mirror of
https://github.com/hashcat/hashcat.git
synced 2024-11-26 09:58:16 +00:00
626 lines
22 KiB
Plaintext
626 lines
22 KiB
Plaintext
|
/**
|
||
|
* Author......: Jens Steube <jens.steube@gmail.com>
|
||
|
* License.....: MIT
|
||
|
*/
|
||
|
|
||
|
#define _OLDOFFICE34_
|
||
|
|
||
|
#include "include/constants.h"
|
||
|
#include "include/kernel_vendor.h"
|
||
|
|
||
|
#ifdef VLIW1
|
||
|
#define VECT_SIZE1
|
||
|
#endif
|
||
|
|
||
|
#ifdef VLIW2
|
||
|
#define VECT_SIZE1
|
||
|
#endif
|
||
|
|
||
|
#define DGST_R0 0
|
||
|
#define DGST_R1 1
|
||
|
#define DGST_R2 2
|
||
|
#define DGST_R3 3
|
||
|
|
||
|
#include "include/kernel_functions.c"
|
||
|
#include "types_nv.c"
|
||
|
#include "common_nv.c"
|
||
|
|
||
|
#ifdef VECT_SIZE1
|
||
|
#define VECT_COMPARE_S "check_single_vect1_comp4.c"
|
||
|
#define VECT_COMPARE_M "check_multi_vect1_comp4.c"
|
||
|
#endif
|
||
|
|
||
|
#ifdef VECT_SIZE2
|
||
|
#define VECT_COMPARE_S "check_single_vect2_comp4.c"
|
||
|
#define VECT_COMPARE_M "check_multi_vect2_comp4.c"
|
||
|
#endif
|
||
|
|
||
|
#ifdef VECT_SIZE4
|
||
|
#define VECT_COMPARE_S "check_single_vect4_comp4.c"
|
||
|
#define VECT_COMPARE_M "check_multi_vect4_comp4.c"
|
||
|
#endif
|
||
|
|
||
|
__device__ static void sha1_transform (const u32x w0[4], const u32x w1[4], const u32x w2[4], const u32x w3[4], u32x digest[5])
|
||
|
{
|
||
|
u32x A = digest[0];
|
||
|
u32x B = digest[1];
|
||
|
u32x C = digest[2];
|
||
|
u32x D = digest[3];
|
||
|
u32x E = digest[4];
|
||
|
|
||
|
u32x w0_t = w0[0];
|
||
|
u32x w1_t = w0[1];
|
||
|
u32x w2_t = w0[2];
|
||
|
u32x w3_t = w0[3];
|
||
|
u32x w4_t = w1[0];
|
||
|
u32x w5_t = w1[1];
|
||
|
u32x w6_t = w1[2];
|
||
|
u32x w7_t = w1[3];
|
||
|
u32x w8_t = w2[0];
|
||
|
u32x w9_t = w2[1];
|
||
|
u32x wa_t = w2[2];
|
||
|
u32x wb_t = w2[3];
|
||
|
u32x wc_t = w3[0];
|
||
|
u32x wd_t = w3[1];
|
||
|
u32x we_t = w3[2];
|
||
|
u32x wf_t = w3[3];
|
||
|
|
||
|
#undef K
|
||
|
#define K SHA1C00
|
||
|
|
||
|
SHA1_STEP (SHA1_F0o, A, B, C, D, E, w0_t);
|
||
|
SHA1_STEP (SHA1_F0o, E, A, B, C, D, w1_t);
|
||
|
SHA1_STEP (SHA1_F0o, D, E, A, B, C, w2_t);
|
||
|
SHA1_STEP (SHA1_F0o, C, D, E, A, B, w3_t);
|
||
|
SHA1_STEP (SHA1_F0o, B, C, D, E, A, w4_t);
|
||
|
SHA1_STEP (SHA1_F0o, A, B, C, D, E, w5_t);
|
||
|
SHA1_STEP (SHA1_F0o, E, A, B, C, D, w6_t);
|
||
|
SHA1_STEP (SHA1_F0o, D, E, A, B, C, w7_t);
|
||
|
SHA1_STEP (SHA1_F0o, C, D, E, A, B, w8_t);
|
||
|
SHA1_STEP (SHA1_F0o, B, C, D, E, A, w9_t);
|
||
|
SHA1_STEP (SHA1_F0o, A, B, C, D, E, wa_t);
|
||
|
SHA1_STEP (SHA1_F0o, E, A, B, C, D, wb_t);
|
||
|
SHA1_STEP (SHA1_F0o, D, E, A, B, C, wc_t);
|
||
|
