hashcat/nv/m10410_a1.cu

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

#define _MD5_

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

#ifdef  VLIW1
#define VECT_SIZE1
#endif

#ifdef  VLIW2
#define VECT_SIZE1
#endif

#define DGST_R0 0
#define DGST_R1 1
#define DGST_R2 2
#define DGST_R3 3

#include "include/kernel_functions.c"
#include "types_nv.c"
#include "common_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

__device__ __constant__ u32 padding[8] =
{
  0x5e4ebf28,
  0x418a754e,
  0x564e0064,
  0x0801faff,
  0xb6002e2e,
  0x803e68d0,
  0xfea90c2f,
  0x7a695364
};

typedef struct
{
  u8 S[256];

  u32 wtf_its_faster;

} RC4_KEY;

__device__ static void swap (RC4_KEY *rc4_key, const u8 i, const u8 j)
{
  u8 tmp;

  tmp           = rc4_key->S[i];
  rc4_key->S[i] = rc4_key->S[j];
  rc4_key->S[j] = tmp;
}

__device__ static void rc4_init_16 (RC4_KEY *rc4_key, const u32 data[4])
{
  u32 v = 0x03020100;
  u32 a = 0x04040404;

  u32 *ptr = (u32 *) rc4_key->S;

  #pragma unroll 64
  for (u32 i = 0; i < 64; i++)
  {
    *ptr++ = v; v += a;
  }

  const u32 d0 = data[0] >>  0;
  const u32 d1 = data[0] >>  8;
  const u32 d2 = data[0] >> 16;
  const u32 d3 = data[0] >> 24;
  const u32 d4 = data[1] >>  0;

  u32 i = 0;
  u32 j = 0;

  #pragma unroll 52
  for (i = 0; i < 255; i += 5)
  {
    j += rc4_key->S[i + 0]; j += d0; swap (rc4_key, i + 0, j);
    j += rc4_key->S[i + 1]; j += d1; swap (rc4_key, i + 1, j);
    j += rc4_key->S[i + 2]; j += d2; swap (rc4_key, i + 2, j);
    j += rc4_key->S[i + 3]; j += d3; swap (rc4_key, i + 3, j);
    j += rc4_key->S[i + 4]; j += d4; swap (rc4_key, i + 4, j);
  }

  j += rc4_key->S[i + 0]; j += d0; swap (rc4_key, i + 0, j);
}

__device__ static u8 rc4_next_16 (RC4_KEY *rc4_key, u8 i, u8 j, const u32 in[4], u32 out[4])
{
  for (u32 k = 0; k < 4; k++)
  {
    u32 xor4 = 0;

    u8 idx;

    i += 1;
    j += rc4_key->S[i];

    swap (rc4_key, i, j);

    idx = rc4_key->S[i] + rc4_key->S[j];

    xor4 |= rc4_key->S[idx] <<  0;

    i += 1;
    j += rc4_key->S[i];

    swap (rc4_key, i, j);

    idx = rc4_key->S[i] + rc4_key->S[j];

    xor4 |= rc4_key->S[idx] <<  8;

    i += 1;
    j += rc4_key->S[i];

    swap (rc4_key, i, j);

    idx = rc4_key->S[i] + rc4_key->S[j];

    xor4 |= rc4_key->S[idx] << 16;

    i += 1;
    j += rc4_key->S[i];

    swap (rc4_key, i, j);

    idx = rc4_key->S[i] + rc4_key->S[j];

    xor4 |= rc4_key->S[idx] << 24;

    out[k] = in[k] ^ xor4;
  }

  return j;
}

__device__ __constant__ comb_t c_combs[1024];

extern "C" __global__ void __launch_bounds__ (64, 1) m10410_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 pdf_t *pdf_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 combs_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 wordl0[4];

  wordl0[0] = pws[gid].i[ 0];
  wordl0[1] = pws[gid].i[ 1];
  wordl0[2] = pws[gid].i[ 2];
  wordl0[3] = pws[gid].i[ 3];

  u32x wordl1[4];

  wordl1[0] = pws[gid].i[ 4];
  wordl1[1] = pws[gid].i[ 5];
  wordl1[2] = pws[gid].i[ 6];
  wordl1[3] = pws[gid].i[ 7];

  u32x wordl2[4];

  wordl2[0] = 0;
  wordl2[1] = 0;
  wordl2[2] = 0;
  wordl2[3] = 0;

  u32x wordl3[4];

  wordl3[0] = 0;
  wordl3[1] = 0;
  wordl3[2] = 0;
  wordl3[3] = 0;

  const u32 pw_l_len = pws[gid].pw_len;

  if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
  {
    switch_buffer_by_offset (wordl0, wordl1, wordl2, wordl3, c_combs[0].pw_len);
  }

