2 * Author......: Jens Steube <jens.steube@gmail.com>
8 #include "include/constants.h"
9 #include "include/kernel_vendor.h"
28 #include "include/kernel_functions.c"
29 #include "types_amd.c"
30 #include "common_amd.c"
33 #define VECT_COMPARE_M "check_multi_vect1_comp4.c"
37 #define VECT_COMPARE_M "check_multi_vect2_comp4.c"
41 #define VECT_COMPARE_M "check_multi_vect4_comp4.c"
44 __constant u32 k_sha256[64] =
46 SHA256C00, SHA256C01, SHA256C02, SHA256C03,
47 SHA256C04, SHA256C05, SHA256C06, SHA256C07,
48 SHA256C08, SHA256C09, SHA256C0a, SHA256C0b,
49 SHA256C0c, SHA256C0d, SHA256C0e, SHA256C0f,
50 SHA256C10, SHA256C11, SHA256C12, SHA256C13,
51 SHA256C14, SHA256C15, SHA256C16, SHA256C17,
52 SHA256C18, SHA256C19, SHA256C1a, SHA256C1b,
53 SHA256C1c, SHA256C1d, SHA256C1e, SHA256C1f,
54 SHA256C20, SHA256C21, SHA256C22, SHA256C23,
55 SHA256C24, SHA256C25, SHA256C26, SHA256C27,
56 SHA256C28, SHA256C29, SHA256C2a, SHA256C2b,
57 SHA256C2c, SHA256C2d, SHA256C2e, SHA256C2f,
58 SHA256C30, SHA256C31, SHA256C32, SHA256C33,
59 SHA256C34, SHA256C35, SHA256C36, SHA256C37,
60 SHA256C38, SHA256C39, SHA256C3a, SHA256C3b,
61 SHA256C3c, SHA256C3d, SHA256C3e, SHA256C3f,
64 static void sha256_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[8])
92 #define ROUND_EXPAND() \
94 w0_t = SHA256_EXPAND (we_t, w9_t, w1_t, w0_t); \
95 w1_t = SHA256_EXPAND (wf_t, wa_t, w2_t, w1_t); \
96 w2_t = SHA256_EXPAND (w0_t, wb_t, w3_t, w2_t); \
97 w3_t = SHA256_EXPAND (w1_t, wc_t, w4_t, w3_t); \
98 w4_t = SHA256_EXPAND (w2_t, wd_t, w5_t, w4_t); \
99 w5_t = SHA256_EXPAND (w3_t, we_t, w6_t, w5_t); \
100 w6_t = SHA256_EXPAND (w4_t, wf_t, w7_t, w6_t); \
101 w7_t = SHA256_EXPAND (w5_t, w0_t, w8_t, w7_t); \
102 w8_t = SHA256_EXPAND (w6_t, w1_t, w9_t, w8_t); \
103 w9_t = SHA256_EXPAND (w7_t, w2_t, wa_t, w9_t); \
104 wa_t = SHA256_EXPAND (w8_t, w3_t, wb_t, wa_t); \
105 wb_t = SHA256_EXPAND (w9_t, w4_t, wc_t, wb_t); \
106 wc_t = SHA256_EXPAND (wa_t, w5_t, wd_t, wc_t); \
107 wd_t = SHA256_EXPAND (wb_t, w6_t, we_t, wd_t); \
108 we_t = SHA256_EXPAND (wc_t, w7_t, wf_t, we_t); \
109 wf_t = SHA256_EXPAND (wd_t, w8_t, w0_t, wf_t); \
112 #define ROUND_STEP(i) \
114 SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha256[i + 0]); \
115 SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha256[i + 1]); \
116 SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha256[i + 2]); \
117 SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha256[i + 3]); \
118 SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha256[i + 4]); \
119 SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha256[i + 5]); \
120 SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha256[i + 6]); \
121 SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha256[i + 7]); \
122 SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha256[i + 8]); \
123 SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha256[i + 9]); \
124 SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha256[i + 10]); \
125 SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha256[i + 11]); \
126 SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha256[i + 12]); \
127 SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha256[i + 13]); \
128 SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, k_sha256[i + 14]); \
129 SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha256[i + 15]); \
135 for (int i = 16; i < 64; i += 16)
137 ROUND_EXPAND (); ROUND_STEP (i);
150 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m05200_init (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global pwsafe3_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 void *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)
156 const u32 gid = get_global_id (0);
158 if (gid >= gid_max) return;
162 w0[0] = pws[gid].i[ 0];
163 w0[1] = pws[gid].i[ 1];
164 w0[2] = pws[gid].i[ 2];
165 w0[3] = pws[gid].i[ 3];
169 w1[0] = pws[gid].i[ 4];
170 w1[1] = pws[gid].i[ 5];
188 const u32 pw_len = pws[gid].pw_len;
196 salt_buf0[0] = salt_bufs[salt_pos].salt_buf[0];
197 salt_buf0[1] = salt_bufs[salt_pos].salt_buf[1];
