2 * Authors.....: Jens Steube <jens.steube@gmail.com>
3 * Gabriele Gristina <matrix@hashcat.net>
10 //too much register pressure
11 //#define NEW_SIMD_CODE
13 #include "inc_vendor.cl"
14 #include "inc_hash_constants.h"
15 #include "inc_hash_functions.cl"
16 #include "inc_types.cl"
17 #include "inc_common.cl"
20 #include "inc_simd.cl"
22 __constant u32 padding[8] =
42 void swap (__local RC4_KEY *rc4_key, const u8 i, const u8 j)
47 rc4_key->S[i] = rc4_key->S[j];
51 void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 data[4])
56 __local u32 *ptr = (__local u32 *) rc4_key->S;
61 for (u32 i = 0; i < 64; i++)
66 const u32 d0 = data[0] >> 0;
67 const u32 d1 = data[0] >> 8;
68 const u32 d2 = data[0] >> 16;
69 const u32 d3 = data[0] >> 24;
70 const u32 d4 = data[1] >> 0;
77 for (u32 i = 0; i < 255; i += 5)
79 j += rc4_key->S[i + 0] + d0; swap (rc4_key, i + 0, j);
80 j += rc4_key->S[i + 1] + d1; swap (rc4_key, i + 1, j);
81 j += rc4_key->S[i + 2] + d2; swap (rc4_key, i + 2, j);
82 j += rc4_key->S[i + 3] + d3; swap (rc4_key, i + 3, j);
83 j += rc4_key->S[i + 4] + d4; swap (rc4_key, i + 4, j);
86 j += rc4_key->S[255] + d0; swap (rc4_key, 255, j);
89 u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, __constant u32 *in, u32 out[4])
94 for (u32 k = 0; k < 4; k++)
103 swap (rc4_key, i, j);
105 idx = rc4_key->S[i] + rc4_key->S[j];
107 xor4 |= rc4_key->S[idx] << 0;
112 swap (rc4_key, i, j);
114 idx = rc4_key->S[i] + rc4_key->S[j];
116 xor4 |= rc4_key->S[idx] << 8;
121 swap (rc4_key, i, j);
123 idx = rc4_key->S[i] + rc4_key->S[j];
125 xor4 |= rc4_key->S[idx] << 16;
130 swap (rc4_key, i, j);
132 idx = rc4_key->S[i] + rc4_key->S[j];
134 xor4 |= rc4_key->S[idx] << 24;
136 out[k] = in[k] ^ xor4;
142 void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[4])
166 MD5_STEP (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00);
167 MD5_STEP (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01);
168 MD5_STEP (MD5_Fo, c, d, a, b, w2_t, MD5C02, MD5S02);
169 MD5_STEP (MD5_Fo, b, c, d, a, w3_t, MD5C03, MD5S03);
170 MD5_STEP (MD5_Fo, a, b, c, d, w4_t, MD5C04, MD5S00);
171 MD5_STEP (MD5_Fo, d, a, b, c, w5_t, MD5C05, MD5S01);
172 MD5_STEP (MD5_Fo, c, d, a, b, w6_t, MD5C06, MD5S02);
173 MD5_STEP (MD5_Fo, b, c, d, a, w7_t, MD5C07, MD5S03);
174 MD5_STEP (MD5_Fo, a, b, c, d, w8_t, MD5C08, MD5S00);
175 MD5_STEP (MD5_Fo, d, a, b, c, w9_t, MD5C09, MD5S01);
176 MD5_STEP (MD5_Fo, c, d, a, b, wa_t, MD5C0a, MD5S02);
177 MD5_STEP (MD5_Fo, b, c, d, a, wb_t, MD5C0b, MD5S03);
178 MD5_STEP (MD5_Fo, a, b, c, d, wc_t, MD5C0c, MD5S00);
179 MD5_STEP (MD5_Fo, d, a, b, c, wd_t, MD5C0d, MD5S01);
180 MD5_STEP (MD5_Fo, c, d, a, b, we_t, MD5C0e, MD5S02);
181 MD5_STEP (MD5_Fo, b, c, d, a, wf_t, MD5C0f, MD5S03);
183 MD5_STEP (MD5_Go, a, b, c, d, w1_t, MD5C10, MD5S10);
184 MD5_STEP (MD5_Go, d, a, b, c, w6_t, MD5C11, MD5S11);
185 MD5_STEP (MD5_Go, c, d, a, b, wb_t, MD5C12, MD5S12);
186 MD5_STEP (MD5_Go, b, c, d, a, w0_t, MD5C13, MD5S13);
187 MD5_STEP (MD5_Go, a, b, c, d, w5_t, MD5C14, MD5S10);
188 MD5_STEP (MD5_Go, d, a, b, c, wa_t, MD5C15, MD5S11);
189 MD5_STEP (MD5_Go, c, d, a, b, wf_t, MD5C16, MD5S12);
190 MD5_STEP (MD5_Go, b, c, d, a, w4_t, MD5C17, MD5S13);
191 MD5_STEP (MD5_Go, a, b, c, d, w9_t, MD5C18, MD5S10);
192 MD5_STEP (MD5_Go, d, a, b, c, we_t, MD5C19, MD5S11);
193 MD5_STEP (MD5_Go, c, d, a, b, w3_t, MD5C1a, MD5S12);
194 MD5_STEP (MD5_Go, b, c, d, a, w8_t, MD5C1b, MD5S13);
195 MD5_STEP (MD5_Go, a, b, c, d, wd_t, MD5C1c, MD5S10);
