2 * Author......: Jens Steube <jens.steube@gmail.com>
8 //too much register pressure
9 //#define NEW_SIMD_CODE
11 #include "inc_vendor.cl"
12 #include "inc_hash_constants.h"
13 #include "inc_hash_functions.cl"
14 #include "inc_types.cl"
15 #include "inc_common.cl"
18 #include "inc_simd.cl"
28 void swap (__local RC4_KEY *rc4_key, const u8 i, const u8 j)
33 rc4_key->S[i] = rc4_key->S[j];
37 void rc4_init_16 (__local RC4_KEY *rc4_key, const u32 data[4])
42 __local u32 *ptr = (__local u32 *) rc4_key->S;
47 for (u32 i = 0; i < 64; i++)
54 for (u32 i = 0; i < 16; i++)
62 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
63 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
64 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
65 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
69 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
70 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
71 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
72 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
76 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
77 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
78 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
79 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
83 j += rc4_key->S[idx] + (v >> 0); swap (rc4_key, idx, j); idx++;
84 j += rc4_key->S[idx] + (v >> 8); swap (rc4_key, idx, j); idx++;
85 j += rc4_key->S[idx] + (v >> 16); swap (rc4_key, idx, j); idx++;
86 j += rc4_key->S[idx] + (v >> 24); swap (rc4_key, idx, j); idx++;
90 u8 rc4_next_16 (__local RC4_KEY *rc4_key, u8 i, u8 j, const u32 in[4], u32 out[4])
95 for (u32 k = 0; k < 4; k++)
104 swap (rc4_key, i, j);
106 idx = rc4_key->S[i] + rc4_key->S[j];
108 xor4 |= rc4_key->S[idx] << 0;
113 swap (rc4_key, i, j);
115 idx = rc4_key->S[i] + rc4_key->S[j];
117 xor4 |= rc4_key->S[idx] << 8;
122 swap (rc4_key, i, j);
124 idx = rc4_key->S[i] + rc4_key->S[j];
126 xor4 |= rc4_key->S[idx] << 16;
131 swap (rc4_key, i, j);
133 idx = rc4_key->S[i] + rc4_key->S[j];
135 xor4 |= rc4_key->S[idx] << 24;
137 out[k] = in[k] ^ xor4;
143 void md5_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[4])
167 MD5_STEP (MD5_Fo, a, b, c, d, w0_t, MD5C00, MD5S00);
168 MD5_STEP (MD5_Fo, d, a, b, c, w1_t, MD5C01, MD5S01);
169 MD5_STEP (MD5_Fo, c, d, a, b, w2_t, MD5C02, MD5S02);
170 MD5_STEP (MD5_Fo, b, c, d, a, w3_t, MD5C03, MD5S03);
171 MD5_STEP (MD5_Fo, a, b, c, d, w4_t, MD5C04, MD5S00);
172 MD5_STEP (MD5_Fo, d, a, b, c, w5_t, MD5C05, MD5S01);
173 MD5_STEP (MD5_Fo, c, d, a, b, w6_t, MD5C06, MD5S02);
174 MD5_STEP (MD5_Fo, b, c, d, a, w7_t, MD5C07, MD5S03);
175 MD5_STEP (MD5_Fo, a, b, c, d, w8_t, MD5C08, MD5S00);
176 MD5_STEP (MD5_Fo, d, a, b, c, w9_t, MD5C09, MD5S01);
177 MD5_STEP (MD5_Fo, c, d, a, b, wa_t, MD5C0a, MD5S02);
178 MD5_STEP (MD5_Fo, b, c, d, a, wb_t, MD5C0b, MD5S03);
179 MD5_STEP (MD5_Fo, a, b, c, d, wc_t, MD5C0c, MD5S00);
180 MD5_STEP (MD5_Fo, d, a, b, c, wd_t, MD5C0d, MD5S01);
181 MD5_STEP (MD5_Fo, c, d, a, b, we_t, MD5C0e, MD5S02);
182 MD5_STEP (MD5_Fo, b, c, d, a, wf_t, MD5C0f, MD5S03);
184 MD5_STEP (MD5_Go, a, b, c, d, w1_t, MD5C10, MD5S10);