  /**
   * key
   */

  __shared__ RC4_KEY rc4_keys[64];
  RC4_KEY *rc4_key = &rc4_keys[lid];

  /**
   * loop
   */

  for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
  {
    u32 wordr0[4];

    wordr0[0] = c_combs[il_pos].i[0];
    wordr0[1] = c_combs[il_pos].i[1];
    wordr0[2] = c_combs[il_pos].i[2];
    wordr0[3] = c_combs[il_pos].i[3];

    u32 wordr1[4];

    wordr1[0] = c_combs[il_pos].i[4];
    wordr1[1] = c_combs[il_pos].i[5];
    wordr1[2] = c_combs[il_pos].i[6];
    wordr1[3] = c_combs[il_pos].i[7];

    u32 wordr2[4];

    wordr2[0] = 0;
    wordr2[1] = 0;
    wordr2[2] = 0;
    wordr2[3] = 0;

    u32 wordr3[4];

    wordr3[0] = 0;
    wordr3[1] = 0;
    wordr3[2] = 0;
    wordr3[3] = 0;

    if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
    {
      switch_buffer_by_offset (wordr0, wordr1, wordr2, wordr3, pw_l_len);
    }

    u32x w0[2];

    w0[0] = wordl0[0] | wordr0[0];
    w0[1] = wordl0[1] | wordr0[1];

    // now the RC4 part

    u32x key[4];

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

    rc4_init_16 (rc4_key, key);

    u32x out[4];

    rc4_next_16 (rc4_key, 0, 0, padding, out);

    const u32x r0 = out[0];
    const u32x r1 = out[1];
    const u32x r2 = out[2];
    const u32x r3 = out[3];

    #include VECT_COMPARE_M
  }
}

extern "C" __global__ void __launch_bounds__ (64, 1) m10410_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 pdf_t *pdf_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
}

extern "C" __global__ void __launch_bounds__ (64, 1) m10410_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 pdf_t *pdf_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
}

extern "C" __global__ void __launch_bounds__ (64, 1) m10410_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 pdf_t *pdf_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 combs_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 wordl0[4];

  wordl0[0] = pws[gid].i[ 0];
  wordl0[1] = pws[gid].i[ 1];
  wordl0[2] = pws[gid].i[ 2];
  wordl0[3] = pws[gid].i[ 3];

  u32x wordl1[4];

  wordl1[0] = pws[gid].i[ 4];
  wordl1[1] = pws[gid].i[ 5];
  wordl1[2] = pws[gid].i[ 6];
  wordl1[3] = pws[gid].i[ 7];

  u32x wordl2[4];

  wordl2[0] = 0;
  wordl2[1] = 0;
  wordl2[2] = 0;
  wordl2[3] = 0;

  u32x wordl3[4];

  wordl3[0] = 0;
  wordl3[1] = 0;
  wordl3[2] = 0;
  wordl3[3] = 0;

  const u32 pw_l_len = pws[gid].pw_len;

  if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
  {
    switch_buffer_by_offset (wordl0, wordl1, wordl2, wordl3, c_combs[0].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]
  };

  /**
   * key
   */

  __shared__ RC4_KEY rc4_keys[64];
  RC4_KEY *rc4_key = &rc4_keys[lid];

  /**
   * loop
   */

  for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
  {
    u32 wordr0[4];

    wordr0[0] = c_combs[il_pos].i[0];
    wordr0[1] = c_combs[il_pos].i[1];
    wordr0[2] = c_combs[il_pos].i[2];
    wordr0[3] = c_combs[il_pos].i[3];

    u32 wordr1[4];

    wordr1[0] = c_combs[il_pos].i[4];
    wordr1[1] = c_combs[il_pos].i[5];
    wordr1[2] = c_combs[il_pos].i[6];
    wordr1[3] = c_combs[il_pos].i[7];

    u32 wordr2[4];

    wordr2[0] = 0;
    wordr2[1] = 0;
    wordr2[2] = 0;
    wordr2[3] = 0;

    u32 wordr3[4];

    wordr3[0] = 0;
    wordr3[1] = 0;
    wordr3[2] = 0;
    wordr3[3] = 0;

    if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
    {
      switch_buffer_by_offset (wordr0, wordr1, wordr2, wordr3, pw_l_len);
    }

    u32x w0[2];

    w0[0] = wordl0[0] | wordr0[0];
    w0[1] = wordl0[1] | wordr0[1];

    // now the RC4 part

    u32x key[4];

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

    rc4_init_16 (rc4_key, key);

    u32x out[4];

    rc4_next_16 (rc4_key, 0, 0, padding, out);

    const u32x r0 = out[0];
    const u32x r1 = out[1];
    const u32x r2 = out[2];
    const u32x r3 = out[3];

    #include VECT_COMPARE_S
  }
}

extern "C" __global__ void __launch_bounds__ (64, 1) m10410_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 pdf_t *pdf_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
}

extern "C" __global__ void __launch_bounds__ (64, 1) m10410_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 pdf_t *pdf_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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
{
}