[2];
    w1[3] = wordl1[3] | wordr1[3];
    w2[0] = wordl2[0] | wordr2[0];
    w2[1] = wordl2[1] | wordr2[1];
    w2[2] = wordl2[2] | wordr2[2];
    w2[3] = wordl2[3] | wordr2[3];
    w3[0] = wordl3[0] | wordr3[0];
    w3[1] = wordl3[1] | wordr3[1];
    w3[2] = wordl3[2] | wordr3[2];
    w3[3] = wordl3[3] | wordr3[3];

    /**
     * SHA1($pass)
     */

    w0[0] = swap32 (w0[0]);
    w0[1] = swap32 (w0[1]);
    w0[2] = swap32 (w0[2]);
    w0[3] = swap32 (w0[3]);
    w1[0] = swap32 (w1[0]);
    w1[1] = swap32 (w1[1]);
    w1[2] = swap32 (w1[2]);
    w1[3] = swap32 (w1[3]);
    w2[0] = swap32 (w2[0]);
    w2[1] = swap32 (w2[1]);
    w2[2] = swap32 (w2[2]);
    w2[3] = swap32 (w2[3]);
    w3[0] = swap32 (w3[0]);
    w3[1] = swap32 (w3[1]);
    w3[2] = 0;
    w3[3] = pw_len * 8;

    u32x digest[5];

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

    sha1_transform_vector (w0, w1, w2, w3, digest);

    /**
     * SHA1($salt.SHA1($pass))
     */

    u32x a;
    u32x b;
    u32x c;
    u32x d;
    u32x e;

    a = digest[0];
    b = digest[1];
    c = digest[2];
    d = digest[3];
    e = digest[4];

    w0[0] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
    w0[1] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
    w0[2] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
    w0[3] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
    w1[0] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
    w1[1] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
    w1[2] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
    w1[3] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
    w2[0] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
    w2[1] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
    w2[2] = 0x80000000;
    w2[3] = 0;
    w3[0] = 0;
    w3[1] = 0;
    w3[2] = 0;
    w3[3] = 0;

    switch_buffer_by_offset_be (w0, w1, w2, w3, 9);

    w0[0] |= salt_buf0[0];
    w0[1] |= salt_buf0[1];
    w0[2] |= salt_buf0[2];
    w3[3] = 49 * 8;

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

    sha1_transform_vector (w0, w1, w2, w3, digest);

    /**
     * SHA1($salt.SHA1($salt.SHA1($pass)))
     */

    a = digest[0];
    b = digest[1];
    c = digest[2];
    d = digest[3];
    e = digest[4];

    w0[0] = uint_to_hex_lower8_le ((a >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((a >> 24) & 255) << 16;
    w0[1] = uint_to_hex_lower8_le ((a >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((a >>  8) & 255) << 16;
    w0[2] = uint_to_hex_lower8_le ((b >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((b >> 24) & 255) << 16;
    w0[3] = uint_to_hex_lower8_le ((b >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((b >>  8) & 255) << 16;
    w1[0] = uint_to_hex_lower8_le ((c >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((c >> 24) & 255) << 16;
    w1[1] = uint_to_hex_lower8_le ((c >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((c >>  8) & 255) << 16;
    w1[2] = uint_to_hex_lower8_le ((d >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((d >> 24) & 255) << 16;
    w1[3] = uint_to_hex_lower8_le ((d >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((d >>  8) & 255) << 16;
    w2[0] = uint_to_hex_lower8_le ((e >> 16) & 255) <<  0
          | uint_to_hex_lower8_le ((e >> 24) & 255) << 16;
    w2[1] = uint_to_hex_lower8_le ((e >>  0) & 255) <<  0
          | uint_to_hex_lower8_le ((e >>  8) & 255) << 16;
    w2[2] = 0x80000000;
    w2[3] = 0;
    w3[0] = 0;
    w3[1] = 0;
    w3[2] = 0;
    w3[3] = 0;

    switch_buffer_by_offset_be (w0, w1, w2, w3, 9);

    w0[0] |= salt_buf0[0];
    w0[1] |= salt_buf0[1];
    w0[2] |= salt_buf0[2];
    w3[3] = 49 * 8;

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

    sha1_transform_vector (w0, w1, w2, w3, digest);

    COMPARE_S_SIMD (digest[3], digest[4], digest[2], digest[1]);
  }
}

__kernel void m13900_s08 (KERN_ATTR_BASIC ())
{
}

__kernel void m13900_s16 (KERN_ATTR_BASIC ())
{
}