X-Git-Url: https://www.flypig.org.uk/git/?a=blobdiff_plain;f=OpenCL%2Fm08000_a0.cl;h=4540abfd877c7507eb6d2938e6044202a58e543d;hb=161a6eb4bc643d8e636e96eda613f5137d30da59;hp=87a2748b117e72ad53e9c2ffaac52593404a2e39;hpb=2283d5c843c425aebfdad4a521b1e0ad85fae387;p=hashcat.git diff --git a/OpenCL/m08000_a0.cl b/OpenCL/m08000_a0.cl index 87a2748..4540abf 100644 --- a/OpenCL/m08000_a0.cl +++ b/OpenCL/m08000_a0.cl @@ -1,26 +1,22 @@ /** - * Author......: Jens Steube + * Authors.....: Jens Steube + * Gabriele Gristina + * * License.....: MIT */ #define _SHA256_ -#include "include/constants.h" -#include "include/kernel_vendor.h" +#define NEW_SIMD_CODE -#define DGST_R0 3 -#define DGST_R1 7 -#define DGST_R2 2 -#define DGST_R3 6 - -#include "include/kernel_functions.c" -#include "types_ocl.c" -#include "common.c" -#include "include/rp_gpu.h" -#include "rp.c" - -#define COMPARE_S "check_single_comp4.c" -#define COMPARE_M "check_multi_comp4.c" +#include "inc_vendor.cl" +#include "inc_hash_constants.h" +#include "inc_hash_functions.cl" +#include "inc_types.cl" +#include "inc_common.cl" +#include "inc_rp.h" +#include "inc_rp.cl" +#include "inc_simd.cl" __constant u32 k_sha256[64] = { @@ -42,33 +38,38 @@ __constant u32 k_sha256[64] = SHA256C3c, SHA256C3d, SHA256C3e, SHA256C3f, }; -static void sha256_transform (u32 digest[8], const u32 w[16]) +#define SHA256_S0_S(x) (rotl32_S ((x), 25u) ^ rotl32_S ((x), 14u) ^ SHIFT_RIGHT_32 ((x), 3u)) +#define SHA256_S1_S(x) (rotl32_S ((x), 15u) ^ rotl32_S ((x), 13u) ^ SHIFT_RIGHT_32 ((x), 10u)) + +#define SHA256_EXPAND_S(x,y,z,w) (SHA256_S1_S (x) + y + SHA256_S0_S (z) + w) + +void sha256_transform (u32x digest[8], const u32x w[16]) { - 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 = w[ 0]; - u32 w1_t = w[ 1]; - u32 w2_t = w[ 2]; - u32 w3_t = w[ 3]; - u32 w4_t = w[ 4]; - u32 w5_t = w[ 5]; - u32 w6_t = w[ 6]; - u32 w7_t = w[ 7]; - u32 w8_t = w[ 8]; - u32 w9_t = w[ 9]; - u32 wa_t = w[10]; - u32 wb_t = w[11]; - u32 wc_t = w[12]; - u32 wd_t = w[13]; - u32 we_t = w[14]; - u32 wf_t = w[15]; + u32x a = digest[0]; + u32x b = digest[1]; + u32x c = digest[2]; + u32x d = digest[3]; + u32x e = digest[4]; + u32x f = digest[5]; + u32x g = digest[6]; + u32x h = digest[7]; + + u32x w0_t = w[ 0]; + u32x w1_t = w[ 1]; + u32x w2_t = w[ 2]; + u32x w3_t = w[ 3]; + u32x w4_t = w[ 4]; + u32x w5_t = w[ 5]; + u32x w6_t = w[ 6]; + u32x w7_t = w[ 7]; + u32x w8_t = w[ 8]; + u32x w9_t = w[ 9]; + u32x wa_t = w[10]; + u32x wb_t = w[11]; + u32x wc_t = w[12]; + u32x wd_t = w[13]; + u32x we_t = w[14]; + u32x wf_t = w[15]; #define ROUND_EXPAND() \ { \ @@ -112,7 +113,9 @@ static void sha256_transform (u32 digest[8], const u32 w[16]) ROUND_STEP (0); + #ifdef _unroll #pragma unroll + #endif for (int i = 16; i < 64; i += 16) { ROUND_EXPAND (); ROUND_STEP (i); @@ -128,31 +131,182 @@ static void sha256_transform (u32 digest[8], const u32 w[16]) digest[7] += h; } -__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_m04 (__global pw_t *pws, __global gpu_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 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) +void sha256_transform_z (u32x digest[8]) +{ + u32x a = digest[0]; + u32x b = digest[1]; + u32x c = digest[2]; + u32x d = digest[3]; + u32x e = digest[4]; + u32x f = digest[5]; + u32x g = digest[6]; + u32x h = digest[7]; + + #define ROUND_STEP_Z(i) \ + { \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, 0, k_sha256[i + 0]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, 0, k_sha256[i + 1]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, 