/**
 * 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_rp_optimized.h"
#include "inc_rp_optimized.cl"
#include "inc_simd.cl"
#include "inc_hash_whirlpool.cl"

DECLSPEC void whirlpool_transform_transport_vector (const u32x *w, u32x *digest, __local u32 (*s_Ch)[256], __local u32 (*s_Cl)[256])
{
  whirlpool_transform_vector (w + 0, w + 4, w + 8, w + 12, digest, s_Ch, s_Cl);
}

__kernel void m06100_m04 (KERN_ATTR_RULES ())
{
  /**
   * 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 s_Ch[8][256];
  __local u32 s_Cl[8][256];

  for (u32 i = lid; i < 256; i += lsz)
  {
    s_Ch[0][i] = Ch[0][i];
    s_Ch[1][i] = Ch[1][i];
    s_Ch[2][i] = Ch[2][i];
    s_Ch[3][i] = Ch[3][i];
    s_Ch[4][i] = Ch[4][i];
    s_Ch[5][i] = Ch[5][i];
    s_Ch[6][i] = Ch[6][i];
    s_Ch[7][i] = Ch[7][i];

    s_Cl[0][i] = Cl[0][i];
    s_Cl[1][i] = Cl[1][i];
    s_Cl[2][i] = Cl[2][i];
    s_Cl[3][i] = Cl[3][i];
    s_Cl[4][i] = Cl[4][i];
    s_Cl[5][i] = Cl[5][i];
    s_Cl[6][i] = Cl[6][i];
    s_Cl[7][i] = Cl[7][i];
  }

  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_len = pws[gid].pw_len & 63;

  /**
   * 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);

    append_0x80_2x4_VV (w0, w1, out_len);

    u32x w[16];

    w[ 0] = swap32 (w0[0]);
    w[ 1] = swap32 (w0[1]);
    w[ 2] = swap32 (w0[2]);
    w[ 3] = swap32 (w0[3]);
    w[ 4] = swap32 (w1[0]);
    w[ 5] = swap32 (w1[1]);
    w[ 6] = swap32 (w1[2]);
    w[ 7] = swap32 (w1[3]);
    w[ 8] = 0;
    w[ 9] = 0;
    w[10] = 0;
    w[11] = 0;
    w[12] = 0;
    w[13] = 0;
    w[14] = 0;
    w[15] = out_len * 8;

    /**
     * Whirlpool
     */

    u32x dgst[16];

    dgst[ 0] = 0;
    dgst[ 1] = 0;
    dgst[ 2] = 0;
    dgst[ 3] = 0;
    dgst[ 4] = 0;
    dgst[ 5] = 0;
    dgst[ 6] = 0;
    dgst[ 7] = 0;
    dgst[ 8] = 0;
    dgst[ 9] = 0;
    dgst[10] = 0;
    dgst[11] = 0;
    dgst[12] = 0;
    dgst[13] = 0;
    dgst[14] = 0;
    dgst[15] = 0;

    whirlpool_transform_transport_vector (w, dgst, s_Ch, s_Cl);

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

__kernel void m06100_m08 (KERN_ATTR_RULES ())
{
}

__kernel void m06100_m16 (KERN_ATTR_RULES ())
{
}

__kernel void m06100_s04 (KERN_ATTR_RULES ())
{
  /**
   * 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 s_Ch[8][256];
  __local u32 s_Cl[8][256];

  for (u32 i = lid; i < 256; i += lsz)
  {
    s_Ch[0][i] = Ch[0][i];
    s_Ch[1][i] = Ch[1][i];
    s_Ch[2][i] = Ch[2][i];
    s_Ch[3][i] = Ch[3][i];
    s_Ch[4][i] = Ch[4][i];
    s_Ch[5][i] = Ch[5][i];
    s_Ch[6][i] = Ch[6][i];
    s_Ch[7][i] = Ch[7][i];

    s_Cl[0][i] = Cl[0][i];
    s_Cl[1][i] = Cl[1][i];
    s_Cl[2][i] = Cl[2][i];
    s_Cl[3][i] = Cl[3][i];
    s_Cl[4][i] = Cl[4][i];
    s_Cl[5][i] = Cl[5][i];
    s_Cl[6][i] = Cl[6][i];
    s_Cl[7][i] = Cl[7][i];
  }

  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_len = pws[gid].pw_len & 63;

  /**
   * 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);

    append_0x80_2x4_VV (w0, w1, out_len);

    u32x w[16];

    w[ 0] = swap32 (w0[0]);
    w[ 1] = swap32 (w0[1]);
    w[ 2] = swap32 (w0[2]);
    w[ 3] = swap32 (w0[3]);
    w[ 4] = swap32 (w1[0]);
    w[ 5] = swap32 (w1[1]);
    w[ 6] = swap32 (w1[2]);
    w[ 7] = swap32 (w1[3]);
    w[ 8] = 0;
    w[ 9] = 0;
    w[10] = 0;
    w[11] = 0;
    w[12] = 0;
    w[13] = 0;
    w[14] = 0;
    w[15] = out_len * 8;

    /**
     * Whirlpool
     */

    u32x dgst[16];

    dgst[ 0] = 0;
    dgst[ 1] = 0;
    dgst[ 2] = 0;
    dgst[ 3] = 0;
    dgst[ 4] = 0;
    dgst[ 5] = 0;
    dgst[ 6] = 0;
    dgst[ 7] = 0;
    dgst[ 8] = 0;
    dgst[ 9] = 0;
    dgst[10] = 0;
    dgst[11] = 0;
    dgst[12] = 0;
    dgst[13] = 0;
    dgst[14] = 0;
    dgst[15] = 0;

    whirlpool_transform_transport_vector (w, dgst, s_Ch, s_Cl);

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

__kernel void m06100_s08 (KERN_ATTR_RULES ())
{
}

__kernel void m06100_s16 (KERN_ATTR_RULES ())
{
}