/**
 * Author......: Jens Steube <jens.steube@gmail.com>
 * License.....: MIT
 */

#define _SHA512_

#include "include/constants.h"
#include "include/kernel_vendor.h"

#ifdef  VLIW1
#define VECT_SIZE1
#endif

#ifdef  VLIW2
#define VECT_SIZE1
#endif

#define DGST_R0 14
#define DGST_R1 15
#define DGST_R2 6
#define DGST_R3 7

#include "include/kernel_functions.c"
#include "types_nv.c"
#include "common_nv.c"
#include "include/rp_gpu.h"
#include "rp_nv.c"

#ifdef  VECT_SIZE1
#define VECT_COMPARE_S "check_single_vect1_comp4.c"
#define VECT_COMPARE_M "check_multi_vect1_comp4.c"
#endif

#ifdef  VECT_SIZE2
#define VECT_COMPARE_S "check_single_vect2_comp4.c"
#define VECT_COMPARE_M "check_multi_vect2_comp4.c"
#endif

#ifdef  VECT_SIZE4
#define VECT_COMPARE_S "check_single_vect4_comp4.c"
#define VECT_COMPARE_M "check_multi_vect4_comp4.c"
#endif

__device__ __constant__ gpu_rule_t c_rules[1024];

__device__ __constant__ u64 k_sha512[80] =
{
  SHA512C00, SHA512C01, SHA512C02, SHA512C03,
  SHA512C04, SHA512C05, SHA512C06, SHA512C07,
  SHA512C08, SHA512C09, SHA512C0a, SHA512C0b,
  SHA512C0c, SHA512C0d, SHA512C0e, SHA512C0f,
  SHA512C10, SHA512C11, SHA512C12, SHA512C13,
  SHA512C14, SHA512C15, SHA512C16, SHA512C17,
  SHA512C18, SHA512C19, SHA512C1a, SHA512C1b,
  SHA512C1c, SHA512C1d, SHA512C1e, SHA512C1f,
  SHA512C20, SHA512C21, SHA512C22, SHA512C23,
  SHA512C24, SHA512C25, SHA512C26, SHA512C27,
  SHA512C28, SHA512C29, SHA512C2a, SHA512C2b,
  SHA512C2c, SHA512C2d, SHA512C2e, SHA512C2f,
  SHA512C30, SHA512C31, SHA512C32, SHA512C33,
  SHA512C34, SHA512C35, SHA512C36, SHA512C37,
  SHA512C38, SHA512C39, SHA512C3a, SHA512C3b,
  SHA512C3c, SHA512C3d, SHA512C3e, SHA512C3f,
  SHA512C40, SHA512C41, SHA512C42, SHA512C43,
  SHA512C44, SHA512C45, SHA512C46, SHA512C47,
  SHA512C48, SHA512C49, SHA512C4a, SHA512C4b,
  SHA512C4c, SHA512C4d, SHA512C4e, SHA512C4f,
};

__device__ static void sha512_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u64x digest[8])
{
  u64x w0_t = hl32_to_64 (w0[0], w0[1]);
  u64x w1_t = hl32_to_64 (w0[2], w0[3]);
  u64x w2_t = hl32_to_64 (w1[0], w1[1]);
  u64x w3_t = hl32_to_64 (w1[2], w1[3]);
  u64x w4_t = hl32_to_64 (w2[0], w2[1]);
  u64x w5_t = hl32_to_64 (w2[2], w2[3]);
  u64x w6_t = hl32_to_64 (w3[0], w3[1]);
  u64x w7_t = 0;
  u64x w8_t = 0;
  u64x w9_t = 0;
  u64x wa_t = 0;
  u64x wb_t = 0;
  u64x wc_t = 0;
  u64x wd_t = 0;
  u64x we_t = 0;
  u64x wf_t = hl32_to_64 (w3[2], w3[3]);

  u64x a = digest[0];
  u64x b = digest[1];
  u64x c = digest[2];
  u64x d = digest[3];
  u64x e = digest[4];
  u64x f = digest[5];
  u64x g = digest[6];
  u64x h = digest[7];

