hashcat/OpenCL/m30905_a3-pure.cl

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

//#define NEW_SIMD_CODE

// #define SECP256K1_TMPS_TYPE CONSTANT_AS
#define SECP256K1_TMPS_TYPE PRIVATE_AS

#ifdef KERNEL_STATIC
#include M2S(INCLUDE_PATH/inc_vendor.h)
#include M2S(INCLUDE_PATH/inc_types.h)
#include M2S(INCLUDE_PATH/inc_platform.cl)
#include M2S(INCLUDE_PATH/inc_common.cl)
#include M2S(INCLUDE_PATH/inc_scalar.cl)
#include M2S(INCLUDE_PATH/inc_hash_sha256.cl)
#include M2S(INCLUDE_PATH/inc_hash_ripemd160.cl)
#include M2S(INCLUDE_PATH/inc_ecc_secp256k1.cl)
#endif

DECLSPEC u32 hex_convert_u32 (PRIVATE_AS const u32 c)
{
  return (c & 15) + (c >> 6) * 9;
}

DECLSPEC u32 hex_u32_to_u32 (PRIVATE_AS const u32 hex0, PRIVATE_AS const u32 hex1)
{
  u32 v = 0;

  v |= hex_convert_u32 ((hex0 >>  0) & 0xff) << 28;
  v |= hex_convert_u32 ((hex0 >>  8) & 0xff) << 24;
  v |= hex_convert_u32 ((hex0 >> 16) & 0xff) << 20;
  v |= hex_convert_u32 ((hex0 >> 24) & 0xff) << 16;

  v |= hex_convert_u32 ((hex1 >>  0) & 0xff) << 12;
  v |= hex_convert_u32 ((hex1 >>  8) & 0xff) <<  8;
  v |= hex_convert_u32 ((hex1 >> 16) & 0xff) <<  4;
  v |= hex_convert_u32 ((hex1 >> 24) & 0xff) <<  0;

  return (v);
}

KERNEL_FQ void m30905_mxx (KERN_ATTR_VECTOR ())
{
  /**
   * modifier
   */

  const u64 gid = get_global_id (0);

  if (gid >= GID_CNT) return;


  /**
   * base
   */

  const u32 pw_len = pws[gid].pw_len;

  if (pw_len != 64) return;


  // copy password to w

  u32 w[16];

  for (u32 i = 0; i < 16; i++) // pw_len / 4
  {
    w[i] = pws[gid].i[i];
  }

  for (u32 i = 1; i < 16; i++)
  {
    if (is_valid_hex_32 (w[i]) == 0) return;
  }

  secp256k1_t preG; // need to change SECP256K1_TMPS_TYPE above to: PRIVATE_AS

  set_precomputed_basepoint_g (&preG);


  /**
   * loop
   */

  u32 w0l = w[0];

  for (u32 il_pos = 0; il_pos < IL_CNT; il_pos += VECT_SIZE)
  {
    u32x w0r = words_buf_r[il_pos / VECT_SIZE];

    const u32 w0 = w0l | w0r;

    w[0] = w0;

    if (is_valid_hex_32 (w[0]) == 0) continue;


    // convert password from hex to binary

    u32 tmp[16] = { 0 };

    for (u32 i = 0, j = 0; i < 8; i += 1, j += 2)
    {
      tmp[i] = hex_u32_to_u32 (w[j + 0], w[j + 1]);
    }

    u32 prv_key[9];

    prv_key[0] = tmp[7];
    prv_key[1] = tmp[6];
    prv_key[2] = tmp[5];
    prv_key[3] = tmp[4];
    prv_key[4] = tmp[3];
    prv_key[5] = tmp[2];
    prv_key[6] = tmp[1];
    prv_key[7] = tmp[0];


    // convert: pub_key = G * prv_key

    u32 x[8];
    u32 y[8];

    point_mul_xy (x, y, prv_key, &preG);


    // to public key:

    u32 pub_key[16] = { 0 }; // why is re-using the "tmp" variable here slower ?