SHA1_STEP (SHA1_F0o, C, D, E, A, B, wd_t);
|
||
|
SHA1_STEP (SHA1_F0o, B, C, D, E, A, we_t);
|
||
|
SHA1_STEP (SHA1_F0o, A, B, C, D, E, wf_t);
|
||
|
w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F0o, E, A, B, C, D, w0_t);
|
||
|
w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F0o, D, E, A, B, C, w1_t);
|
||
|
w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F0o, C, D, E, A, B, w2_t);
|
||
|
w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F0o, B, C, D, E, A, w3_t);
|
||
|
|
||
|
#undef K
|
||
|
#define K SHA1C01
|
||
|
|
||
|
w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w4_t);
|
||
|
w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w5_t);
|
||
|
w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w6_t);
|
||
|
w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w7_t);
|
||
|
w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w8_t);
|
||
|
w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w9_t);
|
||
|
wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wa_t);
|
||
|
wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wb_t);
|
||
|
wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wc_t);
|
||
|
wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wd_t);
|
||
|
we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, we_t);
|
||
|
wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wf_t);
|
||
|
w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w0_t);
|
||
|
w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w1_t);
|
||
|
w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w2_t);
|
||
|
w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w3_t);
|
||
|
w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w4_t);
|
||
|
w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w5_t);
|
||
|
w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w6_t);
|
||
|
w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w7_t);
|
||
|
|
||
|
#undef K
|
||
|
#define K SHA1C02
|
||
|
|
||
|
w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w8_t);
|
||
|
w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w9_t);
|
||
|
wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wa_t);
|
||
|
wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wb_t);
|
||
|
wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wc_t);
|
||
|
wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, wd_t);
|
||
|
we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, we_t);
|
||
|
wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, wf_t);
|
||
|
w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w0_t);
|
||
|
w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w1_t);
|
||
|
