Add optimized -m 24800 kernels

OpenCL/m24800_a0-optimized.cl

@@ -0,0 +1,362 @@
+/**
+ * Author......: See docs/credits.txt
+ * License.....: MIT
+ */
+
+#define NEW_SIMD_CODE
+
+#ifdef KERNEL_STATIC
+#include "inc_vendor.h"
+#include "inc_types.h"
+#include "inc_platform.cl"
+#include "inc_common.cl"
+#include "inc_rp_optimized.h"
+#include "inc_rp_optimized.cl"
+#include "inc_simd.cl"
+#include "inc_hash_sha1.cl"
+#endif
+
+DECLSPEC void hmac_sha1_pad (u32x *w0, u32x *w1, u32x *w2, u32x *w3, u32x *ipad, u32x *opad)
+{
+  w0[0] = w0[0] ^ 0x36363636;
+  w0[1] = w0[1] ^ 0x36363636;
+  w0[2] = w0[2] ^ 0x36363636;
+  w0[3] = w0[3] ^ 0x36363636;
+  w1[0] = w1[0] ^ 0x36363636;
+  w1[1] = w1[1] ^ 0x36363636;
+  w1[2] = w1[2] ^ 0x36363636;
+  w1[3] = w1[3] ^ 0x36363636;
+  w2[0] = w2[0] ^ 0x36363636;
+  w2[1] = w2[1] ^ 0x36363636;
+  w2[2] = w2[2] ^ 0x36363636;
+  w2[3] = w2[3] ^ 0x36363636;
+  w3[0] = w3[0] ^ 0x36363636;
+  w3[1] = w3[1] ^ 0x36363636;
+  w3[2] = w3[2] ^ 0x36363636;
+  w3[3] = w3[3] ^ 0x36363636;
+
+  ipad[0] = SHA1M_A;
+  ipad[1] = SHA1M_B;
+  ipad[2] = SHA1M_C;
+  ipad[3] = SHA1M_D;
+  ipad[4] = SHA1M_E;
+
+  sha1_transform_vector (w0, w1, w2, w3, ipad);
+
+  w0[0] = w0[0] ^ 0x6a6a6a6a;
+  w0[1] = w0[1] ^ 0x6a6a6a6a;
+  w0[2] = w0[2] ^ 0x6a6a6a6a;
+  w0[3] = w0[3] ^ 0x6a6a6a6a;
+  w1[0] = w1[0] ^ 0x6a6a6a6a;
+  w1[1] = w1[1] ^ 0x6a6a6a6a;
+  w1[2] = w1[2] ^ 0x6a6a6a6a;
+  w1[3] = w1[3] ^ 0x6a6a6a6a;
+  w2[0] = w2[0] ^ 0x6a6a6a6a;
+  w2[1] = w2[1] ^ 0x6a6a6a6a;
+  w2[2] = w2[2] ^ 0x6a6a6a6a;
+  w2[3] = w2[3] ^ 0x6a6a6a6a;
+  w3[0] = w3[0] ^ 0x6a6a6a6a;
+  w3[1] = w3[1] ^ 0x6a6a6a6a;
+  w3[2] = w3[2] ^ 0x6a6a6a6a;
+  w3[3] = w3[3] ^ 0x6a6a6a6a;
+
+  opad[0] = SHA1M_A;
+  opad[1] = SHA1M_B;
+  opad[2] = SHA1M_C;
+  opad[3] = SHA1M_D;
+  opad[4] = SHA1M_E;
+
+  sha1_transform_vector (w0, w1, w2, w3, opad);
+}
+
+DECLSPEC void hmac_sha1_run (u32x *w0, u32x *w1, u32x *w2, u32x *w3, u32x *ipad, u32x *opad, u32x *digest)
+{
+  digest[0] = ipad[0];
+  digest[1] = ipad[1];
+  digest[2] = ipad[2];
+  digest[3] = ipad[3];
+  digest[4] = ipad[4];
+
+  sha1_transform_vector (w0, w1, w2, w3, digest);
+
+  w0[0] = digest[0];
+  w0[1] = digest[1];
+  w0[2] = digest[2];
+  w0[3] = digest[3];
+  w1[0] = digest[4];
+  w1[1] = 0x80000000;
+  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] = 0;
+  w3[3] = (64 + 20) * 8;
+
+  digest[0] = opad[0];
+  digest[1] = opad[1];
+  digest[2] = opad[2];
+  digest[3] = opad[3];
+  digest[4] = opad[4];
+
+  sha1_transform_vector (w0, w1, w2, w3, digest);
+}
+
+KERNEL_FQ void m24800_m04 (KERN_ATTR_RULES ())
+{
+  /**
+   * 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 & 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_optimized (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
+
+    const u32x out_len2 = out_len * 2;
+
+    w0[0] = hc_swap32 (w0[0]);
+    w0[1] = hc_swap32 (w0[1]);
+    w0[2] = hc_swap32 (w0[2]);
+    w0[3] = hc_swap32 (w0[3]);
+    w1[0] = hc_swap32 (w1[0]);
+    w1[1] = hc_swap32 (w1[1]);
+    w1[2] = hc_swap32 (w1[2]);
+    w1[3] = hc_swap32 (w1[3]);
+
+    make_utf16beN (w1, w2, w3);
+    make_utf16beN (w0, w0, w1);
+
+    u32x x0_t[4];
+    u32x x1_t[4];
+    u32x x2_t[4];
+    u32x x3_t[4];
+
+    x0_t[0] = w0[0];
+    x0_t[1] = w0[1];
+    x0_t[2] = w0[2];
+    x0_t[3] = w0[3];
+    x1_t[0] = w1[0];
+    x1_t[1] = w1[1];
+    x1_t[2] = w1[2];
+    x1_t[3] = w1[3];
+    x2_t[0] = w2[0];
+    x2_t[1] = w2[1];
+    x2_t[2] = w2[2];
+    x2_t[3] = w2[3];
+    x3_t[0] = w3[0];
+    x3_t[1] = w3[1];
+    x3_t[2] = w3[2];
+    x3_t[3] = w3[3];
+
+    u32x ipad[5];
+    u32x opad[5];
+
+    hmac_sha1_pad (x0_t, x1_t, x2_t, x3_t, ipad, opad);
+
+    x0_t[0] = w0[0];
+    x0_t[1] = w0[1];
+    x0_t[2] = w0[2];
+    x0_t[3] = w0[3];
+    x1_t[0] = w1[0];
+    x1_t[1] = w1[1];
+    x1_t[2] = w1[2];
+    x1_t[3] = w1[3];
+    x2_t[0] = w2[0];
+    x2_t[1] = w2[1];
+    x2_t[2] = w2[2];
+    x2_t[3] = w2[3];
+    x3_t[0] = w3[0];
+    x3_t[1] = w3[1];
+    x3_t[2] = w3[2];
+    x3_t[3] = w3[3];
+
+    append_0x80_4x4_VV (x0_t, x1_t, x2_t, x3_t, out_len2 ^ 3);
+
+    x3_t[2] = 0;
+    x3_t[3] = (64 + out_len2) * 8;
+
+    u32x digest[5];
+
+    hmac_sha1_run (x0_t, x1_t, x2_t, x3_t, ipad, opad, digest);
+
+    COMPARE_M_SIMD (digest[3], digest[4], digest[2], digest[1]);
+  }
+}
+
+KERNEL_FQ void m24800_m08 (KERN_ATTR_RULES ())
+{
+}
+
+KERNEL_FQ void m24800_m16 (KERN_ATTR_RULES ())
+{
+}
+
+KERNEL_FQ void m24800_s04 (KERN_ATTR_RULES ())
+{
+  /**
+   * 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 & 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_optimized (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
+
+    const u32x out_len2 = out_len * 2;
+
+    w0[0] = hc_swap32 (w0[0]);
+    w0[1] = hc_swap32 (w0[1]);
+    w0[2] = hc_swap32 (w0[2]);
+    w0[3] = hc_swap32 (w0[3]);
+    w1[0] = hc_swap32 (w1[0]);
+    w1[1] = hc_swap32 (w1[1]);
+    w1[2] = hc_swap32 (w1[2]);
+    w1[3] = hc_swap32 (w1[3]);
+
+    make_utf16beN (w1, w2, w3);
+    make_utf16beN (w0, w0, w1);
+
+    u32x x0_t[4];
+    u32x x1_t[4];
+    u32x x2_t[4];
+    u32x x3_t[4];
+
+    x0_t[0] = w0[0];
+    x0_t[1] = w0[1];
+    x0_t[2] = w0[2];
+    x0_t[3] = w0[3];
+    x1_t[0] = w1[0];
+    x1_t[1] = w1[1];
+    x1_t[2] = w1[2];
+    x1_t[3] = w1[3];
+    x2_t[0] = w2[0];
+    x2_t[1] = w2[1];
+    x2_t[2] = w2[2];
+    x2_t[3] = w2[3];
+    x3_t[0] = w3[0];
+    x3_t[1] = w3[1];
+    x3_t[2] = w3[2];
+    x3_t[3] = w3[3];
+
+    u32x ipad[5];
+    u32x opad[5];
+
+    hmac_sha1_pad (x0_t, x1_t, x2_t, x3_t, ipad, opad);
+
+    x0_t[0] = w0[0];
+    x0_t[1] = w0[1];
+    x0_t[2] = w0[2];
+    x0_t[3] = w0[3];
+    x1_t[0] = w1[0];
+    x1_t[1] = w1[1];
+    x1_t[2] = w1[2];
+    x1_t[3] = w1[3];
+    x2_t[0] = w2[0];
+    x2_t[1] = w2[1];
+    x2_t[2] = w2[2];
+    x2_t[3] = w2[3];
+    x3_t[0] = w3[0];
+    x3_t[1] = w3[1];
+    x3_t[2] = w3[2];
+    x3_t[3] = w3[3];
+
+    append_0x80_4x4_VV (x0_t, x1_t, x2_t, x3_t, out_len2 ^ 3);
+
+    x3_t[2] = 0;
+    x3_t[3] = (64 + out_len2) * 8;
+
+    u32x digest[5];
+
+    hmac_sha1_run (x0_t, x1_t, x2_t, x3_t, ipad, opad, digest);
+
+    COMPARE_S_SIMD (digest[3], digest[4], digest[2], digest[1]);
+  }
+}
+
+KERNEL_FQ void m24800_s08 (KERN_ATTR_RULES ())
+{
+}
+
+KERNEL_FQ void m24800_s16 (KERN_ATTR_RULES ())
+{
+}