196 MD5_STEP (MD5_Go, d, a, b, c, w2_t, MD5C1d, MD5S11);
197 MD5_STEP (MD5_Go, c, d, a, b, w7_t, MD5C1e, MD5S12);
198 MD5_STEP (MD5_Go, b, c, d, a, wc_t, MD5C1f, MD5S13);
200 MD5_STEP (MD5_H , a, b, c, d, w5_t, MD5C20, MD5S20);
201 MD5_STEP (MD5_H , d, a, b, c, w8_t, MD5C21, MD5S21);
202 MD5_STEP (MD5_H , c, d, a, b, wb_t, MD5C22, MD5S22);
203 MD5_STEP (MD5_H , b, c, d, a, we_t, MD5C23, MD5S23);
204 MD5_STEP (MD5_H , a, b, c, d, w1_t, MD5C24, MD5S20);
205 MD5_STEP (MD5_H , d, a, b, c, w4_t, MD5C25, MD5S21);
206 MD5_STEP (MD5_H , c, d, a, b, w7_t, MD5C26, MD5S22);
207 MD5_STEP (MD5_H , b, c, d, a, wa_t, MD5C27, MD5S23);
208 MD5_STEP (MD5_H , a, b, c, d, wd_t, MD5C28, MD5S20);
209 MD5_STEP (MD5_H , d, a, b, c, w0_t, MD5C29, MD5S21);
210 MD5_STEP (MD5_H , c, d, a, b, w3_t, MD5C2a, MD5S22);
211 MD5_STEP (MD5_H , b, c, d, a, w6_t, MD5C2b, MD5S23);
212 MD5_STEP (MD5_H , a, b, c, d, w9_t, MD5C2c, MD5S20);
213 MD5_STEP (MD5_H , d, a, b, c, wc_t, MD5C2d, MD5S21);
214 MD5_STEP (MD5_H , c, d, a, b, wf_t, MD5C2e, MD5S22);
215 MD5_STEP (MD5_H , b, c, d, a, w2_t, MD5C2f, MD5S23);
217 MD5_STEP (MD5_I , a, b, c, d, w0_t, MD5C30, MD5S30);
218 MD5_STEP (MD5_I , d, a, b, c, w7_t, MD5C31, MD5S31);
219 MD5_STEP (MD5_I , c, d, a, b, we_t, MD5C32, MD5S32);
220 MD5_STEP (MD5_I , b, c, d, a, w5_t, MD5C33, MD5S33);
221 MD5_STEP (MD5_I , a, b, c, d, wc_t, MD5C34, MD5S30);
222 MD5_STEP (MD5_I , d, a, b, c, w3_t, MD5C35, MD5S31);
223 MD5_STEP (MD5_I , c, d, a, b, wa_t, MD5C36, MD5S32);
224 MD5_STEP (MD5_I , b, c, d, a, w1_t, MD5C37, MD5S33);
225 MD5_STEP (MD5_I , a, b, c, d, w8_t, MD5C38, MD5S30);
226 MD5_STEP (MD5_I , d, a, b, c, wf_t, MD5C39, MD5S31);
227 MD5_STEP (MD5_I , c, d, a, b, w6_t, MD5C3a, MD5S32);
228 MD5_STEP (MD5_I , b, c, d, a, wd_t, MD5C3b, MD5S33);
229 MD5_STEP (MD5_I , a, b, c, d, w4_t, MD5C3c, MD5S30);
230 MD5_STEP (MD5_I , d, a, b, c, wb_t, MD5C3d, MD5S31);
231 MD5_STEP (MD5_I , c, d, a, b, w2_t, MD5C3e, MD5S32);
232 MD5_STEP (MD5_I , b, c, d, a, w9_t, MD5C3f, MD5S33);
240 __kernel void m10400_m04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 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 u32 gid_max)
246 const u32 lid = get_local_id (0);
252 const u32 gid = get_global_id (0);
254 if (gid >= gid_max) return;
259 pw_buf0[0] = pws[gid].i[ 0];
260 pw_buf0[1] = pws[gid].i[ 1];
261 pw_buf0[2] = pws[gid].i[ 2];
262 pw_buf0[3] = pws[gid].i[ 3];
263 pw_buf1[0] = pws[gid].i[ 4];
264 pw_buf1[1] = pws[gid].i[ 5];
265 pw_buf1[2] = pws[gid].i[ 6];
266 pw_buf1[3] = pws[gid].i[ 7];
268 const u32 pw_len = pws[gid].pw_len;
274 __local RC4_KEY rc4_keys[64];
276 __local RC4_KEY *rc4_key = &rc4_keys[lid];
284 o_buf[0] = pdf_bufs[salt_pos].o_buf[0];
285 o_buf[1] = pdf_bufs[salt_pos].o_buf[1];
286 o_buf[2] = pdf_bufs[salt_pos].o_buf[2];
287 o_buf[3] = pdf_bufs[salt_pos].o_buf[3];
288 o_buf[4] = pdf_bufs[salt_pos].o_buf[4];
289 o_buf[5] = pdf_bufs[salt_pos].o_buf[5];
290 o_buf[6] = pdf_bufs[salt_pos].o_buf[6];
291 o_buf[7] = pdf_bufs[salt_pos].o_buf[7];
293 u32 P = pdf_bufs[salt_pos].P;
297 id_buf[0] = pdf_bufs[salt_pos].id_buf[0];
298 id_buf[1] = pdf_bufs[salt_pos].id_buf[1];
299 id_buf[2] = pdf_bufs[salt_pos].id_buf[2];
300 id_buf[3] = pdf_bufs[salt_pos].id_buf[3];
306 for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
313 const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
341 switch_buffer_by_offset_le (p0, p1, p2, p3, out_len);
344 // truncate at 32 is wanted, not a bug!