w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w2_t);
|
||
|
w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w3_t);
|
||
|
w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w4_t);
|
||
|
w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, w5_t);
|
||
|
w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, w6_t);
|
||
|
w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F2o, A, B, C, D, E, w7_t);
|
||
|
w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F2o, E, A, B, C, D, w8_t);
|
||
|
w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F2o, D, E, A, B, C, w9_t);
|
||
|
wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F2o, C, D, E, A, B, wa_t);
|
||
|
wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F2o, B, C, D, E, A, wb_t);
|
||
|
|
||
|
#undef K
|
||
|
#define K SHA1C03
|
||
|
|
||
|
wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wc_t);
|
||
|
wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wd_t);
|
||
|
we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, we_t);
|
||
|
wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, wf_t);
|
||
|
w0_t = rotl32 ((wd_t ^ w8_t ^ w2_t ^ w0_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w0_t);
|
||
|
w1_t = rotl32 ((we_t ^ w9_t ^ w3_t ^ w1_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w1_t);
|
||
|
w2_t = rotl32 ((wf_t ^ wa_t ^ w4_t ^ w2_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w2_t);
|
||
|
w3_t = rotl32 ((w0_t ^ wb_t ^ w5_t ^ w3_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w3_t);
|
||
|
w4_t = rotl32 ((w1_t ^ wc_t ^ w6_t ^ w4_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w4_t);
|
||
|
w5_t = rotl32 ((w2_t ^ wd_t ^ w7_t ^ w5_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, w5_t);
|
||
|
w6_t = rotl32 ((w3_t ^ we_t ^ w8_t ^ w6_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, w6_t);
|
||
|
w7_t = rotl32 ((w4_t ^ wf_t ^ w9_t ^ w7_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, w7_t);
|
||
|
w8_t = rotl32 ((w5_t ^ w0_t ^ wa_t ^ w8_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, w8_t);
|
||
|
w9_t = rotl32 ((w6_t ^ w1_t ^ wb_t ^ w9_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, w9_t);
|
||
|
wa_t = rotl32 ((w7_t ^ w2_t ^ wc_t ^ wa_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wa_t);
|
||
|
wb_t = rotl32 ((w8_t ^ w3_t ^ wd_t ^ wb_t), 1u); SHA1_STEP (SHA1_F1, A, B, C, D, E, wb_t);
|
||
|
wc_t = rotl32 ((w9_t ^ w4_t ^ we_t ^ wc_t), 1u); SHA1_STEP (SHA1_F1, E, A, B, C, D, wc_t);
|
||
|
wd_t = rotl32 ((wa_t ^ w5_t ^ wf_t ^ wd_t), 1u); SHA1_STEP (SHA1_F1, D, E, A, B, C, wd_t);
|
||
|
