hashcat/OpenCL/m10400_a0-optimized.cl

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

//too much register pressure
//#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_rp_optimized.h"
#include "inc_rp_optimized.cl"
#include "inc_simd.cl"
#include "inc_hash_md5.cl"

__constant u32a padding[8] =
{
  0x5e4ebf28,
  0x418a754e,
  0x564e0064,
  0x0801faff,
  0xb6002e2e,
  0x803e68d0,
  0xfea90c2f,
  0x7a695364
};

typedef struct
{
  u8 S[256];

  u32 wtf_its_faster;

} RC4_KEY;

DECLSPEC void swap (__local 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;
}

DECLSPEC void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 *data)
{
  u32 v = 0x03020100;
  u32 a = 0x04040404;

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

  #ifdef _unroll
  #pragma unroll
  #endif
  for (u32 i = 0; i < 64; i++)
  {
    ptr[i] = 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 j = 0;

  #ifdef _unroll
  #pragma unroll
  #endif
  for (u32 i = 0; i < 255; i += 5)
  {
    j += rc4_key->S[i + 0] + d0; swap (rc4_key, i + 0, j);
    j += rc4_key->S[i + 1] + d1; swap (rc4_key, i + 1, j);
    j += rc4_key->S[i + 2] + d2; swap (rc4_key, i + 2, j);
    j += rc4_key->S[i + 3] + d3; swap (rc4_key, i + 3, j);
    j += rc4_key->S[i + 4] + d4; swap (rc4_key, i + 4, j);
  }

  j += rc4_key->S[255] + d0; swap (rc4_key, 255, j);
}

DECLSPEC u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, __constant u32 *in, u32 *out)
{
  #ifdef _unroll
  #pragma unroll
  #endif
  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;
}

__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m10400_m04 (__global pw_t *pws, __constant const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global void *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 pdf_t *pdf_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 u64 gid_max)
{
  /**
   * modifier
   */

  const u64 lid = get_local_id (0);

  /**
   * base
   */

  const u64 gid = get_global_id (0);

  if (gid >= gid_max) return;

  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_len = pws[gid].pw_len;

  /**
   * shared
   */

  __local RC4_KEY rc4_keys[64];

  __local RC4_KEY *rc4_key = &rc4_keys[lid];

  /**
   * U_buf
   */

  u32 o_buf[8];

  o_buf[0] = pdf_bufs[digests_offset].o_buf[0];
  o_buf[1] = pdf_bufs[digests_offset].o_buf[1];
  o_buf[2] = pdf_bufs[digests_offset].o_buf[2];
  o_buf[3] = pdf_bufs[digests_offset].o_buf[3];
  o_buf[4] = pdf_bufs[digests_offset].o_buf[4];
  o_buf[5] = pdf_bufs[digests_offset].o_buf[5];
  o_buf[6] = pdf_bufs[digests_offset].o_buf[6];
  o_buf[7] = pdf_bufs[digests_offset].o_buf[7];

  u32 P = pdf_bufs[digests_offset].P;

  u32 id_buf[4];

  id_buf[0] = pdf_bufs[digests_offset].id_buf[0];
  id_buf[1] = pdf_bufs[digests_offset].id_buf[1];
  id_buf[2] = pdf_bufs[digests_offset].id_buf[2];
  id_buf[3] = pdf_bufs[digests_offset].id_buf[3];

  /**
   * loop
   */

  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
  {
    u32x w0[4] = { 0 };
    u32x w1[4] = { 0 };
    u32x w2[4] = { 0 };
    u32x w3[4] = { 0 };

    const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);

    /**
     * pdf
     */

    u32 p0[4];
    u32 p1[4];
    u32 p2[4];
    u32 p3[4];

    p0[0] = padding[0];
    p0[1] = padding[1];
    p0[2] = padding[2];
    p0[3] = padding[3];
    p1[0] = padding[4];
    p1[1] = padding[5];
    p1[2] = padding[6];
    p1[3] = padding[7];
    p2[0] = 0;
    p2[1] = 0;
    p2[2] = 0;
    p2[3] = 0;
    p3[0] = 0;
    p3[1] = 0;
    p3[2] = 0;
    p3[3] = 0;

    switch_buffer_by_offset_le (p0, p1, p2, p3, out_len);

    // add password
    // truncate at 32 is wanted, not a bug!
    // add o_buf

    w0[0] |= p0[0];
    w0[1] |= p0[1];
    w0[2] |= p0[2];
    w0[3] |= p0[3];
    w1[0] |= p1[0];
    w1[1] |= p1[1];
    w1[2] |= p1[2];
    w1[3] |= p1[3];
    w2[0]  = o_buf[0];
    w2[1]  = o_buf[1];
    w2[2]  = o_buf[2];
    w2[3]  = o_buf[3];
    w3[0]  = o_buf[4];
    w3[1]  = o_buf[5];
    w3[2]  = o_buf[6];
    w3[3]  = o_buf[7];

    u32 digest[4];

    digest[0] = MD5M_A;
    digest[1] = MD5M_B;
    digest[2] = MD5M_C;
    digest[3] = MD5M_D;

    md5_transform (w0, w1, w2, w3, digest);

    w0[0] = P;
    w0[1] = id_buf[0];
    w0[2] = id_buf[1];
    w0[3] = id_buf[2];
    w1[0] = id_buf[3];
    w1[1] = 0x80;
    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] = 84 * 8;
    w3[3] = 0;

    md5_transform (w0, w1, w2, w3, digest);