OpenCL/m24800_a0-pure.cl

@@ -33,9 +33,6 @@ KERNEL_FQ void m24800_mxx (KERN_ATTR_RULES ())
 
   COPY_PW (pws[gid]);
 
-  u32 t[128] = { 0 };
-
-
   /**
    * loop
    */
@@ -46,15 +43,18 @@ KERNEL_FQ void m24800_mxx (KERN_ATTR_RULES ())
 
     tmp.pw_len = apply_rules (rules_buf[il_pos].cmds, tmp.i, tmp.pw_len);
 
-    // we need to swap the endian before we convert to unicode.
+    // swap endian
     for (u32 i = 0, idx = 0; i < tmp.pw_len; i += 4, idx += 1)
     {
-      tmp.i[idx] = hc_swap32(tmp.i[idx]);
+      tmp.i[idx] = hc_swap32 (tmp.i[idx]);
     }
 
+    u32 t[128] = { 0 };
+
     // make it unicode.
-    for(u32 i = 0, idx = 0; idx < tmp.pw_len; i += 2, idx += 1){
+    for (u32 i = 0, idx = 0; idx < tmp.pw_len; i += 2, idx += 1)
-        make_utf16beN(&tmp.i[idx], &t[i], &t[i+1]);
+    {
+      make_utf16beN (&tmp.i[idx], &t[i], &t[i+1]);
     }
 
     // hash time
@@ -104,8 +104,6 @@ KERNEL_FQ void m24800_sxx (KERN_ATTR_RULES ())
 
   COPY_PW (pws[gid]);
 
-  u32 t[128] = { 0 };
-
   /**
    * loop
    */
@@ -119,20 +117,23 @@ KERNEL_FQ void m24800_sxx (KERN_ATTR_RULES ())
     // swap endian
     for (u32 i = 0, idx = 0; i < tmp.pw_len; i += 4, idx += 1)
     {
-      tmp.i[idx] = hc_swap32(tmp.i[idx]);
+      tmp.i[idx] = hc_swap32 (tmp.i[idx]);
     }
 
+    u32 t[128] = { 0 };
+
     // make it unicode.
-    for(u32 i = 0, idx = 0; idx < tmp.pw_len; i += 2, idx += 1){
+    for (u32 i = 0, idx = 0; idx < tmp.pw_len; i += 2, idx += 1)
-        make_utf16beN(&tmp.i[idx], &t[i], &t[i+1]);
+    {
+      make_utf16beN (&tmp.i[idx], &t[i], &t[i+1]);
     }
 
     // hash time
     sha1_hmac_ctx_t ctx;
 
-    sha1_hmac_init (&ctx, t, tmp.pw_len*2);
+    sha1_hmac_init (&ctx, t, tmp.pw_len * 2);
 
-    sha1_hmac_update (&ctx, t, tmp.pw_len*2);
+    sha1_hmac_update (&ctx, t, tmp.pw_len * 2);
 
     sha1_hmac_final (&ctx);
 

