From 5eb76ccdde48694336739446c6f5af58855bcd3e Mon Sep 17 00:00:00 2001 From: jsteube Date: Tue, 4 Jul 2017 11:22:48 +0200 Subject: [PATCH] Vectorized Password Safe v3 and added support for long passwords --- OpenCL/m05200.cl | 315 ++++++----------------------------------------- src/interface.c | 11 +- 2 files changed, 50 insertions(+), 276 deletions(-) diff --git a/OpenCL/m05200.cl b/OpenCL/m05200.cl index c8662a622..d99876d04 100644 --- a/OpenCL/m05200.cl +++ b/OpenCL/m05200.cl @@ -3,123 +3,19 @@ * License.....: MIT */ +#define NEW_SIMD_CODE + #include "inc_vendor.cl" #include "inc_hash_constants.h" #include "inc_hash_functions.cl" #include "inc_types.cl" #include "inc_common.cl" +#include "inc_simd.cl" +#include "inc_hash_sha256.cl" #define COMPARE_S "inc_comp_single.cl" #define COMPARE_M "inc_comp_multi.cl" -__constant u32a k_sha256[64] = -{ - SHA256C00, SHA256C01, SHA256C02, SHA256C03, - SHA256C04, SHA256C05, SHA256C06, SHA256C07, - SHA256C08, SHA256C09, SHA256C0a, SHA256C0b, - SHA256C0c, SHA256C0d, SHA256C0e, SHA256C0f, - SHA256C10, SHA256C11, SHA256C12, SHA256C13, - SHA256C14, SHA256C15, SHA256C16, SHA256C17, - SHA256C18, SHA256C19, SHA256C1a, SHA256C1b, - SHA256C1c, SHA256C1d, SHA256C1e, SHA256C1f, - SHA256C20, SHA256C21, SHA256C22, SHA256C23, - SHA256C24, SHA256C25, SHA256C26, SHA256C27, - SHA256C28, SHA256C29, SHA256C2a, SHA256C2b, - SHA256C2c, SHA256C2d, SHA256C2e, SHA256C2f, - SHA256C30, SHA256C31, SHA256C32, SHA256C33, - SHA256C34, SHA256C35, SHA256C36, SHA256C37, - SHA256C38, SHA256C39, SHA256C3a, SHA256C3b, - SHA256C3c, SHA256C3d, SHA256C3e, SHA256C3f, -}; - -void sha256_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[8]) -{ - u32 a = digest[0]; - u32 b = digest[1]; - u32 c = digest[2]; - u32 d = digest[3]; - u32 e = digest[4]; - u32 f = digest[5]; - u32 g = digest[6]; - u32 h = digest[7]; - - u32 w0_t = w0[0]; - u32 w1_t = w0[1]; - u32 w2_t = w0[2]; - u32 w3_t = w0[3]; - u32 w4_t = w1[0]; - u32 w5_t = w1[1]; - u32 w6_t = w1[2]; - u32 w7_t = w1[3]; - u32 w8_t = w2[0]; - u32 w9_t = w2[1]; - u32 wa_t = w2[2]; - u32 wb_t = w2[3]; - u32 wc_t = w3[0]; - u32 wd_t = w3[1]; - u32 we_t = w3[2]; - u32 wf_t = w3[3]; - - #define ROUND_EXPAND() \ - { \ - w0_t = SHA256_EXPAND (we_t, w9_t, w1_t, w0_t); \ - w1_t = SHA256_EXPAND (wf_t, wa_t, w2_t, w1_t); \ - w2_t = SHA256_EXPAND (w0_t, wb_t, w3_t, w2_t); \ - w3_t = SHA256_EXPAND (w1_t, wc_t, w4_t, w3_t); \ - w4_t = SHA256_EXPAND (w2_t, wd_t, w5_t, w4_t); \ - w5_t = SHA256_EXPAND (w3_t, we_t, w6_t, w5_t); \ - w6_t = SHA256_EXPAND (w4_t, wf_t, w7_t, w6_t); \ - w7_t = SHA256_EXPAND (w5_t, w0_t, w8_t, w7_t); \ - w8_t = SHA256_EXPAND (w6_t, w1_t, w9_t, w8_t); \ - w9_t = SHA256_EXPAND (w7_t, w2_t, wa_t, w9_t); \ - wa_t = SHA256_EXPAND (w8_t, w3_t, wb_t, wa_t); \ - wb_t = SHA256_EXPAND (w9_t, w4_t, wc_t, wb_t); \ - wc_t = SHA256_EXPAND (wa_t, w5_t, wd_t, wc_t); \ - wd_t = SHA256_EXPAND (wb_t, w6_t, we_t, wd_t); \ - we_t = SHA256_EXPAND (wc_t, w7_t, wf_t, we_t); \ - wf_t = SHA256_EXPAND (wd_t, w8_t, w0_t, wf_t); \ - } - - #define ROUND_STEP(i) \ - { \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w0_t, k_sha256[i + 0]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w1_t, k_sha256[i + 1]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w2_t, k_sha256[i + 2]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w3_t, k_sha256[i + 3]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w4_t, k_sha256[i + 4]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w5_t, k_sha256[i + 5]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w6_t, k_sha256[i + 6]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w7_t, k_sha256[i + 7]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w8_t, k_sha256[i + 8]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w9_t, k_sha256[i + 9]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, wa_t, k_sha256[i + 10]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, wb_t, k_sha256[i + 11]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, wc_t, k_sha256[i + 12]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, wd_t, k_sha256[i + 13]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, we_t, k_sha256[i + 14]); \ - SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, wf_t, k_sha256[i + 15]); \ - } - - ROUND_STEP (0); - - #ifdef _unroll - #pragma unroll - #endif - for (int i = 16; i < 64; i += 16) - { - ROUND_EXPAND (); ROUND_STEP (i); - } - - digest[0] += a; - digest[1] += b; - digest[2] += c; - digest[3] += d; - digest[4] += e; - digest[5] += f; - digest[6] += g; - digest[7] += h; -} - __kernel void m05200_init (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global pwsafe3_tmp_t *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global const void *esalt_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) { /** @@ -130,167 +26,42 @@ __kernel void m05200_init (__global pw_t *pws, __global const kernel_rule_t *rul if (gid >= gid_max) return; - u32 w0[4]; + sha256_ctx_t ctx; - w0[0] = pws[gid].i[ 0]; - w0[1] = pws[gid].i[ 1]; - w0[2] = pws[gid].i[ 2]; - w0[3] = pws[gid].i[ 3]; + sha256_init (&ctx); - u32 w1[4]; + sha256_update_global_swap (&ctx, pws[gid].i, pws[gid].pw_len); - w1[0] = pws[gid].i[ 4]; - w1[1] = pws[gid].i[ 5]; - w1[2] = 0; - w1[3] = 0; + sha256_update_global_swap (&ctx, salt_bufs[salt_pos].salt_buf, salt_bufs[salt_pos].salt_len); - u32 w2[4]; + sha256_final (&ctx); - w2[0] = 0; - w2[1] = 0; - w2[2] = 0; - w2[3] = 0; - - u32 w3[4]; - - w3[0] = 0; - w3[1] = 0; - w3[2] = 0; - w3[3] = 0; - - const u32 pw_len = pws[gid].pw_len; - - /** - * salt - */ - - u32 salt_buf0[4]; - - salt_buf0[0] = salt_bufs[salt_pos].salt_buf[0]; - salt_buf0[1] = salt_bufs[salt_pos].salt_buf[1]; - salt_buf0[2] = salt_bufs[salt_pos].salt_buf[2]; - salt_buf0[3] = salt_bufs[salt_pos].salt_buf[3]; - - u32 salt_buf1[4]; - - salt_buf1[0] = salt_bufs[salt_pos].salt_buf[4]; - salt_buf1[1] = salt_bufs[salt_pos].salt_buf[5]; - salt_buf1[2] = salt_bufs[salt_pos].salt_buf[6]; - salt_buf1[3] = salt_bufs[salt_pos].salt_buf[7]; - - u32 salt_buf2[4]; - - salt_buf2[0] = 0; - salt_buf2[1] = 0; - salt_buf2[2] = 0; - salt_buf2[3] = 0; - - u32 salt_buf3[4]; - - salt_buf3[0] = 0; - salt_buf3[1] = 0; - salt_buf3[2] = 