we_t = rotl32 ((wb_t ^ w6_t ^ w0_t ^ we_t), 1u); SHA1_STEP (SHA1_F1, C, D, E, A, B, we_t);
|
||
|
wf_t = rotl32 ((wc_t ^ w7_t ^ w1_t ^ wf_t), 1u); SHA1_STEP (SHA1_F1, B, C, D, E, A, wf_t);
|
||
|
|
||
|
digest[0] += A;
|
||
|
digest[1] += B;
|
||
|
digest[2] += C;
|
||
|
digest[3] += D;
|
||
|
digest[4] += E;
|
||
|
}
|
||
|
|
||
|
__device__ __constant__ comb_t c_combs[1024];
|
||
|
|
||
|
extern "C" __global__ void __launch_bounds__ (256, 1) m09820_m04 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const oldoffice34_t *oldoffice34_bufs, u32 *d_return_buf, 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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
|
||
|
{
|
||
|
/**
|
||
|
* modifier
|
||
|
*/
|
||
|
|
||
|
const u32 lid = threadIdx.x;
|
||
|
|
||
|
/**
|
||
|
* base
|
||
|
*/
|
||
|
|
||
|
const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
|
||
|
|
||
|
if (gid >= gid_max) return;
|
||
|
|
||
|
u32x wordl0[4];
|
||
|
|
||
|
wordl0[0] = pws[gid].i[ 0];
|
||
|
wordl0[1] = pws[gid].i[ 1];
|
||
|
wordl0[2] = pws[gid].i[ 2];
|
||
|
wordl0[3] = pws[gid].i[ 3];
|
||
|
|
||
|
u32x wordl1[4];
|
||
|
|
||
|
wordl1[0] = pws[gid].i[ 4];
|
||
|
wordl1[1] = pws[gid].i[ 5];
|
||
|
wordl1[2] = pws[gid].i[ 6];
|
||
|
wordl1[3] = pws[gid].i[ 7];
|
||
|
|
||
|
u32x wordl2[4];
|
||
|
|
||
|
wordl2[0] = 0;
|
||
|
wordl2[1] = 0;
|
||
|
wordl2[2] = 0;
|
||
|
wordl2[3] = 0;
|
||
|
|
||
|
u32x wordl3[4];
|
||
|
|
||
|
wordl3[0] = 0;
|
||
|
wordl3[1] = 0;
|
||
|
wordl3[2] = 0;
|
||
|
wordl3[3] = 0;
|
||
|
|
||
|
const u32 pw_l_len = pws[gid].pw_len;
|
||
|
|
||
|
if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
|
||
|
{
|
||
|
switch_buffer_by_offset (wordl0, wordl1, wordl2, wordl3, c_combs[0].pw_len);
|
||
|
}
|
||
|
|
||
|
/**
|
||
|
* salt
|
||
|
*/
|
||
|
|
||
|
u32 salt_buf[4];
|
||
|
|
||
|
salt_buf[0] = salt_bufs[salt_pos].salt_buf[0];
|
||
|
salt_buf[1] = salt_bufs[salt_pos].salt_buf[1];
|
||
|
salt_buf[2] = salt_bufs[salt_pos].salt_buf[2];
|
||
|
salt_buf[3] = salt_bufs[salt_pos].salt_buf[3];
|
||
|
|
||
|
const u32 salt_len = 16;
|
||
|
|
||
|
/**
|
||
|
* loop
|
||
|
*/
|
||
|
|
||
|
for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
|
||
|
{
|
||
|
const u32 pw_r_len = c_combs[il_pos].pw_len;
|
||
|
|
||
|
const u32 pw_len = pw_l_len + pw_r_len;
|
||
|
|
||
|
const u32 pw_salt_len = (pw_len * 2) + salt_len;
|
||
|
|
||
|
u32 wordr0[4];
|
||
|
|
||
|
wordr0[0] = c_combs[il_pos].i[0];
|
||
|
wordr0[1] = c_combs[il_pos].i[1];
|
||
|
wordr0[2] = c_combs[il_pos].i[2];
|
||
|
wordr0[3] = c_combs[il_pos].i[3];
|
||
|
|
||
|
u32 wordr1[4];
|
||
|
|
||
|
wordr1[0] = c_combs[il_pos].i[4];
|
||
|
wordr1[1] = c_combs[il_pos].i[5];