OpenCL/m24800_a1-optimized.cl

@@ -0,0 +1,464 @@
+/**
+ * Author......: See docs/credits.txt
+ * License.....: MIT
+ */
+
+#define NEW_SIMD_CODE
+
+#ifdef KERNEL_STATIC
+#include "inc_vendor.h"
+#include "inc_types.h"
+#include "inc_platform.cl"
+#include "inc_common.cl"
+#include "inc_simd.cl"
+#include "inc_hash_sha1.cl"
+#endif
+
+DECLSPEC void hmac_sha1_pad (u32x *w0, u32x *w1, u32x *w2, u32x *w3, u32x *ipad, u32x *opad)
+{
+  w0[0] = w0[0] ^ 0x36363636;
+  w0[1] = w0[1] ^ 0x36363636;
+  w0[2] = w0[2] ^ 0x36363636;
+  w0[3] = w0[3] ^ 0x36363636;
+  w1[0] = w1[0] ^ 0x36363636;
+  w1[1] = w1[1] ^ 0x36363636;
+  w1[2] = w1[2] ^ 0x36363636;
+  w1[3] = w1[3] ^ 0x36363636;
+  w2[0] = w2[0] ^ 0x36363636;
+  w2[1] = w2[1] ^ 0x36363636;
+  w2[2] = w2[2] ^ 0x36363636;
+  w2[3] = w2[3] ^ 0x36363636;
+  w3[0] = w3[0] ^ 0x36363636;
+  w3[1] = w3[1] ^ 0x36363636;
+  w3[2] = w3[2] ^ 0x36363636;
+  w3[3] = w3[3] ^ 0x36363636;
+
+  ipad[0] = SHA1M_A;
+  ipad[1] = SHA1M_B;
+  ipad[2] = SHA1M_C;
+  ipad[3] = SHA1M_D;
+  ipad[4] = SHA1M_E;
+
+  sha1_transform_vector (w0, w1, w2, w3, ipad);
+
+  w0[0] = w0[0] ^ 0x6a6a6a6a;
+  w0[1] = w0[1] ^ 0x6a6a6a6a;
+  w0[2] = w0[2] ^ 0x6a6a6a6a;
+  w0[3] = w0[3] ^ 0x6a6a6a6a;
+  w1[0] = w1[0] ^ 0x6a6a6a6a;
+  w1[1] = w1[1] ^ 0x6a6a6a6a;
+  w1[2] = w1[2] ^ 0x6a6a6a6a;
+  w1[3] = w1[3] ^ 0x6a6a6a6a;
+  w2[0] = w2[0] ^ 0x6a6a6a6a;
+  w2[1] = w2[1] ^ 0x6a6a6a6a;
+  w2[2] = w2[2] ^ 0x6a6a6a6a;
+  w2[3] = w2[3] ^ 0x6a6a6a6a;
+  w3[0] = w3[0] ^ 0x6a6a6a6a;
+  w3[1] = w3[1] ^ 0x6a6a6a6a;
+  w3[2] = w3[2] ^ 0x6a6a6a6a;
+  w3[3] = w3[3] ^ 0x6a6a6a6a;
+
+  opad[0] = SHA1M_A;
+  opad[1] = SHA1M_B;
+  opad[2] = SHA1M_C;
+  opad[3] = SHA1M_D;
+  opad[4] = SHA1M_E;
+
+  sha1_transform_vector (w0, w1, w2, w3, opad);
+}
+
+DECLSPEC void hmac_sha1_run (u32x *w0, u32x *w1, u32x *w2, u32x *w3, u32x *ipad, u32x *opad, u32x *digest)
+{
+  digest[0] = ipad[0];
+  digest[1] = ipad[1];
+  digest[2] = ipad[2];
+  digest[3] = ipad[3];
+  digest[4] = ipad[4];
+
+  sha1_transform_vector (w0, w1, w2, w3, digest);
+
+  w0[0] = digest[0];
+  w0[1] = digest[1];
+  w0[2] = digest[2];
+  w0[3] = digest[3];
+  w1[0] = digest[4];
+  w1[1] = 0x80000000;
+  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] = 0;
+  w3[3] = (64 + 20) * 8;
+
+  digest[0] = opad[0];
+  digest[1] = opad[1];
+  digest[2] = opad[2];
+  digest[3] = opad[3];
+  digest[4] = opad[4];
+
+  sha1_transform_vector (w0, w1, w2, w3, digest);
+}
+
+KERNEL_FQ void m24800_m04 (KERN_ATTR_BASIC ())
+{
+  /**
+   * 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_l_len = pws[gid].pw_len & 63;
+
+  /**
+   * loop
+   */
+
+  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
+  {
+    const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos) & 63;
+
+    const u32x pw_len = (pw_l_len + pw_r_len) & 63;
+
+    const u32x pw_len2 = pw_len * 2;
+
+    /**
+     * concat password candidate
+     */
+
+    u32x wordl0[4] = { 0 };
+    u32x wordl1[4] = { 0 };
+    u32x wordl2[4] = { 0 };
+    u32x wordl3[4] = { 0 };
+
+    wordl0[0] = pw_buf0[0];
+    wordl0[1] = pw_buf0[1];
+    wordl0[2] = pw_buf0[2];
+    wordl0[3] = pw_buf0[3];
+    wordl1[0] = pw_buf1[0];
+    wordl1[1] = pw_buf1[1];
+    wordl1[2] = pw_buf1[2];
+    wordl1[3] = pw_buf1[3];
+
+    u32x wordr0[4] = { 0 };
+    u32x wordr1[4] = { 0 };
+    u32x wordr2[4] = { 0 };
+    u32x wordr3[4] = { 0 };
+
+    wordr0[0] = ix_create_combt (combs_buf, il_pos, 0);
+    wordr0[1] = ix_create_combt (combs_buf, il_pos, 1);
+    wordr0[2] = ix_create_combt (combs_buf, il_pos, 2);
+    wordr0[3] = ix_create_combt (combs_buf, il_pos, 3);
+    wordr1[0] = ix_create_combt (combs_buf, il_pos, 4);
+    wordr1[1] = ix_create_combt (combs_buf, il_pos, 5);
+    wordr1[2] = ix_create_combt (combs_buf, il_pos, 6);
+    wordr1[3] = ix_create_combt (combs_buf, il_pos, 7);
+
+    if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
+    {
+      switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len);
+    }
+    else
+    {
+      switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len);
+    }
+
+    u32x w0[4];
+    u32x w1[4];
+    u32x w2[4];
+    u32x w3[4];
+
+    w0[0] = wordl0[0] | wordr0[0];
+    w0[1] = wordl0[1] | wordr0[1];
+    w0[2] = wordl0[2] | wordr0[2];
+    w0[3] = wordl0[3] | wordr0[3];
+    w1[0] = wordl1[0] | wordr1[0];
+    w1[1] = wordl1[1] | wordr1[1];
+    w1[2] = wordl1[2] | wordr1[2];
+    w1[3] = wordl1[3] | wordr1[3];
+
+    w0[0] = hc_swap32 (w0[0]);
+    w0[1] = hc_swap32 (w0[1]);
+    w0[2] = hc_swap32 (w0[2]);
+    w0[3] = hc_swap32 (w0[3]);
+    w1[0] = hc_swap32 (w1[0]);
+    w1[1] = hc_swap32 (w1[1]);
+    w1[2] = hc_swap32 (w1[2]);
+    w1[3] = hc_swap32 (w1[3]);
+
+    make_utf16beN (w1, w2, w3);
+    make_utf16beN (w0, w0, w1);
+
+    u32x x0_t[4];
+    u32x x1_t[4];
+    u32x x2_t[4];
+    u32x x3_t[4];
+
+    x0_t[0] = w0[0];
+    x0_t[1] = w0[1];
+    x0_t[2] = w0[2];
+    x0_t[3] = w0[3];
+    x1_t[0] = w1[0];
+    x1_t[1] = w1[1];
+    x1_t[2] = w1[2];
+    x1_t[3] = w1[3];
+    x2_t[0] = w2[0];
+    x2_t[1] = w2[1];
+    x2_t[2] = w2[2];
+    x2_t[3] = w2[3];
+    x3_t[0] = w3[0];
+    x3_t[1] = w3[1];
+    x3_t[2] = w3[2];
+    x3_t[3] = w3[3];
+
+    u32x ipad[5];
+    u32x opad[5];
+
+    hmac_sha1_pad (x0_t, x1_t, x2_t, x3_t, ipad, opad);
+
+    x0_t[0] = w0[0];
+    x0_t[1] = w0[1];
+    x0_t[2] = w0[2];
+    x0_t[3] = w0[3];
+    x1_t[0] = w1[0];
+    x1_t[1] = w1[1];
+    x1_t[2] = w1[2];
+    x1_t[3] = w1[3];
+    x2_t[0] = w2[0];
+    x2_t[1] = w2[1];
+    x2_t[2] = w2[2];
+    x2_t[3] = w2[3];
+    x3_t[0] = w3[0];
+    x3_t[1] = w3[1];
+    x3_t[2] = w3[2];
+    x3_t[3] = w3[3];
+
+    append_0x80_4x4_VV (x0_t, x1_t, x2_t, x3_t, pw_len2 ^ 3);
+
+    x3_t[2] = 0;
+    x3_t[3] = (64 + pw_len2) * 8;
+
+    u32x digest[5];
+
+    hmac_sha1_run (x0_t, x1_t, x2_t, x3_t, ipad, opad, digest);
+
+    COMPARE_M_SIMD (digest[3], digest[4], digest[2], digest[1]);
+  }
+}
+
+KERNEL_FQ void m24800_m08 (KERN_ATTR_BASIC ())
+{
+}
+
+KERNEL_FQ void m24800_m16 (KERN_ATTR_BASIC ())
+{
+}
+
+KERNEL_FQ void m24800_s04 (KERN_ATTR_BASIC ())
+{
+  /**
+   * 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_l_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)
+  {
+    const u32x pw_r_len = pwlenx_create_combt (combs_buf, il_pos) & 63;
+
+    const u32x pw_len = (pw_l_len + pw_r_len) & 63;
+
+    const u32x pw_len2 = pw_len * 2;
+
+    /**
+     * concat password candidate
+     */
+
+    u32x wordl0[4] = { 0 };
+    u32x wordl1[4] = { 0 };
+    u32x wordl2[4] = { 0 };
+    u32x wordl3[4] = { 0 };
+
+    wordl0[0] = pw_buf0[0];
+    wordl0[1] = pw_buf0[1];
+    wordl0[2] = pw_buf0[2];
+    wordl0[3] = pw_buf0[3];
+    wordl1[0] = pw_buf1[0];
+    wordl1[1] = pw_buf1[1];
+    wordl1[2] = pw_buf1[2];
+    wordl1[3] = pw_buf1[3];
+
+    u32x wordr0[4] = { 0 };
+    u32x wordr1[4] = { 0 };
+    u32x wordr2[4] = { 0 };
+    u32x wordr3[4] = { 0 };
+
+    wordr0[0] = ix_create_combt (combs_buf, il_pos, 0);
+    wordr0[1] = ix_create_combt (combs_buf, il_pos, 1);
+    wordr0[2] = ix_create_combt (combs_buf, il_pos, 2);
+    wordr0[3] = ix_create_combt (combs_buf, il_pos, 3);
+    wordr1[0] = ix_create_combt (combs_buf, il_pos, 4);
+    wordr1[1] = ix_create_combt (combs_buf, il_pos, 5);
+    wordr1[2] = ix_create_combt (combs_buf, il_pos, 6);
+    wordr1[3] = ix_create_combt (combs_buf, il_pos, 7);
+
+    if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
+    {
+      switch_buffer_by_offset_le_VV (wordr0, wordr1, wordr2, wordr3, pw_l_len);
+    }
+    else
+    {
+      switch_buffer_by_offset_le_VV (wordl0, wordl1, wordl2, wordl3, pw_r_len);
+    }
+
+    u32x w0[4];
+    u32x w1[4];
+    u32x w2[4];
+    u32x w3[4];
+
+    w0[0] = wordl0[0] | wordr0[0];
+    w0[1] = wordl0[1] | wordr0[1];
+    w0[2] = wordl0[2] | wordr0[2];
+    w0[3] = wordl0[3] | wordr0[3];
+    w1[0] = wordl1[0] | wordr1[0];
+    w1[1] = wordl1[1] | wordr1[1];
+    w1[2] = wordl1[2] | wordr1[2];
+    w1[3] = wordl1[3] | wordr1[3];
+
+    w0[0] = hc_swap32 (w0[0]);
+    w0[1] = hc_swap32 (w0[1]);
+    w0[2] = hc_swap32 (w0[2]);
+    w0[3] = hc_swap32 (w0[3]);
+    w1[0] = hc_swap32 (w1[0]);
+    w1[1] = hc_swap32 (w1[1]);
+    w1[2] = hc_swap32 (w1[2]);
+    w1[3] = hc_swap32 (w1[3]);
+
+    make_utf16beN (w1, w2, w3);
+    make_utf16beN (w0, w0, w1);
+
+    u32x x0_t[4];
+    u32x x1_t[4];
+    u32x x2_t[4];
+    u32x x3_t[4];
+
+    x0_t[0] = w0[0];
+    x0_t[1] = w0[1];
+    x0_t[2] = w0[2];
+    x0_t[3] = w0[3];
+    x1_t[0] = w1[0];
+    x1_t[1] = w1[1];
+    x1_t[2] = w1[2];
+    x1_t[3] = w1[3];
+    x2_t[0] = w2[0];
+    x2_t[1] = w2[1];
+    x2_t[2] = w2[2];
+    x2_t[3] = w2[3];
+    x3_t[0] = w3[0];
+    x3_t[1] = w3[1];
+    x3_t[2] = w3[2];
+    x3_t[3] = w3[3];
+
+    u32x ipad[5];
+    u32x opad[5];
+
+    hmac_sha1_pad (x0_t, x1_t, x2_t, x3_t, ipad, opad);
+
+    x0_t[0] = w0[0];
+    x0_t[1] = w0[1];
+    x0_t[2] = w0[2];
+    x0_t[3] = w0[3];
+    x1_t[0] = w1[0];
+    x1_t[1] = w1[1];
+    x1_t[2] = w1[2];
+    x1_t[3] = w1[3];
+    x2_t[0] = w2[0];
+    x2_t[1] = w2[1];
+    x2_t[2] = w2[2];
+    x2_t[3] = w2[3];
+    x3_t[0] = w3[0];
+    x3_t[1] = w3[1];
+    x3_t[2] = w3[2];
+    x3_t[3] = w3[3];
+
+    append_0x80_4x4_VV (x0_t, x1_t, x2_t, x3_t, pw_len2 ^ 3);
+
+    x3_t[2] = 0;
+    x3_t[3] = (64 + pw_len2) * 8;
+
+    u32x digest[5];
+
+    hmac_sha1_run (x0_t, x1_t, x2_t, x3_t, ipad, opad, digest);
+
+    COMPARE_S_SIMD (digest[3], digest[4], digest[2], digest[1]);
+  }
+}
+
+KERNEL_FQ void m24800_s08 (KERN_ATTR_BASIC ())
+{
+}
+
+KERNEL_FQ void m24800_s16 (KERN_ATTR_BASIC ())
+{
+}