198 salt_buf0[2] = salt_bufs[salt_pos].salt_buf[2];
199 salt_buf0[3] = salt_bufs[salt_pos].salt_buf[3];
203 salt_buf1[0] = salt_bufs[salt_pos].salt_buf[4];
204 salt_buf1[1] = salt_bufs[salt_pos].salt_buf[5];
205 salt_buf1[2] = salt_bufs[salt_pos].salt_buf[6];
206 salt_buf1[3] = salt_bufs[salt_pos].salt_buf[7];
222 u32 salt_len = salt_bufs[salt_pos].salt_len;
224 switch_buffer_by_offset (salt_buf0, salt_buf1, salt_buf2, salt_buf3, pw_len);
226 w0[0] |= salt_buf0[0];
227 w0[1] |= salt_buf0[1];
228 w0[2] |= salt_buf0[2];
229 w0[3] |= salt_buf0[3];
231 w1[0] |= salt_buf1[0];
232 w1[1] |= salt_buf1[1];
233 w1[2] |= salt_buf1[2];
234 w1[3] |= salt_buf1[3];
236 w2[0] |= salt_buf2[0];
237 w2[1] |= salt_buf2[1];
238 w2[2] |= salt_buf2[2];
239 w2[3] |= salt_buf2[3];
241 w3[0] |= salt_buf3[0];
242 w3[1] |= salt_buf3[1];
243 w3[2] |= salt_buf3[2];
244 w3[3] |= salt_buf3[3];
246 const u32 block_len = pw_len + salt_len;
248 append_0x80_4 (w0, w1, w2, w3, block_len);
254 w0[0] = swap_workaround (w0[0]);
255 w0[1] = swap_workaround (w0[1]);
256 w0[2] = swap_workaround (w0[2]);
257 w0[3] = swap_workaround (w0[3]);
259 w1[0] = swap_workaround (w1[0]);
260 w1[1] = swap_workaround (w1[1]);
261 w1[2] = swap_workaround (w1[2]);
262 w1[3] = swap_workaround (w1[3]);
264 w2[0] = swap_workaround (w2[0]);
265 w2[1] = swap_workaround (w2[1]);
266 w2[2] = swap_workaround (w2[2]);
267 w2[3] = swap_workaround (w2[3]);
269 w3[0] = swap_workaround (w3[0]);
270 w3[1] = swap_workaround (w3[1]);
272 w3[3] = block_len * 8;
280 digest[0] = SHA256M_A;
281 digest[1] = SHA256M_B;
282 digest[2] = SHA256M_C;
283 digest[3] = SHA256M_D;
284 digest[4] = SHA256M_E;
285 digest[5] = SHA256M_F;
286 digest[6] = SHA256M_G;
287 digest[7] = SHA256M_H;
289 sha256_transform (w0, w1, w2, w3, digest);
295 tmps[gid].digest_buf[0] = digest[0];
296 tmps[gid].digest_buf[1] = digest[1];
297 tmps[gid].digest_buf[2] = digest[2];
298 tmps[gid].digest_buf[3] = digest[3];
299 tmps[gid].digest_buf[4] = digest[4];
300 tmps[gid].digest_buf[5] = digest[5];
301 tmps[gid].digest_buf[6] = digest[6];
302 tmps[gid].digest_buf[7] = digest[7];
305 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m05200_loop (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global pwsafe3_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 void *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)
307 const u32 gid = get_global_id (0);
309 if (gid >= gid_max) return;
313 digest[0] = tmps[gid].digest_buf[0];
314 digest[1] = tmps[gid].digest_buf[1];
315 digest[2] = tmps[gid].digest_buf[2];
316 digest[3] = tmps[gid].digest_buf[3];
317 digest[4] = tmps[gid].digest_buf[4];
318 digest[5] = tmps[gid].digest_buf[5];
319 digest[6] = tmps[gid].digest_buf[6];
320 digest[7] = tmps[gid].digest_buf[7];
326 for (u32 i = 0; i < loop_cnt; i++)
356 digest[0] = SHA256M_A;
357 digest[1] = SHA256M_B;
358 digest[2] = SHA256M_C;
359 digest[3] = SHA256M_D;
360 digest[4] = SHA256M_E;
361 digest[5] = SHA256M_F;
362 digest[6] = SHA256M_G;
363 digest[7] = SHA256M_H;
365 sha256_transform (w0, w1, w2, w3, digest);
368 tmps[gid].digest_buf[0] = digest[0];
369 tmps[gid].digest_buf[1] = digest[1];
370 tmps[gid].digest_buf[2] = digest[2];
371 tmps[gid].digest_buf[3] = digest[3];
372 tmps[gid].digest_buf[4] = digest[4];
373 tmps[gid].digest_buf[5] = digest[5];
374 tmps[gid].digest_buf[6] = digest[6];
375 tmps[gid].digest_buf[7] = digest[7];
378 __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m05200_comp (__global pw_t *pws, __global gpu_rule_t *rules_buf, __global comb_t *combs_buf, __global bf_t *bfs_buf, __global pwsafe3_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 void *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)
384 const u32 gid = get_global_id (0);
386 if (gid >= gid_max) return;
388 const u32 lid = get_local_id (0);
394 const u32x r0 = tmps[gid].digest_buf[DGST_R0];
395 const u32x r1 = tmps[gid].digest_buf[DGST_R1];
396 const u32x r2 = tmps[gid].digest_buf[DGST_R2];
397 const u32x r3 = tmps[gid].digest_buf[DGST_R3];
401 #include VECT_COMPARE_M