|
||
|
wordr1[2] = c_combs[il_pos].i[6];
|
||
|
wordr1[3] = c_combs[il_pos].i[7];
|
||
|
|
||
|
u32 wordr2[4];
|
||
|
|
||
|
wordr2[0] = 0;
|
||
|
wordr2[1] = 0;
|
||
|
wordr2[2] = 0;
|
||
|
wordr2[3] = 0;
|
||
|
|
||
|
u32 wordr3[4];
|
||
|
|
||
|
wordr3[0] = 0;
|
||
|
wordr3[1] = 0;
|
||
|
wordr3[2] = 0;
|
||
|
wordr3[3] = 0;
|
||
|
|
||
|
if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
|
||
|
{
|
||
|
switch_buffer_by_offset (wordr0, wordr1, wordr2, wordr3, pw_l_len);
|
||
|
}
|
||
|
|
||
|
u32x w0[4];
|
||
|
|
||
|
w0[0] = wordl0[0] | wordr0[0];
|
||
|
w0[1] = wordl0[1] | wordr0[1];
|
||
|
w0[2] = wordl0[2] | wordr0[2];
|
||
|
w0[3] = wordl0[3] | wordr0[3];
|
||
|
|
||
|
u32x w1[4];
|
||
|
|
||
|
w1[0] = wordl1[0] | wordr1[0];
|
||
|
w1[1] = wordl1[1] | wordr1[1];
|
||
|
w1[2] = wordl1[2] | wordr1[2];
|
||
|
w1[3] = wordl1[3] | wordr1[3];
|
||
|
|
||
|
u32x w2[4];
|
||
|
|
||
|
w2[0] = wordl2[0] | wordr2[0];
|
||
|
w2[1] = wordl2[1] | wordr2[1];
|
||
|
w2[2] = wordl2[2] | wordr2[2];
|
||
|
w2[3] = wordl2[3] | wordr2[3];
|
||
|
|
||
|
u32x w3[4];
|
||
|
|
||
|
w3[0] = wordl3[0] | wordr3[0];
|
||
|
w3[1] = wordl3[1] | wordr3[1];
|
||
|
w3[2] = 0;
|
||
|
w3[3] = 0;
|
||
|
|
||
|
append_0x80_2 (w0, w1, pw_len);
|
||
|
|
||
|
u32x w0_t[4];
|
||
|
u32x w1_t[4];
|
||
|
u32x w2_t[4];
|
||
|
u32x w3_t[4];
|
||
|
|
||
|
make_unicode (w0, w0_t, w1_t);
|
||
|
make_unicode (w1, w2_t, w3_t);
|
||
|
|
||
|
switch_buffer_by_offset (w0_t, w1_t, w2_t, w3_t, salt_len);
|
||
|
|
||
|
w0_t[0] = salt_buf[0];
|
||
|
w0_t[1] = salt_buf[1];
|
||
|
w0_t[2] = salt_buf[2];
|
||
|
w0_t[3] = salt_buf[3];
|
||
|
w1_t[0] = swap_workaround (w1_t[0]);
|
||
|
w1_t[1] = swap_workaround (w1_t[1]);
|
||
|
w1_t[2] = swap_workaround (w1_t[2]);
|
||
|
w1_t[3] = swap_workaround (w1_t[3]);
|
||
|
w2_t[0] = swap_workaround (w2_t[0]);
|
||
|
w2_t[1] = swap_workaround (w2_t[1]);
|
||
|
w2_t[2] = swap_workaround (w2_t[2]);
|
||
|
w2_t[3] = swap_workaround (w2_t[3]);
|
||
|
w3_t[0] = swap_workaround (w3_t[0]);
|
||
|
w3_t[1] = swap_workaround (w3_t[1]);
|
||
|
w3_t[2] = 0;
|
||
|
w3_t[3] = pw_salt_len * 8;
|
||
|
|
||
|
u32x digest[5];
|
||
|
|
||
|
digest[0] = SHA1M_A;
|
||
|
digest[1] = SHA1M_B;
|
||
|
digest[2] = SHA1M_C;
|
||
|
digest[3] = SHA1M_D;
|
||
|
digest[4] = SHA1M_E;
|
||
|
|
||
|
sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
|
||
|
|
||
|
w0_t[0] = digest[0];
|
||
|
w0_t[1] = digest[1];
|
||
|
w0_t[2] = digest[2];
|
||
|
w0_t[3] = digest[3];
|
||
|
w1_t[0] = digest[4];
|
||
|
w1_t[1] = 0;
|
||
|
w1_t[2] = 0x80000000;
|
||
|
w1_t[3] = 0;
|
||
|
w2_t[0] = 0;
|
||
|
w2_t[1] = 0;
|
||
|
w2_t[2] = 0;
|