    const u32 type = 0x02 | (y[0] & 1);

    pub_key[8] =               (x[0] << 24);
    pub_key[7] = (x[0] >> 8) | (x[1] << 24);
    pub_key[6] = (x[1] >> 8) | (x[2] << 24);
    pub_key[5] = (x[2] >> 8) | (x[3] << 24);
    pub_key[4] = (x[3] >> 8) | (x[4] << 24);
    pub_key[3] = (x[4] >> 8) | (x[5] << 24);
    pub_key[2] = (x[5] >> 8) | (x[6] << 24);
    pub_key[1] = (x[6] >> 8) | (x[7] << 24);
    pub_key[0] = (x[7] >> 8) | (type << 24);


    // calculate HASH160 for pub key

    sha256_ctx_t ctx;

    sha256_init   (&ctx);
    sha256_update (&ctx, pub_key, 33); // length of public key: 33
    sha256_final  (&ctx);

    for (u32 i = 0; i < 8; i++) tmp[i] = ctx.h[i];

    // tmp[ 8] = 0; tmp[ 9] = 0; tmp[10] = 0; tmp[11] = 0;
    // tmp[12] = 0; tmp[13] = 0; tmp[14] = 0; tmp[15] = 0;

    for (u32 i = 8; i < 16; i++) tmp[i] = 0;


    // now let's do RIPEMD-160 on the sha256sum

    ripemd160_ctx_t rctx;

    ripemd160_init        (&rctx);
    ripemd160_update_swap (&rctx, tmp, 32);
    ripemd160_final       (&rctx);


    /*
     * 2nd RIPEMD160 (SHA256 ()):
     */

    tmp[0] = (rctx.h[0] << 16) | (         0x1400); // (swapped) OP_0 operation (0x00),
    tmp[1] = (rctx.h[1] << 16) | (rctx.h[0] >> 16); // 0x14 == 20, this indicates the
    tmp[2] = (rctx.h[2] << 16) | (rctx.h[1] >> 16); // data len
    tmp[3] = (rctx.h[3] << 16) | (rctx.h[2] >> 16);
    tmp[4] = (rctx.h[4] << 16) | (rctx.h[3] >> 16);
    tmp[5] =                     (rctx.h[4] >> 16);

    for (u32 i = 6; i < 16; i++) tmp[i] = 0;

    sha256_init        (&ctx);
    sha256_update_swap (&ctx, tmp, 22);
    sha256_final       (&ctx);

    for (u32 i = 0; i <  8; i++) tmp[i] = ctx.h[i];

    ripemd160_init        (&rctx);
    ripemd160_update_swap (&rctx, tmp, 32);
    ripemd160_final       (&rctx);

    const u32 r0 = rctx.h[0];
    const u32 r1 = rctx.h[1];
    const u32 r2 = rctx.h[2];
    const u32 r3 = rctx.h[3];

    COMPARE_M_SCALAR (r0, r1, r2, r3);
  }
}

KERNEL_FQ void m30905_sxx (KERN_ATTR_VECTOR ())
{
  /**
   * modifier
   */

  const u64 gid = get_global_id (0);

  if (gid >= GID_CNT) return;


  /**
   * digest
   */

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

  /**
   * base
   */

  const u32 pw_len = pws[gid].pw_len;

  if (pw_len != 64) return;


  // copy password to w

  u32 w[16];

  for (u32 i = 0; i < 16; i++) // pw_len / 4
  {
    w[i] = pws[gid].i[i];
  }

  for (u32 i = 1; i < 16; i++)
  {
    if (is_valid_hex_32 (w[i]) == 0) return;
  }

  secp256k1_t preG; // need to change SECP256K1_TMPS_TYPE above to: PRIVATE_AS

  set_precomputed_basepoint_g (&preG);


  /**
   * loop
   */

  u32 w0l = w[0];

  for (u32 il_pos = 0; il_pos < IL_CNT; il_pos += VECT_SIZE)
  {
    u32x w0r = words_buf_r[il_pos / VECT_SIZE];

    const u32 w0 = w0l | w0r;

    w[0] = w0;

    if (is_valid_hex_32 (w[0]) == 0) continue;