OpenCL/m24800_a1-pure.cl

@@ -38,8 +38,6 @@ KERNEL_FQ void m24800_mxx (KERN_ATTR_BASIC ())
     w[idx] = hc_swap32_S (pws[gid].i[idx]);
   }
 
-  u32 t[128] = { 0 };
-
   /**
    * loop
    */
@@ -68,16 +66,19 @@ KERNEL_FQ void m24800_mxx (KERN_ATTR_BASIC ())
       c[i] |= w[i];
     }
 
+    u32 t[128] = { 0 };
+
     // make it unicode.
-    for(u32 i = 0, idx = 0; idx < pw_len + comb_len; i += 2, idx += 1){
+    for (u32 i = 0, idx = 0; idx < pw_len + comb_len; i += 2, idx += 1)
-        make_utf16beN(&c[idx], &t[i], &t[i+1]);
+    {
+        make_utf16beN (&c[idx], &t[i], &t[i+1]);
     }
 
     sha1_hmac_ctx_t ctx;
 
-    sha1_hmac_init (&ctx, t, (pw_len + comb_len)*2);
+    sha1_hmac_init (&ctx, t, (pw_len + comb_len) * 2);
 
-    sha1_hmac_update (&ctx, t, (pw_len + comb_len)*2);
+    sha1_hmac_update (&ctx, t, (pw_len + comb_len) * 2);
 
     sha1_hmac_final (&ctx);
 
@@ -120,7 +121,6 @@ KERNEL_FQ void m24800_sxx (KERN_ATTR_BASIC ())
   const u32 pw_len = pws[gid].pw_len;
 
   u32 w[64] = { 0 };
-  u32 t[128] = { 0 };
 
   for (u32 i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
   {
@@ -155,16 +155,19 @@ KERNEL_FQ void m24800_sxx (KERN_ATTR_BASIC ())
       c[i] |= w[i];
     }
 
+    u32 t[128] = { 0 };
+
     // make it unicode.
-    for(u32 i = 0, idx = 0; idx < pw_len + comb_len; i += 2, idx += 1){
+    for (u32 i = 0, idx = 0; idx < pw_len + comb_len; i += 2, idx += 1)
-        make_utf16beN(&c[idx], &t[i], &t[i+1]);
+    {
+        make_utf16beN (&c[idx], &t[i], &t[i+1]);
     }
 
     sha1_hmac_ctx_t ctx;
 
-    sha1_hmac_init (&ctx, t, (pw_len + comb_len)*2);
+    sha1_hmac_init (&ctx, t, (pw_len + comb_len) * 2);
 