0, k_sha256[i + 2]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, 0, k_sha256[i + 3]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, 0, k_sha256[i + 4]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, 0, k_sha256[i + 5]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, 0, k_sha256[i + 6]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, 0, k_sha256[i + 7]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, 0, k_sha256[i + 8]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, 0, k_sha256[i + 9]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, 0, k_sha256[i + 10]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, 0, k_sha256[i + 11]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, 0, k_sha256[i + 12]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, 0, k_sha256[i + 13]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, 0, k_sha256[i + 14]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, 0, k_sha256[i + 15]); \ + } + + ROUND_STEP_Z (0); + + #ifdef _unroll + #pragma unroll + #endif + for (int i = 16; i < 64; i += 16) + { + ROUND_STEP_Z (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; +} + +void sha256_transform_s (u32x digest[8], __local u32 *w) +{ + u32x a = digest[0]; + u32x b = digest[1]; + u32x c = digest[2]; + u32x d = digest[3]; + u32x e = digest[4]; + u32x f = digest[5]; + u32x g = digest[6]; + u32x h = digest[7]; + + #define ROUND_STEP_S(i) \ + { \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w[i + 0], k_sha256[i + 0]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w[i + 1], k_sha256[i + 1]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w[i + 2], k_sha256[i + 2]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w[i + 3], k_sha256[i + 3]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w[i + 4], k_sha256[i + 4]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w[i + 5], k_sha256[i + 5]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w[i + 6], k_sha256[i + 6]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w[i + 7], k_sha256[i + 7]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, a, b, c, d, e, f, g, h, w[i + 8], k_sha256[i + 8]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, h, a, b, c, d, e, f, g, w[i + 9], k_sha256[i + 9]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, g, h, a, b, c, d, e, f, w[i + 10], k_sha256[i + 10]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, f, g, h, a, b, c, d, e, w[i + 11], k_sha256[i + 11]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, e, f, g, h, a, b, c, d, w[i + 12], k_sha256[i + 12]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, d, e, f, g, h, a, b, c, w[i + 13], k_sha256[i + 13]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, c, d, e, f, g, h, a, b, w[i + 14], k_sha256[i + 14]); \ + SHA256_STEP (SHA256_F0o, SHA256_F1o, b, c, d, e, f, g, h, a, w[i + 15], k_sha256[i + 15]); \ + } + + ROUND_STEP_S (0); + + #ifdef _unroll + #pragma unroll + #endif + for (int i = 16; i < 64; i += 16) + { + ROUND_STEP_S (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 m08000_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 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) { /** * modifier */ + const u32 gid = get_global_id (0); const u32 lid = get_local_id (0); + const u32 lsz = get_local_size (0); /** - * base + * salt */ - const u32 gid = get_global_id (0); + const u32 salt_buf0 = swap32_S (salt_bufs[salt_pos].salt_buf[ 0]); + const u32 salt_buf1 = swap32_S (salt_bufs[salt_pos].salt_buf[ 1]); + const u32 salt_buf2 = swap32_S (salt_bufs[salt_pos].salt_buf[ 2]); // 0x80 + + /** + * precompute final msg blocks + */ + + __local u32 w_s1[64]; + __local u32 w_s2[64]; + + for (u32 i = lid; i < 64; i += lsz) + { + w_s1[i] = 0; + w_s2[i] = 0; + } + + barrier (CLK_LOCAL_MEM_FENCE); + + if (lid == 0) + { + w_s1[15] = 0 | salt_buf0 >> 16; + + #ifdef _unroll + #pragma unroll + #endif + for (int i = 16; i < 64; i++) + { + w_s1[i] = SHA256_EXPAND_S (w_s1[i - 2], w_s1[i - 7], w_s1[i - 15], w_s1[i - 16]); + } + + w_s2[ 0] = salt_buf0 << 16 | salt_buf1 >> 16; + w_s2[ 1] = salt_buf1 << 16 | salt_buf2 >> 16; + w_s2[ 2] = salt_buf2 << 16 | 0; + w_s2[15] = (510 + 8) * 8; + + #ifdef _unroll + #pragma unroll + #endif + for (int i = 16; i < 64; i++) + { + w_s2[i] = SHA256_EXPAND_S (w_s2[i - 2], w_s2[i - 7], w_s2[i - 15], w_s2[i - 16]); + } + } + + barrier (CLK_LOCAL_MEM_FENCE); if (gid >= gid_max) return; + /** + * base + */ + u32 pw_buf0[4]; + u32 pw_buf1[4]; pw_buf0[0] = pws[gid].i[ 0]; pw_buf0[1] = pws[gid].i[ 1]; pw_buf0[2] = pws[gid].i[ 2]; pw_buf0[3] = pws[gid].i[ 3]; - - u32 pw_buf1[4]; - pw_buf1[0] = pws[gid].i[ 4]; pw_buf1[1] = pws[gid].i[ 5]; pw_buf1[2] = pws[gid].i[ 6]; @@ -160,54 +314,28 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_m04 (__glo const u32 pw_len = pws[gid].pw_len; - /** - * salt - */ - - const u32 salt_buf0 = swap32 (salt_bufs[salt_pos].salt_buf[ 0]); - const u32 salt_buf1 = swap32 (salt_bufs[salt_pos].salt_buf[ 1]); - const u32 salt_buf2 = swap32 (salt_bufs[salt_pos].salt_buf[ 2]); // 0x80 - /** * loop */ - for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++) + for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE) { - u32 w0[4]; - u32 w1[4]; - u32 w2[4]; - u32 w3[4]; - - w0[0] = pw_buf0[0]; - w0[1] = pw_buf0[1]; - w0[2] = pw_buf0[2]; - w0[3] = pw_buf0[3]; - w1[0] = pw_buf1[0]; - w1[1] = pw_buf1[1]; - w1[2] = pw_buf1[2]; - w1[3] = pw_buf1[3]; - w2[0] = 0; - w2[1] = 0; - w2[2] = 0; - w2[3] = 0; - w3[0] = 0; - w3[1] = 0; - w3[2] = 0; - w3[3] = 0; - - const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len); - - u32 w0_t[4]; - u32 w1_t[4]; - u32 w2_t[4]; - u32 w3_t[4]; + u32x w0[4] = { 0 }; + u32x w1[4] = { 0 }; + u32x w2[4] = { 0 }; + u32x w3[4] = { 0 }; - make_unicode (w0, w0_t, w1_t); + apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1); + + 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); - u32 w_t[16]; + u32x w_t[16]; w_t[ 0] = swap32 (w0_t[0]); w_t[ 1] = swap32 (w0_t[1]); @@ -243,7 +371,7 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_m04 (__glo w_t[14] = w_t[14] >> 8; w_t[15] = w_t[15] >> 8; - u32 digest[8]; + u32x digest[8]; digest[0] = SHA256M_A; digest[1] = SHA256M_B; @@ -254,85 +382,102 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_m04 (__glo digest[6] = SHA256M_G; digest[7] = SHA256M_H; - sha256_transform (digest, w_t); // 0 - 64 - - w_t[ 0] = 0; - w_t[ 1] = 0; - w_t[ 2] = 0; - w_t[ 3] = 0; - w_t[ 4] = 0; - w_t[ 5] = 0; - w_t[ 6] = 0; - w_t[ 7] = 0; - w_t[ 8] = 0; - w_t[ 9] = 0; - w_t[10] = 0; - w_t[11] = 0; - w_t[12] = 0; - w_t[13] = 0; - w_t[14] = 0; - w_t[15] = 0; - - sha256_transform (digest, w_t); // 64 - 128 - sha256_transform (digest, w_t); // 128 - 192 - sha256_transform (digest, w_t); // 192 - 256 - sha256_transform (digest, w_t); // 256 - 320 - sha256_transform (digest, w_t); // 320 - 384 - sha256_transform (digest, w_t); // 384 - 448 - - w_t[15] = 0 | salt_buf0 >> 16; - - sha256_transform (digest, w_t); // 448 - 512 - - w_t[ 0] = salt_buf0 << 16 | salt_buf1 >> 16; - w_t[ 1] = salt_buf1 << 16 | salt_buf2 >> 16; - w_t[ 2] = salt_buf2 << 16 | 