    // convert password from hex to binary

    u32 tmp[16] = { 0 };

    for (u32 i = 0, j = 0; i < 8; i += 1, j += 2)
    {
      tmp[i] = hex_u32_to_u32 (w[j + 0], w[j + 1]);
    }

    u32 prv_key[9];

    prv_key[0] = tmp[7];
    prv_key[1] = tmp[6];
    prv_key[2] = tmp[5];
    prv_key[3] = tmp[4];
    prv_key[4] = tmp[3];
    prv_key[5] = tmp[2];
    prv_key[6] = tmp[1];
    prv_key[7] = tmp[0];


    // convert: pub_key = G * prv_key

    u32 x[8];
    u32 y[8];

    point_mul_xy (x, y, prv_key, &preG);


    // to public key:

    u32 pub_key[16] = { 0 }; // why is re-using the "tmp" variable here slower ?

    const u32 type = 0x02 | (y[0] & 1);

    pub_key[8] =               (x[0] << 24);
    pub_key[7] = (x[0] >> 8) | (x[1] << 24);
    pub_key[6] = (x[1] >> 8) | (x[2] << 24);
    pub_key[5] = (x[2] >> 8) | (x[3] << 24);
    pub_key[4] = (x[3] >> 8) | (x[4] << 24);
    pub_key[3] = (x[4] >> 8) | (x[5] << 24);
    pub_key[2] = (x[5] >> 8) | (x[6] << 24);
    pub_key[1] = (x[6] >> 8) | (x[7] << 24);
    pub_key[0] = (x[7] >> 8) | (type << 24);


    // calculate HASH160 for pub key

    sha256_ctx_t ctx;

    sha256_init   (&ctx);
    sha256_update (&ctx, pub_key, 33); // length of public key: 33
    sha256_final  (&ctx);

    for (u32 i = 0; i < 8; i++) tmp[i] = ctx.h[i];

    // tmp[ 8] = 0; tmp[ 9] = 0; tmp[10] = 0; tmp[11] = 0;
    // tmp[12] = 0; tmp[13] = 0; tmp[14] = 0; tmp[15] = 0;

    for (u32 i = 8; i < 16; i++) tmp[i] = 0;


    // now let's do RIPEMD-160 on the sha256sum

    ripemd160_ctx_t rctx;

    ripemd160_init        (&rctx);
    ripemd160_update_swap (&rctx, tmp, 32);
    ripemd160_final       (&rctx);


    /*
     * 2nd RIPEMD160 (SHA256 ()):
     */

    tmp[0] = (rctx.h[0] << 16) | (         0x1400); // (swapped) OP_0 operation (0x00),
    tmp[1] = (rctx.h[1] << 16) | (rctx.h[0] >> 16); // 0x14 == 20, this indicates the
    tmp[2] = (rctx.h[2] << 16) | (rctx.h[1] >> 16); // data len
    tmp[3] = (rctx.h[3] << 16) | (rctx.h[2] >> 16);
    tmp[4] = (rctx.h[4] << 16) | (rctx.h[3] >> 16);
    tmp[5] =                     (rctx.h[4] >> 16);

    for (u32 i = 6; i < 16; i++) tmp[i] = 0;

    sha256_init        (&ctx);
    sha256_update_swap (&ctx, tmp, 22);
    sha256_final       (&ctx);

    for (u32 i = 0; i <  8; i++) tmp[i] = ctx.h[i];

    ripemd160_init        (&rctx);
    ripemd160_update_swap (&rctx, tmp, 32);
    ripemd160_final       (&rctx);

    const u32 r0 = rctx.h[0];
    const u32 r1 = rctx.h[1];
    const u32 r2 = rctx.h[2];
    const u32 r3 = rctx.h[3];

    COMPARE_S_SCALAR (r0, r1, r2, r3);
  }
}