  #define ROUND_EXPAND()                            \
  {                                                 \
    w0_t = SHA512_EXPAND (we_t, w9_t, w1_t, w0_t);  \
    w1_t = SHA512_EXPAND (wf_t, wa_t, w2_t, w1_t);  \
    w2_t = SHA512_EXPAND (w0_t, wb_t, w3_t, w2_t);  \
    w3_t = SHA512_EXPAND (w1_t, wc_t, w4_t, w3_t);  \
    w4_t = SHA512_EXPAND (w2_t, wd_t, w5_t, w4_t);  \
    w5_t = SHA512_EXPAND (w3_t, we_t, w6_t, w5_t);  \
    w6_t = SHA512_EXPAND (w4_t, wf_t, w7_t, w6_t);  \
    w7_t = SHA512_EXPAND (w5_t, w0_t, w8_t, w7_t);  \
    w8_t = SHA512_EXPAND (w6_t, w1_t, w9_t, w8_t);  \
    w9_t = SHA512_EXPAND (w7_t, w2_t, wa_t, w9_t);  \
    wa_t = SHA512_EXPAND (w8_t, w3_t, wb_t, wa_t);  \
    wb_t = SHA512_EXPAND (w9_t, w4_t, wc_t, wb_t);  \
    wc_t = SHA512_EXPAND (wa_t, w5_t, wd_t, wc_t);  \
    wd_t = SHA512_EXPAND (wb_t, w6_t, we_t, wd_t);  \
    we_t = SHA512_EXPAND (wc_t, w7_t, wf_t, we_t);  \
    wf_t = SHA512_EXPAND (wd_t, w8_t, w0_t, wf_t);  \
  }

  #define ROUND_STEP(i)                                                                   \
  {                                                                                       \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha512[i +  0]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha512[i +  1]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha512[i +  2]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha512[i +  3]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha512[i +  4]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha512[i +  5]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha512[i +  6]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha512[i +  7]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha512[i +  8]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha512[i +  9]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha512[i + 10]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha512[i + 11]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha512[i + 12]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha512[i + 13]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, c, d, e, f, g, h, a, b, we_t, k_sha512[i + 14]); \
    SHA512_STEP (SHA512_F0o, SHA512_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha512[i + 15]); \
  }

  ROUND_STEP (0);

  for (int i = 16; i < 80; i += 16)
  {
    ROUND_EXPAND (); ROUND_STEP (i);
  }

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

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

extern "C" __global__ void __launch_bounds__ (256, 1) m01700_m04 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, u32 *d_scryptV_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
  /**
   * modifier
   */

  const u32 lid = threadIdx.x;

  /**
   * base
   */

  const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;

  if (gid >= gid_max) return;

  u32x pw_buf0[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];

  u32x pw_buf1[4];

  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_len = pws[gid].pw_len;

  /**
   * loop
   */

  for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++)
  {
    u32x w0[4];

    w0[0] = pw_buf0[0];
    w0[1] = pw_buf0[1];
    w0[2] = pw_buf0[2];
    w0[3] = pw_buf0[3];

    u32x w1[4];

    w1[0] = pw_buf1[0];
    w1[1] = pw_buf1[1];
    w1[2] = pw_buf1[2];
    w1[3] = pw_buf1[3];

    u32x w2[4];

    w2[0] = 0;
    w2[1] = 0;
    w2[2] = 0;
    w2[3] = 0;

    u32x w3[4];

    w3[0] = 0;
    w3[1] = 0;
    w3[2] = 0;
    w3[3] = 0;

    const u32 out_len = apply_rules (c_rules[il_pos].cmds, w0, w1, pw_len);

    append_0x80_2 (w0, w1, out_len);

    /**
     * SHA512
     */

    u32x w0_t[4];
    u32x w1_t[4];
    u32x w2_t[4];
    u32x w3_t[4];

    w0_t[0] = swap_workaround (w0[0]);
    w0_t[1] = swap_workaround (w0[1]);
    w0_t[2] = swap_workaround (w0[2]);
    w0_t[3] = swap_workaround (w0[3]);
    w1_t[0] = swap_workaround (w1[0]);
    w1_t[1] = swap_workaround (w1[1]);
    w1_t[2] = swap_workaround (w1[2]);
    w1_t[3] = swap_workaround (w1[3]);
    w2_t[0] = swap_workaround (w2[0]);
    w2_t[1] = swap_workaround (w2[1]);
    w2_t[2] = swap_workaround (w2[2]);
    w2_t[3] = swap_workaround (w2[3]);
    w3_t[0] = swap_workaround (w3[0]);
    w3_t[1] = swap_workaround (w3[1]);
    w3_t[2] = 0;
    w3_t[3] = out_len * 8;

