Upgrade kernel to support dynamic local work sizes

[hashcat.git] / OpenCL / m06232.cl

diff --git a/OpenCL/m06232.cl b/OpenCL/m06232.cl
index e2c2e8f..0058bc8 100644 (file)
--- a/OpenCL/m06232.cl
+++ b/OpenCL/m06232.cl
@@ -1465,13 +1465,51 @@ static u32 u8add (const u32 a, const u32 b)
   return r;
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m06232_init (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global tc_tmp_t *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global tc_t *esalt_bufs, __global u32 *d_return_buf, __global 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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m06232_init (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global tc_tmp_t *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global tc_t *esalt_bufs, __global u32 *d_return_buf, __global 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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
 {
   /**
-   * base
+   * modifier
    */
 
   const u32 gid = get_global_id (0);
+  const u32 lid = get_local_id (0);
+  const u32 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 w0[4];
 
@@ -1522,33 +1560,6 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m06232_init (__gl
   w3[2] = u8add (w3[2], esalt_bufs[salt_pos].keyfile_buf[14]);
   w3[3] = u8add (w3[3], esalt_bufs[salt_pos].keyfile_buf[15]);
 
-  /**
-   * shared mem
-   */
-
-  __local u32 s_Ch[8][256];
-  __local u32 s_Cl[8][256];
-
-  const u32 lid = get_local_id (0);
-
-  const u32 lid4 = lid * 4;
-
-  for (u32 i = 0; i < 8; i++)
-  {
-    s_Ch[i][lid4 + 0] = Ch[i][lid4 + 0];
-    s_Ch[i][lid4 + 1] = Ch[i][lid4 + 1];
-    s_Ch[i][lid4 + 2] = Ch[i][lid4 + 2];
-    s_Ch[i][lid4 + 3] = Ch[i][lid4 + 3];
-    s_Cl[i][lid4 + 0] = Cl[i][lid4 + 0];
-    s_Cl[i][lid4 + 1] = Cl[i][lid4 + 1];
-    s_Cl[i][lid4 + 2] = Cl[i][lid4 + 2];
-    s_Cl[i][lid4 + 3] = Cl[i][lid4 + 3];
-  }
-
-  barrier (CLK_LOCAL_MEM_FENCE);
-
-  if (gid >= gid_max) return;
-
   /**
    * salt
    */
@@ -1695,34 +1706,50 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m06232_init (__gl
   }
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m06232_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global tc_tmp_t *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global tc_t *esalt_bufs, __global u32 *d_return_buf, __global 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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m06232_loop (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global tc_tmp_t *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global tc_t *esalt_bufs, __global u32 *d_return_buf, __global 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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
 {
-  const u32 truecrypt_mdlen = salt_bufs[0].truecrypt_mdlen;
-
-  __local u32 s_Ch[8][256];
-  __local u32 s_Cl[8][256];
+  /**
+   * modifier
+   */
 
   const u32 gid = get_global_id (0);
   const u32 lid = get_local_id (0);
+  const u32 lsz = get_local_size (0);
+
+  /**
+   * shared
+   */
 
-  const u32 lid4 = lid * 4;
+  __local u32 s_Ch[8][256];
+  __local u32 s_Cl[8][256];
 
-  for (u32 i = 0; i < 8; i++)
+  for (u32 i = lid; i < 256; i += lsz)
   {
-    s_Ch[i][lid4 + 0] = Ch[i][lid4 + 0];
-    s_Ch[i][lid4 + 1] = Ch[i][lid4 + 1];
-    s_Ch[i][lid4 + 2] = Ch[i][lid4 + 2];
-    s_Ch[i][lid4 + 3] = Ch[i][lid4 + 3];
-    s_Cl[i][lid4 + 0] = Cl[i][lid4 + 0];
-    s_Cl[i][lid4 + 1] = Cl[i][lid4 + 1];
-    s_Cl[i][lid4 + 2] = Cl[i][lid4 + 2];
-    s_Cl[i][lid4 + 3] = Cl[i][lid4 + 3];
+    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;
 
+  const u32 truecrypt_mdlen = salt_bufs[0].truecrypt_mdlen;
+
   u32 ipad[16];
 
   ipad[ 0] = tmps[gid].ipad[ 0];
@@ -1897,7 +1924,7 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m06232_loop (__gl
   }
 }
 
-__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m06232_comp (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global tc_tmp_t *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global tc_t *esalt_bufs, __global u32 *d_return_buf, __global 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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
+__kernel void m06232_comp (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global tc_tmp_t *tmps, __global void *hooks, __global u32 *bitmaps_buf_s1_a, __global u32 *bitmaps_buf_s1_b, __global u32 *bitmaps_buf_s1_c, __global u32 *bitmaps_buf_s1_d, __global u32 *bitmaps_buf_s2_a, __global u32 *bitmaps_buf_s2_b, __global u32 *bitmaps_buf_s2_c, __global u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global digest_t *digests_buf, __global u32 *hashes_shown, __global salt_t *salt_bufs, __global tc_t *esalt_bufs, __global u32 *d_return_buf, __global 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 rules_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
 {
   /**
    * base