185 MD5_STEP (MD5_Go, d, a, b, c, w6_t, MD5C11, MD5S11);
186 MD5_STEP (MD5_Go, c, d, a, b, wb_t, MD5C12, MD5S12);
187 MD5_STEP (MD5_Go, b, c, d, a, w0_t, MD5C13, MD5S13);
188 MD5_STEP (MD5_Go, a, b, c, d, w5_t, MD5C14, MD5S10);
189 MD5_STEP (MD5_Go, d, a, b, c, wa_t, MD5C15, MD5S11);
190 MD5_STEP (MD5_Go, c, d, a, b, wf_t, MD5C16, MD5S12);
191 MD5_STEP (MD5_Go, b, c, d, a, w4_t, MD5C17, MD5S13);
192 MD5_STEP (MD5_Go, a, b, c, d, w9_t, MD5C18, MD5S10);
193 MD5_STEP (MD5_Go, d, a, b, c, we_t, MD5C19, MD5S11);
194 MD5_STEP (MD5_Go, c, d, a, b, w3_t, MD5C1a, MD5S12);
195 MD5_STEP (MD5_Go, b, c, d, a, w8_t, MD5C1b, MD5S13);
196 MD5_STEP (MD5_Go, a, b, c, d, wd_t, MD5C1c, MD5S10);
197 MD5_STEP (MD5_Go, d, a, b, c, w2_t, MD5C1d, MD5S11);
198 MD5_STEP (MD5_Go, c, d, a, b, w7_t, MD5C1e, MD5S12);
199 MD5_STEP (MD5_Go, b, c, d, a, wc_t, MD5C1f, MD5S13);
201 MD5_STEP (MD5_H , a, b, c, d, w5_t, MD5C20, MD5S20);
202 MD5_STEP (MD5_H , d, a, b, c, w8_t, MD5C21, MD5S21);
203 MD5_STEP (MD5_H , c, d, a, b, wb_t, MD5C22, MD5S22);
204 MD5_STEP (MD5_H , b, c, d, a, we_t, MD5C23, MD5S23);
205 MD5_STEP (MD5_H , a, b, c, d, w1_t, MD5C24, MD5S20);
206 MD5_STEP (MD5_H , d, a, b, c, w4_t, MD5C25, MD5S21);
207 MD5_STEP (MD5_H , c, d, a, b, w7_t, MD5C26, MD5S22);
208 MD5_STEP (MD5_H , b, c, d, a, wa_t, MD5C27, MD5S23);
209 MD5_STEP (MD5_H , a, b, c, d, wd_t, MD5C28, MD5S20);
210 MD5_STEP (MD5_H , d, a, b, c, w0_t, MD5C29, MD5S21);
211 MD5_STEP (MD5_H , c, d, a, b, w3_t, MD5C2a, MD5S22);
212 MD5_STEP (MD5_H , b, c, d, a, w6_t, MD5C2b, MD5S23);
213 MD5_STEP (MD5_H , a, b, c, d, w9_t, MD5C2c, MD5S20);
214 MD5_STEP (MD5_H , d, a, b, c, wc_t, MD5C2d, MD5S21);
215 MD5_STEP (MD5_H , c, d, a, b, wf_t, MD5C2e, MD5S22);
216 MD5_STEP (MD5_H , b, c, d, a, w2_t, MD5C2f, MD5S23);
218 MD5_STEP (MD5_I , a, b, c, d, w0_t, MD5C30, MD5S30);
219 MD5_STEP (MD5_I , d, a, b, c, w7_t, MD5C31, MD5S31);
220 MD5_STEP (MD5_I , c, d, a, b, we_t, MD5C32, MD5S32);
221 MD5_STEP (MD5_I , b, c, d, a, w5_t, MD5C33, MD5S33);
222 MD5_STEP (MD5_I , a, b, c, d, wc_t, MD5C34, MD5S30);
223 MD5_STEP (MD5_I , d, a, b, c, w3_t, MD5C35, MD5S31);
224 MD5_STEP (MD5_I , c, d, a, b, wa_t, MD5C36, MD5S32);
225 MD5_STEP (MD5_I , b, c, d, a, w1_t, MD5C37, MD5S33);
226 MD5_STEP (MD5_I , a, b, c, d, w8_t, MD5C38, MD5S30);
227 MD5_STEP (MD5_I , d, a, b, c, wf_t, MD5C39, MD5S31);
228 MD5_STEP (MD5_I , c, d, a, b, w6_t, MD5C3a, MD5S32);
229 MD5_STEP (MD5_I , b, c, d, a, wd_t, MD5C3b, MD5S33);
230 MD5_STEP (MD5_I , a, b, c, d, w4_t, MD5C3c, MD5S30);
231 MD5_STEP (MD5_I , d, a, b, c, wb_t, MD5C3d, MD5S31);
232 MD5_STEP (MD5_I , c, d, a, b, w2_t, MD5C3e, MD5S32);
233 MD5_STEP (MD5_I , b, c, d, a, w9_t, MD5C3f, MD5S33);
241 __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_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)
247 const u32 lid = get_local_id (0);
253 const u32 gid = get_global_id (0);
255 if (gid >= gid_max) return;
260 pw_buf0[0] = pws[gid].i[ 0];
261 pw_buf0[1] = pws[gid].i[ 1];
262 pw_buf0[2] = pws[gid].i[ 2];
263 pw_buf0[3] = pws[gid].i[ 3];
264 pw_buf1[0] = pws[gid].i[ 4];
265 pw_buf1[1] = pws[gid].i[ 5];
266 pw_buf1[2] = pws[gid].i[ 6];
