From c3f0bb77dd21aa2ebad417d4fbaf1ec9257456ec Mon Sep 17 00:00:00 2001 From: jsteube Date: Mon, 3 Jul 2017 13:06:59 +0200 Subject: [PATCH] Vectorized AIX {ssha1} kernel and added support for long passwords --- OpenCL/m06700.cl | 521 +++++++++++++---------------------------------- src/interface.c | 5 +- 2 files changed, 147 insertions(+), 379 deletions(-) diff --git a/OpenCL/m06700.cl b/OpenCL/m06700.cl index 575c3b792..70e14f3eb 100644 --- a/OpenCL/m06700.cl +++ b/OpenCL/m06700.cl @@ -8,192 +8,13 @@ #include "inc_hash_functions.cl" #include "inc_types.cl" #include "inc_common.cl" +#include "inc_simd.cl" +#include "inc_hash_sha1.cl" #define COMPARE_S "inc_comp_single.cl" #define COMPARE_M "inc_comp_multi.cl" -void sha1_transform (const u32 w0[4], const u32 w1[4], const u32 w2[4], const u32 w3[4], u32 digest[5]) -{ - u32 A = digest[0]; - u32 B = digest[1]; - u32 C = digest[2]; - u32 D = digest[3]; - u32 E = digest[4]; - - 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]; - - #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; -} - -void hmac_sha1_pad (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 opad[5]) -{ - w0[0] = w0[0] ^ 0x36363636; - w0[1] = w0[1] ^ 0x36363636; - w0[2] = w0[2] ^ 0x36363636; - w0[3] = w0[3] ^ 0x36363636; - w1[0] = w1[0] ^ 0x36363636; - w1[1] = w1[1] ^ 0x36363636; - w1[2] = w1[2] ^ 0x36363636; - w1[3] = w1[3] ^ 0x36363636; - w2[0] = w2[0] ^ 0x36363636; - w2[1] = w2[1] ^ 0x36363636; - w2[2] = w2[2] ^ 0x36363636; - w2[3] = w2[3] ^ 0x36363636; - w3[0] = w3[0] ^ 0x36363636; - w3[1] = w3[1] ^ 0x36363636; - w3[2] = w3[2] ^ 0x36363636; - w3[3] = w3[3] ^ 0x36363636; - - ipad[0] = SHA1M_A; - ipad[1] = SHA1M_B; - ipad[2] = SHA1M_C; - ipad[3] = SHA1M_D; - ipad[4] = SHA1M_E; - - sha1_transform (w0, w1, w2, w3, ipad); - - w0[0] = w0[0] ^ 0x6a6a6a6a; - w0[1] = w0[1] ^ 0x6a6a6a6a; - w0[2] = w0[2] ^ 0x6a6a6a6a; - w0[3] = w0[3] ^ 0x6a6a6a6a; - w1[0] = w1[0] ^ 0x6a6a6a6a; - w1[1] = w1[1] ^ 0x6a6a6a6a; - w1[2] = w1[2] ^ 0x6a6a6a6a; - w1[3] = w1[3] ^ 0x6a6a6a6a; - w2[0] = w2[0] ^ 0x6a6a6a6a; - w2[1] = w2[1] ^ 0x6a6a6a6a; - w2[2] = w2[2] ^ 0x6a6a6a6a; - w2[3] = w2[3] ^ 0x6a6a6a6a; - w3[0] = w3[0] ^ 0x6a6a6a6a; - w3[1] = w3[1] ^ 0x6a6a6a6a; - w3[2] = w3[2] ^ 0x6a6a6a6a; - w3[3] = w3[3] ^ 0x6a6a6a6a; - - opad[0] = SHA1M_A; - opad[1] = SHA1M_B; - opad[2] = SHA1M_C; - opad[3] = SHA1M_D; - opad[4] = SHA1M_E; - - sha1_transform (w0, w1, w2, w3, opad); -} - -void hmac_sha1_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 opad[5], u32 digest[5]) +void hmac_sha1_run_V (u32x w0[4], u32x w1[4], u32x w2[4], u32x w3[4], u32x ipad[5], u32x opad[5], u32x digest[5]) { digest[0] = ipad[0]; digest[1] = ipad[1]; @@ -201,7 +22,7 @@ void hmac_sha1_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 digest[3] = ipad[3]; digest[4] = ipad[4]; - sha1_transform (w0, w1, w2, w3, digest); + sha1_transform_vector (w0, w1, w2, w3, digest); w0[0] = digest[0]; w0[1] = digest[1]; @@ -226,7 +47,7 @@ void hmac_sha1_run (u32 w0[4], u32 w1[4], u32 w2[4], u32 w3[4], u32 ipad[5], u32 digest[3] = opad[3]; digest[4] = opad[4]; - sha1_transform (w0, w1, w2, w3, digest); + sha1_transform_vector (w0, w1, w2, w3, digest); } __kernel void m06700_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 sha1aix_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 wpa_t *wpa_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) @@ -240,209 +61,72 @@ __kernel void m06700_init (__global pw_t *pws, __global const kernel_rule_t *rul if (gid >= gid_max) return; u32 w0[4]; + u32 w1[4]; + u32 w2[4]; + u32 w3[4]; w0[0] = pws[gid].i[ 0]; w0[1] = pws[gid].i[ 1]; w0[2] = pws[gid].i[ 2]; w0[3] = pws[gid].i[ 3]; - - u32 w1[4]; - w1[0] = pws[gid].i[ 4]; w1[1] = pws[gid].i[ 5]; w1[2] = pws[gid].i[ 6]; w1[3] = pws[gid].i[ 7]; - - u32 w2[4]; - w2[0] = pws[gid].i[ 8]; w2[1] = pws[gid].i[ 9]; w2[2] = pws[gid].i[10]; w2[3] = pws[gid].i[11]; - - u32 w3[4]; - w3[0] = pws[gid].i[12]; w3[1] = pws[gid].i[13]; w3[2] = pws[gid].i[14]; w3[3] = pws[gid].i[15]; - /** - * salt - */ + w0[0] = swap32_S (w0[0]); + w0[1] = swap32_S (w0[1]); + w0[2] = swap32_S (w0[2]); + w0[3] = swap32_S (w0[3]); + w1[0] = swap32_S (w1[0]); + w1[1] = swap32_S (w1[1]); + w1[2] = swap32_S (w1[2]); + w1[3] = swap32_S (w1[3]); + w2[0] = swap32_S (w2[0]); + w2[1] = swap32_S (w2[1]); + w2[2] = swap32_S (w2[2]); + w2[3] = swap32_S (w2[3]); + w3[0] = swap32_S (w3[0]); + w3[1] = swap32_S (w3[1]); + w3[2] = swap32_S (w3[2]); + w3[3] = swap32_S (w3[3]); - u32 salt_len = salt_bufs[salt_pos].salt_len; + sha1_hmac_ctx_t sha1_hmac_ctx; - u32 salt_buf0[4]; + sha1_hmac_init (&sha1_hmac_ctx, w0, w1, w2, w3); - 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]; + tmps[gid].ipad[0] = sha1_hmac_ctx.ipad.h[0]; + tmps[gid].ipad[1] = sha1_hmac_ctx.ipad.h[1]; + tmps[gid].ipad[2] = sha1_hmac_ctx.ipad.h[2]; + tmps[gid].ipad[3] = sha1_hmac_ctx.ipad.h[3]; + tmps[gid].ipad[4] = sha1_hmac_ctx.ipad.h[4]; - u32 salt_buf1[4]; + tmps[gid].opad[0] = sha1_hmac_ctx.opad.h[0]; + tmps[gid].opad[1] = sha1_hmac_ctx.opad.h[1]; + tmps[gid].opad[2] = sha1_hmac_ctx.opad.h[2]; + tmps[gid].opad[3] = sha1_hmac_ctx.opad.h[3]; + tmps[gid].opad[4] = sha1_hmac_ctx.opad.h[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]; + sha1_hmac_update_global_swap (&sha1_hmac_ctx, salt_bufs[salt_pos].salt_buf, salt_bufs[salt_pos].salt_len); - u32 salt_buf2[4]; - - salt_buf2[0] = salt_bufs[salt_pos].salt_buf[ 8]; - salt_buf2[1] = salt_bufs[salt_pos].salt_buf[ 9]; - salt_buf2[2] = salt_bufs[salt_pos].salt_buf[10]; - salt_buf2[3] = salt_bufs[salt_pos].salt_buf[11]; - - u32 salt_buf3[4]; - - salt_buf3[0] = 0; - salt_buf3[1] = 0; - salt_buf3[2] = 0; - salt_buf3[3] = 0; - - append_0x01_4x4 (salt_buf0, salt_buf1, salt_buf2, salt_buf3, salt_len + 3); - - append_0x80_4x4 (salt_buf0, salt_buf1, salt_buf2, salt_buf3, salt_len + 4); - - /** - * pads - */ - - 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] = swap32 (w3[2]); - w3[3] = swap32 (w3[3]); - - u32 ipad[5]; - u32 opad[5]; - - hmac_sha1_pad (w0, w1, w2, w3, ipad, opad); - - tmps[gid].ipad[0] = ipad[0]; - tmps[gid].ipad[1] = ipad[1]; - tmps[gid].ipad[2] = ipad[2]; - tmps[gid].ipad[3] = ipad[3]; - tmps[gid].ipad[4] = ipad[4]; - - tmps[gid].opad[0] = opad[0]; - tmps[gid].opad[1] = opad[1]; - tmps[gid].opad[2] = opad[2]; - tmps[gid].opad[3] = opad[3]; - tmps[gid].opad[4] = opad[4]; - - 