371 md5_transform (w0, w1, w2, w3, digest);
390 md5_transform (w0, w1, w2, w3, digest);
394 digest[1] = digest[1] & 0xff;
398 rc4_init_16 (rc4_key, digest);
402 rc4_next_16 (rc4_key, 0, 0, padding, out);
404 COMPARE_M_SIMD (out[0], out[1], out[2], out[3]);
408 __kernel void m10400_m08 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 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 u32 gid_max)
412 __kernel void m10400_m16 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 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 u32 gid_max)
416 __kernel void m10400_s04 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 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 u32 gid_max)
422 const u32 lid = get_local_id (0);
428 const u32 gid = get_global_id (0);
430 if (gid >= gid_max) return;
435 pw_buf0[0] = pws[gid].i[ 0];
436 pw_buf0[1] = pws[gid].i[ 1];
437 pw_buf0[2] = pws[gid].i[ 2];
438 pw_buf0[3] = pws[gid].i[ 3];
439 pw_buf1[0] = pws[gid].i[ 4];
440 pw_buf1[1] = pws[gid].i[ 5];
441 pw_buf1[2] = pws[gid].i[ 6];
442 pw_buf1[3] = pws[gid].i[ 7];
444 const u32 pw_len = pws[gid].pw_len;
450 __local RC4_KEY rc4_keys[64];
452 __local RC4_KEY *rc4_key = &rc4_keys[lid];
460 o_buf[0] = pdf_bufs[salt_pos].o_buf[0];
461 o_buf[1] = pdf_bufs[salt_pos].o_buf[1];
462 o_buf[2] = pdf_bufs[salt_pos].o_buf[2];
463 o_buf[3] = pdf_bufs[salt_pos].o_buf[3];
464 o_buf[4] = pdf_bufs[salt_pos].o_buf[4];
465 o_buf[5] = pdf_bufs[salt_pos].o_buf[5];
466 o_buf[6] = pdf_bufs[salt_pos].o_buf[6];
467 o_buf[7] = pdf_bufs[salt_pos].o_buf[7];
469 u32 P = pdf_bufs[salt_pos].P;
473 id_buf[0] = pdf_bufs[salt_pos].id_buf[0];
474 id_buf[1] = pdf_bufs[salt_pos].id_buf[1];
475 id_buf[2] = pdf_bufs[salt_pos].id_buf[2];
476 id_buf[3] = pdf_bufs[salt_pos].id_buf[3];
482 const u32 search[4] =
484 digests_buf[digests_offset].digest_buf[DGST_R0],
485 digests_buf[digests_offset].digest_buf[DGST_R1],
486 digests_buf[digests_offset].digest_buf[DGST_R2],
487 digests_buf[digests_offset].digest_buf[DGST_R3]
494 for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
501 const u32x out_len = apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
529 switch_buffer_by_offset_le (p0, p1, p2, p3, out_len);
532 // truncate at 32 is wanted, not a bug!
559 md5_transform (w0, w1, w2, w3, digest);
578 md5_transform (w0, w1, w2, w3, digest);
582 digest[1] = digest[1] & 0xff;
586 rc4_init_16 (rc4_key, digest);
590 rc4_next_16 (rc4_key, 0, 0, padding, out);
592 COMPARE_S_SIMD (out[0], out[1], out[2], out[3]);
596 __kernel void m10400_s08 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 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 u32 gid_max)
600 __kernel void m10400_s16 (__global pw_t *pws, __global kernel_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global void *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 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 u32 gid_max)