267 pw_buf1[3] = pws[gid].i[ 7];
269 const u32 pw_len = pws[gid].pw_len;
275 __local RC4_KEY rc4_keys[64];
277 __local RC4_KEY *rc4_key = &rc4_keys[lid];
283 const u32 version = oldoffice01_bufs[salt_pos].version;
285 u32 encryptedVerifier[4];
287 encryptedVerifier[0] = oldoffice01_bufs[salt_pos].encryptedVerifier[0];
288 encryptedVerifier[1] = oldoffice01_bufs[salt_pos].encryptedVerifier[1];
289 encryptedVerifier[2] = oldoffice01_bufs[salt_pos].encryptedVerifier[2];
290 encryptedVerifier[3] = oldoffice01_bufs[salt_pos].encryptedVerifier[3];
296 for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
303 apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
310 w0[1] = w0[1] & 0xff;
333 md5_transform (w0, w1, w2, w3, digest);
337 rc4_init_16 (rc4_key, digest);
341 u8 j = rc4_next_16 (rc4_key, 0, 0, encryptedVerifier, out);
365 md5_transform (w0, w1, w2, w3, digest);
367 rc4_next_16 (rc4_key, 16, j, digest, out);
369 COMPARE_M_SIMD (out[0], out[1], out[2], out[3]);
373 __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_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)
377 __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_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)
381 __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_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)
387 const u32 lid = get_local_id (0);
393 const u32 gid = get_global_id (0);
395 if (gid >= gid_max) return;
400 pw_buf0[0] = pws[gid].i[ 0];
401 pw_buf0[1] = pws[gid].i[ 1];
402 pw_buf0[2] = pws[gid].i[ 2];
403 pw_buf0[3] = pws[gid].i[ 3];
404 pw_buf1[0] = pws[gid].i[ 4];
405 pw_buf1[1] = pws[gid].i[ 5];
406 pw_buf1[2] = pws[gid].i[ 6];
407 pw_buf1[3] = pws[gid].i[ 7];
409 const u32 pw_len = pws[gid].pw_len;
415 __local RC4_KEY rc4_keys[64];
417 __local RC4_KEY *rc4_key = &rc4_keys[lid];
423 const u32 version = oldoffice01_bufs[salt_pos].version;
425 u32 encryptedVerifier[4];
427 encryptedVerifier[0] = oldoffice01_bufs[salt_pos].encryptedVerifier[0];
428 encryptedVerifier[1] = oldoffice01_bufs[salt_pos].encryptedVerifier[1];
429 encryptedVerifier[2] = oldoffice01_bufs[salt_pos].encryptedVerifier[2];
430 encryptedVerifier[3] = oldoffice01_bufs[salt_pos].encryptedVerifier[3];
436 const u32 search[4] =
438 digests_buf[digests_offset].digest_buf[DGST_R0],
439 digests_buf[digests_offset].digest_buf[DGST_R1],
440 digests_buf[digests_offset].digest_buf[DGST_R2],
441 digests_buf[digests_offset].digest_buf[DGST_R3]
448 for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE)
455 apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1);
462 w0[1] = w0[1] & 0xff;
485 md5_transform (w0, w1, w2, w3, digest);
489 rc4_init_16 (rc4_key, digest);
493 u8 j = rc4_next_16 (rc4_key, 0, 0, encryptedVerifier, out);
517 md5_transform (w0, w1, w2, w3, digest);
519 rc4_next_16 (rc4_key, 16, j, digest, out);
521 COMPARE_S_SIMD (out[0], out[1], out[2], out[3]);
525 __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_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)
529 __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_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)