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]; - - 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] = swap32 (w3[2]); - w3[3] = (64 + salt_len + 4) * 8; - - u32 dgst[5]; - - hmac_sha1_run (w0, w1, w2, w3, ipad, opad, dgst); - - tmps[gid].dgst[0] = dgst[0]; - tmps[gid].dgst[1] = dgst[1]; - tmps[gid].dgst[2] = dgst[2]; - tmps[gid].dgst[3] = dgst[3]; - tmps[gid].dgst[4] = dgst[4]; - - tmps[gid].out[0] = dgst[0]; - tmps[gid].out[1] = dgst[1]; - tmps[gid].out[2] = dgst[2]; - tmps[gid].out[3] = dgst[3]; - tmps[gid].out[4] = dgst[4]; -} - -__kernel void m06700_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 sha1aix_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 wpa_t *wpa_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; - - u32 ipad[5]; - u32 opad[5]; - - ipad[0] = tmps[gid].ipad[0]; - ipad[1] = tmps[gid].ipad[1]; - ipad[2] = tmps[gid].ipad[2]; - ipad[3] = tmps[gid].ipad[3]; - ipad[4] = tmps[gid].ipad[4]; - - opad[0] = tmps[gid].opad[0]; - opad[1] = tmps[gid].opad[1]; - opad[2] = tmps[gid].opad[2]; - opad[3] = tmps[gid].opad[3]; - opad[4] = tmps[gid].opad[4]; - - u32 dgst[5]; - u32 out[5]; - - dgst[0] = tmps[gid].dgst[0]; - dgst[1] = tmps[gid].dgst[1]; - dgst[2] = tmps[gid].dgst[2]; - dgst[3] = tmps[gid].dgst[3]; - dgst[4] = tmps[gid].dgst[4]; - - out[0] = tmps[gid].out[0]; - out[1] = tmps[gid].out[1]; - out[2] = tmps[gid].out[2]; - out[3] = tmps[gid].out[3]; - out[4] = tmps[gid].out[4]; - - for (u32 j = 0; j < loop_cnt; j++) + for (u32 i = 0, j = 1; i < 4; i += 5, j += 1) { - u32 w0[4]; - u32 w1[4]; - u32 w2[4]; - u32 w3[4]; + sha1_hmac_ctx_t sha1_hmac_ctx2 = sha1_hmac_ctx; - w0[0] = dgst[0]; - w0[1] = dgst[1]; - w0[2] = dgst[2]; - w0[3] = dgst[3]; - w1[0] = dgst[4]; - w1[1] = 0x80000000; + w0[0] = j; + w0[1] = 0; + w0[2] = 0; + w0[3] = 0; + w1[0] = 0; + w1[1] = 0; w1[2] = 0; w1[3] = 0; w2[0] = 0; @@ -452,28 +136,109 @@ __kernel void m06700_loop (__global pw_t *pws, __global const kernel_rule_t *rul w3[0] = 0; w3[1] = 0; w3[2] = 0; - w3[3] = (64 + 20) * 8; + w3[3] = 0; - hmac_sha1_run (w0, w1, w2, w3, ipad, opad, dgst); + sha1_hmac_update_64 (&sha1_hmac_ctx2, w0, w1, w2, w3, 4); - out[0] ^= dgst[0]; - out[1] ^= dgst[1]; - out[2] ^= dgst[2]; - out[3] ^= dgst[3]; - out[4] ^= dgst[4]; + sha1_hmac_final (&sha1_hmac_ctx2); + + tmps[gid].dgst[i + 0] = sha1_hmac_ctx2.opad.h[0]; + tmps[gid].dgst[i + 1] = sha1_hmac_ctx2.opad.h[1]; + tmps[gid].dgst[i + 2] = sha1_hmac_ctx2.opad.h[2]; + tmps[gid].dgst[i + 3] = sha1_hmac_ctx2.opad.h[3]; + tmps[gid].dgst[i + 4] = sha1_hmac_ctx2.opad.h[4]; + + tmps[gid].out[i + 0] = tmps[gid].dgst[i + 0]; + tmps[gid].out[i + 1] = tmps[gid].dgst[i + 1]; + tmps[gid].out[i + 2] = tmps[gid].dgst[i + 2]; + tmps[gid].out[i + 3] = tmps[gid].dgst[i + 3]; + tmps[gid].out[i + 4] = tmps[gid].dgst[i + 4]; } +} - tmps[gid].dgst[0] = dgst[0]; - tmps[gid].dgst[1] = dgst[1]; - tmps[gid].dgst[2] = dgst[2]; - tmps[gid].dgst[3] = dgst[3]; - tmps[gid].dgst[4] = dgst[4]; +__kernel void m06700_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 