    u64x digest[8];

    digest[0] = SHA512M_A;
    digest[1] = SHA512M_B;
    digest[2] = SHA512M_C;
    digest[3] = SHA512M_D;
    digest[4] = SHA512M_E;
    digest[5] = SHA512M_F;
    digest[6] = SHA512M_G;
    digest[7] = SHA512M_H;

    sha512_transform (w0_t, w1_t, w2_t, w3_t, digest);


    const u32x r0 = l32_from_64 (digest[7]);
    const u32x r1 = h32_from_64 (digest[7]);
    const u32x r2 = l32_from_64 (digest[3]);
    const u32x r3 = h32_from_64 (digest[3]);

    #include VECT_COMPARE_M
  }
}

extern "C" __global__ void __launch_bounds__ (256, 1) m01700_m08 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, u32 *d_scryptV_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
}

extern "C" __global__ void __launch_bounds__ (256, 1) m01700_m16 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, u32 *d_scryptV_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
}

extern "C" __global__ void __launch_bounds__ (256, 1) m01700_s04 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, u32 *d_scryptV_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
  /**
   * modifier
   */

  const u32 lid = threadIdx.x;

  /**
   * base
   */

  const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;

  if (gid >= gid_max) return;

  u32x pw_buf0[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];

  u32x pw_buf1[4];

  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_len = pws[gid].pw_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 < rules_cnt; il_pos++)
  {
    u32x w0[4];

    w0[0] = pw_buf0[0];
    w0[1] = pw_buf0[1];
    w0[2] = pw_buf0[2];
    w0[3] = pw_buf0[3];

    u32x w1[4];

    w1[0] = pw_buf1[0];
    w1[1] = pw_buf1[1];
    w1[2] = pw_buf1[2];
    w1[3] = pw_buf1[3];

    u32x w2[4];

    w2[0] = 0;
    w2[1] = 0;
    w2[2] = 0;
    w2[3] = 0;

    u32x w3[4];

    w3[0] = 0;
    w3[1] = 0;
    w3[2] = 0;
    w3[3] = 0;

    const u32 out_len = apply_rules (c_rules[il_pos].cmds, w0, w1, pw_len);

    append_0x80_2 (w0, w1, out_len);

    /**
     * SHA512
     */

    u32x w0_t[4];
    u32x w1_t[4];
    u32x w2_t[4];
    u32x w3_t[4];

    w0_t[0] = swap_workaround (w0[0]);
    w0_t[1] = swap_workaround (w0[1]);
    w0_t[2] = swap_workaround (w0[2]);
    w0_t[3] = swap_workaround (w0[3]);
    w1_t[0] = swap_workaround (w1[0]);
    w1_t[1] = swap_workaround (w1[1]);
    w1_t[2] = swap_workaround (w1[2]);
    w1_t[3] = swap_workaround (w1[3]);
    w2_t[0] = swap_workaround (w2[0]);
    w2_t[1] = swap_workaround (w2[1]);
    w2_t[2] = swap_workaround (w2[2]);
    w2_t[3] = swap_workaround (w2[3]);
    w3_t[0] = swap_workaround (w3[0]);
    w3_t[1] = swap_workaround (w3[1]);
    w3_t[2] = 0;
    w3_t[3] = out_len * 8;

    u64x digest[8];

    digest[0] = SHA512M_A;
    digest[1] = SHA512M_B;
    digest[2] = SHA512M_C;
    digest[3] = SHA512M_D;
    digest[4] = SHA512M_E;
    digest[5] = SHA512M_F;
    digest[6] = SHA512M_G;
    digest[7] = SHA512M_H;

    sha512_transform (w0_t, w1_t, w2_t, w3_t, digest);


    const u32x r0 = l32_from_64 (digest[7]);
    const u32x r1 = h32_from_64 (digest[7]);
    const u32x r2 = l32_from_64 (digest[3]);
    const u32x r3 = h32_from_64 (digest[3]);

    #include VECT_COMPARE_S
  }
}

extern "C" __global__ void __launch_bounds__ (256, 1) m01700_s08 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, u32 *d_scryptV_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
}

extern "C" __global__ void __launch_bounds__ (256, 1) m01700_s16 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const void *esalt_bufs, u32 *d_return_buf, u32 *d_scryptV_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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
}