-    sha1_hmac_update (&ctx, t, (pw_len + comb_len)*2);
+    sha1_hmac_update (&ctx, t, (pw_len + comb_len) * 2);
 
     sha1_hmac_final (&ctx);
 

OpenCL/m24800_a3-optimized.cl

@@ -0,0 +1,612 @@
+/**
+ * Author......: See docs/credits.txt
+ * License.....: MIT
+ */
+
+#define NEW_SIMD_CODE
+
+#ifdef KERNEL_STATIC
+#include "inc_vendor.h"
+#include "inc_types.h"
+#include "inc_platform.cl"
+#include "inc_common.cl"
+#include "inc_simd.cl"
+#include "inc_hash_sha1.cl"
+#endif
+
+DECLSPEC void hmac_sha1_pad (u32x *w0, u32x *w1, u32x *w2, u32x *w3, u32x *ipad, u32x *opad)
+{
+  w0[0] = w0[0] ^ 0x36363636;
+  w0[1] = w0[1] ^ 0x36363636;
+  w0[2] = w0[2] ^ 0x36363636;
+  w0[3] = w0[3] ^ 0x36363636;
+  w1[0] = w1[0] ^ 0x36363636;
+  w1[1] = w1[1] ^ 0x36363636;
+  w1[2] = w1[2] ^ 0x36363636;
+  w1[3] = w1[3] ^ 0x36363636;
+  w2[0] = w2[0] ^ 0x36363636;
+  w2[1] = w2[1] ^ 0x36363636;
+  w2[2] = w2[2] ^ 0x36363636;
+  w2[3] = w2[3] ^ 0x36363636;
+  w3[0] = w3[0] ^ 0x36363636;
+  w3[1] = w3[1] ^ 0x36363636;
+  w3[2] = w3[2] ^ 0x36363636;
+  w3[3] = w3[3] ^ 0x36363636;
+
+  ipad[0] = SHA1M_A;
+  ipad[1] = SHA1M_B;
+  ipad[2] = SHA1M_C;
+  ipad[3] = SHA1M_D;
+  ipad[4] = SHA1M_E;
+
+  sha1_transform_vector (w0, w1, w2, w3, ipad);
+
+  w0[0] = w0[0] ^ 0x6a6a6a6a;
+  w0[1] = w0[1] ^ 0x6a6a6a6a;
+  w0[2] = w0[2] ^ 0x6a6a6a6a;
+  w0[3] = w0[3] ^ 0x6a6a6a6a;
+  w1[0] = w1[0] ^ 0x6a6a6a6a;
+  w1[1] = w1[1] ^ 0x6a6a6a6a;
+  w1[2] = w1[2] ^ 0x6a6a6a6a;
+  w1[3] = w1[3] ^ 0x6a6a6a6a;
+  w2[0] = w2[0] ^ 0x6a6a6a6a;
+  w2[1] = w2[1] ^ 0x6a6a6a6a;
+  w2[2] = w2[2] ^ 0x6a6a6a6a;
+  w2[3] = w2[3] ^ 0x6a6a6a6a;
+  w3[0] = w3[0] ^ 0x6a6a6a6a;
+  w3[1] = w3[1] ^ 0x6a6a6a6a;
+  w3[2] = w3[2] ^ 0x6a6a6a6a;
+  w3[3] = w3[3] ^ 0x6a6a6a6a;
+
+  opad[0] = SHA1M_A;
+  opad[1] = SHA1M_B;
+  opad[2] = SHA1M_C;
+  opad[3] = SHA1M_D;
+  opad[4] = SHA1M_E;
+
+  sha1_transform_vector (w0, w1, w2, w3, opad);
+}
+
+DECLSPEC void hmac_sha1_run (u32x *w0, u32x *w1, u32x *w2, u32x *w3, u32x *ipad, u32x *opad, u32x *digest)
+{
+  digest[0] = ipad[0];
+  digest[1] = ipad[1];
+  digest[2] = ipad[2];
+  digest[3] = ipad[3];
+  digest[4] = ipad[4];
+
+  sha1_transform_vector (w0, w1, w2, w3, digest);
+
+  w0[0] = digest[0];
+  w0[1] = digest[1];
+  w0[2] = digest[2];
+  w0[3] = digest[3];
+  w1[0] = digest[4];
+  w1[1] = 0x80000000;
+  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] = 0;
+  w3[3] = (64 + 20) * 8;
+
+  digest[0] = opad[0];
+  digest[1] = opad[1];
+  digest[2] = opad[2];
+  digest[3] = opad[3];
+  digest[4] = opad[4];
+
+  sha1_transform_vector (w0, w1, w2, w3, digest);
+}
+
+DECLSPEC void m24800m (u32 *w0, u32 *w1, u32 *w2, u32 *w3, const u32 pw_len, KERN_ATTR_BASIC ())
+{
+  /**
+   * modifier
+   */
+
+  const u64 gid = get_global_id (0);
+  const u64 lid = get_local_id (0);
+
+  /**
+   * loop
+   */
+
+  u32 w0l = w0[0];
+
+  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
+  {
+    const u32x w0r = ix_create_bft (bfs_buf, il_pos);
+
+    const u32x w0lr = w0l | w0r;
+
+    /**
+     * pads
+     */
+
+    u32x w0_t[4];
+    u32x w1_t[4];
+    u32x w2_t[4];
+    u32x w3_t[4];
+
+    w0_t[0] = w0lr;
+    w0_t[1] = w0[1];
+    w0_t[2] = w0[2];
+    w0_t[3] = w0[3];
+    w1_t[0] = w1[0];
+    w1_t[1] = w1[1];
+    w1_t[2] = w1[2];
+    w1_t[3] = w1[3];
+    w2_t[0] = w2[0];
+    w2_t[1] = w2[1];
+    w2_t[2] = w2[2];
+    w2_t[3] = w2[3];
+    w3_t[0] = w3[0];
+    w3_t[1] = w3[1];
+    w3_t[2] = w3[2];
+    w3_t[3] = w3[3];
+
+    //make_utf16beN (w1_t, w2_t, w3_t);
+    //make_utf16beN (w0_t, w0_t, w1_t);
+
+    u32x x0_t[4];
+    u32x x1_t[4];
+    u32x x2_t[4];
+    u32x x3_t[4];
+
+    x0_t[0] = w0_t[0];
+    x0_t[1] = w0_t[1];
+    x0_t[2] = w0_t[2];
+    x0_t[3] = w0_t[3];
+    x1_t[0] = w1_t[0];
+    x1_t[1] = w1_t[1];
+    x1_t[2] = w1_t[2];
+    x1_t[3] = w1_t[3];
+    x2_t[0] = w2_t[0];
+    x2_t[1] = w2_t[1];
+    x2_t[2] = w2_t[2];
+    x2_t[3] = w2_t[3];
+    x3_t[0] = w3_t[0];
+    x3_t[1] = w3_t[1];
+    x3_t[2] = w3_t[2];
+    x3_t[3] = w3_t[3];
+
+    u32x ipad[5];
+    u32x opad[5];
+
+    hmac_sha1_pad (x0_t, x1_t, x2_t, x3_t, ipad, opad);
+
+    x0_t[0] = w0_t[0];
+    x0_t[1] = w0_t[1];
+    x0_t[2] = w0_t[2];
+    x0_t[3] = w0_t[3];
+    x1_t[0] = w1_t[0];
+    x1_t[1] = w1_t[1];
+    x1_t[2] = w1_t[2];
+    x1_t[3] = w1_t[3];
+    x2_t[0] = w2_t[0];
+    x2_t[1] = w2_t[1];
+    x2_t[2] = w2_t[2];
+    x2_t[3] = w2_t[3];
+    x3_t[0] = w3_t[0];
+    x3_t[1] = w3_t[1];
+    x3_t[2] = w3_t[2];
+    x3_t[3] = w3_t[3];
+
+    append_0x80_4x4 (x0_t, x1_t, x2_t, x3_t, pw_len ^ 3);
+
+    x3_t[2] = 0;
+    x3_t[3] = (64 + pw_len) * 8;
+