sha1aix_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 wpa_t *wpa_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); - tmps[gid].out[0] = out[0]; - tmps[gid].out[1] = out[1]; - tmps[gid].out[2] = out[2]; - tmps[gid].out[3] = out[3]; - tmps[gid].out[4] = out[4]; + if ((gid * VECT_SIZE) >= gid_max) return; + + u32x ipad[5]; + u32x opad[5]; + + ipad[0] = packv (tmps, ipad, gid, 0); + ipad[1] = packv (tmps, ipad, gid, 1); + ipad[2] = packv (tmps, ipad, gid, 2); + ipad[3] = packv (tmps, ipad, gid, 3); + ipad[4] = packv (tmps, ipad, gid, 4); + + opad[0] = packv (tmps, opad, gid, 0); + opad[1] = packv (tmps, opad, gid, 1); + opad[2] = packv (tmps, opad, gid, 2); + opad[3] = packv (tmps, opad, gid, 3); + opad[4] = packv (tmps, opad, gid, 4); + + for (u32 i = 0; i < 4; i += 5) + { + u32x dgst[5]; + u32x out[5]; + + dgst[0] = packv (tmps, dgst, gid, i + 0); + dgst[1] = packv (tmps, dgst, gid, i + 1); + dgst[2] = packv (tmps, dgst, gid, i + 2); + dgst[3] = packv (tmps, dgst, gid, i + 3); + dgst[4] = packv (tmps, dgst, gid, i + 4); + + out[0] = packv (tmps, out, gid, i + 0); + out[1] = packv (tmps, out, gid, i + 1); + out[2] = packv (tmps, out, gid, i + 2); + out[3] = packv (tmps, out, gid, i + 3); + out[4] = packv (tmps, out, gid, i + 4); + + for (u32 j = 0; j < loop_cnt; j++) + { + u32x w0[4]; + u32x w1[4]; + u32x w2[4]; + u32x w3[4]; + + w0[0] = dgst[0]; + w0[1] = dgst[1]; + w0[2] = dgst[2]; + w0[3] = dgst[3]; + w1[0] = dgst[4]; + w1[1] = 0x80000000; + w1[2] = 0; + w1[3] = 0; + w2[0] = 0; + w2[1] = 0; + w2[2] = 0; + w2[3] = 0; + w3[0] = 0; + w3[1] = 0; + w3[2] = 0; + w3[3] = (64 + 20) * 8; + + hmac_sha1_run_V (w0, w1, w2, w3, ipad, opad, dgst); + + out[0] ^= dgst[0]; + out[1] ^= dgst[1]; + out[2] ^= dgst[2]; + out[3] ^= dgst[3]; + out[4] ^= dgst[4]; + } + + unpackv (tmps, dgst, gid, i + 0, dgst[0]); + unpackv (tmps, dgst, gid, i + 1, dgst[1]); + unpackv (tmps, dgst, gid, i + 2, dgst[2]); + unpackv (tmps, dgst, gid, i + 3, dgst[3]); + unpackv (tmps, dgst, gid, i + 4, dgst[4]); + + unpackv (tmps, out, gid, i + 0, out[0]); + unpackv (tmps, out, gid, i + 1, out[1]); + unpackv (tmps, out, gid, i + 2, out[2]); + unpackv (tmps, out, gid, i + 3, out[3]); + unpackv (tmps, out, gid, i + 4, out[4]); + } } __kernel void m06700_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 sha1aix_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 wpa_t *wpa_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 b1e589f44..bbe938354 100644 --- a/src/interface.c +++ b/src/interface.c @@ -22233,7 +22233,8 @@ int hashconfig_init (hashcat_ctx_t *hashcat_ctx) hashconfig->kern_type = KERN_TYPE_SHA1AIX; hashconfig->dgst_size = DGST_SIZE_4_5; hashconfig->parse_func = sha1aix_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; @@ -24622,6 +24623,8 @@ int hashconfig_init (hashcat_ctx_t *hashcat_ctx) break; case 6600: hashconfig->pw_max = 64; // PBKDF2-HMAC-SHA1 max break; + case 6700: hashconfig->pw_max = 64; // PBKDF2-HMAC-SHA1 max + break; case 8500: hashconfig->pw_max = 8; // DES max break; case 8800: hashconfig->pw_max = 64; // PBKDF2-HMAC-SHA1 max