0; - w_t[15] = (510 + 8) * 8; - - sha256_transform (digest, w_t); // 512 - 576 - - const u32 r0 = digest[3]; - const u32 r1 = digest[7]; - const u32 r2 = digest[2]; - const u32 r3 = digest[6]; - - #include COMPARE_M + sha256_transform (digest, w_t); // 0 - 64 + sha256_transform_z (digest); // 64 - 128 + sha256_transform_z (digest); // 128 - 192 + sha256_transform_z (digest); // 192 - 256 + sha256_transform_z (digest); // 256 - 320 + sha256_transform_z (digest); // 320 - 384 + sha256_transform_z (digest); // 384 - 448 + sha256_transform_s (digest, w_s1); // 448 - 512 + sha256_transform_s (digest, w_s2); // 512 - 576 + + COMPARE_M_SIMD (digest[3], digest[7], digest[2], digest[6]); } } -__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_m08 (__global pw_t *pws, __global gpu_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 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) +__kernel void m08000_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 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) { } -__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_m16 (__global pw_t *pws, __global gpu_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 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) +__kernel void m08000_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 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) { } -__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_s04 (__global pw_t *pws, __global gpu_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 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) +__kernel void m08000_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 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) { /** * modifier */ + const u32 gid = get_global_id (0); const u32 lid = get_local_id (0); + const u32 lsz = get_local_size (0); /** - * base + * salt */ - const u32 gid = get_global_id (0); + const u32 salt_buf0 = swap32_S (salt_bufs[salt_pos].salt_buf[ 0]); + const u32 salt_buf1 = swap32_S (salt_bufs[salt_pos].salt_buf[ 1]); + const u32 salt_buf2 = swap32_S (salt_bufs[salt_pos].salt_buf[ 2]); // 0x80 + + /** + * precompute final msg blocks + */ + + __local u32 w_s1[64]; + __local u32 w_s2[64]; + + for (u32 i = lid; i < 64; i += lsz) + { + w_s1[i] = 0; + w_s2[i] = 0; + } + + barrier (CLK_LOCAL_MEM_FENCE); + + if (lid == 0) + { + w_s1[15] = 0 | salt_buf0 >> 16; + + #ifdef _unroll + #pragma unroll + #endif + for (int i = 16; i < 64; i++) + { + w_s1[i] = SHA256_EXPAND_S (w_s1[i - 2], w_s1[i - 7], w_s1[i - 15], w_s1[i - 16]); + } + + w_s2[ 0] = salt_buf0 << 16 | salt_buf1 >> 16; + w_s2[ 1] = salt_buf1 << 16 | salt_buf2 >> 16; + w_s2[ 2] = salt_buf2 << 16 | 0; + w_s2[15] = (510 + 8) * 8; + + #ifdef _unroll + #pragma unroll + #endif + for (int i = 16; i < 64; i++) + { + w_s2[i] = SHA256_EXPAND_S (w_s2[i - 2], w_s2[i - 7], w_s2[i - 15], w_s2[i - 16]); + } + } + + barrier (CLK_LOCAL_MEM_FENCE); if (gid >= gid_max) return; + /** + * base + */ + u32 pw_buf0[4]; + u32 pw_buf1[4]; pw_buf0[0] = pws[gid].i[ 0]; pw_buf0[1] = pws[gid].i[ 1]; pw_buf0[2] = pws[gid].i[ 2]; pw_buf0[3] = pws[gid].i[ 3]; - - u32 pw_buf1[4]; - pw_buf1[0] = pws[gid].i[ 4]; pw_buf1[1] = pws[gid].i[ 5]; pw_buf1[2] = pws[gid].i[ 6]; @@ -340,14 +485,6 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_s04 (__glo const u32 pw_len = pws[gid].pw_len; - /** - * salt - */ - - const u32 salt_buf0 = swap32 (salt_bufs[salt_pos].salt_buf[ 0]); - const u32 salt_buf1 = swap32 (salt_bufs[salt_pos].salt_buf[ 1]); - const u32 salt_buf2 = swap32 (salt_bufs[salt_pos].salt_buf[ 2]); // 