+    u32x digest[5];
+
+    hmac_sha1_run (x0_t, x1_t, x2_t, x3_t, ipad, opad, digest);
+
+    COMPARE_M_SIMD (digest[3], digest[4], digest[2], digest[1]);
+  }
+}
+
+DECLSPEC void m24800s (u32 *w0, u32 *w1, u32 *w2, u32 *w3, const u32 pw_len, KERN_ATTR_BASIC ())
+{
+  /**
+   * modifier
+   */
+
+  const u64 gid = get_global_id (0);
+  const u64 lid = get_local_id (0);
+
+  /**
+   * 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
+   */
+
+  u32 w0l = w0[0];
+
+  for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
+  {
+    const u32x w0r = ix_create_bft (bfs_buf, il_pos);
+
+    const u32x w0lr = w0l | w0r;
+
+    /**
+     * pads
+     */
+
+    u32x w0_t[4];
+    u32x w1_t[4];
+    u32x w2_t[4];
+    u32x w3_t[4];
+
+    w0_t[0] = w0lr;
+    w0_t[1] = w0[1];
+    w0_t[2] = w0[2];
+    w0_t[3] = w0[3];
+    w1_t[0] = w1[0];
+    w1_t[1] = w1[1];
+    w1_t[2] = w1[2];
+    w1_t[3] = w1[3];
+    w2_t[0] = w2[0];
+    w2_t[1] = w2[1];
+    w2_t[2] = w2[2];
+    w2_t[3] = w2[3];
+    w3_t[0] = w3[0];
+    w3_t[1] = w3[1];
+    w3_t[2] = w3[2];
+    w3_t[3] = w3[3];
+
+    //make_utf16beN (w1_t, w2_t, w3_t);
+    //make_utf16beN (w0_t, w0_t, w1_t);
+
+    u32x x0_t[4];
+    u32x x1_t[4];
+    u32x x2_t[4];
+    u32x x3_t[4];
+
+    x0_t[0] = w0_t[0];
+    x0_t[1] = w0_t[1];
+    x0_t[2] = w0_t[2];
+    x0_t[3] = w0_t[3];
+    x1_t[0] = w1_t[0];
+    x1_t[1] = w1_t[1];
+    x1_t[2] = w1_t[2];
+    x1_t[3] = w1_t[3];
+    x2_t[0] = w2_t[0];
+    x2_t[1] = w2_t[1];
+    x2_t[2] = w2_t[2];
+    x2_t[3] = w2_t[3];
+    x3_t[0] = w3_t[0];
+    x3_t[1] = w3_t[1];
+    x3_t[2] = w3_t[2];
+    x3_t[3] = w3_t[3];
+
+    u32x ipad[5];
+    u32x opad[5];
+
+    hmac_sha1_pad (x0_t, x1_t, x2_t, x3_t, ipad, opad);
+
+    x0_t[0] = w0_t[0];
+    x0_t[1] = w0_t[1];
+    x0_t[2] = w0_t[2];
+    x0_t[3] = w0_t[3];
+    x1_t[0] = w1_t[0];
+    x1_t[1] = w1_t[1];
+    x1_t[2] = w1_t[2];
+    x1_t[3] = w1_t[3];
+    x2_t[0] = w2_t[0];
+    x2_t[1] = w2_t[1];
+    x2_t[2] = w2_t[2];
+    x2_t[3] = w2_t[3];
+    x3_t[0] = w3_t[0];
+    x3_t[1] = w3_t[1];
+    x3_t[2] = w3_t[2];
+    x3_t[3] = w3_t[3];
+
+    append_0x80_4x4 (x0_t, x1_t, x2_t, x3_t, pw_len ^ 3);
+
+    x3_t[2] = 0;
+    x3_t[3] = (64 + pw_len) * 8;
+
+    u32x digest[5];
+
+    hmac_sha1_run (x0_t, x1_t, x2_t, x3_t, ipad, opad, digest);
+
+    COMPARE_S_SIMD (digest[3], digest[4], digest[2], digest[1]);
+  }
+}
+
+KERNEL_FQ void m24800_m04 (KERN_ATTR_BASIC ())
+{
+  /**
+   * base
+   */
+
+  const u64 gid = get_global_id (0);
+
+  if (gid >= gid_max) return;
+
+  u32 w0[4];
+
+  w0[0] = pws[gid].i[ 0];
+  w0[1] = pws[gid].i[ 1];
+  w0[2] = pws[gid].i[ 2];
+  w0[3] = pws[gid].i[ 3];
+
+  u32 w1[4];
+
+  w1[0] = 0;
+  w1[1] = 0;
+  w1[2] = 0;
+  w1[3] = 0;
+
+  u32 w2[4];
+
+  w2[0] = 0;
+  w2[1] = 0;
+  w2[2] = 0;
+  w2[3] = 0;
+
+  u32 w3[4];
+
+  w3[0] = 0;
+  w3[1] = 0;
+  w3[2] = 0;
+  w3[3] = 0;
+
+  const u32 pw_len = pws[gid].pw_len & 63;
+
+  /**
+   * main
+   */
+
+  m24800m (w0, w1, w2, w3, pw_len, pws, rules_buf, combs_buf, bfs_buf, tmps, hooks, bitmaps_buf_s1_a, bitmaps_buf_s1_b, bitmaps_buf_s1_c, bitmaps_buf_s1_d, bitmaps_buf_s2_a, bitmaps_buf_s2_b, bitmaps_buf_s2_c, bitmaps_buf_s2_d, plains_buf, digests_buf, hashes_shown, salt_bufs, esalt_bufs, d_return_buf, d_extra0_buf, d_extra1_buf, d_extra2_buf, d_extra3_buf, bitmap_mask, bitmap_shift1, bitmap_shift2, SALT_POS, loop_pos, loop_cnt, il_cnt, digests_cnt, DIGESTS_OFFSET, combs_mode, pws_pos, gid_max);
+}
+
+KERNEL_FQ void m24800_m08 (KERN_ATTR_BASIC ())
+{
+  /**
+   * base
+   */
+
+  const u64 gid = get_global_id (0);
+
+  if (gid >= gid_max) return;
+
+  u32 w0[4];
+
+  w0[0] = pws[gid].i[ 0];
+  w0[1] = pws[gid].i[ 1];
+  w0[2] = pws[gid].i[ 2];
+  w0[3] = pws[gid].i[ 3];
+
+  u32 w1[4];
+
+  w1[0] = pws[gid].i[ 4];
+  w1[1] = pws[gid].i[ 5];
+  w1[2] = pws[gid].i[ 6];
+  w1[3] = pws[gid].i[ 7];
+
+  u32 w2[4];
+
+  w2[0] = 0;
+  w2[1] = 0;
+  w2[2] = 0;
+  w2[3] = 0;
+
+  u32 w3[4];
+
+  w3[0] = 0;
+  w3[1] = 0;
+  w3[2] = 0;
+  w3[3] = 0;
+
+  const u32 pw_len = pws[gid].pw_len & 63;
+
+  /**
+   * main
+   */
+
+  m24800m (w0, w1, w2, w3, pw_len, pws, rules_buf, combs_buf, bfs_buf, tmps, hooks, bitmaps_buf_s1_a, bitmaps_buf_s1_b, bitmaps_buf_s1_c, bitmaps_buf_s1_d, bitmaps_buf_s2_a, bitmaps_buf_s2_b, bitmaps_buf_s2_c, bitmaps_buf_s2_d, plains_buf, digests_buf, hashes_shown, salt_bufs, esalt_bufs, d_return_buf, d_extra0_buf, d_extra1_buf, d_extra2_buf, d_extra3_buf, bitmap_mask, bitmap_shift1, bitmap_shift2, SALT_POS, loop_pos, loop_cnt, il_cnt, digests_cnt, DIGESTS_OFFSET, combs_mode, pws_pos, gid_max);
+}
+
