2 * Author......: Jens Steube <jens.steube@gmail.com>
8 //too much register pressure
9 //#define NEW_SIMD_CODE
11 #include "include/constants.h"
12 #include "include/kernel_vendor.h"
19 #include "include/kernel_functions.c"
20 #include "OpenCL/types_ocl.c"
21 #include "OpenCL/common.c"
22 #include "include/rp_kernel.h"
23 #include "OpenCL/rp.c"
24 #include "OpenCL/simd.c"
34 void swap (__local RC4_KEY *rc4_key, const u8 i, const u8 j)
39 rc4_key->S[i] = rc4_key->S[j];
43 void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 data[4])
48 __local u32 *ptr = (__local u32 *) rc4_key->S;
53 for (u32 i = 0; i < 64; i++)
60 for (u32 i = 0; i < 16; i++)
68 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
69 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
70 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
71 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
75 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
76 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
77 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
78 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
82 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
83 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
84 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
85 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
89 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
90 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
91 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
92 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
96 u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, const u32 in[4], u32 out[4])
101 for (u32 k = 0; k < 4; k++)
110 swap (rc4_key, i, j);
112 idx = rc4_key->S[i] + rc4_key->S[j];
114 xor4 |= rc4_key->S[idx] << 0;
119 swap (rc4_key, i, j);
121 idx = rc4_key->S[i] + rc4_key->S[j];
123 xor4 |= rc4_key->S[idx] << 8;
128 swap (rc4_key, i, j);
130 idx = rc4_key->S[i] + rc4_key->S[j];
132 xor4 |= rc4_key->S[idx] << 16;
137 swap (rc4_key, i, j);
139 idx = rc4_key->S[i] + rc4_key->S[j];
141 xor4 |= rc4_key->S[idx] << 24;
143 out[k] = in[k] ^ xor4;
149 void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[4])
173 MD5_STEP (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00);
174 MD5_STEP (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01);
175 MD5_STEP (MD5_Fo, c, d, a, b, w2_t, MD5C02, MD5S02);
176 MD5_STEP (MD5_Fo, b, c, d, a, w3_t, MD5C03, MD5S03);
177 MD5_STEP (MD5_Fo, a, b, c, d, w4_t, MD5C04, MD5S00);
178 MD5_STEP (MD5_Fo, d, a, b, c, w5_t, MD5C05, MD5S01);
179 MD5_STEP (MD5_Fo, c, d, a, b, w6_t, MD5C06, MD5S02);
180 MD5_STEP (MD5_Fo, b, c, d, a, w7_t, MD5C07, MD5S03);
181 MD5_STEP (MD5_Fo, a, b, c, d, w8_t, MD5C08, MD5S00);
182 MD5_STEP (MD5_Fo, d, a, b, c, w9_t, MD5C09, MD5S01);
183 MD5_STEP (MD5_Fo, c, d, a, b, wa_t, MD5C0a, MD5S02);
184 MD5_STEP (MD5_Fo, b, c, d, a, wb_t, MD5C0b, MD5S03);
185 MD5_STEP (MD5_Fo, a, b, c, d, wc_t, MD5C0c, MD5S00);
186 MD5_STEP (MD5_Fo, d, a, b, c, wd_t, MD5C0d, MD5S01);
187 MD5_STEP (MD5_Fo, c, d, a, b, we_t, MD5C0e, MD5S02);
188 MD5_STEP (MD5_Fo, b, c, d, a, wf_t, MD5C0f, MD5S03);
190 MD5_STEP (MD5_Go, a, b, c, d, w1_t, MD5C10, MD5S10);
191 MD5_STEP (MD5_Go, d, a, b, c, w6_t, MD5C11, MD5S11);