    // now the RC4 part

    digest[1] = digest[1] & 0xff;
    digest[2] = 0;
    digest[3] = 0;

    rc4_init_16 (rc4_key, digest);

    u32 out[4];

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

    COMPARE_M_SIMD (out[0], out[1], out[2], out[3]);
  }
}

__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m10400_m08 (__global pw_t *pws, __constant const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global void *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 pdf_t *pdf_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 u64 gid_max)
{
}

__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m10400_m16 (__global pw_t *pws, __constant const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global void *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 pdf_t *pdf_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 u64 gid_max)
{
}

__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m10400_s04 (__global pw_t *pws, __constant const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global void *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 pdf_t *pdf_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 u64 gid_max)
{
  /**
   * modifier
   */

  const u64 lid = get_local_id (0);

  /**
   * base
   */

  const u64 gid = get_global_id (0);

  if (gid >= gid_max) return;

  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_len = pws[gid].pw_len;

  /**
   * shared
   */

  __local RC4_KEY rc4_keys[64];

  __local RC4_KEY *rc4_key = &rc4_keys[lid];

  /**
   * U_buf
   */

  u32 o_buf[8];

  o_buf[0] = pdf_bufs[digests_offset].o_buf[0];
  o_buf[1] = pdf_bufs[digests_offset].o_buf[1];
  o_buf[2] = pdf_bufs[digests_offset].o_buf[2];
  o_buf[3] = pdf_bufs[digests_offset].o_buf[3];
  o_buf[4] = pdf_bufs[digests_offset].o_buf[4];
  o_buf[5] = pdf_bufs[digests_offset].o_buf[5];
  o_buf[6] = pdf_bufs[digests_offset].o_buf[6];
  o_buf[7] = pdf_bufs[digests_offset].o_buf[7];

  u32 P = pdf_bufs[digests_offset].P;

  u32 id_buf[4];

  id_buf[0] = pdf_bufs[digests_offset].id_buf[0];
  id_buf[1] = pdf_bufs[digests_offset].id_buf[1];
  id_buf[2] = pdf_bufs[digests_offset].id_buf[2];
  id_buf[3] = pdf_bufs[digests_offset].id_buf[3];

  /**
   * 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)
  {
    u32x w0[4] = { 0 };
    u32x w1[4] = { 0 };
    u32x w2[4] = { 0 };
    u32x w3[4] = { 0 };

    const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);

    /**
     * pdf
     */

    u32 p0[4];
    u32 p1[4];
    u32 p2[4];
    u32 p3[4];

    p0[0] = padding[0];
    p0[1] = padding[1];
    p0[2] = padding[2];
    p0[3] = padding[3];
    p1[0] = padding[4];
    p1[1] = padding[5];
    p1[2] = padding[6];
    p1[3] = padding[7];
    p2[0] = 0;
    p2[1] = 0;
    p2[2] = 0;
    p2[3] = 0;
    p3[0] = 0;
    p3[1] = 0;
    p3[2] = 0;
    p3[3] = 0;

    switch_buffer_by_offset_le (p0, p1, p2, p3, out_len);

    // add password
    // truncate at 32 is wanted, not a bug!
    // add o_buf

    w0[0] |= p0[0];
    w0[1] |= p0[1];
    w0[2] |= p0[2];
    w0[3] |= p0[3];
    w1[0] |= p1[0];
    w1[1] |= p1[1];
    w1[2] |= p1[2];
    w1[3] |= p1[3];
    w2[0]  = o_buf[0];
    w2[1]  = o_buf[1];
    w2[2]  = o_buf[2];
    w2[3]  = o_buf[3];
    w3[0]  = o_buf[4];
    w3[1]  = o_buf[5];
    w3[2]  = o_buf[6];
    w3[3]  = o_buf[7];

    u32 digest[4];

    digest[0] = MD5M_A;
    digest[1] = MD5M_B;
    digest[2] = MD5M_C;
    digest[3] = MD5M_D;

    md5_transform (w0, w1, w2, w3, digest);

    w0[0] = P;
    w0[1] = id_buf[0];
    w0[2] = id_buf[1];
    w0[3] = id_buf[2];
    w1[0] = id_buf[3];
    w1[1] = 0x80;
    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] = 84 * 8;
    w3[3] = 0;

    md5_transform (w0, w1, w2, w3, digest);

    // now the RC4 part

    digest[1] = digest[1] & 0xff;
    digest[2] = 0;
    digest[3] = 0;

    rc4_init_16 (rc4_key, digest);

    u32 out[4];

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

    COMPARE_S_SIMD (out[0], out[1], out[2], out[3]);
  }
}

__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m10400_s08 (__global pw_t *pws, __constant const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global void *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 pdf_t *pdf_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 u64 gid_max)
{
}

__kernel void __attribute__((reqd_work_group_size(64, 1, 1))) m10400_s16 (__global pw_t *pws, __constant const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global void *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 pdf_t *pdf_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 u64 gid_max)
{
}