+KERNEL_FQ void m24800_m16 (KERN_ATTR_BASIC ())
+{
+  /**
+   * base
+   */
+
+  const u64 gid = get_global_id (0);
+
+  if (gid >= gid_max) return;
+
+  u32 w0[4];
+
+  w0[0] = pws[gid].i[ 0];
+  w0[1] = pws[gid].i[ 1];
+  w0[2] = pws[gid].i[ 2];
+  w0[3] = pws[gid].i[ 3];
+
+  u32 w1[4];
+
+  w1[0] = pws[gid].i[ 4];
+  w1[1] = pws[gid].i[ 5];
+  w1[2] = pws[gid].i[ 6];
+  w1[3] = pws[gid].i[ 7];
+
+  u32 w2[4];
+
+  w2[0] = pws[gid].i[ 8];
+  w2[1] = pws[gid].i[ 9];
+  w2[2] = pws[gid].i[10];
+  w2[3] = pws[gid].i[11];
+
+  u32 w3[4];
+
+  w3[0] = pws[gid].i[12];
+  w3[1] = pws[gid].i[13];
+  w3[2] = 0;
+  w3[3] = 0;
+
+  const u32 pw_len = pws[gid].pw_len & 63;
+
+  /**
+   * main
+   */
+
+  m24800m (w0, w1, w2, w3, pw_len, pws, rules_buf, combs_buf, bfs_buf, tmps, hooks, bitmaps_buf_s1_a, bitmaps_buf_s1_b, bitmaps_buf_s1_c, bitmaps_buf_s1_d, bitmaps_buf_s2_a, bitmaps_buf_s2_b, bitmaps_buf_s2_c, bitmaps_buf_s2_d, plains_buf, digests_buf, hashes_shown, salt_bufs, esalt_bufs, d_return_buf, d_extra0_buf, d_extra1_buf, d_extra2_buf, d_extra3_buf, bitmap_mask, bitmap_shift1, bitmap_shift2, SALT_POS, loop_pos, loop_cnt, il_cnt, digests_cnt, DIGESTS_OFFSET, combs_mode, pws_pos, gid_max);
+}
+
+KERNEL_FQ void m24800_s04 (KERN_ATTR_BASIC ())
+{
+  /**
+   * base
+   */
+
+  const u64 gid = get_global_id (0);
+
+  if (gid >= gid_max) return;
+
+  u32 w0[4];
+
+  w0[0] = pws[gid].i[ 0];
+  w0[1] = pws[gid].i[ 1];
+  w0[2] = pws[gid].i[ 2];
+  w0[3] = pws[gid].i[ 3];
+
+  u32 w1[4];
+
+  w1[0] = 0;
+  w1[1] = 0;
+  w1[2] = 0;
+  w1[3] = 0;
+
+  u32 w2[4];
+
+  w2[0] = 0;
+  w2[1] = 0;
+  w2[2] = 0;
+  w2[3] = 0;
+
+  u32 w3[4];
+
+  w3[0] = 0;
+  w3[1] = 0;
+  w3[2] = 0;
+  w3[3] = 0;
+
+  const u32 pw_len = pws[gid].pw_len & 63;
+
+  /**
+   * main
+   */
+
+  m24800s (w0, w1, w2, w3, pw_len, pws, rules_buf, combs_buf, bfs_buf, tmps, hooks, bitmaps_buf_s1_a, bitmaps_buf_s1_b, bitmaps_buf_s1_c, bitmaps_buf_s1_d, bitmaps_buf_s2_a, bitmaps_buf_s2_b, bitmaps_buf_s2_c, bitmaps_buf_s2_d, plains_buf, digests_buf, hashes_shown, salt_bufs, esalt_bufs, d_return_buf, d_extra0_buf, d_extra1_buf, d_extra2_buf, d_extra3_buf, bitmap_mask, bitmap_shift1, bitmap_shift2, SALT_POS, loop_pos, loop_cnt, il_cnt, digests_cnt, DIGESTS_OFFSET, combs_mode, pws_pos, gid_max);
+}
+
+KERNEL_FQ void m24800_s08 (KERN_ATTR_BASIC ())
+{
+  /**
+   * base
+   */
+
+  const u64 gid = get_global_id (0);
+
+  if (gid >= gid_max) return;
+
+  u32 w0[4];
+
+  w0[0] = pws[gid].i[ 0];
+  w0[1] = pws[gid].i[ 1];
+  w0[2] = pws[gid].i[ 2];
+  w0[3] = pws[gid].i[ 3];
+
+  u32 w1[4];
+
+  w1[0] = pws[gid].i[ 4];
+  w1[1] = pws[gid].i[ 5];
+  w1[2] = pws[gid].i[ 6];
+  w1[3] = pws[gid].i[ 7];
+
+  u32 w2[4];
+
+  w2[0] = 0;
+  w2[1] = 0;
+  w2[2] = 0;
+  w2[3] = 0;
+
+  u32 w3[4];
+
+  w3[0] = 0;
+  w3[1] = 0;
+  w3[2] = 0;
+  w3[3] = 0;
+
+  const u32 pw_len = pws[gid].pw_len & 63;
+
+  /**
+   * main
+   */
+
+  m24800s (w0, w1, w2, w3, pw_len, pws, rules_buf, combs_buf, bfs_buf, tmps, hooks, bitmaps_buf_s1_a, bitmaps_buf_s1_b, bitmaps_buf_s1_c, bitmaps_buf_s1_d, bitmaps_buf_s2_a, bitmaps_buf_s2_b, bitmaps_buf_s2_c, bitmaps_buf_s2_d, plains_buf, digests_buf, hashes_shown, salt_bufs, esalt_bufs, d_return_buf, d_extra0_buf, d_extra1_buf, d_extra2_buf, d_extra3_buf, bitmap_mask, bitmap_shift1, bitmap_shift2, SALT_POS, loop_pos, loop_cnt, il_cnt, digests_cnt, DIGESTS_OFFSET, combs_mode, pws_pos, gid_max);
+}
+
+KERNEL_FQ void m24800_s16 (KERN_ATTR_BASIC ())
+{
+  /**
+   * base
+   */
+
+  const u64 gid = get_global_id (0);
+
+  if (gid >= gid_max) return;
+
+  u32 w0[4];
+
+  w0[0] = pws[gid].i[ 0];
+  w0[1] = pws[gid].i[ 1];
+  w0[2] = pws[gid].i[ 2];
+  w0[3] = pws[gid].i[ 3];
+
+  u32 w1[4];
+
+  w1[0] = pws[gid].i[ 4];
+  w1[1] = pws[gid].i[ 5];
+  w1[2] = pws[gid].i[ 6];
+  w1[3] = pws[gid].i[ 7];
+
+  u32 w2[4];
+
+  w2[0] = pws[gid].i[ 8];
+  w2[1] = pws[gid].i[ 9];
+  w2[2] = pws[gid].i[10];
+  w2[3] = pws[gid].i[11];
+
+  u32 w3[4];
+
+  w3[0] = pws[gid].i[12];
+  w3[1] = pws[gid].i[13];
+  w3[2] = 0;
+  w3[3] = 0;
+
+  const u32 pw_len = pws[gid].pw_len & 63;
+
+  /**
+   * main
+   */
+
+  m24800s (w0, w1, w2, w3, pw_len, pws, rules_buf, combs_buf, bfs_buf, tmps, hooks, bitmaps_buf_s1_a, bitmaps_buf_s1_b, bitmaps_buf_s1_c, bitmaps_buf_s1_d, bitmaps_buf_s2_a, bitmaps_buf_s2_b, bitmaps_buf_s2_c, bitmaps_buf_s2_d, plains_buf, digests_buf, hashes_shown, salt_bufs, esalt_bufs, d_return_buf, d_extra0_buf, d_extra1_buf, d_extra2_buf, d_extra3_buf, bitmap_mask, bitmap_shift1, bitmap_shift2, SALT_POS, loop_pos, loop_cnt, il_cnt, digests_cnt, DIGESTS_OFFSET, combs_mode, pws_pos, gid_max);
+}