||
|
w2_t[3] = 0;
|
||
|
w3_t[0] = 0;
|
||
|
w3_t[1] = 0;
|
||
|
w3_t[2] = 0;
|
||
|
w3_t[3] = (20 + 4) * 8;
|
||
|
|
||
|
digest[0] = SHA1M_A;
|
||
|
digest[1] = SHA1M_B;
|
||
|
digest[2] = SHA1M_C;
|
||
|
digest[3] = SHA1M_D;
|
||
|
digest[4] = SHA1M_E;
|
||
|
|
||
|
sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
|
||
|
|
||
|
u32x a = swap_workaround (digest[0]);
|
||
|
u32x b = swap_workaround (digest[1]) & 0xff;
|
||
|
|
||
|
const u32x r0 = a;
|
||
|
const u32x r1 = b;
|
||
|
const u32x r2 = 0;
|
||
|
const u32x r3 = 0;
|
||
|
|
||
|
#include VECT_COMPARE_M
|
||
|
}
|
||
|
}
|
||
|
|
||
|
extern "C" __global__ void __launch_bounds__ (256, 1) m09820_m08 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const oldoffice34_t *oldoffice34_bufs, u32 *d_return_buf, 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)
|
||
|
{
|
||
|
}
|
||
|
|
||
|
extern "C" __global__ void __launch_bounds__ (256, 1) m09820_m16 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const oldoffice34_t *oldoffice34_bufs, u32 *d_return_buf, 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)
|
||
|
{
|
||
|
}
|
||
|
|
||
|
extern "C" __global__ void __launch_bounds__ (256, 1) m09820_s04 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const oldoffice34_t *oldoffice34_bufs, u32 *d_return_buf, 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 combs_cnt, const u32 digests_cnt, const u32 digests_offset, const u32 combs_mode, const u32 gid_max)
|
||
|
{
|
||
|
/**
|
||
|
* modifier
|
||
|
*/
|
||
|
|
||
|
const u32 lid = threadIdx.x;
|
||
|
|
||
|
/**
|
||
|
* base
|
||
|
*/
|
||
|
|
||
|
const u32 gid = (blockIdx.x * blockDim.x) + threadIdx.x;
|
||
|
|
||
|
if (gid >= gid_max) return;
|
||
|
|
||
|
u32x wordl0[4];
|
||
|
|
||
|
wordl0[0] = pws[gid].i[ 0];
|
||
|
wordl0[1] = pws[gid].i[ 1];
|
||
|
wordl0[2] = pws[gid].i[ 2];
|
||
|
wordl0[3] = pws[gid].i[ 3];
|
||
|
|
||
|
u32x wordl1[4];
|
||
|
|
||
|
wordl1[0] = pws[gid].i[ 4];
|
||
|
wordl1[1] = pws[gid].i[ 5];
|
||
|
wordl1[2] = pws[gid].i[ 6];
|
||
|
wordl1[3] = pws[gid].i[ 7];
|
||
|
|
||
|
u32x wordl2[4];
|
||
|
|
||
|
wordl2[0] = 0;
|
||
|
wordl2[1] = 0;
|
||
|
wordl2[2] = 0;
|
||
|
wordl2[3] = 0;
|
||
|
|
||
|
u32x wordl3[4];
|
||
|
|
||
|
wordl3[0] = 0;
|
||
|
wordl3[1] = 0;
|
||
|
wordl3[2] = 0;
|
||
|
wordl3[3] = 0;
|
||
|
|
||
|
const u32 pw_l_len = pws[gid].pw_len;
|
||
|
|
||
|
if (combs_mode == COMBINATOR_MODE_BASE_RIGHT)
|
||
|
{
|
||
|
switch_buffer_by_offset (wordl0, wordl1, wordl2, wordl3, c_combs[0].pw_len);
|
||
|
}
|
||
|
|
||
|
/**
|
||
|
* digest
|
||
|
*/
|
||
|
|
||
|
const u32 search[4] =
|
||
|
{
|
||
|