0; - salt_buf3[3] = 0; - - u32 salt_len = salt_bufs[salt_pos].salt_len; - - switch_buffer_by_offset_le (salt_buf0, salt_buf1, salt_buf2, salt_buf3, pw_len); - - w0[0] |= salt_buf0[0]; - w0[1] |= salt_buf0[1]; - w0[2] |= salt_buf0[2]; - w0[3] |= salt_buf0[3]; - - w1[0] |= salt_buf1[0]; - w1[1] |= salt_buf1[1]; - w1[2] |= salt_buf1[2]; - w1[3] |= salt_buf1[3]; - - w2[0] |= salt_buf2[0]; - w2[1] |= salt_buf2[1]; - w2[2] |= salt_buf2[2]; - w2[3] |= salt_buf2[3]; - - w3[0] |= salt_buf3[0]; - w3[1] |= salt_buf3[1]; - w3[2] |= salt_buf3[2]; - w3[3] |= salt_buf3[3]; - - const u32 block_len = pw_len + salt_len; - - append_0x80_4x4 (w0, w1, w2, w3, block_len); - - /** - * init - */ - - w0[0] = swap32 (w0[0]); - w0[1] = swap32 (w0[1]); - w0[2] = swap32 (w0[2]); - w0[3] = swap32 (w0[3]); - - w1[0] = swap32 (w1[0]); - w1[1] = swap32 (w1[1]); - w1[2] = swap32 (w1[2]); - w1[3] = swap32 (w1[3]); - - w2[0] = swap32 (w2[0]); - w2[1] = swap32 (w2[1]); - w2[2] = swap32 (w2[2]); - w2[3] = swap32 (w2[3]); - - w3[0] = swap32 (w3[0]); - w3[1] = swap32 (w3[1]); - w3[2] = 0; - w3[3] = block_len * 8; - - /** - * main - */ - - u32 digest[8]; - - digest[0] = SHA256M_A; - digest[1] = SHA256M_B; - digest[2] = SHA256M_C; - digest[3] = SHA256M_D; - digest[4] = SHA256M_E; - digest[5] = SHA256M_F; - digest[6] = SHA256M_G; - digest[7] = SHA256M_H; - - sha256_transform (w0, w1, w2, w3, digest); - - /** - * store - */ - - tmps[gid].digest_buf[0] = digest[0]; - tmps[gid].digest_buf[1] = digest[1]; - tmps[gid].digest_buf[2] = digest[2]; - tmps[gid].digest_buf[3] = digest[3]; - tmps[gid].digest_buf[4] = digest[4]; - tmps[gid].digest_buf[5] = digest[5]; - tmps[gid].digest_buf[6] = digest[6]; - tmps[gid].digest_buf[7] = digest[7]; + tmps[gid].digest_buf[0] = ctx.h[0]; + tmps[gid].digest_buf[1] = ctx.h[1]; + tmps[gid].digest_buf[2] = ctx.h[2]; + tmps[gid].digest_buf[3] = ctx.h[3]; + tmps[gid].digest_buf[4] = ctx.h[4]; + tmps[gid].digest_buf[5] = ctx.h[5]; + tmps[gid].digest_buf[6] = ctx.h[6]; + tmps[gid].digest_buf[7] = ctx.h[7]; } __kernel void m05200_loop (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global pwsafe3_tmp_t *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global const void *esalt_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) { const u32 gid = get_global_id (0); - if (gid >= gid_max) return; + if ((gid * VECT_SIZE) >= gid_max) return; - u32 digest[8]; + u32x digest[8]; - digest[0] = tmps[gid].digest_buf[0]; - digest[1] = tmps[gid].digest_buf[1]; - digest[2] = tmps[gid].digest_buf[2]; - digest[3] = tmps[gid].digest_buf[3]; - digest[4] = tmps[gid].digest_buf[4]; - digest[5] = tmps[gid].digest_buf[5]; - digest[6] = tmps[gid].digest_buf[6]; - digest[7] = tmps[gid].digest_buf[7]; + digest[0] = packv (tmps, digest_buf, gid, 0); + digest[1] = packv (tmps, digest_buf, gid, 1); + digest[2] = packv (tmps, digest_buf, gid, 2); + digest[3] = packv (tmps, digest_buf, gid, 3); + digest[4] = packv (tmps, digest_buf, gid, 4); + digest[5] = packv (tmps, digest_buf, gid, 5); + digest[6] = packv (tmps, digest_buf, gid, 6); + digest[7] = packv (tmps, digest_buf, gid, 