192 MD5_STEP (MD5_Go, c, d, a, b, wb_t, MD5C12, MD5S12);
193 MD5_STEP (MD5_Go, b, c, d, a, w0_t, MD5C13, MD5S13);
194 MD5_STEP (MD5_Go, a, b, c, d, w5_t, MD5C14, MD5S10);
195 MD5_STEP (MD5_Go, d, a, b, c, wa_t, MD5C15, MD5S11);
196 MD5_STEP (MD5_Go, c, d, a, b, wf_t, MD5C16, MD5S12);
197 MD5_STEP (MD5_Go, b, c, d, a, w4_t, MD5C17, MD5S13);
198 MD5_STEP (MD5_Go, a, b, c, d, w9_t, MD5C18, MD5S10);
199 MD5_STEP (MD5_Go, d, a, b, c, we_t, MD5C19, MD5S11);
200 MD5_STEP (MD5_Go, c, d, a, b, w3_t, MD5C1a, MD5S12);
201 MD5_STEP (MD5_Go, b, c, d, a, w8_t, MD5C1b, MD5S13);
202 MD5_STEP (MD5_Go, a, b, c, d, wd_t, MD5C1c, MD5S10);
203 MD5_STEP (MD5_Go, d, a, b, c, w2_t, MD5C1d, MD5S11);
204 MD5_STEP (MD5_Go, c, d, a, b, w7_t, MD5C1e, MD5S12);
205 MD5_STEP (MD5_Go, b, c, d, a, wc_t, MD5C1f, MD5S13);
207 MD5_STEP (MD5_H , a, b, c, d, w5_t, MD5C20, MD5S20);
208 MD5_STEP (MD5_H , d, a, b, c, w8_t, MD5C21, MD5S21);
209 MD5_STEP (MD5_H , c, d, a, b, wb_t, MD5C22, MD5S22);
210 MD5_STEP (MD5_H , b, c, d, a, we_t, MD5C23, MD5S23);
211 MD5_STEP (MD5_H , a, b, c, d, w1_t, MD5C24, MD5S20);
212 MD5_STEP (MD5_H , d, a, b, c, w4_t, MD5C25, MD5S21);
213 MD5_STEP (MD5_H , c, d, a, b, w7_t, MD5C26, MD5S22);
214 MD5_STEP (MD5_H , b, c, d, a, wa_t, MD5C27, MD5S23);
215 MD5_STEP (MD5_H , a, b, c, d, wd_t, MD5C28, MD5S20);
216 MD5_STEP (MD5_H , d, a, b, c, w0_t, MD5C29, MD5S21);
217 MD5_STEP (MD5_H , c, d, a, b, w3_t, MD5C2a, MD5S22);
218 MD5_STEP (MD5_H , b, c, d, a, w6_t, MD5C2b, MD5S23);
219 MD5_STEP (MD5_H , a, b, c, d, w9_t, MD5C2c, MD5S20);
220 MD5_STEP (MD5_H , d, a, b, c, wc_t, MD5C2d, MD5S21);
221 MD5_STEP (MD5_H , c, d, a, b, wf_t, MD5C2e, MD5S22);
222 MD5_STEP (MD5_H , b, c, d, a, w2_t, MD5C2f, MD5S23);
224 MD5_STEP (MD5_I , a, b, c, d, w0_t, MD5C30, MD5S30);
225 MD5_STEP (MD5_I , d, a, b, c, w7_t, MD5C31, MD5S31);
226 MD5_STEP (MD5_I , c, d, a, b, we_t, MD5C32, MD5S32);
227 MD5_STEP (MD5_I , b, c, d, a, w5_t, MD5C33, MD5S33);
228 MD5_STEP (MD5_I , a, b, c, d, wc_t, MD5C34, MD5S30);
229 MD5_STEP (MD5_I , d, a, b, c, w3_t, MD5C35, MD5S31);
230 MD5_STEP (MD5_I , c, d, a, b, wa_t, MD5C36, MD5S32);
231 MD5_STEP (MD5_I , b, c, d, a, w1_t, MD5C37, MD5S33);
232 MD5_STEP (MD5_I , a, b, c, d, w8_t, MD5C38, MD5S30);
233 MD5_STEP (MD5_I , d, a, b, c, wf_t, MD5C39, MD5S31);
234 MD5_STEP (MD5_I , c, d, a, b, w6_t, MD5C3a, MD5S32);
235 MD5_STEP (MD5_I , b, c, d, a, wd_t, MD5C3b, MD5S33);
236 MD5_STEP (MD5_I , a, b, c, d, w4_t, MD5C3c, MD5S30);
237 MD5_STEP (MD5_I , d, a, b, c, wb_t, MD5C3d, MD5S31);
238 MD5_STEP (MD5_I , c, d, a, b, w2_t, MD5C3e, MD5S32);
239 MD5_STEP (MD5_I , b, c, d, a, w9_t, MD5C3f, MD5S33);
247 __kernel void m09710_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 oldoffice01_t *oldoffice01_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
253 const u32 lid = get_local_id (0);
259 const u32 gid = get_global_id (0);
261 if (gid >= gid_max) return;
266 pw_buf0[0] = pws[gid].i[ 0];
267 pw_buf0[1] = pws[gid].i[ 1];
268 pw_buf0[2] = pws[gid].i[ 2];
269 pw_buf0[3] = pws[gid].i[ 3];
270 pw_buf1[0] = pws[gid].i[ 4];
271 pw_buf1[1] = pws[gid].i[ 5];
272 pw_buf1[2] = pws[gid].i[ 6];
273 pw_buf1[3] = pws[gid].i[ 7];
275 const u32 pw_len = pws[gid].pw_len;
281 __local RC4_KEY rc4_keys[64];
283 __local RC4_KEY *rc4_key = &rc4_keys[lid];