digests_buf[digests_offset].digest_buf[DGST_R0],
|
||
|
digests_buf[digests_offset].digest_buf[DGST_R1],
|
||
|
digests_buf[digests_offset].digest_buf[DGST_R2],
|
||
|
digests_buf[digests_offset].digest_buf[DGST_R3]
|
||
|
};
|
||
|
|
||
|
/**
|
||
|
* salt
|
||
|
*/
|
||
|
|
||
|
u32 salt_buf[4];
|
||
|
|
||
|
salt_buf[0] = salt_bufs[salt_pos].salt_buf[0];
|
||
|
salt_buf[1] = salt_bufs[salt_pos].salt_buf[1];
|
||
|
salt_buf[2] = salt_bufs[salt_pos].salt_buf[2];
|
||
|
salt_buf[3] = salt_bufs[salt_pos].salt_buf[3];
|
||
|
|
||
|
const u32 salt_len = 16;
|
||
|
|
||
|
/**
|
||
|
* loop
|
||
|
*/
|
||
|
|
||
|
for (u32 il_pos = 0; il_pos < combs_cnt; il_pos++)
|
||
|
{
|
||
|
const u32 pw_r_len = c_combs[il_pos].pw_len;
|
||
|
|
||
|
const u32 pw_len = pw_l_len + pw_r_len;
|
||
|
|
||
|
const u32 pw_salt_len = (pw_len * 2) + salt_len;
|
||
|
|
||
|
u32 wordr0[4];
|
||
|
|
||
|
wordr0[0] = c_combs[il_pos].i[0];
|
||
|
wordr0[1] = c_combs[il_pos].i[1];
|
||
|
wordr0[2] = c_combs[il_pos].i[2];
|
||
|
wordr0[3] = c_combs[il_pos].i[3];
|
||
|
|
||
|
u32 wordr1[4];
|
||
|
|
||
|
wordr1[0] = c_combs[il_pos].i[4];
|
||
|
wordr1[1] = c_combs[il_pos].i[5];
|
||
|
wordr1[2] = c_combs[il_pos].i[6];
|
||
|
wordr1[3] = c_combs[il_pos].i[7];
|
||
|
|
||
|
u32 wordr2[4];
|
||
|
|
||
|
wordr2[0] = 0;
|
||
|
wordr2[1] = 0;
|
||
|
wordr2[2] = 0;
|
||
|
wordr2[3] = 0;
|
||
|
|
||
|
u32 wordr3[4];
|
||
|
|
||
|
wordr3[0] = 0;
|
||
|
wordr3[1] = 0;
|
||
|
wordr3[2] = 0;
|
||
|
wordr3[3] = 0;
|
||
|
|
||
|
if (combs_mode == COMBINATOR_MODE_BASE_LEFT)
|
||
|
{
|
||
|
switch_buffer_by_offset (wordr0, wordr1, wordr2, wordr3, pw_l_len);
|
||
|
}
|
||
|
|
||
|
u32x w0[4];
|
||
|
|
||
|
w0[0] = wordl0[0] | wordr0[0];
|
||
|
w0[1] = wordl0[1] | wordr0[1];
|
||
|
w0[2] = wordl0[2] | wordr0[2];
|
||
|
w0[3] = wordl0[3] | wordr0[3];
|
||
|
|
||
|
u32x w1[4];
|
||
|
|
||
|
w1[0] = wordl1[0] | wordr1[0];
|
||
|
w1[1] = wordl1[1] | wordr1[1];
|
||
|
w1[2] = wordl1[2] | wordr1[2];
|
||
|
w1[3] = wordl1[3] | wordr1[3];
|
||
|
|
||
|
u32x w2[4];
|
||
|
|
||
|
w2[0] = wordl2[0] | wordr2[0];
|
||
|
w2[1] = wordl2[1] | wordr2[1];
|
||
|
w2[2] = wordl2[2] | wordr2[2];
|
||
|
w2[3] = wordl2[3] | wordr2[3];
|
||
|
|
||
|
u32x w3[4];
|
||
|
|
||
|
w3[0] = wordl3[0] | wordr3[0];
|
||
|
w3[1] = wordl3[1] | wordr3[1];
|
||
|
w3[2] = 0;
|
||
|
w3[3] = 0;
|
||
|
|
||
|
append_0x80_2 (w0, w1, pw_len);
|
||
|
|
||
|
u32x w0_t[4];
|
||
|
u32x w1_t[4];
|
||
|
u32x w2_t[4];
|
||
|
u32x w3_t[4];
|
||
|
|
||
|
make_unicode (w0, w0_t, w1_t);
|
||
|
make_unicode (w1, w2_t, w3_t);
|
||
|
|
||
|
switch_buffer_by_offset (w0_t, w1_t, w2_t, w3_t, salt_len);
|
||
|
|
||
|