0x80 - /** * digest */ @@ -364,42 +501,24 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_s04 (__glo * loop */ - for (u32 il_pos = 0; il_pos < rules_cnt; il_pos++) + for (u32 il_pos = 0; il_pos < il_cnt; il_pos += VECT_SIZE) { - u32 w0[4]; - u32 w1[4]; - u32 w2[4]; - u32 w3[4]; - - w0[0] = pw_buf0[0]; - w0[1] = pw_buf0[1]; - w0[2] = pw_buf0[2]; - w0[3] = pw_buf0[3]; - w1[0] = pw_buf1[0]; - w1[1] = pw_buf1[1]; - w1[2] = pw_buf1[2]; - w1[3] = pw_buf1[3]; - w2[0] = 0; - w2[1] = 0; - w2[2] = 0; - w2[3] = 0; - w3[0] = 0; - w3[1] = 0; - w3[2] = 0; - w3[3] = 0; - - const u32 out_len = apply_rules (rules_buf[il_pos].cmds, w0, w1, pw_len); - - u32 w0_t[4]; - u32 w1_t[4]; - u32 w2_t[4]; - u32 w3_t[4]; + u32x w0[4] = { 0 }; + u32x w1[4] = { 0 }; + u32x w2[4] = { 0 }; + u32x w3[4] = { 0 }; - make_unicode (w0, w0_t, w1_t); + apply_rules_vect (pw_buf0, pw_buf1, pw_len, rules_buf, il_pos, w0, w1); + + 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); - u32 w_t[16]; + u32x w_t[16]; w_t[ 0] = swap32 (w0_t[0]); w_t[ 1] = swap32 (w0_t[1]); @@ -435,7 +554,7 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_s04 (__glo w_t[14] = w_t[14] >> 8; w_t[15] = w_t[15] >> 8; - u32 digest[8]; + u32x digest[8]; digest[0] = SHA256M_A; digest[1] = SHA256M_B; @@ -446,56 +565,24 @@ __kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_s04 (__glo digest[6] = SHA256M_G; digest[7] = SHA256M_H; - sha256_transform (digest, w_t); // 0 - 64 - - w_t[ 0] = 0; - w_t[ 1] = 0; - w_t[ 2] = 0; - w_t[ 3] = 0; - w_t[ 4] = 0; - w_t[ 5] = 0; - w_t[ 6] = 0; - w_t[ 7] = 0; - w_t[ 8] = 0; - w_t[ 9] = 0; - w_t[10] = 0; - w_t[11] = 0; - w_t[12] = 0; - w_t[13] = 0; - w_t[14] = 0; - w_t[15] = 0; - - sha256_transform (digest, w_t); // 64 - 128 - sha256_transform (digest, w_t); // 128 - 192 - sha256_transform (digest, w_t); // 192 - 256 - sha256_transform (digest, w_t); // 256 - 320 - sha256_transform (digest, w_t); // 320 - 384 - sha256_transform (digest, w_t); // 384 - 448 - - w_t[15] = 0 | salt_buf0 >> 16; - - sha256_transform (digest, w_t); // 448 - 512 - - w_t[ 0] = salt_buf0 << 16 | salt_buf1 >> 16; - w_t[ 1] = salt_buf1 << 16 | salt_buf2 >> 16; - w_t[ 2] = salt_buf2 << 16 | 0; - w_t[15] = (510 + 8) * 8; - - sha256_transform (digest, w_t); // 512 - 576 - - const u32 r0 = digest[3]; - const u32 r1 = digest[7]; - const u32 r2 = digest[2]; - const u32 r3 = digest[6]; - - #include COMPARE_S + sha256_transform (digest, w_t); // 0 - 64 + sha256_transform_z (digest); // 64 - 128 + sha256_transform_z (digest); // 128 - 192 + sha256_transform_z (digest); // 192 - 256 + sha256_transform_z (digest); // 256 - 320 + sha256_transform_z (digest); // 320 - 384 + sha256_transform_z (digest); // 384 - 448 + sha256_transform_s (digest, w_s1); // 448 - 512 + sha256_transform_s (digest, w_s2); // 512 - 576 + + COMPARE_S_SIMD (digest[3], digest[7], digest[2], digest[6]); } } -__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_s08 (__global pw_t *pws, __global gpu_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 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) +__kernel void m08000_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 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) { } -__kernel void __attribute__((reqd_work_group_size (64, 1, 1))) m08000_s16 (__global pw_t *pws, __global gpu_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 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) +__kernel void m08000_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 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) { }