289 const u32 version = oldoffice01_bufs[salt_pos].version;
291 u32 encryptedVerifier[4];
293 encryptedVerifier[0] = oldoffice01_bufs[salt_pos].encryptedVerifier[0];
294 encryptedVerifier[1] = oldoffice01_bufs[salt_pos].encryptedVerifier[1];
295 encryptedVerifier[2] = oldoffice01_bufs[salt_pos].encryptedVerifier[2];
296 encryptedVerifier[3] = oldoffice01_bufs[salt_pos].encryptedVerifier[3];
302 for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
309 apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
316 w0[1] = w0[1] & 0xff;
339 md5_transform (w0, w1, w2, w3, digest);
343 rc4_init_16 (rc4_key, digest);
347 u8 j = rc4_next_16 (rc4_key, 0, 0, encryptedVerifier, out);
371 md5_transform (w0, w1, w2, w3, digest);
373 rc4_next_16 (rc4_key, 16, j, digest, out);
375 COMPARE_M_SIMD (out[0], out[1], out[2], out[3]);
379 __kernel void m09710_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 oldoffice01_t *oldoffice01_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
383 __kernel void m09710_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 oldoffice01_t *oldoffice01_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
387 __kernel void m09710_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 oldoffice01_t *oldoffice01_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
393 const u32 lid = get_local_id (0);
399 const u32 gid = get_global_id (0);
401 if (gid >= gid_max) return;
406 pw_buf0[0] = pws[gid].i[ 0];
407 pw_buf0[1] = pws[gid].i[ 1];
408 pw_buf0[2] = pws[gid].i[ 2];
409 pw_buf0[3] = pws[gid].i[ 3];
410 pw_buf1[0] = pws[gid].i[ 4];
411 pw_buf1[1] = pws[gid].i[ 5];
412 pw_buf1[2] = pws[gid].i[ 6];
413 pw_buf1[3] = pws[gid].i[ 7];
415 const u32 pw_len = pws[gid].pw_len;
421 __local RC4_KEY rc4_keys[64];
423 __local RC4_KEY *rc4_key = &rc4_keys[lid];
429 const u32 version = oldoffice01_bufs[salt_pos].version;
431 u32 encryptedVerifier[4];
433 encryptedVerifier[0] = oldoffice01_bufs[salt_pos].encryptedVerifier[0];
434 encryptedVerifier[1] = oldoffice01_bufs[salt_pos].encryptedVerifier[1];
435 encryptedVerifier[2] = oldoffice01_bufs[salt_pos].encryptedVerifier[2];
436 encryptedVerifier[3] = oldoffice01_bufs[salt_pos].encryptedVerifier[3];
442 const u32 search[4] =
444 digests_buf[digests_offset].digest_buf[DGST_R0],
445 digests_buf[digests_offset].digest_buf[DGST_R1],
446 digests_buf[digests_offset].digest_buf[DGST_R2],
447 digests_buf[digests_offset].digest_buf[DGST_R3]
454 for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
461 apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
468 w0[1] = w0[1] & 0xff;
491 md5_transform (w0, w1, w2, w3, digest);
495 rc4_init_16 (rc4_key, digest);
499 u8 j = rc4_next_16 (rc4_key, 0, 0, encryptedVerifier, out);
523 md5_transform (w0, w1, w2, w3, digest);
525 rc4_next_16 (rc4_key, 16, j, digest, out);
527 COMPARE_S_SIMD (out[0], out[1], out[2], out[3]);
531 __kernel void m09710_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 oldoffice01_t *oldoffice01_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
535 __kernel void m09710_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 oldoffice01_t *oldoffice01_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 il_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)