OpenCL/m24800_a3-pure.cl

@@ -32,15 +32,12 @@ KERNEL_FQ void m24800_mxx (KERN_ATTR_VECTOR ())
   const u32 pw_len = pws[gid].pw_len;
 
   u32x w[64] = { 0 };
-  u32x t[128] = { 0 };
 
   for (u32 i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
   {
     w[idx] = pws[gid].i[idx];
   }
 
-  /* The password is the salt too */
-
   /**
    * loop
    */
@@ -55,15 +52,18 @@ KERNEL_FQ void m24800_mxx (KERN_ATTR_VECTOR ())
 
     w[0] = w0;
 
-    for(u32 i = 0, idx = 0; idx < pw_len; i += 2, idx += 1){
+    u32x t[128] = { 0 };
-        make_utf16beN(&w[idx], &t[i], &t[i+1]);
+
+    for (u32 i = 0, idx = 0; idx < pw_len; i += 2, idx += 1)
+    {
+      make_utf16beN (&w[idx], &t[i + 0], &t[i + 1]);
     }
 
     sha1_hmac_ctx_vector_t ctx;
 
-    sha1_hmac_init_vector (&ctx, t, pw_len*2);
+    sha1_hmac_init_vector (&ctx, t, pw_len * 2);
 
-    sha1_hmac_update_vector (&ctx, t, pw_len*2);
+    sha1_hmac_update_vector (&ctx, t, pw_len * 2);
 
     sha1_hmac_final_vector (&ctx);
 
@@ -106,7 +106,6 @@ KERNEL_FQ void m24800_sxx (KERN_ATTR_VECTOR ())
   const u32 pw_len = pws[gid].pw_len;
 
   u32x w[64] = { 0 };
-  u32x t[128] = { 0 };
 
   for (u32 i = 0, idx = 0; i < pw_len; i += 4, idx += 1)
   {
@@ -126,15 +125,19 @@ KERNEL_FQ void m24800_sxx (KERN_ATTR_VECTOR ())
     const u32x w0 = w0l | w0r;
 
     w[0] = w0;
-    for(u32 i = 0, idx = 0; idx < pw_len; i += 2, idx += 1){
+
-        make_utf16beN(&w[idx], &t[i], &t[i+1]);
+    u32x t[128] = { 0 };
+
+    for (u32 i = 0, idx = 0; idx < pw_len; i += 2, idx += 1)
+    {
+      make_utf16beN (&w[idx], &t[i + 0], &t[i + 1]);
     }
 
     sha1_hmac_ctx_vector_t ctx;
 
-    sha1_hmac_init_vector (&ctx, t, pw_len*2);
+    sha1_hmac_init_vector (&ctx, t, pw_len * 2);
 
-    sha1_hmac_update_vector (&ctx, t, pw_len*2);
+    sha1_hmac_update_vector (&ctx, t, pw_len * 2);
 
     sha1_hmac_final_vector (&ctx);
 

docs/changes.txt

@@ -15,6 +15,7 @@
 - Added hash-mode: RSA/DSA/EC/OPENSSH Private Keys
 - Added hash-mode: SQLCipher
 - Added hash-mode: Dahua Authentication MD5
+- Added hash-mode: Umbraco HMAC-SHA1
 - Added hash-mode: sha1(sha1($pass).$salt)
 
 ##

docs/readme.txt

@@ -325,6 +325,7 @@ NVIDIA GPUs require "NVIDIA Driver" (440.64 or later) and "CUDA Toolkit" (9.0 or
 - SMF (Simple Machines Forum) > v1.1
 - MediaWiki B type
 - Redmine
+- Umbraco HMAC-SHA1
 - Joomla < 2.5.18
 - OpenCart
 - PrestaShop

src/modules/module_24800.c

@@ -22,8 +22,6 @@ static const u64   KERN_TYPE      = 24800;
 static const u32   OPTI_TYPE      = OPTI_TYPE_ZERO_BYTE
                                   | OPTI_TYPE_NOT_ITERATED;
 static const u64   OPTS_TYPE      = OPTS_TYPE_PT_GENERATE_BE
-                                  | OPTS_TYPE_ST_ADD80
-                                  | OPTS_TYPE_ST_ADDBITS15
                                   | OPTS_TYPE_PT_UTF16LE;
 static const u32   SALT_TYPE      = SALT_TYPE_EMBEDDED;
 static const char *ST_PASS        = "hashcat";
@@ -47,13 +45,15 @@ const char *module_st_pass        (MAYBE_UNUSED const hashconfig_t *hashconfig,
 int module_hash_decode (MAYBE_UNUSED const hashconfig_t *hashconfig, MAYBE_UNUSED void *digest_buf, MAYBE_UNUSED salt_t *salt, MAYBE_UNUSED void *esalt_buf, MAYBE_UNUSED void *hook_salt_buf, MAYBE_UNUSED hashinfo_t *hash_info, const char *line_buf, MAYBE_UNUSED const int line_len)
 {
   u32 *digest = (u32 *) digest_buf;
+
   token_t token;
 
   token.token_cnt  = 1;
 
   token.len_min[0] = 28;
   token.len_max[0] = 28;
-  token.attr[0]    = TOKEN_ATTR_VERIFY_LENGTH;
+  token.attr[0]    = TOKEN_ATTR_VERIFY_LENGTH
+                   | TOKEN_ATTR_VERIFY_BASE64A;
 
   const int rc_tokenizer = input_tokenizer ((const u8 *) line_buf, line_len, &token);
 
@@ -61,7 +61,8 @@ int module_hash_decode (MAYBE_UNUSED const hashconfig_t *hashconfig, MAYBE_UNUSE
 
   const u8 *hash_pos = token.buf[0];
   const int hash_len = token.len[0];
-  u8 tmp_buf[20] = { 0 };
+
+  u8 tmp_buf[32] = { 0 };
 
   const int decoded_len = base64_decode (base64_to_int, hash_pos, hash_len, tmp_buf);
 
@@ -99,12 +100,13 @@ int module_hash_encode (MAYBE_UNUSED const hashconfig_t *hashconfig, MAYBE_UNUSE
   tmp[3] = byte_swap_32 (tmp[3]);
   tmp[4] = byte_swap_32 (tmp[4]);
 
-
+  u8 ptr_plain[100] = { 0 };
-  char ptr_plain[28];
 
   base64_encode (int_to_base64, (const u8 *) tmp, 20, (u8 *) ptr_plain);
 
-  return snprintf (line_buf, line_size, "%s", ptr_plain);
+  const int out_len = snprintf (line_buf, line_size, "%s", (char *) ptr_plain);
+
+  return out_len;
 }
 
 void module_init (module_ctx_t *module_ctx)

tools/test_modules/m24800.pm

@@ -10,21 +10,20 @@ use warnings;
 
 use Digest::SHA1 qw (sha1);
 use Digest::HMAC qw (hmac hmac_hex);
-use Encode qw/encode decode/;
+use Encode qw (encode decode);
 use MIME::Base64;
 
-sub module_constraints { [[0, 256], [0, 256], [0, 55], [0, 55], [-1, -1]] }
+sub module_constraints { [[0, 256], [0, 256], [0, 27], [0, 27], [0, 27]] }
 
 sub module_generate_hash
 {
   my $word = shift;
-  my $unicode_word;
 
-  $unicode_word = encode("UTF-16LE", $word);  
+  my $unicode_word = encode ("UTF-16LE", $word);
 
   my $digest = hmac ($unicode_word, $unicode_word, \&sha1, 64);
 
-  my $hash = sprintf ("%s", encode_base64($digest));
+  my $hash = sprintf ("%s", encode_base64 ($digest, ""));
 
   return $hash;
 }