w0_t[0] = salt_buf[0];
|
||
|
w0_t[1] = salt_buf[1];
|
||
|
w0_t[2] = salt_buf[2];
|
||
|
w0_t[3] = salt_buf[3];
|
||
|
w1_t[0] = swap_workaround (w1_t[0]);
|
||
|
w1_t[1] = swap_workaround (w1_t[1]);
|
||
|
w1_t[2] = swap_workaround (w1_t[2]);
|
||
|
w1_t[3] = swap_workaround (w1_t[3]);
|
||
|
w2_t[0] = swap_workaround (w2_t[0]);
|
||
|
w2_t[1] = swap_workaround (w2_t[1]);
|
||
|
w2_t[2] = swap_workaround (w2_t[2]);
|
||
|
w2_t[3] = swap_workaround (w2_t[3]);
|
||
|
w3_t[0] = swap_workaround (w3_t[0]);
|
||
|
w3_t[1] = swap_workaround (w3_t[1]);
|
||
|
w3_t[2] = 0;
|
||
|
w3_t[3] = pw_salt_len * 8;
|
||
|
|
||
|
u32x digest[5];
|
||
|
|
||
|
digest[0] = SHA1M_A;
|
||
|
digest[1] = SHA1M_B;
|
||
|
digest[2] = SHA1M_C;
|
||
|
digest[3] = SHA1M_D;
|
||
|
digest[4] = SHA1M_E;
|
||
|
|
||
|
sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
|
||
|
|
||
|
w0_t[0] = digest[0];
|
||
|
w0_t[1] = digest[1];
|
||
|
w0_t[2] = digest[2];
|
||
|
w0_t[3] = digest[3];
|
||
|
w1_t[0] = digest[4];
|
||
|
w1_t[1] = 0;
|
||
|
w1_t[2] = 0x80000000;
|
||
|
w1_t[3] = 0;
|
||
|
w2_t[0] = 0;
|
||
|
w2_t[1] = 0;
|
||
|
w2_t[2] = 0;
|
||
|
w2_t[3] = 0;
|
||
|
w3_t[0] = 0;
|
||
|
w3_t[1] = 0;
|
||
|
w3_t[2] = 0;
|
||
|
w3_t[3] = (20 + 4) * 8;
|
||
|
|
||
|
digest[0] = SHA1M_A;
|
||
|
digest[1] = SHA1M_B;
|
||
|
digest[2] = SHA1M_C;
|
||
|
digest[3] = SHA1M_D;
|
||
|
digest[4] = SHA1M_E;
|
||
|
|
||
|
sha1_transform (w0_t, w1_t, w2_t, w3_t, digest);
|
||
|
|
||
|
u32x a = swap_workaround (digest[0]);
|
||
|
u32x b = swap_workaround (digest[1]) & 0xff;
|
||
|
|
||
|
const u32x r0 = a;
|
||
|
const u32x r1 = b;
|
||
|
const u32x r2 = 0;
|
||
|
const u32x r3 = 0;
|
||
|
|
||
|
#include VECT_COMPARE_S
|
||
|
}
|
||
|
}
|
||
|
|
||
|
extern "C" __global__ void __launch_bounds__ (256, 1) m09820_s08 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const oldoffice34_t *oldoffice34_bufs, u32 *d_return_buf, 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)
|
||
|
{
|
||
|
}
|
||
|
|
||
|
extern "C" __global__ void __launch_bounds__ (256, 1) m09820_s16 (const pw_t *pws, const gpu_rule_t *rules_buf, const comb_t *combs_buf, const bf_t *bfs_buf, const void *tmps, void *hooks, const u32 *bitmaps_buf_s1_a, const u32 *bitmaps_buf_s1_b, const u32 *bitmaps_buf_s1_c, const u32 *bitmaps_buf_s1_d, const u32 *bitmaps_buf_s2_a, const u32 *bitmaps_buf_s2_b, const u32 *bitmaps_buf_s2_c, const u32 *bitmaps_buf_s2_d, plain_t *plains_buf, const digest_t *digests_buf, u32 *hashes_shown, const salt_t *salt_bufs, const oldoffice34_t *oldoffice34_bufs, u32 *d_return_buf, 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)
|
||
|
{
|
||
|
}
|