7); /** * init @@ -298,29 +69,23 @@ __kernel void m05200_loop (__global pw_t *pws, __global const kernel_rule_t *rul for (u32 i = 0; i < loop_cnt; i++) { - u32 w0[4]; + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; w0[0] = digest[0]; w0[1] = digest[1]; w0[2] = digest[2]; w0[3] = digest[3]; - - u32 w1[4]; - w1[0] = digest[4]; w1[1] = digest[5]; w1[2] = digest[6]; w1[3] = digest[7]; - - u32 w2[4]; - w2[0] = 0x80000000; w2[1] = 0; w2[2] = 0; w2[3] = 0; - - u32 w3[4]; - w3[0] = 0; w3[1] = 0; w3[2] = 0; @@ -335,17 +100,17 @@ __kernel void m05200_loop (__global pw_t *pws, __global const kernel_rule_t *rul digest[6] = SHA256M_G; digest[7] = SHA256M_H; - sha256_transform (w0, w1, w2, w3, digest); + sha256_transform_vector (w0, w1, w2, w3, digest); } - tmps[gid].digest_buf[0] = digest[0]; - tmps[gid].digest_buf[1] = digest[1]; - tmps[gid].digest_buf[2] = digest[2]; - tmps[gid].digest_buf[3] = digest[3]; - tmps[gid].digest_buf[4] = digest[4]; - tmps[gid].digest_buf[5] = digest[5]; - tmps[gid].digest_buf[6] = digest[6]; - tmps[gid].digest_buf[7] = digest[7]; + unpackv (tmps, digest_buf, gid, 0, digest[0]); + unpackv (tmps, digest_buf, gid, 1, digest[1]); + unpackv (tmps, digest_buf, gid, 2, digest[2]); + unpackv (tmps, digest_buf, gid, 3, digest[3]); + unpackv (tmps, digest_buf, gid, 4, digest[4]); + unpackv (tmps, digest_buf, gid, 5, digest[5]); + unpackv (tmps, digest_buf, gid, 6, digest[6]); + unpackv (tmps, digest_buf, gid, 7, digest[7]); } __kernel void m05200_comp (__global pw_t *pws, __global const kernel_rule_t *rules_buf, __global const pw_t *combs_buf, __global const bf_t *bfs_buf, __global pwsafe3_tmp_t *tmps, __global void *hooks, __global const u32 *bitmaps_buf_s1_a, __global const u32 *bitmaps_buf_s1_b, __global const u32 *bitmaps_buf_s1_c, __global const u32 *bitmaps_buf_s1_d, __global const u32 *bitmaps_buf_s2_a, __global const u32 *bitmaps_buf_s2_b, __global const u32 *bitmaps_buf_s2_c, __global const u32 *bitmaps_buf_s2_d, __global plain_t *plains_buf, __global const digest_t *digests_buf, __global u32 *hashes_shown, __global const salt_t *salt_bufs, __global const void *esalt_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) diff --git a/src/interface.c b/src/interface.c index d92b14c90..5dc53e12e 100644 --- a/src/interface.c +++ b/src/interface.c @@ -21815,7 +21815,8 @@ int hashconfig_init (hashcat_ctx_t *hashcat_ctx) hashconfig->kern_type = KERN_TYPE_PSAFE3; hashconfig->dgst_size = DGST_SIZE_4_8; hashconfig->parse_func = psafe3_parse_hash; - hashconfig->opti_type = OPTI_TYPE_ZERO_BYTE; + hashconfig->opti_type = OPTI_TYPE_ZERO_BYTE + | OPTI_TYPE_SLOW_HASH_SIMD_LOOP; hashconfig->dgst_pos0 = 0; hashconfig->dgst_pos1 = 1; hashconfig->dgst_pos2 = 2; @@ -24628,6 +24629,14 @@ int hashconfig_init (hashcat_ctx_t *hashcat_ctx) case 15500: hashconfig->pw_max = 16; break; } + + // some mode are converted to long passwords but without dropping performance, so they don't need the -L option set by user + + switch (hashconfig->hash_mode) + { + case 5200: hashconfig->pw_max = PW_MAX; + break; + } } // pw_max : algo specific hard max length